cvpr2015papers

@http://www-cs-faculty.stanford.edu/people/karpathy/cvpr2015papers/

CVPR 2015 papers

(in nicer format than  this)
maintained by  @karpathy
cvpr2015papers_第1张图片 NEW: This year I also embedded the (1,2-gram) tfidf vectors of all papers with t-sne and placed them in an interface where you can navigate them visually. I'm not sure if it's useful but it's really cool.
Below every paper are TOP 100 most-occuring words in that paper and their color is based on LDA topic model with k = 7.
(It looks like 0 = datasets?, 1 = deep learning, 2 = videos , 3 = 3D Computer Vision , 4 = optimization?, 5 = low-level Computer Vision?, 6 = descriptors?)
Toggle LDA topics to sort by:  TOPIC0  TOPIC1  TOPIC2  TOPIC3  TOPIC4  TOPIC5  TOPIC6
DeepID-Net: Deformable Deep Convolutional Neural Networks for Object Detection
Ouyang, Wanli, Wang, Xiaogang, Zeng, Xingyu, Qiu, Shi, Luo, Ping, Tian, Yonglong, Li, Hongsheng, Yang, Shuo, Wang, Zhe, Loy, Chen-Change, Tang, Xiaoou

cvpr2015papers_第2张图片
[learning, visual, annotation, baseline, human, task, rejection, example, test, semantic, horse] [deep, layer, map, convolutional, pretraining, imagenet, learned, training, improves, rcnn, table, feature, size, output, learn, context, pooling, pretrain, contextual, pedestrian, cnn, net, better, class, neural, input, shared, train, arxiv, preprint, representation, generic, global, max, network] [object, detection, bounding, box, multiple, based, allowed] [model, deformation, deformable, approach, location, localization] [data, penalty, proposed, experimental, upper, det, denoted] [circular, figure, pattern, averaging, paper] [image, scheme, large]
Convolutional Neural Networks at Constrained Time Cost
He, Kaiming, Sun, Jian

cvpr2015papers_第3张图片
[involved, baseline, understanding] [convolutional, layer, stage, table, pooling, feature, training, accuracy, size, actual, stride, testing, neural, width, map, deep, network, imagenet, investigate, replacement, cnns, sppnet, deeper, previous, increased, smaller, replace, max, better, increase, gpu, augmentation, pool, increasing, replaced, pad, recognition, issue, subsampling, reported, trained] [cost, based, three, memory, series, detection] [model, depth, error, constrained, single, controlled, structure] [time, complexity, running, number, data, reduces, reduced, consider, theoretical, faster] [accurate, paper, commercial, second, pixel] [image, fast, compare, spatial, subsequent, preserve]
Modeling Local and Global Deformations in Deep Learning: Epitomic Convolution, Multiple Instance Learning, and Sliding Window Detection
Papandreou, George, Kokkinos, Iasonas, Savalle, Pierre-Andre

cvpr2015papers_第4张图片
[search, learning, work, set, system, baseline, instance, positive] [epitomic, network, deep, training, convolutional, layer, convolution, input, window, sliding, dcnn, imagenet, size, performance, output, stride, table, global, architecture, neural, epitome, map, response, dcnns, conv, loss, train, max, cnn, pascal, better, fully, accuracy, trous, trained, feature, standard, testing, pyramid, alternative, validation] [object, detection, multiple, score, bounding, box] [position, model, error, aspect, modeling, single, allows, full] [proposed, maximum, computation, data, algorithm] [ratio, small, figure] [image, scale, best, computing, region, invariance, spatial, large, local, rate]
End-to-End Integration of a Convolution Network, Deformable Parts Model and Non-Maximum Suppression
Wan, Li, Eigen, David, Fergus, Rob

cvpr2015papers_第5张图片
[set, system, learning, human, positive, prediction, perform, green, predicted] [loss, training, convolutional, convnet, dpm, network, root, input, layer, performance, train, feature, pascal, response, negative, trained, apply, def, hog, max, output, map, size, table, pyramid, voc, convolution, class, improves, deep, integrating, imagenet, structured, achieves, neural, discriminative, activation] [object, bounding, detection, box, appearance, cost, suppression, entire, online, three] [model, deformation, ground, deformable, location, joint] [note, assignment, function, form, machine] [figure, red, computer, ieee, relative] [image, view, spatial, applied, scale, person, well, region]
Taking a Deeper Look at Pedestrians
Hosang, Jan, Omran, Mohamed, Benenson, Rodrigo, Schiele, Bernt

cvpr2015papers_第6张图片
[test, set, learning, task, default, work, datasets] [pedestrian, training, caltech, alexnet, cifarnet, table, convnets, network, performance, convnet, convolutional, neural, squareschnftrs, trained, layer, previous, input, imagenet, validation, size, deep, decision, top, detector, scf, improves, additional, sdn, stride, better, architecture, acf, spatialpooling, window, reach, improvement, ldcf, feature, katamari, jointdeep, improve, pool, competitive, recognition, published, negative, max, pascal] [detection, proposal, object, iou, indicates, edgeboxes, recall] [kitti, model, provide] [data, method, number, consider] [quality, small, figure, ratio] [best, large, good, despite, image]
Is Object Localization for Free? - Weakly-Supervised Learning With Convolutional Neural Networks
Oquab, Maxime, Bottou, Leon, Laptev, Ivan, Sivic, Josef

cvpr2015papers_第7张图片
[supervised, weakly, learning, prediction, coco, visual, test, predict, search, work, dataset, unsupervised, label, indicating, instance, introduce] [network, training, convolutional, pascal, voc, input, output, window, table, neural, performance, architecture, class, recognition, supervision, map, fully, eak, adding, cnn, global, microsoft, layer, deep, train, sliding, extent, validation, size, trained, connected] [object, bounding, score, multiple, detection, box, complex, segmentation, localize, entire] [location, localization, cluttered, single, position, model, scene, procedure] [method, operation, note, form] [figure, presence, setup] [image, limited, illustrated, scale, area]
Convolutional Feature Masking for Joint Object and Stuff Segmentation
Dai, Jifeng, He, Kaiming, Sun, Jian

cvpr2015papers_第8张图片
[semantic, set, process, baseline, search, positive, domain, category] [feature, convolutional, stuff, cfm, table, pascal, training, voc, masking, design, network, layer, raw, sppnet, input, stochastic, spp, vgg, grass, apr, sgd, extracted, accuracy, pyramid, svm, validation, reported, selective, pooling, train, map, neural, cnn, deep, apply, fully, arxiv, rectangular, generic] [segment, segmentation, object, proposal, detection, based, candidate, combination, score, pursuit, iou, bounding, deterministic, generate, entire, mcg] [joint, single, model] [method, compact, binary, time] [figure, small, measured] [image, region, person, sampled, masked, spatial, impact]
Fully Convolutional Networks for Semantic Segmentation
Long, Jonathan, Shelhamer, Evan, Darrell, Trevor

cvpr2015papers_第9张图片
[semantic, prediction, learning, test, work, visual, vision, set] [convolutional, fully, training, output, net, layer, stride, input, pascal, deep, class, coarse, performance, loss, receptive, feature, architecture, voc, neural, learn, size, table, patchwise, learned, skip, convolution, trained, train, fcn, improvement, convnets, factor, tighe, effective, nii, convnet] [segmentation, detection, object] [model, inference, described, single, full, pixelwise] [computation, data, number, time, nonlinear, random, fusion, convergence, faster, bilinear] [figure, pixel, upsampling, produce, relative, computer, gupta] [image, dense, spatial, sampling, local, sift, upsampled, rate, gradient, fast]
Deformable Part Models are Convolutional Neural Networks
Girshick, Ross, Iandola, Forrest, Darrell, Trevor, Malik, Jitendra

cvpr2015papers_第10张图片
[work, visual, learning, baseline, describe, human, report, set] [dpm, pyramid, feature, level, pooling, map, max, cnn, deeppyramid, convolutional, hog, root, dpms, network, voc, deep, training, input, layer, convolution, performance, negative, neural, maxout, pascal, standard, output, cnns, implemented, trained, train, stride, table, learned, top] [object, detection, score] [model, deformable, inference, face, mixture] [method, equivalent, algorithm, function, construction, number, idea, case, sparse, call] [component, geometry, figure, second] [image, distance, transform, outperforms, implementation, computed, spatial, filter]
Pedestrian Detection Aided by Deep Learning Semantic Tasks
Tian, Yonglong, Luo, Ping, Wang, Xiaogang, Tang, Xiaoou

cvpr2015papers_第11张图片
[attribute, learning, hard, datasets, task, positive, semantic, existing, work, human, dataset] [pedestrian, deep, log, negative, shared, feature, learned, learn, channel, representation, effectiveness, eth, jointly, unshared, achieved, percent, hog, loss, cnn, jointdeep, improve, layer, tacnn, table, acf, caltech, trained, convolutional, network, training, performance, context, previous, road] [detection, multiple, background, including, segmentation, object] [scene, model, single, structure, tree, body, challenging, mixture, occlusion, deformable] [main, data, number, sample, binary, proposed, vector, reduces, denoted, function, matrix, projection] [figure, motion, second, vehicle, carefully, result, vertical, sky] [rate, outperforms, large, corresponding, image]
Deeply Learned Face Representations Are Sparse, Selective, and Robust
Sun, Yi, Wang, Xiaogang, Tang, Xiaoou

cvpr2015papers_第12张图片
[attribute, learning, people, selected, evaluate, test] [neural, deep, identity, activation, net, recognition, accuracy, layer, feature, activated, training, performance, learned, neuron, convolutional, representation, higher, standard, supervisory, discriminative, hidden, combining, inhibitory, excited, constantly, compared, excitatory, belonging, extracted, achieve, achieved, dropout] [three, moderate, distinguish, youtube, entire] [face, single, joint, occluded, occlusion] [binary, sparsity, block, original, random, sparse, code, number] [figure, robust, high, half, second, red, left, paper] [lbp, lfw, best, bayesian, comparison, histogram, image, remaining, large, distance, partial]
Fisher Vectors Meet Neural Networks: A Hybrid Classification Architecture
Perronnin, Florent, Larlus, Diane

cvpr2015papers_第13张图片
[supervised, unsupervised, learning, set, report, visual, perform, test, study, work, system] [hidden, architecture, layer, cnn, training, output, deep, feature, standard, fisher, bagging, convolutional, fvs, learned, hybrid, involves, intermediate, table, pascal, neural, recognition, augmentation, validation, reported, train, map, competitive, cnns, trained, simonyan, better, accuracy, challenge, learn, imagenet, itp] [multiple, object, cost] [error, model, single] [linear, number, kernel, proposed, dimensionality, data, case, random, reduction, vector, encoding, gmm, consider] [figure] [image, large, local, best, pca, rate, compare, retrieval, comparison, scale]
segDeepM: Exploiting Segmentation and Context in Deep Neural Networks for Object Detection
Zhu, Yukun, Urtasun, Raquel, Salakhutdinov, Ruslan, Fidler, Sanja

cvpr2015papers_第14张图片
[set, car, dog, bird, test, horse, scoring, evaluate, bottle, prediction, cat, semantic, exploit, learning, cow, boat] [segdeepm, feature, contextual, performance, pascal, table, voc, val, rcnn, class, additional, improvement, chair, train, map, network, detector, context, deep, training, achieves, segdpm, expanded, neural, trained, convolutional, better, top, layer, input, cnn, bus] [object, box, segmentation, detection, segment, bounding, candidate, score, recall, precision, appearance, cpmc, boost, indicates, hypothesis] [approach, model, current, inference, full] [curve, method, random, number, proposed, variable, order, iterative, note] [figure, small, accurate, plane, result] [person, image, best, grid, computed, well, choose, place, compute]
Saliency Detection by Multi-Context Deep Learning
Zhao, Rui, Ouyang, Wanli, Li, Hongsheng, Wang, Xiaogang

cvpr2015papers_第15张图片
[learning, visual, strategy, dataset, evaluation, prediction, task, work, datasets] [deep, context, global, xlc, convolutional, xgc, training, contemporary, neural, imagenet, network, input, cnn, feature, performance, arxiv, layer, preprint, jointly, map, connected, pcas, table, fully, clarifai, googlenet, learned, pretraining, padded] [saliency, object, detection, salient, contrast, segmentation, framework, based, superpixel, score, background, groundtruth, bounding, drfi, gbmr, objectness, ecssd, three, including] [model, modeling, approach, full, initialization, provide] [proposed, function] [figure, ieee] [image, local, scheme, region, outperforms, large]
Efficient Object Localization Using Convolutional Networks
Tompson, Jonathan, Goroshin, Ross, Jain, Arjun, LeCun, Yann, Bregler, Christoph

cvpr2015papers_第16张图片
[work, human, dataset, learning, set, perform, test, predicted, prediction, instance] [coarse, network, flic, pool, performance, input, convolution, pooling, architecture, convnet, relu, dropout, convolutional, feature, cascade, training, output, standard, layer, improve, table, trained, crop, heatmap, size, cropped, shoulder, maxpool, generalization, detector, accuracy, tompson, shared, sapp, train, toshev, spatialdropout] [detection, mpii, target, average] [model, joint, pose, wrist, location, localization, error, regression, cascaded, estimation, single, body, face, pck, elbow, yang] [normalized, time, number] [figure, high, small, pixel] [spatial, image, distance, scale, resolution, comparison]
Hypercolumns for Object Segmentation and Fine-Grained Localization
Hariharan, Bharath, Arbelaez, Pablo, Girshick, Ross, Malik, Jitendra

cvpr2015papers_第17张图片
[system, semantic, baseline, work, set, task, category, prediction, predict, probability, instance, evaluate, positive] [feature, hypercolumn, network, convolutional, heatmap, training, top, layer, pipeline, representation, apr, better, cnn, train, hypercolumns, architecture, output, table, map, performance, neural, cropped, expanded, upsample, trained, fully, hyp, deep, higher, combine, gain, apk] [bounding, box, detection, segmentation, score, object, overlap, multiple, based, segment, threshold, action] [location, pose, localization, estimation, reasoning, ground, point, described, starting, allows] [linear, vector, original] [figure, pixel, separate, optical, produce] [keypoint, grid, person, region, large, best, image, metric, neighborhood]
A Discriminative CNN Video Representation for Event Detection
Xu, Zhongwen, Yang, Yi, Hauptmann, Alex G.

cvpr2015papers_第18张图片
[latent, concept, med, prediction, positive, visual, collection, competition] [cnn, pooling, performance, medtest, fisher, representation, trecvid, convolutional, map, improved, table, feature, encoded, level, apply, layer, deep, better, training, achieves, spp, network, size, medeval, discriminative, standard, extracted, achieve, cosine, neural] [video, event, average, detection, frame, generate, utilize, complex, action, propose, multiple, multimedia, idt] [] [vector, encoding, dimension, original, proposed, linear, number, fusion, product, data, quantization, analysis] [figure, motion] [vlad, dense, local, best, large, spatial, descriptor, image, similarity, pca, good, comparison, nearest, impact]
Deep Roto-Translation Scattering for Object Classification
Oyallon, Edouard, Mallat, Stephane

cvpr2015papers_第19张图片
[supervised, unsupervised, learning, set, type] [network, deep, feature, training, layer, convolution, representation, caltech, convolutional, accuracy, class, neural, output, table, applying, size, computes, morlet, apply, cascade, build, larger, trained, pooling] [complex, object, performs, art, state, multiple] [prior, indexed, depth, rigid] [order, linear, orthogonal, data, gaussian, number, cifar, random, operator, algorithm, vector] [scattering, second, modulus, square, rotation, color, averaging, translation, removed, figure, angular, angle, small, computer, cascading] [image, wavelet, spatial, transform, computed, scale, geometric, invariant, dict, sift, separable, dictionary, stable, adapted, transforms, invariance, well]
The Treasure Beneath Convolutional Layers: Cross-Convolutional-Layer Pooling for Image Classification
Liu, Lingqiao, Shen, Chunhua, van den Hengel, Anton

cvpr2015papers_第20张图片
[visual, dataset, attribute, existing, indicator, report, create, describe, perform, style] [convolutional, layer, feature, pooling, performance, dcnn, table, representation, input, applying, pretrained, scfv, extracted, map, deep, better, training, neural, network, extracting, pooled, discriminative, achieves, regional, unit, combining, cnn, power, fully, connected, comparable, dcnns, pascal, issue] [object, multiple, extract, based] [scene] [method, proposed, vector, note, sign, encoding, computational, data, quantization, time, dimensionality, experiment, number] [ieee, directly, figure] [image, local, spatial, applied, best, comparison, corresponding, resolution, outperforms]
MatchNet: Unifying Feature and Metric Learning for Patch-Based Matching
Han, Xufeng, Leung, Thomas, Jia, Yangqing, Sukthankar, Rahul, Berg, Alexander C.

cvpr2015papers_第21张图片
[learning, set, evaluate, evaluation, dataset, work, positive, labeled] [feature, network, matchnet, training, convolutional, bottleneck, neural, representation, layer, performance, deep, better, size, input, output, accuracy, learned, trained, jointly, table, nsift, previous, pooling, standard, train, learn, fewer, notredame, negative, simonyan, achieves, learns, tower, recognition, convolution, response, fully] [average, generate, object, based, group] [error, approach, model, structure, point, quantized] [dimension, batch, number] [figure, stereo, robust, absolute] [metric, patch, matching, similarity, image, local, sift, descriptor, distance, rate, pair, difference, buffer, sampling, comparison, liberty, reservoir, gradient, best, sampler]
Recurrent Convolutional Neural Network for Object Recognition
Liang, Ming, Hu, Xiaolin

cvpr2015papers_第22张图片
[learning, visual, set, processing, existing, unsupervised, baseline] [recurrent, neural, rcnn, layer, cnn, convolutional, deep, feature, recognition, training, network, input, maxout, pooling, testing, context, table, unit, max, performance, architecture, rcl, fewer, better, dsn, size, connection, augmentation, global, increasing, output, receptive, achieved, dropout, grant, outperformed] [object, hierarchical, three, weight, dynamic, path] [model, error, depth, structure, current] [number, data, nin, machine, proposed, time, journal, mnist, note, lower, sparse, function] [conference, international, figure, computer, ieee, pattern] [comparison, image, recursive, local, benchmark, rate, large]
Scene Labeling With LSTM Recurrent Neural Networks
Byeon, Wonmin, Breuel, Thomas M., Raue, Federico, Liwicki, Marcus

cvpr2015papers_第23张图片
[labeling, learning, dataset, label, include, vision, processing, predicted, natural] [lstm, input, layer, class, neural, accuracy, recurrent, performance, size, feedforward, global, cnns, output, training, hidden, stanford, convolutional, network, context, deep, reported, learn, window, compared, feature, architecture, testing, road, grass, unbalanced, gate, higher, rcnns, contextual, improve, improved, table] [memory, long, target, background, segmentation, three, object] [scene, approach] [time, number, machine, third, block, note, random] [figure, pixel, computer, conference, ieee, frequency, pattern, international, flow, second, water, small] [image, large, sift, local, spatial, texture]
Beyond Short Snippets: Deep Networks for Video Classification
Yue-Hei Ng, Joe, Hausknecht, Matthew, Vijayanarasimhan, Sudheendra, Vinyals, Oriol, Monga, Rajat, Toderici, George

cvpr2015papers_第24张图片
[learning, dataset, work, process, human] [pooling, lstm, convolutional, network, layer, cnn, neural, feature, conv, recognition, googlenet, table, recurrent, better, training, performance, input, output, global, alexnet, learn, raw, accuracy, connected, architecture, deep, cnns, trained, fully, previous, compared, level, larger, gate, hidden, softmax, achieve] [video, frame, temporal, action, memory, slow, short, long, clip, late, average, capable, sequence, entire] [model, single, approach, performed] [method, number, time, order, proposed, data, lower, fusion] [optical, motion, flow, small, figure] [image, local, applied, rate, best]
Designing Deep Networks for Surface Normal Estimation
Wang, Xiaolong, Fouhey, David, Gupta, Abhinav

cvpr2015papers_第25张图片
[learning, work, task, understanding, dataset, set, prediction, predicting, vision, evaluate, label] [network, global, output, convolutional, input, coarse, performance, fully, connected, window, size, deep, table, apply, training, top, architecture, improve, layer, cnn, learn, better, sliding, structured, loss, cnns, neural, feature, ablative, conv] [edge, object, box, art, detection, state, complementary] [surface, normal, layout, room, scene, single, estimation, full, reasoning, model, approach, predicts, structure, depth, qualitative, capture, regression, indoor, occlusion] [fusion, data, note, form, problem] [figure, vanishing, computer, captured] [local, image, rate, well]
DEEP-CARVING: Discovering Visual Attributes by Carving Deep Neural Nets
Shankar, Sukrit, Garg, Vikas K., Cipolla, Roberto

cvpr2015papers_第26张图片
[attribute, label, instance, visual, learning, supervised, sad, weakly, test, search, set, dataset, prediction, correct, natural, predicting, fig, noun, entropy, sun, positive, task, predicted, weak, scenario, carve, ranking, vision] [training, deep, net, feature, loss, convolutional, layer, recognition, cnns, alexnet, softmax, sigmoid, neural, learn, setting, connected, better, class, architecture, input, cnn, fully] [multiple, object, indicates, precision, average, based] [provide, scene] [number, problem, consider, note, vector] [figure, computer, generally, presence, color] [image, best, cross]
Learning to Generate Chairs With Convolutional Neural Networks
Dosovitskiy, Alexey, Tobias Springenberg, Jost, Brox, Thomas

cvpr2015papers_第27张图片
[learning, set, knowledge, dataset, transfer, source, interesting, work, example, generation, unsupervised, label, manually] [network, feature, training, chair, deep, neural, convolutional, fully, connected, activating, intermediate, layer, performance, class, representation, size, top, stochastic, hidden, morphings, input, output, additional, activation, trained, rbms, architecture, neuron] [generate, generated, target, object, average, segmentation, combination, three] [generative, single, rgb, model, face, viewpoint, approach, allows, inference] [row, number, data, simply, method, gaussian] [figure, small, quality, second, optical, interpolation, interpolate, conference] [image, nearest, pair, large, keypoints, corresponding, spatial, variation, morphing, sift]
Learning to Compare Image Patches via Convolutional Neural Networks
Zagoruyko, Sergey, Komodakis, Nikos

cvpr2015papers_第28张图片
[positive, learning, dataset, test, set, baseline, evaluation, vision, variety] [network, layer, convolutional, neural, performance, input, training, size, architecture, decision, top, trained, feature, spp, simonyan, better, pooling, learn, deep, fully, connected, output, etal, outperform, relu, train, shared, increasing, notredame] [branch, cost] [error, depth, approach] [function, siam, case, mrf, proposed, number, basic] [stereo, figure, computer, ieee, pattern, central, directly, photometric, account] [image, siamese, patch, rate, similarity, false, comparing, descriptor, matching, spatial, local, corresponding, sift, true, consists, wide, computed, mser, computing, liberty, compare, yosemite, applied, benchmark]
Object Detection by Labeling Superpixels
Yan, Junjie, Yu, Yinan, Zhu, Xiangyu, Lei, Zhen, Li, Stan Z.

cvpr2015papers_第29张图片
[labeling, label, generation, set, semantic, category, system, learning, search, visual] [cnn, pascal, rcnn, performance, voc, imagenet, feature, better, deep, convolutional, neural, output, learned, sliding, training, pedestrian, arxiv, preprint, window, testing, belongs, map, cnns, trained, top, selective, layer, googlenet, improve, context, table, improvement] [detection, object, superpixel, cost, based, proposal, superpixels, appearance, bounding, recall, partition, segmentation, structural, weight, truth, generate, leading, score, overlap] [model, term, inference, face, ground, infer, current, scene, single] [method, data, proposed, number, energy, function, compact, problem] [result, smooth, small, ratio, paper, pixel, robust] [image, large, neighborhood]
Visual Saliency Based on Multiscale Deep Features
Li, Guanbin, Yu, Yizhou

cvpr2015papers_第30张图片
[visual, dataset, set, label, learning, existing, labeled, human] [cnn, feature, neural, deep, trained, map, network, training, connected, fully, convolutional, performance, extracted, output, achieves, rectangular, fed, imagenet, architecture, layer, recognition] [saliency, salient, recall, three, precision, multiscale, contrast, object, mdf, segmentation, based, drfi, multiple, score, attention, edge, hierarchical, bounding, box, detection, strength, entire, surrounding, including, truth, generate, msra, boundary, capable, extraction, background] [model, pixelwise, challenging, ground] [method, vector, number, respect, binary, proposed] [coherence, second, pixel] [image, region, spatial, well, considered, large, best, distance, comparison]
Scene Classification With Semantic Fisher Vectors
Dixit, Mandar, Chen, Si, Gao, Dashan, Rasiwasia, Nikhil, Vasconcelos, Nuno

cvpr2015papers_第31张图片
[semantic, embedding, natural, probability, semantics, vision, visual, sun, bag, embeddings, learning] [cnn, fisher, imagenet, feature, bos, convolutional, representation, extracted, mit, layer, recognition, retinotopic, table, multinomial, bof, log, network, class, trained, fully, performance, neural, global, better, indirect, implemented, highly, effective, map, connected, architecture, learned, smn] [object, based, localized] [scene, indoor, mixture, posterior, modeling] [space, vector, mapping, parameter, euclidean, denoted, encoding, linear, distribution, gaussian, dimensional, proposed, projection, problem] [figure, computer, ieee, conference, pattern, directly, direct] [image, patch, local, implementation, comparison, invariant, discriminant, invariance, best, pca, scale]
The Application of Two-Level Attention Models in Deep Convolutional Neural Network for Fine-Grained Image Classification
Xiao, Tianjun, Xu, Yichong, Yang, Kuiyuan, Zhang, Jiaxing, Peng, Yuxin, Zhang, Zheng

cvpr2015papers_第32张图片
[bird, dog, relevant, set, selected, work, dataset, search, visual, process, prediction, domain, learning, select] [cnn, feature, domainnet, level, training, pipeline, performance, accuracy, deep, convolutional, detector, trained, discriminative, network, weakest, advantage, better, class, testing, filternet, train, supervision, extractor, softmax, dpd, neural, selective, arxiv, achieves, preprint, learn] [attention, object, bounding, detection, score, box, based, group, extract, foreground, propose, three, utilize] [error, pose, step, provide, deformable, detailed] [method, cluster, proposed, data, clustering, selects] [figure, second, high, phase] [image, patch, best, large]
Deep LAC: Deep Localization, Alignment and Classification for Fine-Grained Recognition
Lin, Di, Shen, Xiaoyong, Lu, Cewu, Jia, Jiaya

cvpr2015papers_第33张图片
[bird, work, dataset, set, evaluate, category, system, semantic, label] [lac, accuracy, table, deep, vlf, neural, cnn, convolutional, network, performance, recognition, valve, training, previous, input, output, layer, subnetwork, connected, stage, feature, fully, achieved, module, follow, imagenet, representation] [bounding, object, localized, propose, three, box, extract, truth, lgt, framework, based, detection] [alignment, localization, head, body, template, model, location, linkage, error, ground, pose, aligned, term, joint] [function, method, expressed, form, update, case, denoted, objective, analysis] [figure, center, zhang, rotation] [image, large, comparison, similarity, region, major, best]
Deep Neural Networks Are Easily Fooled: High Confidence Predictions for Unrecognizable Images
Nguyen, Anh, Yosinski, Jason, Clune, Jeff

cvpr2015papers_第34张图片
[human, dataset, natural, test, recognize, label, vision, diversity, set, learning, visual, prediction, tend, perform] [dnns, dnn, imagenet, class, cppn, unrecognizable, evolutionary, neural, trained, recognizable, training, fooling, fooled, deep, discriminative, evolved, fool, input, evolving, learn, recognition, learned, evolve, imperceptible, convolutional, arxiv, preprint, encoded, suggests, network, performance, negative, larger, belonging] [evolution, target, score, three, generated] [model, independent, generative] [mnist, easily, ascent, encoding, optimization, repeated] [produce, high, figure, produced, median, computer, small, directly, result, pattern, ieee, conference, producing] [image, gradient, regular, well, match]
From Image-Level to Pixel-Level Labeling With Convolutional Networks
Pinheiro, Pedro O., Collobert, Ronan

cvpr2015papers_第35张图片
[weakly, supervised, vision, learning, dataset, test, label, set, semantic, task, probability, system, instance, datasets, bird] [class, pascal, training, voc, smoothing, convolutional, trained, feature, cnn, output, imagenet, recognition, table, network, layer, overfeat, compared, neural, performance, input, fully, size, architecture, achieved, train, lse, cnns] [segmentation, object, score, bounding, box, detection, three, weight, average, background, based, multiple, precision] [model, prior, approach, inference, rgb] [consider, data, proposed, number, algorithm, extra, problem] [conference, computer, pixel, ieee, pattern, figure, international] [image, aggregation, large, considered, compare, false, well]
Material Recognition in the Wild With the Materials in Context Database
Bell, Sean, Upchurch, Paul, Snavely, Noah, Bala, Kavita

cvpr2015papers_第36张图片
[dataset, minc, test, stone, click, learning, wood, set, foliage, category, work, example, plastic, tile, predict, flickr, evaluate, carpet, convert] [cnn, accuracy, training, alexnet, recognition, opensurfaces, class, googlenet, convolutional, train, fmd, crf, table, cnns, context, trained, sliding, fully, smaller, size, input, larger, network, neural, connected, wallpaper, better, houzz, improve, polished, deep] [segmentation, average, recognizing, object] [full, scene, prior, single, mirror] [data, range, experiment, number, pairwise, method] [material, figure, water, fabric, paper, sky, high, total, glass] [patch, image, best, database, scale, large, dense, spatial, sift]
Going Deeper With Convolutions
Szegedy, Christian, Liu, Wei, Jia, Yangqing, Sermanet, Pierre, Reed, Scott, Anguelov, Dragomir, Erhan, Dumitru, Vanhoucke, Vincent, Rabinovich, Andrew

cvpr2015papers_第37张图片
[learning, work, vision, visual] [conv, inception, network, deep, convolutional, layer, architecture, neural, googlenet, maxpool, increase, ilsvrc, imagenet, size, performance, table, pooling, output, depthconcat, larger, suggests, challenge, training, higher, increasing, ensemble, max, top, pool, trained, accuracy, connected, improved, compared, previous, increased, stage, expensive, dropout, auxiliary, validation, design, highly] [detection, object, bounding, box, based, average] [model, approach, current, single, localization, error, inference, structure, depth, allows] [number, sparse, computational, data, linear, order, main, reduction, optimal, matrix, problem, construction] [computer, figure, quality, result] [image, large, well, dense, external, rate, region]
A Dynamic Convolutional Layer for Short Range Weather Prediction
Klein, Benjamin, Wolf, Lior, Afek, Yehuda

cvpr2015papers_第38张图片
[learning, dataset, prediction, vision, set] [layer, convolutional, convolution, network, input, feature, cnn, radar, output, kij, previous, size, passed, architecture, neural, deep, learned, map, training, loss, computes, applying, validation, performance, tel, arxiv, preprint, recognition, multiplicative, global, max, aviv, softmax, stride, connected, generalization, rain, fully, compared] [dynamic, based, sequence, three, short, frame, forward, contrast] [error, regression, model] [weather, function, sample, order, linear, method, vector, pass, range, euclidean] [conventional, figure, computer, motion, conference, center, color, result, second, ieee, pattern] [image, patch, computed, gradient, synthesis, applied, represents, best, performing, comparison]
Deep Filter Banks for Texture Recognition and Segmentation
Cimpoi, Mircea, Maji, Subhransu, Vedaldi, Andrea

cvpr2015papers_第39张图片
[dataset, datasets, example, msrc, evaluation, domain, learning, visual, vision, set, contribution, collected, cub] [recognition, convolutional, cnn, performance, pooling, better, mit, deep, accuracy, pascal, fmd, dtd, orderless, extracted, trained, improves, cnns, network, size, layer, voc, input, crisp, imagenet, table, categorisation, advantage, pooled, technical, feature, fully, comparable] [segmentation, object, recognizing, performs] [indoor, scene, approach, shape, model, fact] [method, number, data, acc, vector, analysis, order] [material, computer, paper] [texture, image, region, best, descriptor, local, applied, sift, corresponding, well, large, computed]
DeepEdge: A Multi-Scale Bifurcated Deep Network for Top-Down Contour Detection
Bertasius, Gedas, Shi, Jianbo, Torresani, Lorenzo

cvpr2015papers_第40张图片
[human, predict, learning, vision, prediction, perform, dataset] [contour, convolutional, architecture, feature, network, layer, pooling, deep, knet, input, table, trained, bifurcated, training, accuracy, achieves, train, labelers, negative, deepedge, neural, size, higher, recognition, fully, combining] [detection, object, branch, edge, candidate, average, multiple, extract, generate, considering, three, precision, truth] [regression, approach, point, ground, single] [method, number, note] [center, high, figure, pattern, computer, produced, small, second, ieee, presence, produce] [image, fraction, computed, patch, false, scale, marked]
FaceNet: A Unified Embedding for Face Recognition and Clustering
Schroff, Florian, Kalenichenko, Dmitry, Philbin, James

cvpr2015papers_第41张图片
[set, embedding, learning, test, personal, hard, select, positive, work, labeled, embeddings] [inception, loss, training, performance, network, table, triplet, deep, recognition, accuracy, size, convolutional, standard, architecture, input, negative, pooling, val, smaller, trained, layer, reported, representation, reduce, margin, identity, output, bottleneck, achieve, achieves, improve, cnn] [based, online, employ] [face, model, error, approach] [method, number, clustering, batch, data, order, space, dimensionality, reduced, compact] [figure, directly, squared, small, paper, clear] [distance, image, rate, large, lfw, false, well, accept, pair, anchor, person]
Rotating Your Face Using Multi-Task Deep Neural Network
Yim, Junho, Jung, Heechul, Yoo, ByungIn, Choi, Changkyu, Park, Dusik, Kim, Junmo

cvpr2015papers_第42张图片
[task, set, learning, vision, novel, written, test, dataset] [layer, input, recognition, connected, output, fully, identity, dnn, deep, multitask, feature, setting, pooling, frontal, table, cpf, performance, neural, training, auxiliary, multipie, size, extracted, cpi, better, network, designed, previous, convolutional, prepared] [target, path, weight, cost, generated] [face, pose, model, single, described, controlled] [code, main, experiment, proposed, arbitrary, function, method, original, parameter, space, avg] [figure, illumination, second, ieee, computer, conference, preserving, result, desired, pattern, changed, high, produced, international] [image, remote, locally, represents, change, best, equal, rate, preserve, demonstrate, construct, gallery, invariant]
Web-Scale Training for Face Identification
Taigman, Yaniv, Yang, Ming, Ranzato, Marc'Aurelio, Wolf, Lior

cvpr2015papers_第43张图片
[set, learning, dataset, open, vision, transfer, closed, system, bootstrapping, link, datasets, labeled, randomly, baseline, task, protocol, perform, selected, domain, report] [representation, training, performance, recognition, deep, network, layer, trained, convolutional, accuracy, neural, dir, table, larger, input, feature, bottleneck, deepface, train, learn, softmax, architecture, better] [three, bootstrapped, based] [face, initial, single, model] [norm, order, linear, random, vector, method, compressed, data, number, lower, underlying, dimensionality, note, selection] [computer, figure, commercial, ieee, directly, measured, clear] [lfw, large, image, well, benchmark, gallery, scale, similarity, best, internal, unconstrained]
Long-Term Recurrent Convolutional Networks for Visual Recognition and Description
Donahue, Jeffrey, Anne Hendricks, Lisa, Guadarrama, Sergio, Rohrbach, Marcus, Venugopalan, Subhashini, Saenko, Kate, Darrell, Trevor

cvpr2015papers_第44张图片
[visual, description, learning, language, sentence, vision, natural, generation, caption, human, semantic, evaluation, system, prediction] [lstm, input, convolutional, deep, lrcn, recognition, recurrent, neural, architecture, hidden, crf, learn, network, decoder, table, cnn, lstms, training, representation, output, arxiv, previous, preprint, gate, max, unit, rnn, convnet, timestep, feature, layer] [video, temporal, sequence, activity, state, memory, complex, average, based, action, frame, three, perceptual] [model, sequential, described, rgb, single, approach, current] [time, machine, proposed, base, data, number, variable, vector] [length, figure, translation, directly, stack] [image, cell, outperforms, retrieval, encoder]
Improving Object Detection With Deep Convolutional Networks via Bayesian Optimization and Structured Prediction
Zhang, Yuting, Sohn, Kihyuk, Villegas, Ruben, Pan, Gang, Lee, Honglak

cvpr2015papers_第45张图片
[search, set, learning, visual, baseline, test, evaluation, report, positive] [structured, fgs, cnn, performance, voc, selective, map, deep, svm, pascal, training, structobj, convolutional, neural, trained, optimum, table, improvement, train, max, loss, global, feature, better, standard, layer, higher, output, recognition, negative, network, detector] [bounding, detection, object, iou, box, score, framework, truth, based, multiple, entire, propose] [localization, ground, regression, model, supplementary, initial, single] [proposed, optimization, algorithm, function, method, number, objective, random, machine, problem, arg, analysis, computational, gaussian, maximum, distribution, solution] [figure, high, accurate] [local, image, region, bayesian, large, good]
Feedforward Semantic Segmentation With Zoom-Out Features
Mostajabi, Mohammadreza, Yadollahpour, Payman, Shakhnarovich, Gregory

cvpr2015papers_第46张图片
[semantic, work, set, learning, label, labeling, category, include, common, vision, nature, dataset] [feature, voc, accuracy, convolutional, level, loss, table, neural, network, class, architecture, training, structured, performance, arxiv, preprint, global, pooling, feedforward, train, size, better, layer, subscene, fully, additional, val, convnet, context, distant, extracted, recognition, receptive, deep, compared] [segmentation, superpixel, superpixels, object, background, complex, detection, iou, state, con, evidence] [scene, approach, capture, single, model, full, ground] [main, vector, data, machine, number, original, method, asymmetric] [figure, color, small, pixel, computer] [image, computed, local, region, expect, large, despite, applied]
Beyond Frontal Faces: Improving Person Recognition Using Multiple Cues
Zhang, Ning, Paluri, Manohar, Taigman, Yaniv, Fergus, Rob, Bourdev, Lubomir

cvpr2015papers_第47张图片
[set, people, test, dataset, photo, learning, task, vision, instance, example, subset, clothing, facebook, system, prediction, label, flickr, existing] [recognition, identity, performance, training, global, frontal, poselet, deep, train, trained, validation, feature, piper, accuracy, cnn, table, poselets, deepface, convolutional, top, svm, public, combining, album, network, krizhevsky, layer, neural] [three, multiple, bounding, recognizing, score] [face, pose, model, split, body, full, head] [method, number, sparsity, proposed, vector, problem] [figure, computer, conference, pattern, second, high, equation, robust] [person, image, compute, large, metric, consists, invariant, well, unconstrained]
Deep Multiple Instance Learning for Image Classification and Auto-Annotation
Wu, Jiajun, Yu, Yinan, Huang, Chang, Yu, Kai

cvpr2015papers_第48张图片
[learning, instance, keywords, visual, dmil, supervised, annotation, weakly, dataset, keyword, system, vision, search, mil, bag, label, web, popular, returned, perform, unsupervised, relevant, task] [deep, pascal, table, voc, neural, convolutional, mit, learn, representation, network, layer, feature, fully, performance, training, recognition, jointly, connected, log, dnn, hidden, accuracy, achieves, baidu, apply, train, effective] [multiple, object, framework, including, segmentation, average, extract] [scene, indoor, joint, provide] [number, data, note, proposed, exp, vector, method, sample] [figure, assumption, traditional, high, property] [image]
Predicting Eye Fixations Using Convolutional Neural Networks
Liu, Nian, Han, Junwei, Zhang, Dingwen, Wen, Shifeng, Liu, Tianming

cvpr2015papers_第49张图片
[visual, dataset, learning, set, datasets, human, predict, novel, positive, natural, vision, predicting] [layer, fixation, col, learned, convolutional, training, network, learn, feature, mit, performance, testing, deep, size, toronto, input, edn, neural, cerf, trained, shuffled, map, architecture, activation, raw, learns, better, rescaled, convolution] [saliency, eye, three, based, hierarchical, auc, detection, contrast, video, average, extract, multiple, score, considering, judd] [model] [proposed, gaussian, number, optimal, data, linear, coding, sparse, original, third] [figure, ieee, center, second, blur, pattern] [image, early, best, corresponding, benchmark]
Sparse Convolutional Neural Networks
Liu, Baoyuan, Wang, Min, Foroosh, Hassan, Tappen, Marshall, Pensky, Marianna

cvpr2015papers_第50张图片
[learning, work] [convolutional, network, layer, input, neural, accuracy, fully, convolution, performance, connected, reduce, output, feature, achieve, table, redundancy, higher, cnn, smaller, trained, size, deep, drop, training, previous, cpu, actual, cascade, achieved, loss, highly, apply] [object, detection, average, based, memory, divided, propose, decompose] [model, percentage, performed] [sparse, matrix, multiplication, time, sparsity, number, speedup, kernel, method, original, running, algorithm, theoretical, decomposition, row, cache, loading, column, code, scnn, linear, tensor, basis, avx, complexity, data] [figure, small, second, international, high, conference, total, computer] [dense, large, comparison, overhead, pca, image]
From Captions to Visual Concepts and Back
Fang, Hao, Gupta, Saurabh, Iandola, Forrest, Srivastava, Rupesh K., Deng, Li, Dollar, Piotr, Gao, Jianfeng, He, Xiaodong, Mitchell, Margaret, Platt, John C., Lawrence Zitnick, C., Zweig, Geoffrey

cvpr2015papers_第51张图片
[word, set, human, language, caption, system, sentence, visual, generation, evaluation, text, coco, bleu, learning, test, semantic, dmsm, mil, generating, multimodal, novel, dataset, probability, automatic, meteor, common, work, instance, predicted, entropy, selected, search, process, attribute, pplx, randomly, report] [training, neural, deep, table, microsoft, convolutional, map, representation, performance, feature, cnn, better, network, top, validation, global, train, response, fully, detector, arxiv, speech, previous, preprint, alexnet] [score, precision, multiple, based, bounding, generated, likelihood, object, detection, three] [model, approach, detected, conditioned] [data, maximum, vector, machine] [figure, quality, length] [image, similarity, corresponding, compare, compute]
Recognize Complex Events From Static Images by Fusing Deep Channels
Xiong, Yuanjun, Zhu, Kai, Lin, Dahua, Tang, Xiaoou


[visual, vision, human, dataset, semantic, learning, work, web, people, understanding, search, process, dedicated, set, exploit, strategy, novel] [recognition, deep, convolutional, channel, representation, accuracy, cnn, training, map, performance, layer, compared, size, class, feature, network, table, neural, capturing, capability] [event, object, detection, bounding, recognizing, complex, framework, multiple, key, appearance, three, action, propose, static] [scene, face, detected, capture, model, fused, represent, project] [method, analysis, machine, proposed] [computer, ieee, figure, pattern, conference, international, directly, second, comprised] [spatial, large, image, dense, june]
Detector Discovery in the Wild: Joint Multiple Instance and Representation Learning
Hoffman, Judy, Pathak, Deepak, Darrell, Trevor, Saenko, Kate

cvpr2015papers_第52张图片
[labeled, weakly, learning, set, weak, category, positive, annotated, instance, mined, discovered, latent, visual, mil, label, domain, discovery, dataset, supervised, transfer, lsda, source, include, example, scoring] [representation, feature, auxiliary, training, detector, previous, learn, jointly, performance, cnn, train, standard, generic, top, learned, negative, convolutional, imagenet, neural, recognition, map, deep, table, effective, better, network] [detection, strong, object, bounding, box, multiple, precision, interest] [model, joint, initial, localization, initialization] [data, algorithm, optimize, optimization, method, space, problem, min, alternating] [figure, produce, equation] [image, large, outperforms, region, initializing, patch, well]
Matching-CNN Meets KNN: Quasi-Parametric Human Parsing
Liu, Si, Liang, Xiaodan, Liu, Luoqi, Shen, Xiaohui, Yang, Jianchao, Xu, Changsheng, Lin, Liang, Cao, Xiaochun, Yan, Shuicheng

cvpr2015papers_第53张图片
[human, label, semantic, annotated, probability, learning, clothing, work, processing, dataset, test, corpus, bag] [convolutional, knn, parsing, feature, input, paperdoll, training, testing, map, gkl, layer, performance, network, cnn, neural, deep, yamaguchi, skirt, designed, better, transferred, architecture, tkl, achieve, hat, loss, smoothing, train, nonparametric, table, arxiv] [based, path, superpixel, segmentation, object, foreground, framework, detection, score, three, average, generated] [single, model, pose, capture, scene, face] [proposed, number, method, data, kernel] [parametric, estimate, component, figure, absolute] [image, matching, cross, region, matched, calculated, pair, rate, retrieved, corresponding, well]
Deep Hierarchical Parsing for Semantic Segmentation
Sharma, Abhishek, Tuzel, Oncel, Jacobs, David W.

cvpr2015papers_第54张图片
[semantic, parse, visual, label, learning, labeled, dataset, help, labeling] [rcpn, network, training, feature, bypass, loss, neural, combiner, contextual, daimler, context, cnn, layer, previous, parsing, stanford, deep, better, performance, table, categorizer, mapper, wdec, gcom, decombiner, stronger, refers, ppa, learn, root, referred] [segmentation, node, three, iou, background, utilize, mca, hierarchical, based, entire, propose] [error, tree, model, scene, structure, initial, provide, urban] [random, mrf, proposed, problem, number, order, binary, original] [ieee, figure, enhanced, signal, flow, pixel] [image, internal, sift, propagation, gradient, recursive, parent]
DevNet: A Deep Event Network for Multimedia Event Detection and Evidence Recounting
Gan, Chuang, Wang, Naiyan, Yang, Yi, Yeung, Dit-Yan, Hauptmann, Alex G.

cvpr2015papers_第55张图片
[learning, dataset, evaluation, positive, label, med, task, human, visual, work] [deep, convolutional, feature, training, layer, apply, recognition, cnn, shallow, cnns, imagenet, network, class, level, trecvid, discriminative, improved, table, map, representation, higher, neural, compared] [event, key, detection, video, recounting, minndc, devnet, saliency, evidence, multimedia, temporal, frame, action, localize, score, average, based, complex, trajectory, object, framework, detecting] [approach, model, single, scene, described, localization] [vector, data, number, rank, pass, proposed] [figure, pixel, informative, vehicle, quality, computer] [spatial, image, large, rate, dense]
Understanding Image Representations by Measuring Their Equivariance and Equivalence
Lenc, Karel, Vedaldi, Andrea

cvpr2015papers_第56张图片
[learning, visual, set, task, cat, example] [equivariant, hog, feature, convolutional, representation, learned, deep, cnn, layer, input, performance, equivariance, structured, lexn, training, regulariser, table, neighbourhood, learn, map, lcs, loss, trained, better, size, neural, shallow, output, compensated, generic] [] [regression, error, reconstruction, pose, structure, fact] [mapping, equivalence, data, sparsity, permutation, linear, number, case, sparse, experiment, method, original, objective, problem, empirically] [rotation, figure, vertical, horizontal, small, second, translation, stitching, direct] [image, transformation, invariant, invariance, local, corresponding, good, geometric, distance]
Shape Driven Kernel Adaptation in Convolutional Neural Network for Robust Facial Traits Recognition
Li, Shaoxin, Xing, Junliang, Niu, Zhiheng, Shan, Shiguang, Yan, Shuicheng

cvpr2015papers_第57张图片
[adaptation, learning, trait, evaluation, indicated, automatically] [convolutional, cnn, subnetworks, recognition, deep, network, age, feature, multipie, architecture, webface, conduct, subnetwork, learned, neural, learn, morph, convolve, stage, trained, performance, alexnet, representation, layer, input, additional, pool, better, setting, global, deepface, convolution, accuracy] [adaptive, multiple, propose, appearance, generate, based, three] [face, facial, shape, gender, pose, estimation, structure, expression, driven, landmark, model] [kernel, method, proposed, function, fusion, optimization, note, normalized, data, element, idea, denoted, random] [robust, figure, real] [local, image, database, patch, corresponding, comprehensive, variation, large, texture]
Neuroaesthetics in Fashion: Modeling the Perception of Fashionability
Simo-Serra, Edgar, Fidler, Sanja, Moreno-Noguer, Francesc, Urtasun, Raquel

cvpr2015papers_第58张图片
[fashionability, post, user, clothing, fashion, dataset, type, latent, work, predicting, fashionable, pastel, semantic, prediction, casual, enp, style, refer, predict, dress, sentiment, los, analyze, perform, personal, going, country, ethnicity, people, garment, example] [table, feature, crf, learn, network, deep, output, class, improve, setting, map, hidden, performance, neural, input] [online, segmentation, individual, strong] [model, scene, current, social, dominant, unary, location, face, reader] [number, order, analysis, pairwise, random, note, cluster, time] [black, figure, high, photograph, small, blue] [image, large, well, person, visualization, scale, compute, consists]
Show and Tell: A Neural Image Caption Generator
Vinyals, Oriol, Toshev, Alexander, Bengio, Samy, Erhan, Dumitru

cvpr2015papers_第59张图片
[word, description, language, sentence, human, bleu, set, natural, datasets, report, nic, ranking, evaluation, visual, work, generation, vision, task, learning, mscoco, embeddings, embedding, system, text, generating, transfer, dataset, describing, test, example, sbu, probability, flickr, novel, evaluating] [training, neural, lstm, input, recurrent, trained, size, network, rnn, table, gate, log, cnn, deep, pascal, better, performance, output, advantage, fed, convolutional, representation] [generated, memory, generate, score, sequence, based, likelihood] [model, approach, current, provide] [data, machine, time, case, number] [translation, figure, computer, generates, quality] [image, well, large, best, cell, metric]
Integrating Parametric and Non-Parametric Models For Scene Labeling
Shuai, Bing, Wang, Gang, Zuo, Zhen, Wang, Bing, Zhao, Lifan

cvpr2015papers_第60张图片
[labeling, label, learning, semantic, vision, transfer, set, visually, natural, work] [global, cnn, feature, belief, class, convolutional, context, neural, network, table, recognition, better, discriminative, learned, training, stanford, cnns, size, performance, accuracy, achieved, fed, integrating, nonparametric, input, road, deep, siftflow, sand, reported, learn, parsing, achieve, layer, pooling, softmax] [based, likelihood, background, integration, average, object, final] [scene, model, potential, estimation, building, sea, pose] [energy, function, method, clique, denotes, data] [computer, pixel, ieee, conference, pattern, figure, result, parametric, presented, equation, international, frequency, sky] [local, image, nearest, metric, large, sift, distance, spatial, comparison]
On the Relationship Between Visual Attributes and Convolutional Networks
Escorcia, Victor, Carlos Niebles, Juan, Ghanem, Bernard

cvpr2015papers_第61张图片
[attribute, visual, work, relationship, vision, learning, recognize, tend, semantic, encode, existence, furry, dataset, focus, understanding, predict, set, centric, shiny, pink, green, orange, european] [acns, recognition, convolutional, trained, layer, acn, performance, deep, imagenet, network, neural, discriminative, representation, accuracy, ablated, learn, top, conv, wet, brown, activation, wooden, learned, feature, metallic, striped] [object, average, correlation, group, long, based] [percentage, model, rough] [sparse, number, sparsity, normalized, matrix, solution, solving, round] [figure, computer, conference, pattern, high, yellow, small, blue, ieee, red, white, black] [image, well, roc, large]
Evaluation of Output Embeddings for Fine-Grained Image Classification
Akata, Zeynep, Reed, Scott, Walter, Daniel, Lee, Honglak, Schiele, Bernt

cvpr2015papers_第62张图片
[embeddings, learning, embedding, text, unsupervised, awa, semantic, word, label, supervised, soa, sje, cub, wikipedia, corpus, bird, human, attribute, test, wordnet, highest, cmb, visual, labeled, ranking, unlabeled, evaluate, glove, prediction, appear, work, annotation, correspond, perform, goog, set] [output, input, class, combining, training, accuracy, structured, context, learn, deep, learned, build, improve, improves, table, performance, recognition, window, better, size, neural, cnn, network, convolutional, standard] [hierarchical, combination, framework, obtains, target, cost, multiple] [joint, model, continuous] [data, vector, binary, objective, method, number] [computer] [image, large, similarity, compatibility, best, vocabulary, hierarchy, distance]
Deep Hashing for Compact Binary Codes Learning
Erin Liong, Venice, Lu, Jiwen, Wang, Gang, Moulin, Pierre, Zhou, Jie

cvpr2015papers_第63张图片
[learning, unsupervised, supervised, set, semantic, visual, search, dataset, existing] [deep, learn, learned, top, training, layer, network, feature, map, compared, discriminative, neural, table, loss, performance, output, convolutional, grant] [recall, precision, three, object, multiple, hierarchical] [approach, term, model, independent] [hashing, binary, sdh, method, proposed, kmh, hamming, pcah, splh, compact, bre, sample, lsh, itq, mlh, projection, vector, nonlinear, minimizing, function, objective, time, original, number, seek, singapore, quantization, basic, mnist, algorithm, row] [spherical, figure, second, presented] [image, large, computed, query, retrieved, gradient, outperforms, gallery]
An Improved Deep Learning Architecture for Person Re-Identification
Ahmed, Ejaz, Jones, Michael, Marks, Tim K.

cvpr2015papers_第64张图片
[set, learning, positive, test, summary, novel, labeled, report, visual, work, hard] [network, feature, layer, deep, input, architecture, negative, size, training, previous, convolution, convolutional, tied, performance, max, fully, connected, learned, passed, fpnn, trained, learn, map, smaller, identification, response, accuracy, representation, softmax] [state, art, performs] [model, approach, body] [data, method, rank, analysis, problem] [figure, small, ieee, computer, pattern, conference, presented] [neighborhood, person, image, metric, difference, patch, large, pair, local, rate, compare, viper, similarity, esdc, computed, sdalf, outperforms, itml, compute, kissme]
Understanding Deep Image Representations by Inverting Them
Mahendran, Aravindh, Vedaldi, Andrea

cvpr2015papers_第65张图片
[visual, natural, understanding, amount, interesting, learning, study, subset] [hog, deep, neural, cnn, inversion, dsift, representation, network, cnns, standard, invert, table, regulariser, layer, loss, inverting, convolutional, imagenet, shallow, compared, implemented, hoggle, feature, max, binning, larger, learned, increasing, apply] [average, based, viewed, apparent] [reconstruction, error, model, qualitative, prior] [method, euclidean, analysis, function, problem, code, bilinear, norm, linear, algorithm, note, objective, original, normalized, regularization] [figure, inverse, technique, paper, reconstructing, requires, direct, ieee, derivative] [image, gradient, local, best, sift, applied, computing, contained, cell, reference, large]
Combining Local Appearance and Holistic View: Dual-Source Deep Neural Networks for Human Pose Estimation
Fan, Xiaochuan, Zheng, Kang, Lin, Yuewei, Wang, Song

cvpr2015papers_第66张图片
[human, set, selected, learning, dataset, include, select] [testing, training, convolutional, input, flic, cnn, performance, heatmap, layer, tompson, deep, better, sliding, neural, output, size, combine, loss, network] [object, detection, appearance, based, likelihood, average, three, indicates, multiple, including, performs, proposal] [joint, pose, body, estimation, lsp, model, holistic, graphical, localization, location, pdj, detected, full, deeppose, visible, limb, single, pictorial, located, pcp, johnson, capture] [method, proposed, normalized, number, upper] [figure, left, developed, estimating, computer] [image, patch, local, comparison, distance, construct, view, spatial, large, rate]
DeepContour: A Deep Convolutional Feature Learned by Positive-Sharing Loss for Contour Detection
Shen, Wei, Wang, Xinggang, Wang, Yan, Bai, Xiang, Zhang, Zhijiang

cvpr2015papers_第67张图片
[learning, dataset, positive, set, natural, label, work, selected, versus, human, supervised, semantic, labeled, probability] [contour, deep, feature, loss, cnn, convolutional, learned, class, training, layer, neural, softmax, learn, performance, nyud, discriminative, negative, standard, output, architecture, network, generalization, structured, log, setting, apply, better, belongs, table, explore, unit, effective, size] [detection, object, edge, segmentation, truth, partition, based, background, contrast] [shape, ground, model, forest] [data, function, parameter, proposed, method, problem, binary, clustering, number, sparse, code] [pattern, ieee, figure, result, computer, color] [image, patch, local, gradient, cross, sketch]
Mid-Level Deep Pattern Mining
Li, Yao, Liu, Lingqiao, Shen, Chunhua, van den Hengel, Anton

cvpr2015papers_第68张图片
[visual, mining, association, set, rule, discovered, discovery, dataset, discover, process, category, work, natural, baseline, representativeness, positive, test, outperforming, item] [cnn, activation, training, discriminative, pooling, feature, mdpm, table, neural, convolutional, performance, recognition, deep, max, representation, appealing, mit, pascal, voc, conf, detector, accuracy, suppmin, scfvc, ensemble, extracted, class, previous, apply, train, frequent] [object, target, based, bounding, box, detection, merging] [approach, indoor, scene, current] [element, number, proposed, algorithm, encoding, min, largest, method, linear] [pattern, ieee, transaction, figure, high] [image, large, patch, corresponding, vlad, local]
Deep Transfer Metric Learning
Hu, Junlin, Lu, Jiwen, Tan, Yap-Peng

cvpr2015papers_第69张图片
[learning, transfer, domain, source, dataset, set, knowledge, visual, labeled, supervised, label, datasets, unlabeled, subject, exploit, positive] [deep, top, layer, training, discriminative, network, recognition, feature, performance, neural, hidden, better, output, table, accuracy, testing, loss, learned, learn, activation, convolutional, designed] [target, hierarchical, qij, deeply] [face] [method, proposed, function, nonlinear, data, distribution, linear, objective, vector, sample, experimental, subspace, problem, machine] [captured, figure, assume] [metric, dtml, person, caviar, viper, distance, lij, lji, dstml, lfw, image, lsi, lti, rate, mth, wdref, computed, stml, compute, pij, applied, large, ddml, similarity, difference, htj]
Viewpoints and Keypoints
Tulsiani, Shubham, Malik, Jitendra

cvpr2015papers_第70张图片
[prediction, task, predicted, vision, predict, instance, set, human, evaluate, annotated, evaluation, correct, car, system] [performance, convolutional, table, global, cnn, training, train, network, architecture, setting, pascal, neural, output, generic, response, layer, voc, restricted, size, class, improve, trained, feature, xiang, loss, recognition] [object, detection, based, appearance, likelihood, bounding, multiscale, box, propose, truth] [viewpoint, pose, estimation, error, model, approach, conditioned, localization, capture, euler, ground, occluded, described] [method, analysis, denote, note, number, problem] [computer, ieee, pattern, conference, rotation, small, figure] [keypoint, keypoints, local, corresponding, well, image, metric]
Hyper-Class Augmented and Regularized Deep Learning for Fine-Grained Image Classification
Xie, Saining, Yang, Tianbao, Wang, Xiaoyu, Lin, Yuanqing

cvpr2015papers_第71张图片
[learning, vision, dataset, car, labeled, annotated, search, visual, dog, set, label, work, human, novel, baseline, probability, task, report] [deep, recognition, cnn, training, class, neural, performance, fgic, stanford, level, train, imagenet, network, auxiliary, convolutional, feature, improve, augmentation, baidu, table, multitask, accuracy, sharing, learn, generic, activation] [framework, object, propose, multiple, detection, success] [model, approach, pose] [data, proposed, augmented, regularization, number, regularized, easily, method, denote, note, exploiting] [computer, conference, ieee, pattern, international, figure, small, high] [image, large, inherent, view, variance, well]
A Large-Scale Car Dataset for Fine-Grained Categorization and Verification
Yang, Linjie, Luo, Ping, Change Loy, Chen, Tang, Xiaoou

cvpr2015papers_第72张图片
[car, dataset, attribute, benz, door, audi, test, prediction, labeled, bwm, vision, seat, volkswagen, compcars, rich, hatchback, type, interesting, side, convertible, categorization, datasets, predict, subset, learning, novel, baseline, suzuki, positive] [cnn, feature, table, recognition, deep, neural, training, train, observe, capturing, accuracy, class] [object, including, detection, three, series, appearance, utilized, entire, based, truth] [model, joint, face, viewpoint, challenging, ground, continuous] [number, data, maximum, sample, analysis, space, row, implicit] [figure, speed, computer, unique, suv, total, vehicle] [image, bayesian, large, displacement, pair, best]
Mind's Eye: A Recurrent Visual Representation for Image Caption Generation
Chen, Xinlei, Lawrence Zitnick, C.

cvpr2015papers_第73张图片
[visual, sentence, language, word, coco, flickr, generating, novel, human, learning, set, generation, meteor, evaluation, text, bleu, caption, cider, datasets, description, probability, automatic, ability, task, evaluate, dataset, report, ranking, natural, perplexity, vision] [neural, rnn, recurrent, hidden, network, better, learn, representation, training, previous, pascal, feature, arxiv, layer, preprint, deep, table, performance, validation, rnns, standard, microsoft] [generated, generate, based, memory, likelihood, long, truth, numerous, capable] [model, approach, error, provide, scene, ground] [time, random, machine, data, lower] [figure, paper, directly, computer] [image, retrieval, reference]
Action Recognition With Trajectory-Pooled Deep-Convolutional Descriptors
Wang, Limin, Qiao, Yu, Tang, Xiaoou

cvpr2015papers_第74张图片
[dataset, learning, human, datasets, set, visual, describe, evaluation] [feature, convolutional, convnets, recognition, deep, tdd, tdds, improved, performance, pooling, map, net, training, input, size, channel, table, effective, layer, extracting, conduct, discriminative, fisher, convnet, representation, trained, learn, accuracy, architecture, neural] [action, video, temporal, trajectory, spatiotemporal, based, extraction, extract, hof, complementary] [point, model, single] [dimension, number, method, proposed, original, vector, fusion] [optical, motion, figure, high] [spatial, normalization, local, large, dense, descriptor, choose, sampling, image, called, best]
Simultaneous Feature Learning and Hash Coding With Deep Neural Networks
Lai, Hanjiang, Pan, Yan, Liu, Ye, Yan, Shuicheng

cvpr2015papers_第75张图片
[learning, supervised, ranking, returned, search, unsupervised, existing, visual, vision] [input, convolution, deep, architecture, shared, triplet, cnnh, alternative, loss, module, training, top, feature, output, table, representation, designed, size, intermediate, layer, neural, learn, map, sigmoid, encoded, learned, network, class, stage] [precision, three, based, threshold, propose, generate, multiple] [independent] [hash, proposed, hashing, binary, method, hamming, data, number, function, approximate, code, quantization, form, vector, machine, lsh, coding, mlh] [figure, conference, relative, ieee, international, computer, pattern] [image, comparison, preserve, corresponding]
Interleaved Text/Image Deep Mining on a Very Large-Scale Radiology Database
Shin, Hoo-Chang, Lu, Le, Kim, Lauren, Seff, Ari, Yao, Jianhua, Summers, Ronald M.

cvpr2015papers_第76张图片
[topic, learning, radiology, semantic, perplexity, document, latent, language, text, lda, visual, categorization, sentence, word, processing, report, set, datasets, work, natural, diagnostic, mining, collection, describing, dataset, unsupervised, label, referring, system] [deep, cnn, neural, imagenet, output, training, extracted, table, convolutional, level, cnns, trained, representation, learned, learn, map, better] [key, based, object, score, three, cost] [model, supplementary, body, adjacent] [vector, number, space, data, national, dirichlet, machine] [figure, imaging, computer, conference, high] [image, medical, large, disease, good, reference, database, matching]
Deep Visual-Semantic Alignments for Generating Image Descriptions
Karpathy, Andrej, Fei-Fei, Li

cvpr2015papers_第77张图片
[word, sentence, set, multimodal, visual, language, test, text, ranking, generating, work, evaluation, mscoco, learning, embedding, example, description, dataset, karpathy, describe, goal, natural, novel, datasets, annotation, aligns] [neural, rnn, recurrent, network, training, arxiv, preprint, context, brnn, size, input, learn, representation, performance, table, trained, convolutional, powerful, deep, previous, cnn, challenge] [generate, based, score, bounding, sequence, generated, object] [model, full, approach, alignment, single, tree, scene, generative] [dependency, function, data, vector, consider, note, time, method, space] [figure, generates, developed] [image, region, retrieval, outperforms, compute, large]
Deep Convolutional Neural Fields for Depth Estimation From a Single Image
Liu, Fayao, Shen, Chunhua, Lin, Guosheng

cvpr2015papers_第78张图片
[learning, work, conditional, predicting, semantic, prediction, baseline, dataset, rel, set, report] [deep, cnn, network, neural, convolutional, training, crf, map, jointly, input, better, conv, table, composed, train, layer, relu, trained, learn, structured, compared, shared, negative, eigen] [superpixel, partition, performs, based, superpixels] [depth, model, unary, continuous, single, estimation, scene, potential, rpq, error, svr, nyu, regression, neighbouring, inference, smoothness, analytically, indoor, term, spq] [method, pairwise, function, consider, problem, note, solve, random, general, solving, optimization, proposed, matrix, vector, extra] [ieee, outdoor, directly, rms] [image, outperforms]
Efficient and Accurate Approximations of Nonlinear Convolutional Networks
Zhang, Xiangyu, Zou, Jianhua, Ming, Xiang, He, Kaiming, Sun, Jian

cvpr2015papers_第79张图片
[evaluate, work, set] [layer, convolutional, increase, accuracy, table, increased, deep, feature, imagenet, reduce, previous, neural, version, response, training, network, apply, accelerating, trained, size, input, output] [based] [error, model, reconstruction, single, evaluated] [speedup, nonlinear, linear, asymmetric, method, rank, complexity, solution, approximated, problem, approximation, matrix, zij, peedup, energy, decomposition, case, accelerated, selection, jaderberg, number, accumulative, accumulated, original, computation, approximate, min, symmetric, optimization, consider, theoretical, vector, largest, time, reduction, solve] [ratio, solver, figure, accurate, small, adopt] [spatial, pca, rate, image, large, computed, compute, degradation]
Multi-Objective Convolutional Learning for Face Labeling
Liu, Sifei, Yang, Jimei, Huang, Chang, Yang, Ming-Hsuan

cvpr2015papers_第80张图片
[labeling, label, set, learning, semantic, test, dataset, evaluation, existing, work] [nonparametric, network, training, cnn, convolutional, cnns, input, loss, size, hair, deep, output, train, table, crf, trained, structured, accuracy, testing, compared, map, class, apply, parsing, additional, combining, performance, neural, smaller, learned] [based, edge, generated, truth, propose, generate] [face, prior, facial, unary, model, term, approach, ground, inference, helen, single, xij] [pairwise, proposed, method, algorithm, number, crfs, regularization, random, energy, note, addressed] [figure, generates, pixel, skin, second, component] [image, patch, sampling, comparison]
Finding Action Tubes
Gkioxari, Georgia, Malik, Jitendra

cvpr2015papers_第81张图片
[task, human, label, visual, positive, predicted, work, report, dataset, study, predict, correct] [convolutional, train, svm, feature, trained, neural, layer, network, build, accuracy, extracted, recognition, ablation, improvement, table, level, cnns, classify, jain, pascal, better, training, inspired] [action, detection, video, object, frame, ucf, auc, threshold, tube, average, score, appearance, wang, three, static, based, overlap, temporal, truth, performs, linked, localize] [approach, model, shape, rgb, ground] [problem, running] [motion, figure, optical, high, camera, produce, red, signal] [spatial, image, region, roc, rate]
Deep Domain Adaptation for Describing People Based on Fine-Grained Clothing Attributes
Chen, Qiang, Huang, Junshi, Feris, Rogerio, Brown, Lisa M., Dong, Jian, Yan, Shuicheng

cvpr2015papers_第82张图片
[domain, attribute, clothing, adaptation, dataset, learning, source, people, surveillance, shopping, visual, describing, unsupervised, set, search, work, annotated, fashion, supervised, semantic, amount, evaluation, label, human, datasets, annotation, ddan, prediction] [deep, feature, network, layer, training, convolutional, neural, cnn, imagenet, output, fully, standard, bridge, recognition, setting, learned] [target, online, based, detection, cost, bounding, hierarchical, including, merging, candidate, path, correlation] [model, alignment, body, approach, detailed] [data, proposed, problem, consider, number, method, function, linear, analysis, application] [directly, figure, captured, color, small] [street, large, unconstrained, image, similarity, retrieval, consists]
Filtered Feature Channels for Pedestrian Detection
Zhang, Shanshan, Benenson, Rodrigo, Schiele, Bernt

cvpr2015papers_第83张图片
[set, test, baseline, dataset, learning, evaluation, human, work] [training, caltech, pedestrian, feature, channel, decision, top, validation, ldcf, trained, size, observe, additional, learned, informedhaar, detector, acf, informedfilters, randomfilters, performance, compared, reported, table, improvement, design, larger, pooling, hog, context, pcaforeground, squareschnftrs, explore, negative, train, effective, adaboost, rectangular, spatialpooling, improving, inria, standard, katamari, realboost] [detection, based, object, bank, strong] [kitti, model, single, provide, step, discrete] [number, data, method, consider, original, random, vector] [optical, quality, figure, paper, uniform] [best, image, applied, change, local, comparison, integral, considered, gap, lbp]
A Convolutional Neural Network Cascade for Face Detection
Li, Haoxiang, Lin, Zhe, Shen, Xiaohui, Brandt, Jonathan, Hua, Gang

cvpr2015papers_第84张图片
[vision, evaluation, set, positive, visual, test, annotated, work, dataset] [cascade, cnn, layer, detector, training, convolutional, neural, stride, input, performance, cnns, boosted, stage, apply, network, window, achieve, trained, size, cpu, better, vga, headhunter, discriminative, gpu] [detection, bounding, object, score, recall, detect] [face, structure, fps, scan, afw, single, reject, joint] [number, proposed, data, method, kernel, binary, problem, time, faster, discontinuous] [calibration, figure, computer, pattern, ieee, conference, simple, high, accurate, international] [image, false, fast, large, benchmark, resolution]
Sketch-Based 3D Shape Retrieval Using Convolutional Neural Networks
Wang, Fang, Kang, Le, Li, Yi

cvpr2015papers_第85张图片
[learning, domain, dataset, datasets, correct, set, semantic, select, labeled] [network, feature, training, cnn, convolutional, learn, table, loss, input, performance, layer, neural, class, learned, architecture, standard, top, reported, map, size] [based, three, state, recall, propose, art, generate, object, precision, multiple] [shape, model, approach] [method, number, proposed, denotes, function, note, time, sample] [figure, computer, acm, rendering] [retrieval, sketch, siamese, view, furuya, matching, similarity, best, metric, cross, well, comparison, benchmark, large, psb, illustrated, image, demonstrate, query, cyan, dcg, saavedra, pair]
Learning From Massive Noisy Labeled Data for Image Classification
Xiao, Tong, Xia, Tian, Yang, Yi, Huang, Chang, Wang, Xiaogang

cvpr2015papers_第86张图片
[noisy, label, learning, clean, dataset, labeled, confusing, massive, probability, clothing, windbreaker, set, mislabeled, correct, amount, test, visual, hoodie, latent, type, work, unlabeled, supervised, jacket, coat, knitwear, sweater, predict, share, hard, content] [training, deep, better, class, log, cnn, neural, cnns, train, convolutional, layer, accuracy, alexnet, pretrained, validation, size, table, imagenet] [surrounding, truth, based, framework, integrate, confusion, utilize] [model, ground, probabilistic, free, independent, predicts, posterior, wrong] [data, random, method, matrix, problem, proposed, optimal, distribution, theorem, row, number] [noise, figure, small, pure, simple] [image, true, upsampled, limited]
From Categories to Subcategories: Large-Scale Image Classification With Partial Class Label Refinement
Ristin, Marko, Gall, Juergen, Guillaumin, Matthieu, Van Gool, Luc

cvpr2015papers_第87张图片
[set, ncm, learning, test, knowledge, category, baseline, labeled, label, evaluate, subset, annotated, transfer, semantic] [training, accuracy, class, rncmf, trained, scoarse, coarse, table, subcategories, performance, subcategory, improves, additional, kcoarse, achieves, improve, ucoarse, validation, stacked, learned, train, classify, multiclass, fisher, stacking, feature, coarser, outperform, improvement, gain] [based, average, hierarchical, object, node] [model, forest, approach, tree, full] [data, number, random, regularized, function, sample, proposed, objective, consider] [second, relative] [image, metric, nearest, splitting, large, outperforms, impact, best, hierarchy, well, neighbor]
Deep Networks for Saliency Detection via Local Estimation and Global Search
Wang, Lijun, Lu, Huchuan, Ruan, Xiang, Yang, Ming-Hsuan

cvpr2015papers_第88张图片
[set, search, label, learning, visual, svo, predict] [global, feature, training, deep, map, network, accuracy, neural, input, convolutional, dnns, level, generic, table, trained, size, layer] [saliency, object, salient, candidate, detection, based, contrast, background, score, generated, foreground, complex, coverage, detect, boundary, propose, truth, objectness, precision, wctr, entire, recall, drfi, hdct, hrgb, three, overlap, ufo, generate] [estimation, rgb, ground, model, scene, mae, shape] [data, method, proposed, low, distribution, vector, number] [figure, color, high, pixel] [local, image, region, well, geometric, rate, patch, spatial, pca]
Learning Deep Representations for Ground-to-Aerial Geolocalization
Lin, Tsung-Yi, Cui, Yin, Belongie, Serge, Hays, James

cvpr2015papers_第89张图片
[test, learning, set, novel, embedding, task, visual, generalize, search, hard, evaluation, correct, baseline] [feature, deep, training, learned, top, representation, network, convolutional, trained, performance, learn, architecture, layer, loss, margin, input, build] [average, based, object, correctly, publicly] [face, depth, aligned, ground, location, orientation, model, building, scene] [data, method, number, function, dimensional, euclidean, proposed] [figure, traditional, paper, direction] [image, aerial, matching, view, matched, query, distance, unmatched, pair, match, street, imagery, geolocalization, san, corresponding, scale, large, crossview, database, similarity, reference]
Deep Semantic Ranking Based Hashing for Multi-Label Image Retrieval
Zhao, Fang, Huang, Yongzhen, Wang, Liang, Tan, Tieniu

cvpr2015papers_第90张图片
[ranking, semantic, learning, set, dataset, visual, evaluation, relevant, common, supervised] [deep, feature, loss, learn, layer, cnn, representation, neural, convolutional, performance, compared, activation, supervision, level, training, triplet, cnns, map, effective, better, network, connected, larger, learned] [based, framework, measure, multiple] [model, structure, point] [hash, hashing, binary, data, function, multilevel, method, proposed, surrogate, hamming, order, problem, optimization, compact, acg, optimizing, dsrh, number, pairwise, objective, original, multilabel, ndcg, multivariate] [ieee, pattern, figure, preserving, acm] [image, similarity, preserve, database, retrieval, weighted, large, distance, query, list, metric, gradient, construct, well]
Depth and Surface Normal Estimation From Monocular Images Using Regression on Deep Features and Hierarchical CRFs
Li, Bo, Shen, Chunhua, Dai, Yuchao, van den Hengel, Anton, He, Mingyi

cvpr2015papers_第91张图片
[set, work, transfer, learning, predict, predicted, semantic, common, rel] [deep, training, map, size, performance, level, network, table, cnn, convolutional, crf, transferred, neural, context, layer, learned, trained, generalization, input, feature] [framework, hierarchical, extract, truth, formulate] [depth, estimation, regression, surface, model, normal, single, nyu, scene, error, structure, inference, qualitative, ground, term, represent, monocular, provide] [data, method, problem, number, experimental, function, random, proposed, solution, energy, denotes] [pixel, ieee, color, estimated, figure, estimate, presented, liu, blue] [image, local, large, patch, corresponding]
Multi-Instance Object Segmentation With Occlusion Handling
Chen, Yi-Ting, Liu, Xiaokai, Yang, Ming-Hsuan

cvpr2015papers_第92张图片
[semantic, category, set, predictor, task] [cnn, voc, contour, exemplar, pascal, top, performance, validation, map, class, log, train, better, apply, feature, hariharan, network, size] [segmentation, object, detection, based, foreground, likelihood, tackle, framework, bounding, propose, generate, handling, sfj, score, proposal, mcg, iou, superpixel, pcj, box, appearance, overshooting, favorable, categorized] [shape, occlusion, occluding, joint, model, prior, chamfer, reasoning, challenging] [algorithm, proposed, energy, number, note, minimization, parameter] [figure, high, quality, inconsistent, handle, pixel, remove] [image, region, person, distance, scale, large, choose, matched]
Learning Lightness From Human Judgement on Relative Reflectance
Narihira, Takuya, Maire, Michael, Yu, Stella X.

cvpr2015papers_第93张图片
[human, dataset, vision, work, set] [crf, cnn, feature, neural, performance, global, training, convolutional, deep, layer, learn, representation, log, perceived, learned, network, input, output, trained] [hierarchical, multiple] [model, approach, potential, point, error, cij] [sparse, pairwise, coding, linear, decomposition, vector, solution, method, data, note, function, algorithm, order, machine, denotes, encoding] [figure, intrinsic, relative, intensity, computer, conference, pattern, ieee, illumination, shading, direct, total, pixel, recovering, directly] [lightness, image, patch, hsc, local, query, jij, iiw, judgement, difference, rate, well, large, dense, darker, initialized, benchmark, dictionary, implementation]
Towards 3D Object Detection With Bimodal Deep Boltzmann Machines Over RGBD Imagery
Liu, Wei, Ji, Rongrong, Li, Shaozi

cvpr2015papers_第94张图片
[learning, positive, dataset, semantic, labeled, labeling, existing, domain, set, focus, hard, label, european, task, evaluate] [bimodal, training, representation, hidden, feature, deep, raw, train, sliding, detector, layer, stochastic, trained, negative, input, learn, testing, wjl, imagenet, performance, exemplar, pretrained] [detection, object, bounding, based, truth, box, precision, average, framework, extract, score] [rgbd, depth, dbm, rgb, point, model, joint, boltzmann, ground, cloud, rmrc, visible, cad, fused, scene, single, performed, modeling] [data, proposed, binary, energy, gaussian, space] [figure, computer, conference, international, ieee, pattern] [image, reference, corresponding, well, leveraged]
What do 15,000 Object Categories Tell Us About Classifying and Localizing Actions?
Jain, Mihir, van Gemert, Jan C., Snoek, Cees G. M.

cvpr2015papers_第95张图片
[set, dataset, study, human, test, category, evaluation, learning, preference, datasets, localizing, semantically, selected, include, work, relevant] [representation, recognition, accuracy, adding, table, fisher, validation, better, improve, improvement, conduct, learned, classifying, kth, performance, jain, class, discriminative, response, top, improves, follow, convolutional, train, training] [action, object, video, average, responsive, matter, combined, weight, avoidance, ucf, wang, key, peng, interest, tubelet, barbell, groundtruth, three, thumos] [localization, approach, modeling, qualitative] [encoding, number, experiment, empirical, selection, consider, vector] [motion, figure, preferred, high] [best, large, local, compare]
Situational Object Boundary Detection
Uijlings, Jasper R. R., Ferrari, Vittorio

cvpr2015papers_第96张图片
[semantic, set, test, coco, prediction, visual, work, evaluation, learning, predict, annotated, dataset, variety, label, baseline, occur] [class, cnn, global, specific, training, detector, table, imagenet, fisher, trained, contour, pascal, microsoft, apply, voc, train, performance, context, convolutional, better, structured, validation] [object, boundary, monolithic, detection, subclass, situational, agnostic, appearance, precision, situation, recall, segmentation, average, yield, three, specialized, based, edge, truth, sbd] [single, approach, ground, forest] [random, number, method, cluster, row] [figure, caused, water, relative, paper] [image, local, true, choose, outperforms, false, applied, good, fast]
An Active Search Strategy for Efficient Object Class Detection
Gonzalez-Garcia, Abel, Vezhnevets, Alexander, Ferrari, Vittorio

cvpr2015papers_第97张图片
[search, set, evaluating, test, strategy, learning, evaluate, visual, work, active, car, dataset] [window, context, training, belief, extractor, pascal, map, cnn, performance, uva, detector, van, fewer, convolutional, sliding, input, deep, train, subsampling, hyperparameters, decision, highly, actual] [object, detection, score, based, appearance, multiple, average, proposal, node] [force, location, forest, approach, challenging, point, model, cluttered] [method, random, number, runtime, effectively, kernel, reduces] [observation, figure, observed, relative, observing, high, ieee, small] [image, displacement, distance, internal, large, overhead]
Associating Neural Word Embeddings With Deep Image Representations Using Fisher Vectors
Klein, Benjamin, Lev, Guy, Sadeh, Gil, Wolf, Lior

cvpr2015papers_第98张图片
[sentence, set, vision, annotation, dog, woman, learning, test, search, association, language, text, maximization, word, work, common, cca, vec, probability] [fisher, neural, hybrid, previous, deep, recognition, representation, table, arxiv, preprint, validation, convolutional, feature, network, size] [based, three, object, truth] [mixture, model, single, step, volume, current, ground] [vector, rank, laplacian, hglmm, distribution, gmm, gaussian, lmm, parameter, multivariate, diagonal, computational, data, order, respect, represented, matrix, normalized] [computer, conference, black, median, component, ieee, international, equation, pattern] [image, query, geometric, weighted, normalization, derived, retrieval, expectation, benchmark]
Transformation-Invariant Convolutional Jungles
Laptev, Dmitry, Buhmann, Joachim M.

cvpr2015papers_第99张图片
[dataset, set, learning, vision, select, task, include, selected] [feature, convolutional, training, recognition, neural, decision, class, neuronal, size, learn, performance, learned, trained, better, global, layer, train, generalization, achieve, ensemble, max, width, accuracy, identity] [segmentation, incorporate, art, state, propose, object] [allows, face, prior, split] [algorithm, maximum, method, proposed, regularization, consider, data, clustering, original, parameter, number, optimization, idea, yale, machine, lemma, main, kernel, case] [computer, pixel, ieee, conference, pattern, presented, figure, paper] [image, ticj, transformation, local, large, tict, considered, applied, discussed, includes, gradient, transformed]
Multi-Manifold Deep Metric Learning for Image Set Classification
Lu, Jiwen, Wang, Gang, Deng, Weihong, Moulin, Pierre, Zhou, Jie

cvpr2015papers_第100张图片
[set, learning, dataset, datasets, visual, randomly, versus, existing, selected, work] [deep, mmdml, discriminative, training, recognition, feature, layer, compared, network, testing, learned, margin, table, performance, learn, standard, top, dfrv, neural, class, nonlinearity, grant, map] [object, multiple, based, average, video, propose, employed, key, appearance] [face, model, approach, single] [method, manifold, number, ytc, nonlinear, proposed, sample, function, mobo, algorithm, time, china, problem, singapore, analysis, mmsml, vector, linear, nonlinearly, effectively] [figure, captured, second, varying] [image, distance, compute, rate, metric, pubfig, computed, large, similarity, nearest]
Towards Unified Depth and Semantic Prediction From a Single Image
Wang, Peng, Shen, Xiaohui, Lin, Zhe, Cohen, Scott, Price, Brian, Yuille, Alan L.

cvpr2015papers_第101张图片
[semantic, prediction, predicted, predict, consistency, set, label, learning] [global, cnn, map, jointly, trained, network, training, better, feature, learned, output, train, achieves, coarse, previous, loss, convolutional, parsing] [segment, segmentation, edge, object, hierarchical, framework, formulate, generate, appearance, bounding, propose, considering, truth] [depth, joint, inference, potential, ground, hcrf, scene, estimation, unary, single, model, infer, represent, canonical, nyu, neighboring, approach, quantitative, structure] [normalized, method, data, problem, consider, denote, regularize] [pixel, figure, vertical, absolute, relative, accurate, inside, center, balancing] [local, image, scale, geometric, limited, regressor, comparison, difference, large]
Salient Object Subitizing
Zhang, Jianming, Ma, Shugao, Sameki, Mehrnoosh, Sclaroff, Stan, Betke, Margrit, Lin, Zhe, Shen, Xiaohui, Price, Brian, Mech, Radomir

cvpr2015papers_第102张图片
[subitizing, vision, category, dataset, test, amt, existence, labeled, set, visual, human, predicting, mso, baseline, task, labeling, label, resorting, processing, sun, predicted, learning, european, consistency, evaluate, people] [recognition, cnn, accuracy, global, feature, performance, table, hog, neural, imagenet, training, output, convolutional, reported] [salient, object, detection, saliency, bounding, three, average, proposal, based, counting, threshold, detecting, precision] [model, single, provide, scene, localization, approach] [number, method, time, proposed, sample, consistently, analysis, gist, row] [conference, computer, ieee, pattern, international, figure, simple, small, viewing, technique, tested] [image, cell, implementation, large, proportion]
Cross-Scene Crowd Counting via Deep Convolutional Neural Networks
Zhang, Cong, Li, Hongsheng, Wang, Xiaogang, Yang, Xiaokang

cvpr2015papers_第103张图片
[dataset, test, learning, unseen, randomly, set, selected, labeled, treated, surveillance, task, predict, predicted, example] [training, map, cnn, pedestrian, loss, table, deep, global, better, ucsd, performance, trained, convolutional, neural, input, compared, learn, learned, nonparametric] [crowd, density, counting, target, truth, based, count, video, detection, segmentation, foreground, ucf, candidate, three, propose, framework, object, including] [scene, model, regression, ground, estimation, switchable, head, introduced] [method, number, proposed, distribution, data, objective, kernel, normalized] [figure, perspective, estimated, total, absolute] [image, local, patch, retrieved, ridge, scheme, view, best, scale]
Learning Coarse-to-Fine Sparselets for Efficient Object Detection and Scene Classification
Cheng, Gong, Han, Junwei, Guo, Lei, Liu, Tianming

cvpr2015papers_第104张图片
[dataset, set, visual, learning, existing, work, baseline, exploit, unsupervised, aca, supervised] [sparselets, activation, discriminative, training, classification, coarse, train, hidden, shared, learned, pascal, mit, fine, sparselet, trained, factor, voc, learn, table, merced, size, accuracy, layer, neural, land, convolution, discriminatively, performance, detector, achieved, input, hog, effective, dpm, dpms, unified, recognition] [object, detection, based, average, framework, propose, cost] [scene, reconstruction, deformable, term] [method, number, speedup, sparse, sparsity, proposed, objective, function, algorithm, constraint, computational, simultaneously, solution, vector] [total, second] [dictionary, large, image, compute, spatial]
Weakly Supervised Semantic Segmentation for Social Images
Zhang, Wei, Zeng, Sheng, Wang, Dequan, Xue, Xiangyang

cvpr2015papers_第105张图片
[semantic, label, supervised, weakly, noisy, latent, visual, concept, dataset, learning, consistency, existing, conditional, labeled, annotated, probability, task] [cnn, training, feature, global, learn, contextual, fully, performance, convolutional, learned, pascal, network, extracted, accuracy, table, neural, comparable] [segmentation, superpixels, superpixel, object, lsc, correlation, background, multiple, based, overlap, appearance, framework, failure] [model, approach, potential, social, tree, joint, inference, challenging, adjacent, ground, building] [energy, pairwise, proposed, problem, quantization, random, algorithm, graph, vector, function, distribution, number, missing] [figure, sky, field] [image, spatial, comparison, large]
Hierarchical Recurrent Neural Network for Skeleton Based Action Recognition
Du, Yong, Wang, Wei, Wang, Liang

cvpr2015papers_第106张图片
[human, dataset, work, vision, learning] [layer, recognition, recurrent, rnn, neural, brnn, hidden, network, lstm, input, deep, representation, output, training, connected, fully, connection, activation, accuracy, brnns, tanh, gate, effectiveness, contextual, table, subnets, chaudhry] [action, skeleton, temporal, based, three, hierarchical, sequence, forward, msr, propose, berkeley, vantigodi, memory, backward] [model, body, capture, depth, joint, performed, trunk, represent] [proposed, experimental, time, number, problem, function, method] [conference, ieee, bidirectional, computer, pattern, motion, international, figure, adopt, left] [compare, derived, corresponding, spatial, gradient]
Deep Correlation for Matching Images and Text
Yan, Fei, Mikolajczyk, Krystian

cvpr2015papers_第107张图片
[text, dcca, protocol, visual, evaluation, learning, caption, cca, description, natural, latent, work, language, learnt, dataset, test, girl, annotation, automatic, gold, ranked, cula, generation, set, iapr, address, vision, generating, standing, example, content, dog, task, sentence, describing] [deep, table, training, performance, gpu, neural, layer, network, top, cpu, structured, convolutional, architecture, validation] [correlation, based, three, video, art, state, memory, average] [canonical] [objective, proposed, data, trace, batch, norm, matrix, space, linear, rank, analysis, time, row, singular, order, number] [figure, white, tno, total, paper] [image, matching, scheme, retrieval, large, magnitude, implementation, gradient]
Beyond Gaussian Pyramid: Multi-Skip Feature Stacking for Action Recognition
Lan, Zhengzhong, Lin, Ming, Li, Xuanchong, Hauptmann, Alex G., Raj, Bhiksha

cvpr2015papers_第108张图片
[datasets, human, dataset, visual, task, evaluation] [feature, mifs, performance, recognition, representation, convolutional, improved, pyramid, level, learnability, extracted, table, stacking, training, higher, neural, deep, standard, simonyan, improvement, skip, reported, improves, better, arxiv, preprint, fisher, smaller, accuracy, macc] [action, video, trajectory, wang, multiple, event, cost, based, peng, olympics, detection, youtube, dynamic] [single] [number, time, condition, gaussian, differential, singular, computational, method, matrix, theorem, vector, range, proof, linear] [frequency, signal, conventional, figure, motion, generates, small, bias, optical, recover] [large, dense, image, scale, comparison]
Exemplar SVMs as Visual Feature Encoders
Zepeda, Joaquin, Perez, Patrick


[set, positive, vision, learning, visual, dataset, search, european, distractor, evaluate, work, task, system, baseline] [feature, exemplar, negative, generic, map, performance, fisher, sgd, cnn, deep, svms, pool, convolutional, size, extracted, learned, neural, improvement, svm, table, representation, training, generalization, larger] [based, object, propose, average, precision] [approach, single] [encoding, number, proposed, method, vector, linear, base, machine, kernel, close, function, order, analysis] [computer, pattern, relative, conference, international, figure, ieee] [image, local, large, retrieval, encoder, query, recursive, vlad, oxford, scale, consists, computed, well]
Fine-Grained Classification of Pedestrians in Video: Benchmark and State of the Art
Hall, David, Perona, Pietro

cvpr2015papers_第109张图片
[dataset, clothing, people, gait, human, sex, set, labelled, style, work, annotation, label, datasets, entropy, visual, attribute, baseline, learning, collected, task, worker, existing, casual] [categorisation, subcategories, age, class, recognition, deep, performance, pedestrian, subcategory, training, convolutional, size, accuracy, extracted, network, top] [bounding, box, video, weight, frame, three, entire, average] [pose, error, estimation, body, model, occlusion, labelling, single, challenging, lsp, viewpoint, height, articulated, normalised, gender] [number, method, data] [figure, left, estimate, total, warped] [image, large, keypoints, person, benchmark, keypoint, region, best, indicate]
Grasp Type Revisited: A Modern Perspective on a Classical Feature for Vision
Yang, Yezhou, Fermuller, Cornelia, Li, Yi, Aloimonos, Yiannis

cvpr2015papers_第110张图片
[type, human, vision, set, dataset, system, category, learning, recognize, labeled, visual, natural, correct, test, prediction, work] [recognition, cnn, training, convolutional, feature, neural, layer, power, belief, level, input, testing, deep, performance, nan, apply, size, achieved] [action, based, precision, segmentation, three, truth, average, key, cooking, evolution, static, tracking, video, recall, segment, cognitive, robot, detection, object] [grasp, hand, intention, manipulation, grasping, infer, ground, inference, pgraspt, model, knife, current, scene] [data, consider, number, experimental, journal, experiment, third, main, method] [computer, figure, ieee, conference, intelligent, pattern, international, small, cylindrical] [image, unconstrained, patch, large, considered]
Deeply Learned Attributes for Crowded Scene Understanding
Shao, Jing, Kang, Kai, Change Loy, Chen, Wang, Xiaogang

cvpr2015papers_第111张图片
[attribute, dataset, learning, stand, dlsf, prediction, dlmf, user, understanding, set, human, describe, study, subject, sto, existing, example, tag, evaluation, keywords, common, label, audience, probability] [deep, learned, table, performance, input, response, higher, pedestrian, convolutional, feature, compared, network, size, negative, learn, generic, design] [crowd, video, appearance, auc, crowded, walk, three, tracklet, deeply, multiple, frame, background, cue, static, average, action, sit, indicates, recognizing] [scene, model, single, indoor, structure] [row, number, time, proposed] [motion, figure, outdoor, high, second, stability, blue] [well, descriptor, image, compare, cuhk, good]
Learning a Convolutional Neural Network for Non-Uniform Motion Blur Removal
Sun, Jian, Cao, Wenfei, Xu, Zongben, Ponce, Jean

cvpr2015papers_第112张图片
[set, predict, learning, predicting, example, natural, predicted, work, probability] [cnn, convolutional, neural, layer, network, better, table, learned, learn, effective, training, deep, fully] [candidate, based, generate, complex, interval, propose] [approach, estimation, model, challenging, term, single, error] [kernel, distribution, mrf, vector, method, optimize, extension, effectively, note, markov, problem] [motion, blur, blurry, estimate, deblurring, figure, pixel, rotated, camera, estimated, sharp, mse, estimating, uniform, spatially, extend, nonuniform, synthetic, sext, remove, varying, rotation, deconvolution, result, removing, caused, removal, designing, blind, length, directly] [image, patch, dense, local, well, psnr, compare]
Pooled Motion Features for First-Person Videos
Ryoo, Michael S., Rothrock, Brandon, Matthies, Larry

cvpr2015papers_第113张图片
[dataset, set, human, learning, datasets] [feature, representation, pooling, cnn, recognition, ifv, performance, pyramid, advantage, compared, designed, previous, overfeat, max, combining, improved, caffe, convolutional, better, pooled, hog, class, table, public] [pot, temporal, activity, video, series, dogcentric, itf, multiple, uec, hof, appearance, park, based, entire, framework, including, combined, egocentric, mbh, stip, sequence] [capture, single, structure, performed, approach, described, cuboid] [time, vector, superior, number, idea, explicitly] [motion, figure, conventional, camera, optical, result, tested] [bow, descriptor, image, best, local, histogram, well, representing]
3D ShapeNets: A Deep Representation for Volumetric Shapes
Wu, Zhirong, Song, Shuran, Khosla, Aditya, Yu, Fisher, Zhang, Linguang, Tang, Xiaoou, Xiao, Jianxiong

cvpr2015papers_第114张图片
[learning, category, active, dataset, label, test, recognize, evaluate, probability] [recognition, deep, representation, layer, convolutional, map, table, belief, training, testing, convolution, train, hidden, trained, feature, reduce, chair, stride, size, top, build, accuracy] [object, based, complex, propose, average] [shape, model, depth, shapenets, cad, voxels, surface, voxel, planning, mesh, free, completion, visible, nyu, reconstruction, full, represent, point, generative, princeton, potential, probabilistic, joint] [data, distribution, space, algorithm, binary, method, sample, selection, missing] [figure, observed, result, camera, acm, unique, unknown] [view, large, retrieval, choose, geometric, image, limited, sampling, construct, best]
ActivityNet: A Large-Scale Video Benchmark for Human Activity Understanding
Caba Heilbron, Fabian, Escorcia, Victor, Ghanem, Bernard, Carlos Niebles, Juan

cvpr2015papers_第115张图片
[activitynet, human, untrimmed, dataset, datasets, category, existing, household, collection, annotation, vision, set, test, semantic, personal, encode, visual, goal, taxonomy, variety, benchmarking, report, cleaning, hardest, search, introduce, hard, annotated, rich] [recognition, performance, feature, level, top, deep, table, class, validation, challenge, map, structured, training, size] [activity, video, action, trimmed, temporal, detection, three, ing, object, multiple, average, thumos, organization, static, measure] [current, provide, scene] [number, data, order, note, largest, analysis, time, third] [figure, computer, motion, second, international, care] [large, benchmark, scale, compare, false, best, limited, tier, hierarchy]
A Coarse-to-Fine Model for 3D Pose Estimation and Sub-Category Recognition
Mottaghi, Roozbeh, Xiang, Yu, Savarese, Silvio

cvpr2015papers_第116张图片
[set, learning, car, perform, work, subset, sedan, category] [training, layer, table, level, better, recognition, performance, coarse, representation, pascal, dpm, loss, feature, increase, standard, hog] [object, box, hierarchical, bounding, detection, three, proposal, based, segmentation, appearance, groundtruth] [viewpoint, continuous, model, discrete, cad, pose, potential, estimation, inference, joint, represent, alignment, modeling, detailed, glb, reason, typically, mixture, height] [consider, note, method, function, variable, projection, random, number, case] [figure, azimuth, estimated, estimate, result, elevation] [hierarchy, distance, compute, computed, corresponding, sampling, image, compare, represents]
Fine-Grained Recognition Without Part Annotations
Krause, Jonathan, Jin, Hailin, Yang, Jianchao, Fei-Fei, Li

cvpr2015papers_第117张图片
[vision, test, set, visual, work, annotation, european, generating, dataset] [recognition, bbox, training, class, discriminative, performance, cnn, feature, decision, trained, improves, convolutional, learned, learn, better, standard, size, top, combining, caffenet, table, setting] [bounding, segmentation, foreground, object, box, detection, combination, grabcut, based, background, generated, truth, appearance] [pose, alignment, approach, model, aligning, prior, align, single, term, ground, point, shape, represent] [method, graph, time, analysis, minimum, spanning, machine, scaling, number] [computer, pattern, conference, international, figure] [image, difference, large, comparison, propagated]
Picture: A Probabilistic Programming Language for Scene Perception
Kulkarni, Tejas D., Kohli, Pushmeet, Tenenbaum, Joshua B., Mansinghka, Vikash

cvpr2015papers_第118张图片
[vision, human, language, learning, summary, automatically, automatic, set, dataset, latent, work, representative] [neural, representation, deep, layer, stochastic, convolutional, learned, trained, discriminative, network, arxiv] [object, proposal, based, complex, monte, generate, density, likelihood, hypothesized, appearance] [inference, program, scene, probabilistic, picture, model, generative, pose, face, shape, approach, sreal, elliptical, modeling, supplementary, reconstruction, continuous, single, error, earlier, render, mcmc, express, tolerance, quantitative] [random, approximate, programming, function, data, trace, kernel, number, code, gaussian] [figure, observed, rendering, computer, slice, inverse, camera, conference] [image, bayesian, fast, large]
Learning to Segment Moving Objects in Videos
Fragkiadaki, Katerina, Arbelaez, Pablo, Felsen, Panna, Malik, Jitendra

cvpr2015papers_第119张图片
[ranking, set, work, learning, generation] [detector, cnn, performance, previous, network, training, convolutional, trained, combining, architecture, layer, input, increase] [object, moving, trajectory, segmentation, static, frame, video, segment, detection, objectness, boundary, tube, proposal, multiple, supervoxel, background, multiscale, average, moseg, score, mop, bounding, box, based, union, tracking, intersection, hierarchical, boost] [point, challenging, capture, rgb] [method, number, clustering, random, consider, proposed, denote, code, effectively] [motion, optical, figure, pixel, extend] [image, large, best, dense, well, compute, compare, benchmark, displacement, rate, weighted, outperforms]
Multi-Feature Max-Margin Hierarchical Bayesian Model for Action Recognition
Yang, Shuang, Yuan, Chunfeng, Wu, Baoxin, Hu, Weiming, Wang, Fangshi

cvpr2015papers_第120张图片
[learning, latent, dataset, process, topic, perform, task, test, lda, datasets, visual, work, human, label, expected] [hbm, hgm, recognition, representation, feature, learn, class, global, kth, learned, discriminative, table, stp, accuracy, jointly, svm, shared, loss, combining, training, context, capability, achieves] [action, hierarchical, multiple, based, ucf, clip, employ, average, wang, temporal, mbh, video, framework, appearance, including, propose, integrate] [model, generative, inference, probabilistic, graphical, represent] [proposed, gibbs, gaussian, distribution, number, method, dirichlet, parameter] [figure, total] [bayesian, local, sampling, compare, spatial, sift, divergence, image]
Model Recommendation: Generating Object Detectors From Few Samples
Wang, Yu-Xiong, Hebert, Martial

cvpr2015papers_第121张图片
[task, set, transfer, library, collaborative, store, learning, airplane, bed, recommendation, pbc, dataset, system, recommender, unsupervised, category, search, rating, source, select, predict, generating, perform, nmf, test, sun, work, unlabeled, supervised, visual, annotated, randomly, recommended] [input, performance, training, chair, factor, ensemble, size, feature, pascal, map, hidden, voc, shared, better, detector, contextual, dpm] [detection, object, target, based, average, generate, generated, bounding, box, entire] [model, approach, single, rij, procedure] [matrix, data, number, factorization, vector, svd, note, random, space] [direct, figure, setup, relative, directly] [large, probe, best, well, image, region]
Adaptive Region Pooling for Object Detection
Tsai, Yi-Hsuan, Hamsici, Onur C., Yang, Ming-Hsuan

cvpr2015papers_第122张图片
[set, transfer, selected, search, learning, representative, positive, test, esvm, select, evaluate, dataset, car] [pooling, exemplar, feature, training, coarse, table, learn, cnn, testing, pascal, better, performance, achieves, voc, transferred, top, pyramid, discriminative, apply, negative, train, map, size] [object, detection, bounding, recall, adaptive, overlap, segmentation, box, based, target, truth, appearance, propose, performs, mask, segmented, extract, multiple] [structure, detected, model, regression, ground, approach, error] [method, algorithm, proposed, number, note, linear, cluster] [figure, ratio, small, quality] [region, matching, rate, large, sift, well, keypoint, keypoints, image, spatial, compare]
Fusing Subcategory Probabilities for Texture Classification
Song, Yang, Cai, Weidong, Li, Qing, Zhang, Fan, Dagan Feng, David, Huang, Heng

cvpr2015papers_第123张图片
[probability, dataset, test, visual, generation, contribution, datasets, set, representativeness] [subcategory, feature, svm, subcategories, sck, class, performance, caffe, ifv, compared, training, fmd, belonging, design, higher, fck, representation, better, dtd, multiclass, discriminative, convolutional, improvement, cotton, trained, qck, bread, accuracy, improve, table, rck] [based, three, object, segmentation, weight] [model, challenging, performed, fused, represent] [data, method, clustering, vector, subspace, cluster, sparse, number, fusion, matrix, analysis, approximated, encoding, lower, proposed, support] [figure, ieee, paper, pattern, second] [texture, image, computed, distinctiveness, local, difference, descriptor, applied, variation, similarity]
DeepShape: Deep Learned Shape Descriptor for 3D Shape Matching and Retrieval
Xie, Jin, Fang, Yi, Zhu, Fan, Wong, Edward

cvpr2015papers_第124张图片
[noisy, set, clean, human, learning, dataset, green, workshop] [discriminative, layer, hidden, feature, deep, performance, input, output, training, class, train, representation, neural, activation, learned, learn, standard, discrimination, hybrid, fisher, unit] [based, multiscale, multiple, object] [shape, model, represent, vertex, mesh, surface, term, point, pose] [proposed, distribution, method, function, kernel, form, number, order, denote, algorithm, vector, represented, criterion, denotes, denoted, sparse, scatter] [computer, noise, figure, red, small] [descriptor, retrieval, heat, matching, hks, dasd, mcgill, dictionary, geometric, scale, shapegoogle, signature, eurographics, partial, large, local, invariant, insensitive, histogram, crab, benchmark]
Learning Semantic Relationships for Better Action Retrieval in Images
Ramanathan, Vignesh, Li, Congcong, Deng, Jia, Han, Wei, Li, Zhen, Gu, Kunlong, Song, Yang, Bengio, Samy, Rosenberg, Charles, Fei-Fei, Li

cvpr2015papers_第125张图片
[relationship, visual, relation, vision, set, prediction, language, embedding, semantic, learning, dataset, consistency, mutually, work, human, people, positive, exclusive, riding, predicted, rab, test, feeding, textual, holding, newly, ank, knowledge, processing, correct, cab, subset, ang, label, returned, search] [recognition, loss, training, neural, learn, higher, performance, additional, better, layer, improvement, gain, learned, network, stanford, train, compared] [action, object, based, weight, three] [model, prior, full, single] [method, objective, function, vector, data, number, original] [computer, conference, ieee, pattern, international, figure, small] [image, retrieval, person, corresponding, large, similarity]
Joint Calibration of Ensemble of Exemplar SVMs
Modolo, Davide, Vezhnevets, Alexander, Russakovsky, Olga, Ferrari, Vittorio

cvpr2015papers_第126张图片
[positive, test, search, set, evaluate, example, visual, task, subset] [sigmoid, window, training, negative, table, scored, global, better, ilsvrc, loss, pascal, isotonic, trained, voc, cnn, performance, output, map, calibrating, calibrate, class, exemplar, train, root, hog, ensemble, max, rooted, svm] [object, threshold, score, detection, recall, average, node, candidate] [joint, tree, independent, procedure, step, regression, single, subtree] [algorithm, optimal, solution, feasible, positively, number, optimization, method, space, note, pruning, consider, problem, satisfy, data, equivalent, approximate, function] [calibration, observation, figure, presented, small] [false, compare, globally, large, outperforms, image]
SOM: Semantic Obviousness Metric for Image Quality Assessment
Zhang, Peng, Zhou, Wengang, Wu, Lei, Li, Houqiang

cvpr2015papers_第127张图片
[semantic, human, dataset, visual, set, predicted, existing, work, evaluation, type, learning, natural, perception, randomly] [feature, performance, generic, raw, table, training, previous, extracted, pooling, jpeg, cnn, top, global, trained, discriminative, detector] [based, detection, obviousness, iqa, object, assessment, extract, score, srocc, framework, measure, lcc, perceptual, objectness, som, brisque, cornia, dmos, correlation] [live, approach, regression, represent] [codebook, proposed, number, method, experimental, range, machine, linear, encoding, adopted, promising] [quality, distorted, figure, distortion, signal, paper, blind, result] [image, local, reference, metric, region, corresponding, fast, consists, ssim]
Learning a Sequential Search for Landmarks
Singh, Saurabh, Hoiem, Derek, Forsyth, David

cvpr2015papers_第128张图片
[human, learning, set, search, dataset, scoring, tend, test, fashion, visual] [training, learns, context, learned, performance, better, margin, layer, contextual, apply, learn, feature] [detection, appearance, truth, state, candidate, object, detect, art, individual, making] [landmark, model, detected, pose, step, inference, ordering, ground, approach, dependent, location, independent, earlier, body, lsp, articulated, current, error, sequential, modeling, estimation, pictorial, provide] [method, order, note, normalized, fixed, general, consider, random, function, original, close, range] [figure, high, small] [image, spatial, distance, gradient, computed, well, rate, good, best, corresponding, comparison]
Attributes and Categories for Generic Instance Search From One Example
Tao, Ran, Smeulders, Arnold W.M., Chang, Shih-Fu

cvpr2015papers_第129张图片
[instance, search, attribute, shoe, dataset, visual, learning, category, existing, example, evaluate, address, cleanshoes, work, set, automatically, semantic, expvlad, relevant, embedding, aim, manually, triemb, common] [fisher, generic, learned, deep, training, performance, representation, feature, better, combining, improve, table, pascal, top, learn, achieves, cnn, achieved] [object, appearance, propose, three, combination, target] [searching, building, viewpoint] [consider, vector, range, proposed, method, space, problem, arbitrary, locality, number] [figure, second, high, result] [query, image, large, local, descriptor, similarity, retrieval, good, randomized, best, well, matching, scale, wide]
Domain-Size Pooling in Local Descriptors: DSP-SIFT
Dong, Jingming, Soatto, Stefano

cvpr2015papers_第130张图片
[domain, vision, dataset, learning, selected, evaluation, include] [pooling, size, performance, recognition, feature, map, convolutional, neural, arxiv, larger, improvement, discriminative, network, response, standard, detector, increasing] [average, precision, interpretation, including] [scene, point, visibility, performed, detected, single] [number, dimension, complexity, derivation, depends, analysis, method] [computer, conference, pattern, ieee, figure, frequency, small, international, blur, sensitivity, planar, classical] [sift, computed, descriptor, scale, matching, image, spatial, magnitude, oxford, transformation, outperforms, comparison, nuisance, local, large, best, sampling, fischer, gradient, variability, compare, ordinary, distance, corresponding, aggregating]
Zero-Shot Object Recognition by Semantic Manifold Distance
Fu, Zhenyong, Xiang, Tao, Kodirov, Elyor, Gong, Shaogang

cvpr2015papers_第131张图片
[semantic, embedding, unseen, learning, absorbing, test, visual, amp, attribute, zsl, relatedness, word, process, dataset, probability, set, label, existing, strategy, novel, transient, awa, vision, prototype, transductive, knowledge, processing, devise, labelled] [class, feature, imagenet, recognition, connected, neural, cnn, training, deep, compared, performance, achieved, learned, convolutional, accuracy] [based, object, node, state, transition, combination, framework, measure] [model, structure, approach, tree] [space, graph, manifold, vector, chain, markov, method, matrix, proposed, projection, extended, number, represented, data] [computer, result, figure, ieee, pattern, conference] [image, distance, similarity, large, computed, metric, scale, compute, outperforms, nearest]
The S-Hock Dataset: Analyzing Crowds at the Stadium
Conigliaro, Davide, Rota, Paolo, Setti, Francesco, Bassetti, Chiara, Conci, Nicola, Sebe, Nicu, Cristani, Marco

cvpr2015papers_第132张图片
[people, set, game, human, categorization, novel, collection, vision, focus, annotated, annotation, dataset, perform, ice] [neural, standard, training, accuracy, map, size, performance, trained, convolutional, network, dpm, cnn, deep] [crowd, spectator, detection, behavior, action, based, propose, focusing, ock, bounding, crowded, team, supported, sae, rink, frame, segmentation, sport, object, video] [head, pose, social, body, estimation, prior, face, model, single, fact, ground, modeling] [method, order, analysis, consider, proposed, time, upper, literature, computation, data, low, case] [computer, paper, figure, motion, left, second] [person, considered, large, best, patch, image, similarity, computed]
Appearance-Based Gaze Estimation in the Wild
Zhang, Xucong, Sugano, Yusuke, Fritz, Mario, Bulling, Andreas

cvpr2015papers_第133张图片
[dataset, evaluation, datasets, multimodal, work, collected, existing, test, subset, learning, system, evaluate, visual, collection] [training, performance, cnn, input, convolutional, fully, layer, neural, previous, learn, connected, larger] [eye, appearance, tracking, detection, key, three, indicates] [gaze, estimation, head, mpiigaze, pose, face, multiview, model, eyediap, error, coordinate, single, regression, normalisation, facial, normalised, challenging, potential, continuous, approach, controlled, pupil, current] [data, method, number, space, proposed, vector, random, respect, sample, mapping] [figure, camera, angle, intensity, left, illumination, total, handle, lighting, light] [image, large, well, difference, person, region]
Multi-Task Deep Visual-Semantic Embedding for Video Thumbnail Selection
Liu, Wu, Mei, Tao, Zhang, Yongdong, Che, Cherry, Luo, Jiebo

cvpr2015papers_第134张图片
[embedding, thumbnail, semantic, visual, learning, relevance, dataset, search, representative, selected, select, labeled, latent, side, positive, label, representativeness, word, textual, common, keyframes, summarization, content, clicked, strategy, leverage, set, web, freely, evaluate] [deep, trained, training, map, representation, compared, train, performance, layer, loss, improve, input, cnn, highly, achieves, effectiveness, shared, higher] [video, based, candidate, score, employ, frame, online] [model] [method, selection, proposed, fusion, vector, space, data, projection, function, china, order, algorithm] [figure, directly, conventional, equation] [query, image, similarity, compute, choose, gap, well, computing]
Ontological Supervision for Fine Grained Classification of Street View Storefronts
Movshovitz-Attias, Yair, Yu, Qian, Stumpe, Martin C., Shet, Vinay, Arnoud, Sacha, Yatziv, Liron

cvpr2015papers_第135张图片
[business, text, restaurant, food, system, set, store, human, beauty, label, category, visual, shopping, gas, health, vision, prediction, drink, textual, automotive, language, cafe, grained, correct, ontology, salon, listing, task, geographical, google, test, work, labeled, dental, nail, process, ood, word, transcription, containment, search, study, storefront] [training, network, top, accuracy, level, recognition, deep, convolutional, neural, extracted, googlenet, architecture, performance, train] [recall, precision, based, multiple, object, detection] [model, location, described, ground, full, approach] [data, number, method, note, multi] [figure, computer, conference, ieee, real, pattern, high, international] [image, street, large, view, scale, area]
ConceptLearner: Discovering Visual Concepts From Weakly Labeled Image Collections
Zhou, Bolei, Jagadeesh, Vignesh, Piramuthu, Robinson

cvpr2015papers_第136张图片
[concept, visual, set, weakly, labeled, sbu, learning, dataset, supervised, hard, positive, label, sentence, associated, tag, weak, instance, selected, search, describe, sun, description, discovery, relevant, evaluate, baseline, test, discover, work, predicted, discovered, ranked, datasets, fwc, system, knowledge, human, discovering] [learned, recognition, train, pascal, feature, voc, negative, fully, training, detector, deep, pool, learn, supervision, table, top, svm, accuracy, validation] [object, detection, based, multiple, generate, target] [scene] [method, selection, data, vector, number, easy, algorithm, cluster, note, row] [figure] [image, large, database, well, region]
Metric Imitation by Manifold Transfer for Efficient Vision Applications
Dai, Dengxin, Kroeger, Till, Timofte, Radu, Van Gool, Luc

cvpr2015papers_第137张图片
[learning, domain, transfer, embedding, source, human, test, unsupervised, datasets, vision, set, work, knowledge, goal, privileged, popular] [training, learned, table, feature, cnn, learn, trained, convolutional, better, standard, performance, van, learns, improves, expensive, compared, global] [object, imitation, target, three, precision] [evaluated] [manifold, method, data, matrix, lap, dimensionality, space, clustering, linear, mapping, pairwise, computationally, problem, gist, rest, function, projection, euclidean, time] [tested] [metric, sfs, image, tfs, distance, lle, local, good, retrieval, lbp, computed, neighborhood, jor, large, transformation, phog, nearest, cheap, transformed, ukbench, neighbor]
Large-Scale Damage Detection Using Satellite Imagery
Gueguen, Lionel, Hamid, Raffay

cvpr2015papers_第138张图片
[learning, unsupervised, supervised, positive, sun, set, work, labeled, human, novelty, label, indicator, relevant, collected, example] [feature, training, extracted, table, cnn, level, accuracy, learn, size, input, reported] [detection, framework, based, multiple, average, cost] [shape, detected, tree, error, model, canonical] [data, analysis, time, number, problem, note, proposed, linear, lower, consider, vector, upper] [figure, ieee, acquisition, high, elevation] [image, damage, satellite, difference, change, rate, area, damaged, imagery, equal, represents, pair, variability, curation, false, impact, local, computed, gsd, eer, true, sift, opposed]
Joint Patch and Multi-Label Learning for Facial Action Unit Detection
Zhao, Kaili, Chu, Wen-Sheng, De la Torre, Fernando, Cohn, Jeffrey F., Zhang, Honggang

cvpr2015papers_第139张图片
[learning, jpml, apl, positive, selected, gft, relation, automatic, facs, subset, select, work, exploit, describe] [negative, feature, table, performance, training, unit, discriminative, better, standard, compared, jointly, involves, learn, improve, recognition] [action, three, group, correlation, detection, dynamic, truth] [facial, face, joint, spontaneous, expression, model, ground, affective] [matrix, lower, sparse, upper, algorithm, norm, problem, data, vector, number, iteration, min, selection, sparsity, regularizer, spsd, convergence, admm, effectively, coding, linear, column, denote] [figure] [patch, best, skew, image, indicate, computed, local, bold]
Cross-Age Face Verification by Coordinating With Cross-Face Age Verification
Du, Liang, Ling, Haibin

cvpr2015papers_第140张图片
[task, learning, study, visual, dataset, novel, work, share, baseline, versus, strategy, competition, datasets, selected, positive] [age, feature, auxiliary, recognition, loss, training, agcd, coordinating, discriminative, sharing, boosting, morph, cme, previous, exclusion, performance, table, coordination, size, improve, pool] [target, hypothesis, framework, weight, gop, multiple] [face, coordinate, joint, model, facial, descent, modeling, estimation] [algorithm, proposed, greedy, base, experimental, analysis, function, alternating, orthogonal, regularization, regularized, selection, conducted, number, sensitive, binary, update, graph, method] [figure, small, tested, competing] [image, best, aging, eer, gradient, local, invariant, difference, large, cross]
Regularizing Max-Margin Exemplars by Reconstruction and Generative Models
Rubio, Jose C., Ommer, Bjorn

cvpr2015papers_第141张图片
[positive, learning, set, visual, lda, category, vision, mining, hard, esvm, domain, popular, search, process] [training, exemplar, negative, performance, margin, hog, trained, representation, recognition, discriminative, pascal, mlda, voc, train, additional, dpm, class, feature, structured, improve, table, size, generalization] [object, bounding, detection, propose, multiple, average, based, three] [model, reconstruction, generative, approach, single, localization, prior, neighboring] [original, regularization, distribution, covariance, optimization, optimal, maximum, sample, number, matrix, approximation, method, data, order, linear, computationally] [figure, computer, conference, pattern, estimate, quality, ieee] [image, large, good, wavelet, corresponding, well, nearest, patch, region]
Supervised Mid-Level Features for Word Image Representation
Gordo, Albert

cvpr2015papers_第142张图片
[word, character, text, learning, label, annotated, supervised, semantic, encode, atts, visual, dataset, lexicon, set, work, char, perform, midlevel, describe, localizing, spotting, test, correlated, embedding, bbi] [global, training, recognition, representation, fisher, learn, testing, fvs, standard, accuracy, pyramid, deep, level, discriminative, table] [bounding, based, extract, individual, box, framework, propose] [approach, scene, project] [block, dimensionality, space, svt, proposed, projection, normalized, explicitly] [small, produce, figure, second] [image, sift, spatial, retrieval, compute, construct, region, sampled, aggregated, local, densely, aggregate, normalization, computing, best]
Basis Mapping Based Boosting for Object Detection
Ren, Haoyu, Li, Ze-Nian

cvpr2015papers_第143张图片
[weak, vision, dataset, hard, positive, referring, evaluation, select, description, learning, human, evaluate] [training, hog, pedestrian, boosting, xbasis, feature, boosted, logitboost, inria, accuracy, chm, performance, size, train, cascade, pascal, better, effectiveness, achieves, testing, table, voc, trained, learn, detector, improves, compared, caltech] [detection, object, based, propose, intersection] [current, face] [basis, mapping, sample, proposed, kernel, original, linear, space, method, function, machine, algorithm, number, analysis, vector, approximation, problem, dimension, convergence, computation, implicit] [computer, ieee, conference, pattern, speed, international, figure, result, conventional] [mapped, descriptor, histogram, rate, local, best, false, region, image, fast]
Sparse Projections for High-Dimensional Binary Codes
Xia, Yan, He, Kaiming, Kohli, Pushmeet, Sun, Jian

cvpr2015papers_第144张图片
[learning, dataset, work, evaluate, visual] [map, accuracy, deep, effective, table, learned, feature, performance, imagenet, better, neural, raw, convolutional, dnn] [long, based, applicable] [valid, model] [binary, solution, encoding, method, itq, projection, matrix, time, problem, cbe, lsh, number, code, sparse, data, iterative, sparsity, objective, constraint, regularizer, case, function, procrustes, min, note, hamming, faster, orthogonal, thrm, computational, algorithm, thresholding, coding, optimization, circulant, approximate, update, vector, easily] [accurate, length, figure, requires, second] [image, large, nearest, dense, retrieval, degradation, vlad]
Oriented Edge Forests for Boundary Detection
Hallman, Sam, Fowlkes, Charless C.

cvpr2015papers_第145张图片
[label, probability, set, predicted, natural, vision, learning, system] [training, output, detector, decision, performance, trained, contour, input, map, structured, centered, table, learn, feature, class] [edge, boundary, detection, oriented, precision, average, contrast, segmentation, individual, based] [forest, tree, model, orientation, approach, location, posterior, discrete, allows] [space, data, random, parameter, sparse, distribution, number, empirical, analysis, order] [figure, calibration, computer, averaging, ieee, pixel, pattern, requires, conference, simple, small, high, produce] [image, patch, local, sharpening, voting, benchmark, scale, large, distance, computed, monotonic, oef, randomized, outperforms]
Enriching Object Detection With 2D-3D Registration and Continuous Viewpoint Estimation
Bongsoo Choy, Christopher, Stark, Michael, Corbett-Davies, Sam, Savarese, Silvio

cvpr2015papers_第146张图片
[dataset, novel, test, existing, car, work, provided, search, generation, set, par, probability, positive] [hog, performance, exemplar, class, feature, convolution, train, training, detector, output, negative, ensemble, table, discriminative, additional, level, trained] [object, detection, bounding, precision, based, box, average, candidate] [viewpoint, cad, model, pose, template, estimation, rendered, continuous, discrete, single, mcmc, joint, approach, enrich, deformable] [method, conjugate, time, whitening, number, original, parameter, matrix, whitened, iterative, computational, problem, algorithm, condition, data] [figure, focal, rendering, calibration, azimuth, length, result, high] [image, large, gradient, matching, corresponding]
Transferring a Semantic Representation for Person Re-Identification and Search
Shi, Zhiyuan, Hospedales, Timothy M., Xiang, Tao

cvpr2015papers_第147张图片
[attribute, learning, unsupervised, semantic, domain, search, surveillance, supervised, annotated, transfer, fashion, adaptation, clothing, existing, latent, visual, weakly, human, source, colour, datasets, annotation, category, bag, set, description, dasa, evaluate, people, weak, dataset] [representation, auxiliary, factor, learned, performance, table, svm, training, feature, supervision, train, compared, learn, accuracy] [target, appearance, strong, framework, final, generate, based, including, detection] [model, approach, joint, generative, single, face] [data, mrf, vector, method, matrix, number, binary, consider, studied, sample, dimensional] [camera, blue, paper, color] [person, image, query, patch, matching, compute, well, viper]
Hierarchical-PEP Model for Real-World Face Recognition
Li, Haoxiang, Hua, Gang

cvpr2015papers_第148张图片
[vision, set, learning, process, labeled, selected, work, report, baseline, supervised, dataset] [pep, recognition, representation, discriminative, layer, training, feature, accuracy, lde, performance, size, build, table, observe, deep, train, hierarchicalpep, apply, cosine, reduce, previous] [video, propose, based, appearance, hierarchical, multiple] [face, model, joint, pose, elastic, full] [dimensionality, pasc, method, machine, analysis, algorithm, data, experiment, reduction, working, linear, number, selection] [computer, ieee, conference, pattern, figure, simple, international, equation] [image, sift, bayesian, pca, similarity, consists, lfw, patch, pair, matching, comparison, construct, hierarchy, discriminant]
Learning Descriptors for Object Recognition and 3D Pose Estimation
Wohlhart, Paul, Lepetit, Vincent

cvpr2015papers_第149张图片
[set, learning, test, vision, evaluation, dataset, search, work, perform, add, evaluate, representative] [training, hog, recognition, input, train, network, cnn, performance, accuracy, output, learn, fully, class, discriminative, convolutional, neural] [object, detection, cost] [pose, depth, template, rgb, approach, estimation, error, rely, capture, term] [data, method, euclidean, sample, machine, analysis, consider, number, scalable] [computer, conference, pattern, figure, small, ieee, angle, length, ratio, synthetic] [descriptor, linemod, distance, image, similarity, nearest, closest, neighbor, large, dissimilar, corresponding, rate, pair, compute, applied, best, implementation, gradient, mapped, metric, well, ltriplets]
Learning Hypergraph-Regularized Attribute Predictors
Huang, Sheng, Elhoseiny, Mohamed, Elgammal, Ahmed, Yang, Dan

cvpr2015papers_第150张图片
[attribute, hypergraph, learning, hap, prediction, embedding, cshapg, cub, cshaph, usaa, hyperedge, awa, visual, zsl, label, predictor, dataset, test, semantic, side, cshap, collection, provided, categorization, relation, set, depict, supervised, ale, positive, encode, latent] [class, training, feature, accuracy, loss, additional, better, representation, performance, table, jointly, learn] [correlation, three, framework, object, incorporate, based] [approach, model, vertex, capture, aligned] [space, kernel, cut, matrix, proposed, graph, sample, data, exploiting, number, laplacian, problem, normalized, function, consider, denoted, arg, respect, regularized, vector, sign, optimal, mapping] [equation, figure] [well, corresponding]
Beyond Spatial Pooling: Fine-Grained Representation Learning in Multiple Domains
Li, Chi, Reiter, Austin, Hager, Gregory D.

cvpr2015papers_第151张图片
[domain, learning, dataset, visual, work, instance, set, subset] [pooling, feature, recognition, convolutional, representation, better, pooled, testing, discrimination, cshot, deep, table, learned, smaller, achieves, pyramid, performance, achieved, xyz, design, architecture, receptive, max, level, arxiv, context] [object, three, framework, hierarchical, multiple, average, score] [probabilistic, point, pose, depth, current] [proposed, algorithm, method, random, kernel, experiment, space, data, order] [color, viewing, figure, presented] [spatial, invariance, invariant, variance, local, image, bigbird, large, descriptor, distance, choose, region, scale, derived, good, major, view]
Second-Order Constrained Parametric Proposals and Sequential Search-Based Structured Prediction for Semantic Segmentation in RGB-D Images
Banica, Dan, Sminchisescu, Cristian

cvpr2015papers_第152张图片
[semantic, set, test, learning, process, labeling, label, labeled, dataset, selected, prediction, understanding, visual, work, side, generation, category] [table, class, deep, pooling, training, extracted, trained, structured, pooled, feature, performance, recognition] [segmentation, object, based, state, segment, policy, proposal, generated, truth, multiple, action, average, generate, boundary, candidate, score] [depth, scene, rgb, rmrc, inference, nyu, indoor, current, spin, ground, model, sequential, methodology, procedure, point, described, retained, pcf] [order, method, data, number, proposed, problem, algorithm] [intensity, parametric] [image, local, region, spatial, partial, metric, masked, well, overlapping, large]
Learning to Propose Objects
Krahenbuhl, Philipp, Koltun, Vladlen

cvpr2015papers_第153张图片
[set, dataset, type, coco, learning, evaluate, work, add, highest, semantic] [training, ensemble, global, trained, accuracy, table, voc, feature, crf, size, microsoft, train, pascal, performance, optimizes] [object, proposal, segmentation, localized, superpixel, candidate, detection, facility, bounding, average, abo, box, recall, generated, seed, multiple, three, overlap, boundary, edge, gop, combinatorial, koltun, objectness] [approach, model, prior, procedure, location, single, composition] [objective, number, algorithm, optimized, parameter, pairwise, time, vector, method, min, binary, running, approximation, crfs, note, graph] [presented, figure, produced, small, accurate] [image, region, best, distance, outperforms, sampled, well, fast]
Weakly Supervised Object Detection With Convex Clustering
Bilen, Hakan, Pedersoli, Marco, Tuytelaars, Tinne

cvpr2015papers_第154张图片
[learning, latent, supervised, positive, weakly, set, work, probability, search, visual, help] [training, window, discriminative, learned, class, learn, standard, performance, voc, song, table, feature, pascal, additional, loss, svm, global, negative, improve, jointly, neural, learns] [object, detection, target, based, multiple, appearance, background, propose, measure, average] [term, localization, single, model, location, inference, challenging] [method, clustering, convex, formulation, cluster, algorithm, slsvm, optimization, case, parameter, number, problem, regularization, stuck, function, vector, optimal, initialize, enforces, objective] [figure, avoid, smooth, small, enforce] [similarity, image, best, compare, local, comparison, weighted, outperforms]
Video Event Recognition With Deep Hierarchical Context Model
Wang, Xiaoyang, Ji, Qiang

cvpr2015papers_第155张图片
[semantic, vision, learning, dataset, baseline, surveillance, human, set, work, existing, label] [context, level, recognition, feature, deep, hidden, contextual, accuracy, table, combining, performance, learn, improve, layer, previous] [event, three, virat, interaction, based, object, state, appearance, art, video, hierarchical, bounding, utilize, middle, average, box, stip, multiple, key, action, propose, activity, dynamic, wang, including, integrates, integrate] [model, prior, scene, ground, capture, approach, represent, probabilistic] [proposed, vector, method] [conference, figure, computer, ieee, pattern, observation, international] [person, neighborhood, histogram, well, compare, sift, includes, image]
Image Parsing With a Wide Range of Classes and Scene-Level Context
George, Marian

cvpr2015papers_第156张图片
[label, system, baseline, semantic, set, test, labeling, visual, dataset] [parsing, training, balanced, global, context, class, unbalanced, performance, better, nonparametric, recognition, improve, fusing, feature, lmsun, combining, improves, rare, road, bdt, accuracy, achieve, smoothing, size, tighe, output, lsj, previous] [likelihood, superpixels, truth, superpixel, foreground, object, background, cost, score, weight, average, coverage, multiple, combined] [scene, full, ground, inference, step, initial, approach, error, model, probabilistic] [number, method, energy, function, data, mrf, normalized, fusion] [figure, equation, total] [image, retrieval, query, compute, large, rate, computed, lsi, person]
Combination Features and Models for Human Detection
Jiang, Yunsheng, Ma, Jinwen

cvpr2015papers_第157张图片
[human, vision, set, test, work, correspond] [grammar, poselet, hog, feature, performance, voc, recognition, table, effective, hoc, hob, pascal, whl, overlapped, dpm, fuse, conduct, weightednms, competitive, improvement, combine, input, deep, impressive] [detection, object, based, combination, score, framework, integrated, threshold, precision] [model, deformable, structure, procedure, single, volume] [fusion, algorithm, proposed, general, note, method, time, computation, denotes, low, original, order, function] [computer, ieee, color, pattern, figure, direct, hue, calibrated, intensity, saturation] [image, person, large, corresponding, gradient, well, histogram, good, texture]
Computing the Stereo Matching Cost With a Convolutional Neural Network
Zbontar, Jure, LeCun, Yann

cvpr2015papers_第158张图片
[set, learning, vision, dataset, test, work, example, predict, perform, evaluate] [network, training, convolutional, size, neural, output, input, learn, map, yamaguchi, architecture, larger, negative, top, centered, table, recognition] [cost, forward, three, object, combined] [kitti, error, position, neighboring, location, initial, single, term, depth] [method, support, random, machine, number, runtime, pass, denote, analysis, function] [stereo, disparity, left, computer, ieee, figure, pattern, horizontal, conference, vertical, international, camera, pixel] [matching, image, aggregation, computed, patch, semiglobal, computing, arm, ccbca, compute, rate, csgm, differ, pair, well, true]
A Mixed Bag of Emotions: Model, Predict, and Transfer Emotion Distributions
Peng, Kuan-Chuan, Chen, Tsuhan, Sadovnik, Amir, Gallagher, Andrew C.

cvpr2015papers_第159张图片
[emotion, evoked, source, probability, category, transferring, cnnr, predicting, emotional, user, set, transfer, provided, work, predict, amt, positive, dmtr, keywords, subject, joy, semantic, study, hit, randomly, human] [table, previous, better, cnn, input, performance, convolutional, feature, neural, level, training, comparable] [target, truth, framework, multiple, three, average, saliency, propose, based] [ground, affective, dominant, single, model, svr] [distribution, method, basic, random, proposed, experiment, number, changing, algorithm] [color, figure, international, adopt, uniform, ieee] [image, transformed, distance, change, texture, outperforms, closer, tone, compare, database, corresponding, transform, histogram, difference, modify]
Image Retrieval Using Scene Graphs
Johnson, Justin, Krishna, Ranjay, Stark, Michael, Li, Li-Jia, Shamma, David, Bernstein, Michael, Fei-Fei, Li

cvpr2015papers_第160张图片
[test, visual, dataset, vision, man, grounding, set, semantic, attribute, learning, grounded, semantics, novel, example, relationship, text, system, wearing, retrieve, natural, language, tuples, search, work, occur, perform, association, describe] [recognition, performance, representation, table, training, context, class, cnn, structured, crf] [object, recall, iou, entire, based, bounding] [scene, model, full, detailed, localization, represent, modeling] [graph, rank, gist, random, data, number, computational, journal] [computer, conference, ieee, pattern, median, international, figure, simple, detail] [image, retrieval, query, large, sift, outperforms, vocabulary, compute, descriptor, scale]
Scalable Object Detection by Filter Compression With Regularized Sparse Coding
Chao, Ting-Hsuan, Lin, Yen-Liang, Kuo, Yin-Hsi, Hsu, Winston H.

cvpr2015papers_第161张图片
[set, dataset, perform, work, vision, system, process, learning, visual, collected, budget] [map, performance, functionality, convolution, accuracy, achieve, size, feature, better, recognition, conduct, larger, trained, training, loss, stage, learn, dpm, drop, reduce, intermediate, class, higher] [object, detection, score, memory, proposal, based, appearance, illustrate] [reconstruct, model, deformable, error, reconstruction] [sparse, method, coding, codebook, number, speedup, regularized, regularization, original, augmented, huge, computation, experiment, optimization, computational, matrix, codewords, complexity, linear, problem, space, freg, scalability, negligible, minimizing, proposed] [figure, computer, small, ieee, conference, produce, pattern, equation, regard, reconstructed, international, result] [large, difference, computing]
TILDE: A Temporally Invariant Learned DEtector
Verdie, Yannick, Yi, Kwang, Fua, Pascal, Lepetit, Vincent

cvpr2015papers_第162张图片
[learning, vision, dataset, webcam, set, positive, introduce, task, work] [training, detector, feature, performance, learned, layer, loss, trained, learn, response, standard, size, train] [score, average, three, sequence, detection, temporal, extract] [approach, term, detected, shape, location] [method, linear, number, machine, proposed, function, objective, original, time, approximation] [computer, conference, pattern, simple, figure, international, robust, outdoor, ieee, captured, illumination] [keypoints, keypoint, image, regressor, repeatability, sift, oxford, surf, local, scale, applied, separable, sfop, wade, drastic, invariant, mser, sifer, lcf, matching, large, good, wnm, best, fast, chamonix]
Global Refinement of Random Forest
Ren, Shaoqing, Cao, Xudong, Wei, Yichen, Sun, Jian

cvpr2015papers_第163张图片
[learning, prediction, vision, indicator, set, randomly, dataset, human, rule, visual, subset] [global, training, standard, loss, recognition, decision, size, performance, better, accuracy, reduce, boosting, table, version, refined, smaller, age, feature, achieves, power, deep, clearly] [complementary, individual, propose, strong, strength, node, multiple] [forest, leaf, tree, model, error, depth, regression, randomness, face, independent, single, body] [random, number, pruning, vector, machine, data, function, optimization, alternating, optimal, method, parameter, formulation, matrix, mnist, regularization, experiment, reformulation] [conference, computer, figure, ieee, pattern, real, international, ratio, accurate, relative] [large, well, ideal, local, corresponding]
Learning a Non-Linear Knowledge Transfer Model for Cross-View Action Recognition
Rahmani, Hossein, Mian, Ajmal

cvpr2015papers_第164张图片
[learning, knowledge, dataset, transfer, human, existing, set, source, test] [recognition, training, learn, learned, accuracy, input, table, layer, deep, output, feature, representation, shared, discriminative, neural, svm] [action, path, target, trajectory, video, aog, weight, extract, three, average, sequence] [canonical, model, single, performed, viewpoint, joint, approach, point, structure] [proposed, method, data, codebook, general, linear, sparsity, time, number, function, sample, algorithm, decay] [motion, camera, unknown, synthetic, figure, real, requires, technique, captured] [nktm, view, virtual, mocap, ncte, dense, cvp, dvv, ixmas, large, descriptor, respective, transformation]
Finding Distractors In Images
Fried, Ohad, Shechtman, Eli, Goldman, Dan B., Finkelstein, Adam

cvpr2015papers_第165张图片
[distractor, distractors, dataset, vision, user, prediction, work, visual, mturk, mapp, datasets, app, created, mechanical, collected, turk, text, automatic, learning, retargeting, cropping, annotation, future, select, predict, green, probability, interface, photo, predictor] [table, feature, recognition, training, train, top, map, input] [saliency, average, distracting, object, auc, segmentation, salient, segment, score, calculate, threshold, based, detect, detection] [model, rgb, full, hole, described] [main, method, gaussian, number, data, inverted, lasso, random] [computer, ieee, figure, remove, removal, pattern, international, left, removing, attenuated, second, mobile, produce, pixel, quality] [image, good, person]
Co-Saliency Detection via Looking Deep and Wide
Zhang, Dingwen, Han, Junwei, Li, Chao, Wang, Jingdong

cvpr2015papers_第166张图片
[common, transfer, consistency, learning, msrc, dataset, existing, agreement, visual, set, evaluation, unsupervised, visually] [deep, performance, cnn, feature, map, explore, extracted, rbm, table, compared, global, additional, convolutional] [detection, object, saliency, group, spi, based, background, framework, foreground, salient, superpixels, ops, propose, multiple, suppress, video, segmentation, cshs, icoseg, superpixel, generated, recall, extract, contrast, three, candidate] [approach, capture, model, represent, current] [proposed, cluster, order, solve, algorithm, method, effectively, experimental, defined] [figure, ieee, color] [image, wide, similarity, region, bayesian]
Expanding Object Detector's Horizon: Incremental Learning Framework for Object Detection in Videos
Kuznetsova, Alina, Ju Hwang, Sung, Rosenhahn, Bodo, Sigal, Leonid

cvpr2015papers_第167张图片
[domain, lme, learning, embedding, adaptation, dataset, adl, unlabeled, incremental, yto, test, set, category, incrementally, unsupervised, source, existing, work, labeled, select, positive, prototype, datasets, visual, lifelong, instance, baseline, arriving, ucn, newly, adapting] [performance, class, detector, feature, imagenet, learned, training, trained, map, improve, expanded, learn, train] [object, detection, video, target, based, track, bounding, framework, appearance, temporal] [model, initial, probabilistic, approach, continuously, detected, continuous] [data, method, original, complexity, sample, proposed, expansion, problem, space, note, distribution, formulated, number, respect] [figure] [image, similarity, large, corresponding, well]
Don't Just Listen, Use Your Imagination: Leveraging Visual Common Sense for Non-Visual Tasks
Lin, Xiao, Parikh, Devi

cvpr2015papers_第168张图片
[visual, common, sense, fitb, text, human, knowledge, dataset, textual, learning, baseline, understanding, answering, task, clipart, word, description, question, mike, semantic, jenny, describe, learnt, people, describing, test, oij, default, abstract, example, answer, mentioned, vision, evaluate, attribute, language, paraphrasing, generating, study, factual, generation, lpw, set, wrxyd] [performance, learn, training, representation, recognition, input, log] [object, generated, extract, generate, truth, average, three] [scene, approach, model, reasoning, ground, capture, described] [data, solve, base, random] [figure, relative, presence, second, mode] [image, corresponding, pair, well, leveraging]
Multiple Instance Learning for Soft Bags via Top Instances
Li, Weixin, Vasconcelos, Nuno

cvpr2015papers_第169张图片
[positive, bag, mil, soft, instance, separability, learning, supervised, datasets, prediction, expected, source, labeling, set, label, dataset, concept, annotation, separator, indicator, rule, semantic, study, popular, latent, example, evaluation, tlc, weakly, human] [negative, representation, top, svm, performance, feature, training, accuracy, table, setting, larger, max, better, learn, margin] [three, multiple, based, score, combinatorial, object] [fact, composition, procedure] [number, problem, theorem, note, function, proposed, vector, consider, method, linear, optimal, space, bound, support, convex, theoretical, data, solution, programming, empirical] [figure, conventional, small, uniform, generalized] [image, weighted, best, large, separable, weighting, consistent, benchmark, considered]
Query-Adaptive Late Fusion for Image Search and Person Re-Identification
Zheng, Liang, Wang, Shengjin, Tian, Lu, He, Fei, Liu, Ziqiong, Tian, Qi


[search, dataset, learning, datasets, work, test, set, evaluation, relevant, collected, irrelevant] [feature, map, knn, global, effectiveness, table, performance, accuracy, cnn, effective, competitive, recognition, van] [score, three, weight, tail, late, object, multiple, average] [initial, performed, independent, require] [method, fusion, curve, graph, gist, parameter, rank, random, number, normalized, sparse, product, hamming, case, proposed] [figure, result, paper, robust, presented, sensitivity] [image, reference, bow, good, bad, query, area, sift, person, ukbench, database, nearest, neighbor, compare, impact, hsv, comparison, similarity, large, sorted, well, proportion, best, weighting, viper, matching, scale]
Shadow Optimization From Structured Deep Edge Detection
Shen, Li, Wee Chua, Teck, Leman, Karianto

cvpr2015papers_第170张图片
[learning, work, dataset, set, label, novel, prediction, uiuc, probability, colour] [structured, cnn, global, input, output, training, extracted, connected, learned, convolutional, learn, map, dapp, achieves, table] [edge, detection, boundary, groundtruth, framework, ucf, formulate, based, complex, propose, object, measure, moving, superpixel] [shadow, bright, ground, structure, volume, single, approach, labelling, detected, model, shd, cmu, capture, cast, neighbouring] [proposed, method, optimization, number, problem, data, algorithm] [figure, pixel, illumination, recovery, central, computer, robust] [local, image, region, patch, computed, june, compare, connectivity, canny, corresponding, spatial]
3D Deep Shape Descriptor
Fang, Yi, Xie, Jin, Dai, Guoxian, Wang, Meng, Zhu, Fan, Xu, Tiantian, Wong, Edward

cvpr2015papers_第171张图片
[learning, vision, description, set, noisy, dataset, visual, collection] [deep, neural, training, performance, feature, network, discriminative, representation, trained, learn, dnn, margin] [based, structural, target, object, three, group, entire] [shape, model, point, approach, modeling, surface] [kernel, covariance, analysis, machine, data, method, denotes, proposed, isometric, complexity, diffusion, distribution, matrix] [figure, computer, conference, noise, ieee, developed, international, component, robust, informative, result, blue, high, pattern] [descriptor, retrieval, geometric, heat, deepsd, heatsd, signature, incomplete, large, hks, comparison, encoder, heatsds, local, esd, concise, mcgill, dinosaur, query, shapegoogle, image, partial, invariant]
Fine-Grained Histopathological Image Analysis via Robust Segmentation and Large-Scale Retrieval
Zhang, Xiaofan, Su, Hai, Yang, Lin, Zhang, Shaoting

cvpr2015papers_第172张图片
[system, set, supervised, majority, relevant, analyze, category] [accuracy, table, training, performance, designed, extracted, testing, improve, compared, feature, level, design, conduct, van] [segmentation, detection, framework, based, precision, analyzing, truth, seed, segmented, propose, hierarchical, generate, including] [detected, ground, challenging] [hash, hashing, method, analysis, proposed, number, computational, ksh, repulsive, binary, gaussian, denotes, kernel, note, support, main] [robust, traditional, circle, half, accurate, ratio] [image, cell, histopathological, retrieval, scheme, medical, voting, lung, squamous, diagnosis, weighting, weighted, adenocarcinoma, mapped, cancer, cbir, comparing, breast, grading, large, spv, decided, tissue, clinical, microscopic, represents, distance]
Understanding Image Virality
Deza, Arturo, Parikh, Devi

cvpr2015papers_第173张图片
[virality, viral, content, attribute, reddit, human, dataset, understanding, visual, predict, vision, predicting, prediction, asked, study, category, title, popularity, datasets, work, set, sort, selected, annotated, task, people, correlated, highest, chance, familiar, marketing, textual, semantic, beautiful, perception, phenomenon, tend, create, amt, posted, test] [performance, accuracy, context, table, deep, svm, top, level] [score, correlation, multiple, temporal, complex, object, three, role] [social, qualitative] [random, machine, number, studied, experiment, data, explicit, time, consider] [relative, computer, figure, presence, intrinsic, conference, second, synthetically, equation] [image, pair, well, considered, difference, large]
Aligning 3D Models to RGB-D Images of Cluttered Scenes
Gupta, Saurabh, Arbelaez, Pablo, Girshick, Ross, Malik, Jitendra


[instance, set, work, search, system, report, test, semantic, task, evaluate, sofa, understanding, bed, dataset, prediction] [performance, output, table, train, setting, network, coarse, trained, hha, chair, training, top, better, shallow, observe, accuracy, cnn, song] [object, segmentation, box, detection, truth, bounding, mask, recall, plot, placement, average] [model, pose, depth, ground, scene, alignment, tagree, cad, point, inferred, normal, indoor, approach, error, reasoning, align, guo, trainval, occluded] [data, algorithm, number, problem, note, method, proposed, support, consider, missing, upper] [estimate, synthetic, figure, pixel, real, small] [image, region, metric, large]
Classifier Adaptation at Prediction Time
Royer, Amelie, Lampert, Christoph H.

cvpr2015papers_第174张图片
[adaptation, prediction, test, label, feedback, work, domain, vision, natural, rnd, system, adapting, bandit, learning, txt, correct, set, rabbit, adapt, create, semantically, semantic, example, common, visual, occur, differs, probability, unsupervised, watch] [class, training, setting, imagenet, accuracy, recognition, cnn, convolutional, feature, achieve, context, table] [online, object, based, dynamic, situation, walk, sequence, adaptive, three, framework] [error, fact, realistic, probabilistic, current] [data, distribution, time, random, machine, base, underlying, method, update, problem, vector, optimal] [computer, conference, estimate, figure, real, assume, international, ieee, length] [image, change, difference, large, compute, well]
Three Viewpoints Toward Exemplar SVM
Kobayashi, Takumi

cvpr2015papers_第175张图片
[positive, category, novel, vision, dataset, set, learning, irrelevant] [exemplar, svm, feature, negative, performance, margin, recognition, table, training, discriminative, smaller, hog, exsvm, apply, ensemble, larger, input, unit, svms] [object, target, detection, based, three, score, measure, multiple] [detected, scene, viewpoint, described] [sample, vector, method, proposed, original, number, formulation, parameter, support, regularization, linear, min, problem, case, note, kernel, computation, required, pass, matrix, manner, corresponds, simply, reduced, form] [computer, conference, pattern, ieee, second, lagrangian, small, international, figure] [similarity, transformation, applied, image, large, transform, distance, well, transformed]
Curriculum Learning of Multiple Tasks
Pentina, Anastasia, Sharmanska, Viktoriia, Lampert, Christoph H.

cvpr2015papers_第176张图片
[learning, task, seqmt, transfer, vision, human, learner, mergedsvm, study, knowledge, baseline, choosing, test, easiest, attribute, awa, seal, adaptation, automatically, question, multiseqmt, hardest, chimpanzee, work, hyperplane, side, domain, hard, giant, european, expected, learing, diversity, semantic, advantageous] [performance, training, learn, class, better, table, generalization, effective, previous, svm, sharing, feature, achieve] [multiple, weight, average, based, sequence, adaptive, propose, object, performs] [error, model, approach, term, hand, single] [order, algorithm, machine, data, proposed, solving, random, bound, linear, method, solved, solve, case, idea, form] [conference, computer, international, figure, high, assume] [best, corresponding, rate, compare, image]
Learning With Dataset Bias in Latent Subcategory Models
Stamos, Dimitris, Martelli, Samuele, Nabi, Moin, McDonald, Andrew, Murino, Vittorio, Pontil, Massimiliano

cvpr2015papers_第177张图片
[datasets, dataset, visual, latent, learning, set, positive, test, car, task, work, common, perform, provided] [training, lsm, subcategory, lsms, undoing, standard, improvement, trained, lcsm, concatenation, negative, performance, learn, class, loss, setting, multitask, compound, better, dpms, svms, smaller, table, max, testing, previous] [object, weight, average, multiple, framework, detection, based, three] [model, initialization, single, point, term, approach, error, independent, step] [method, function, objective, case, vector, convex, number, problem, linear, clustering, parameter, data, note, solution, machine, simply, cluster, suboptimal, special, optimization, journal] [bias, biased, figure, computer, presented, relative, paper, extend] [good, well, corresponding, large, local]
VisKE: Visual Knowledge Extraction and Question Answering by Visual Verification of Relation Phrases
Sadeghi, Fereshteh, Kumar Divvala, Santosh K., Farhadi, Ali

cvpr2015papers_第178张图片
[relation, visual, knowledge, eating, horse, dog, man, bear, language, phrase, lion, laying, egg, sofa, river, set, icecream, pushing, cat, viske, conceptnet, subject, devouring, bird, relationship, work, text, learning, answering, verifying, salmon, web, question, openie, svo, rabbit, woman, elephant, sheep, involved, snake, system, common, consistency, dataset, semantic, sun, focused, frog, validity] [feature, performance, trained, extracting, grass, context, table, better] [object, appearance, extraction, box, detection, based, precision, action, indicates, bounding] [approach, model, reasoning] [problem, number, method, form, graph, base, case] [figure, pattern, relative, reveals] [spatial, true, large, person, well, difference]
Discovering States and Transformations in Image Collections
Isola, Phillip, Lim, Joseph J., Adelson, Edward H.

cvpr2015papers_第179张图片
[collection, set, noun, relevant, discovering, adjective, vision, test, sliced, visual, system, cooked, fresh, apple, dataset, semantic, generalize, work, melted, unripe, goal, antonym, knowledge, discovery, novel, ripe, understanding, example, transfer, physical, variety, include, selected, door, undergo, wilted, discovered, ranking, moldy, task, chance, discover, scraped, antonymic, chocolate, attribute] [training, performance, raw, recognition, parsing, feature, neural, map, accuracy, class, crf, input, percent] [state, object] [room, scene, fall, depicted] [method, order, problem, data, consider, space, journal] [computer, figure, ieee, conference, pattern, international, material, equation, paper] [image, transformation, similarity, well, query, dissimilar, transformed, large]
Parsing Occluded People by Flexible Compositions
Chen, Xianjie, Yuille, Alan L.

cvpr2015papers_第180张图片
[vision, people, set, human, dataset, type, work, labeled, diagnostic, learning] [recognition, connected, standard, training, parsing, performance, convolutional, learn, max, neural] [score, evidence, object, appearance, wij, detection, node, state, multiple] [model, occlusion, tij, decoupling, visible, full, graphical, pose, body, composition, subtree, dependent, occluded, inference, approach, waf, idod, aop, pcp, absence, tji, prior, location, term, neighboring, modeling, estimation, wji, idpr, mixture, mgn, single] [graph, pairwise, base, number, method, suitable, space, lemma, form] [computer, pattern, figure, bias, relative, conference, presence, international, equation, measurement] [image, spatial, local, best, connectivity, computed]
Modeling Video Evolution for Action Recognition
Fernando, Basura, Gavves, Efstratios, Oramas, Jose M., Ghodrati, Amir, Tuytelaars, Tinne

cvpr2015papers_第181张图片
[ranking, learning, dataset, human, work] [representation, recognition, learn, fisher, training, table, feature, better, map, discriminative, pooling, generalization, improved, observe, applying, capturing, size, class, apply] [video, videodarwin, action, temporal, evolution, appearance, average, frame, based, moving, cooking, forward, reverse, gesture, fdvd, mpii, rfdvd, mbh, propose, score, trajectory, combined, precision] [capture, independent, ordering, modeling, model, approach] [time, function, vector, method, linear, valued, rank, data, order, proposed, kernel, machine, denoted] [figure, varying, motion, stability, relative] [local, compute, well, large, dense, similarity, applied]
Small Instance Detection by Integer Programming on Object Density Maps
Ma, Zheng, Yu, Lei, Chan, Antoni B.

cvpr2015papers_第182张图片
[set, vision, test, people, dataset, instance, baseline, datasets] [map, training, sliding, window, global, feature, input, size, ucsd, performance, table, better, detector, train, output, learned] [density, detection, object, counting, bounding, count, box, integer, crowded, framework, individual, truth, average, crowd, video, segmentation, recall, propose, divided] [ground, location, approach, challenging, model, viewpoint] [number, method, programming, proposed, vector, constraint, problem, function, maximum, mapping, codebook] [small, estimated, computer, estimate, figure, ieee, pattern, total, roi, perspective, pixel, accurate, international, synthetic] [image, local, cell, large, distance, corresponding, matching]
Weakly Supervised Localization of Novel Objects Using Appearance Transfer
Rochan, Mrigank, Wang, Yang

cvpr2015papers_第183张图片
[novel, familiar, learning, word, weakly, labeled, vision, work, set, transfer, knowledge, supervised, semantic, collection, annotated, dataset, goal, car, visual, cat, bird, horse, dog, cow, embedding, transferring, collect, boat, european, processing, datasets, access, subset] [pascal, class, table, voc, learned, feature, train, imagenet, training, learn, apply] [object, appearance, bounding, box, localize, video, propose, combined, based, corloc, edge, combination, proposal] [model, approach, term, localization, reconstruction] [method, vector, data, linear, denote, algorithm, parameter, number, proposed, consider, machine] [computer, conference, ieee, pattern, figure, international, society, assume] [image, large, compare]
Data-Driven 3D Voxel Patterns for Object Category Recognition
Xiang, Yu, Choi, Wongun, Lin, Yuanqing, Savarese, Silvio

cvpr2015papers_第184张图片
[dataset, test, hard, set, car, associated, evaluation, category, evaluate, visual] [recognition, training, dpm, table, train, validation, trained, detector, accuracy, class, testing, compared, jointly, feature, build, acf] [object, detection, segmentation, mask, bounding, average, truth, appearance, moderate, overlap, based, box, detect, occluder, three] [voxel, occlusion, kitti, occluded, model, cad, localization, pose, visibility, orientation, visible, reasoning, ground, approach, depth, ordering, voxels, detected, shape, estimation, outdoorscene, truncation, cuboid] [method, truncated, clustering, easy, data, explicit, number, note, proposed] [figure, estimating, handle] [image, benchmark, similarity, well, large]
A Novel Locally Linear KNN Model for Visual Recognition
Liu, Qingfeng, Liu, Chengjun

cvpr2015papers_第185张图片
[test, dataset, visual, learning, popular, vision, rule, selected, relation, novel, evaluate] [representation, training, recognition, power, class, knn, performance, table, achieve, discriminative, effectiveness, decision, fisher, improve, caltech, global, highly, max, size, accuracy, comparable] [based, density, action, object, grouping, bayes, propose] [model, face, scene, indoor, estimation, error] [proposed, sample, method, linear, vector, llknnc, shifted, llnmc, kernel, parameter, sparse, dimension, experimental, sensitiveness, data, llknn, note, arg, minimum, analysis, number, approximate, largest, sparsity, journal] [pattern, ieee, computer, conference, equation, international] [nearest, locally, dictionary, transformation, image, applied, ideal, tar, derived, demonstrate, spatial]
Classifier Learning With Hidden Information
Wang, Ziheng, Ji, Qiang


[learning, secondary, versus, primary, exploit, lir, existing, relationship, privileged, humpback, cat, side, seal, rat, giant, rpr, baseline, persian, chimpanzee, hippopotamus, panda, set, label, leopard, raccoon, whale, task, predicted, test, statistically, type, predict, student, human] [hidden, training, loss, feature, performance, better, improve, classify, learn, improvement, standard, margin, combine, svm, testing, input, recognition, train] [action, object, target, strong, based, truth] [facial, term, approach, face, ground, expression] [proposed, regularization, method, data, inequality, problem, parameter, function, binary, original, empirical, number, case, represented] [preserving, paper, pig, assumption, generally, figure, total] [image, compare, incomplete]
Best of Both Worlds: Human-Machine Collaboration for Object Annotation
Russakovsky, Olga, Li, Li-Jia, Fei-Fei, Li

cvpr2015papers_第186张图片
[human, vision, annotation, system, user, set, labeling, probability, requester, utility, correct, budget, visual, mdp, question, draw, instance, tight, answer, task, expected, feedback, learning, active, crowdsourcing, labeled, label, correspond, dataset, european, annotated, predicting, engineering, returned] [recognition, class, accuracy, input, table, response, training, effective, decision] [object, box, detection, bounding, precision, action, multiple, transition, average, iou, cost, framework, principled, state, crowd] [model, current, scene, error] [time, data, number, method, machine, analysis] [computer, pattern, figure, conference, international, quality, yellow, accurately, desired, ieee, accurate] [image, compute, computing, computed]
Learning Graph Structure for Multi-Label Image Classification via Clique Generation
Tan, Mingkui, Shi, Qinfeng, van den Hengel, Anton, Shen, Chunhua, Gao, Junbin, Hu, Fuyuan, Zhang, Zhen

cvpr2015papers_第187张图片
[label, set, learning, relevant, active, prediction, knowledge, identify, conditional] [performance, loss, max, feature, table, neural, output, learned, learn, structured, cnn, stochastic, pascal, better, map, input] [propose, based] [scene, structure, tree, model, capture, joint, approach, graphical, inference, probabilistic] [data, graph, problem, cgm, method, solving, hamming, clique, arg, number, dual, random, proposed, objective, chowliu, algorithm, plem, ssvm, vector, multilabel, min, convex, plst, solve, sparse, mmoc, lead, update, solution, binary, mapping, denotes, programming, note, proximal, exploiting, coding, basic, optimization, sparsity, constraint] [figure, synthetic, ieee, adopt] [image, comparison]
Watch and Learn: Semi-Supervised Learning for Object Detectors From Video
Misra, Ishan, Shrivastava, Abhinav, Hebert, Martial

cvpr2015papers_第188张图片
[labeled, set, labeling, learning, label, purity, dataset, test, automatically, ssl, work, diverse, datasets, esvms, annotated, esvm, semantic, subset, decorrelated, unlabeled, positive, visual, weakly, relevance, boot, existing] [training, input, feature, trained, performance, negative, train, detector, eigen] [object, detection, tracking, video, bounding, multiple, box, long, recall, average, reliable, virat, start, truth, track, frame] [approach, kitti, ground, pose, initial, drift] [data, algorithm, iteration, method, selection, sparse, iteratively, lsvm, random, outlier, number] [figure, high, assume, motion] [image, large, good, sparsely, false, scale, compute]
Virtual View Networks for Object Reconstruction
Carreira, Joao, Kar, Abhishek, Tulsiani, Shubham, Malik, Jitendra

cvpr2015papers_第189张图片
[test, vision, collection, set, visual, work, prediction, semantic] [network, training, recognition, class, deep, feature, pascal, better, additional, convolutional, neural, voc] [object, target, multiple, segmentation, based, propose, truth, detection] [reconstruction, shape, viewpoint, alignment, pose, single, structure, rigid, sfm, ground, point, align, extrapolated, mirrored, reconstruct] [factorization, euclidean, data, analysis, missing, proposed, machine, matrix, algorithm, method] [computer, ieee, pattern, conference, synthetic, figure, motion, robust, accurate, international, assume, inliers, assuming, reconstructed, observation] [image, matching, nearest, virtual, distance, neighbor, keypoints, grid, view, docking, dense, vvn, compute, computed, sift, large]
Efficient Label Collection for Unlabeled Image Datasets
Wigness, Maggie, Draper, Bruce A., Ross Beveridge, J.

cvpr2015papers_第190张图片
[labeling, label, effort, visual, hcgl, concept, active, labeled, learning, subset, set, human, supervised, evaluation, mkml, unlabeled, collected, process, collecting, latency, dataset, existing, system, vision, annotator, selected, collection, majority, msrc, discovery] [training, performance, accuracy, class, fully, train, feature, higher] [group, hierarchical, structural, multiple, object, grouping, path, three] [represent, single, ordering, scene, approach, provide, structure, introduced, autonomous] [data, clustering, largest, iteratively, number, selection, cluster, order, binary] [figure, computer, noise, pattern, conference, technique, international, length, high, result] [image, change, large, dominating, benchmark, local, comparison, guided, match, performing]
Learning to Segment Under Various Forms of Weak Supervision
Xu, Jia, Schwing, Alexander G., Urtasun, Raquel

cvpr2015papers_第191张图片
[semantic, weak, set, learning, supervised, labeled, weakly, label, tag, test, work, refer, dataset, unlabeled, evaluate, interactive, report, evaluation, incorporating, ilt, task] [supervision, accuracy, class, training, setting, feature, performance, fully, level, parsing, standard, loss, size, smaller] [bounding, segmentation, box, object, assigned, appearance, foreground, inferring, truth] [inference, full, scene, approach, model, challenging, ground, single, program, fact] [matrix, note, form, data, linear, number, vector, sample, assignment, constraint, algorithm, geodesic, case, optimization, binary, random, method, optimal, respect] [ratio, second, pixel, figure] [image, partial, large, well]
Towards Open World Recognition
Bendale, Abhijit, Boult, Terrance

cvpr2015papers_第192张图片
[open, set, learning, incremental, system, ncm, nno, vision, risk, visual, work, evaluation, incrementally, probability, labeled, closed, novel, process, manage, add, dataset, measurable, protocol, label, existing, novelty] [recognition, training, class, testing, performance, feature, svm, imagenet, accuracy, additional, learned, standard, learn, neural] [object, based, detection, multiple] [model, provide, approach, cap] [space, function, number, algorithm, machine, vector, data, scalable, proposed, problem, support, update, time, consider, solution, thresholding, analysis, theorem] [unknown, computer, pattern, ieee, conference, paper, robust, handle] [nearest, metric, scale, large, image, distance, transformed]
Fast 2D Border Ownership Assignment
Teo, Ching, Fermuller, Cornelia, Aloimonos, Yiannis

cvpr2015papers_第193张图片
[prediction, dataset, vision, set, visual, example, predict, natural, predicted] [structured, contour, training, top, feature, input, accuracy, compared, decision, table, previous, learn] [ownership, border, boundary, extremal, bsds, edge, token, foreground, object, srf, grouping, detection, segmentation, target, gestalt, background, oriented, key, cue, detect] [depth, approach, indoor, orientation, ground, torque, step, closure, shape, structure, split, tree, capture, occlusion] [spectral, basis, random, assignment, principal, note, data, number, tangent, order, proposed] [computer, spiral, figure, simple] [image, local, patch, derived, gradient, corresponding, pca]
Joint Inference of Groups, Events and Human Roles in Aerial Videos
Shu, Tianmin, Xie, Dan, Rothrock, Brandon, Todorovic, Sinisa, Chun Zhu, Song

cvpr2015papers_第194张图片
[human, people, latent, work, dataset, learning, baseline, noisy, existing, set, label, source] [recognition, log, parsing, training, extracted, level, representation, accuracy, jointly] [event, group, tracking, role, video, spatiotemporal, foreground, object, temporal, box, exchange, based, recognizing, trajectory, bounding, activity, dynamic, subtraction, assigned, three, background, grouping, info, assign, viewed, detection, tracklets] [inference, template, social, scene, prior, model, approach, joint, infer, represent] [assignment, time, graph, method, optimal, pairwise] [small, result, ieee, figure, camera] [aerial, large, spatial, illustrated, best, corresponding]
Salient Object Detection via Bootstrap Learning
Tong, Na, Lu, Huchuan, Ruan, Xiang, Yang, Ming-Hsuan

cvpr2015papers_第195张图片
[weak, learning, dataset, set, visual, svo, test, labeled, perform] [map, channel, training, performance, input, feature, boosting, learned] [saliency, salient, detection, strong, based, bootstrap, object, dark, recall, precision, wco, asd, multiple, background, dsr, generate, gmr, cbsal, foreground, lrmr, superpixels, generated, truth, detect, sod, propose, measure, average, threshold, weight] [model, ground, prior, approach] [proposed, method, algorithm, graph, number, binary, note, kernel, cut, constructed] [figure, pixel, color] [image, region, computed, construct, compute, local, well]
Joint Action Recognition and Pose Estimation From Video
Xiaohan Nie, Bruce, Xiong, Caiming, Zhu, Song-Chun

cvpr2015papers_第196张图片
[human, learning, dataset, label, example, relation, set] [recognition, feature, training, accuracy, train, trained, performance, learn, root, representation, discriminative, table] [action, score, video, appearance, detection, three, temporal, penn, frame, truth, terminal, node, framework, based, transition, decomposed, aog, interest, hierarchical] [pose, estimation, model, represent, single, inference, ground, deformation, mixture, adjacent, smoothness, graphical, capture, joint, occlusion] [method, number, represented, graph, data] [motion, component, small, figure, directly] [image, large, best, geometric, distance, dense]
HC-Search for Structured Prediction in Computer Vision
Lam, Michael, Rao Doppa, Janardhan, Todorovic, Sinisa, Dietterich, Thomas G.

cvpr2015papers_第197张图片
[search, learning, label, vision, work, labeling, prediction, semantic, dataset, set, ranking, natural] [performance, training, learned, stanford, loss, structured, previous, connected, learn, accuracy, table, ybest, output, larger, apply, feature, context, input] [heuristic, segmentation, state, cost, dagger, object, background, candidate, nematocyst, berkeley, successor, detection, imitation, superpixel, edge, ucm, truth, generate, superpixels, score, generated, flipbit, threshold, branching, hierarchical] [depth, scene, error, initial, approach, inference, step, single, current, ground, estimation] [function, space, time, bound, number, greedy, exact, upper, algorithm, method, machine, energy, journal, oracle] [computer, figure, robust, relative] [best, image, randomized, good, true]
First-Person Pose Recognition Using Egocentric Workspaces
Rogez, Gregory, Supancic, III, James S., Ramanan, Deva

cvpr2015papers_第198张图片
[set, work, add, human, noisy] [feature, training, map, recognition, svm, global, size, performance, previous, additional, implement, classifying] [egocentric, generate, object, detection, tracking, correctly, weight, score, background, plot] [depth, hand, pose, voxels, volumetric, quantized, workspace, volume, model, joint, estimation, binarized, front, cloud, voxel, approach, coordinate, single, manipulating, visible, interacting, capture, manipulated, fall, body, allows] [data, binary, sparse, linear, tensor, note, vector] [camera, figure, synthetic, perspective, simple, spherical, computer, pixel, varying, orthographic, conference, international, produce, ieee] [image, arm, local, corresponding, grid, outperforms, cvpr, virtual, large, compare, vary]
MUlti-Store Tracker (MUSTer): A Cognitive Psychology Inspired Approach to Object Tracking
Hong, Zhibin, Chen, Zhe, Wang, Chaohui, Mei, Xue, Prokhorov, Danil, Tao, Dacheng

cvpr2015papers_第199张图片
[set, visual, processing, store, learning, evaluation, active, process] [output, feature, performance, input, neural, trained, top, response] [memory, tracking, target, based, object, tracker, muster, appearance, correlation, long, icf, threshold, background, precision, updated, kcf, forgetting, success, retention, integrated, online, employ, survival, plot, meem, strength, employed, struck, tracked, shortterm, bounding, adaptive, ootb, frame] [model, estimation, occlusion, control, detected] [proposed, ransac, number, curve, consider, original, conducted, kernel] [robust, component, figure, relative, result, inside, maintain, second] [keypoints, keypoint, matching, image, database, matched, scale, consists, large, nearest, local, best, corresponding, stable, benchmark, filter]
Symmetry-Based Text Line Detection in Natural Scenes
Zhang, Zheng, Shen, Wei, Yao, Cong, Bai, Xiang

cvpr2015papers_第200张图片
[text, icdar, natural, character, dataset, swt, positive, symmetrical, evaluation, reading, probability, work, novel, neumann, datasets] [feature, training, recognition, level, representation, performance, testing, standard, connected, higher, deep, advantage, achieves] [detection, bounding, based, recall, appearance, stroke, multiple, extraction, truth, detecting, precision, middle, group] [scene, challenging, ground, template, height, localization] [proposed, algorithm, low, method, number] [symmetry, rectangle, property, axis, ieee, figure, robust, conventional, directly, handle, color, removal, component, pixel, computer] [image, false, mser, region, local, compute, texture, scale, difference]
Understanding Image Structure via Hierarchical Shape Parsing
Liu, Xian-Ming, Ji, Rongrong, Wang, Changhu, Liu, Wei, Zhong, Bineng, Huang, Thomas S.

cvpr2015papers_第201张图片
[visual, vision, human, search, learning, collection, perform, exploring, unsupervised, understanding, example, probability, parse] [parsing, feature, pooling, representation, level, structured, selective, table, performance, recognition, larger, compared, top, improve, pyramid, exemplar, convolutional, better, build, inspired, neural] [hierarchical, edge, object, structural, detection, objectness, appearance, based, node, recall, parsed, complex, superpixels, cognitive, indicates, performs, saliency] [tree, structure, shape, scene, building, approach, inference] [proposed, method, number, algorithm, space, analysis, markov, machine, journal] [computer, figure, ieee, pattern, conference, international, paper, component] [image, scale, matching, region, spatial, local, well, corresponding]
Object Proposal by Multi-Branch Hierarchical Segmentation
Wang, Chaoyang, Zhao, Long, Liang, Shuang, Zhang, Liqing, Jia, Jinyuan, Wei, Yichen

cvpr2015papers_第202张图片
[learning, search, process, set, strategy, work, evaluate, automatic, positive, test] [stage, performance, training, train, loss, pascal, top, achieves, effective, generic, validation, negative, map, increase, previous, compared, setting] [object, merging, proposal, segmentation, recall, iou, hierarchical, branch, detection, complementary, branching, gop, threshold, bing, gls, multiple, merged, score, propose, segment, auc, box, truth, start, bounding, weight, svalid, based, strain] [initial, approach, starting, ground, neighboring, composition, structure] [greedy, number, method, algorithm, linear, order, curve, machine, university, journal] [degree, figure, small, computer] [image, rate, area, large, pair, best, comparison]
Discriminative Shape From Shading in Uncalibrated Illumination
Richter, Stephan R., Roth, Stefan

cvpr2015papers_第203张图片
[learning, dataset, test, set, prediction, natural, work, address, evaluate, datasets, aim] [training, map, input, discriminative, mit, performance, contour, output, feature, context, trained] [object, strong, correlation] [surface, shape, silhouette, normal, approach, regression, single, estimation, model, point, realistic, allows, modeling, evaluated, environment, drel, forest, medial, occluding] [data, method, gaussian, time, directional, convex, consider, distribution] [illumination, shading, synthetic, lighting, integrability, estimate, intrinsic, component, color, direction, real, unknown, estimated, estimating, white, relative, second, pixel, texton, uniform, figure, barron, photometric, sphere, light] [image, spatial, distance, local, well, scale]
High-Fidelity Pose and Expression Normalization for Face Recognition in the Wild
Zhu, Xiangyu, Lei, Zhen, Yan, Junjie, Yi, Dong, Li, Stan Z.

cvpr2015papers_第204张图片
[vision, learning, trend, type, evaluate] [recognition, performance, table, identity, accuracy, feature, frontal, deep, contour, level, discriminative] [propose, boundary, based, appearance, adaptive, surrounding] [face, pose, expression, facial, landmark, hpen, model, invisible, shape, morphable, marching, fitting, neutral, depth, robustness, adjusted, yaw, step, editing, visibility] [method, normalized, original, data, analysis, proposed, machine, linear] [ieee, computer, pattern, conference, detail, international, figure, assumption, estimated, poisson, illumination, robust, caused, preserving, estimate] [image, normalization, region, database, unconstrained, large, lfw, corresponding, anchor, ddml, demonstrate, pca, correspondence]
Effective Learning-Based Illuminant Estimation Using Simple Features
Cheng, Dongliang, Price, Brian, Cohen, Scott, Brown, Michael S.


[set, work, evaluation, natural, subset] [training, performance, table, feature, input, ensemble, better, output, trained, van, reported] [based, truth, average, including, three, group, framework] [regression, tree, estimation, rgb, error, scene, ground, approach, surface, evaluated, model] [data, method, proposed, number, time, spectral, space] [color, chromaticity, illuminant, illumination, constancy, camera, gamut, figure, simple, median, josa, estimated, mode, angular, result, laboratory, palette, sfu, corrected, paper, learningbased] [image, best, local, statistical, comparison, bayesian, compare, fast, good]
Image Specificity
Jas, Mainak, Parikh, Devi

cvpr2015papers_第205张图片
[sentence, baseline, vision, description, visual, people, sim, predicted, predict, dataset, human, textual, automated, search, study, word, content, set, describing, describe, association, category, evaluation, specificity, automatically, beat, automatic, tend, learning, predicting, associated, memorability, generating, correlated, semantic] [recognition, training, train, better, neural] [target, correlation, multiple, average, object, score, measure, based] [approach, model, scene, described] [rank, number, note, computational, low] [conference, computer, ieee, figure, pattern, international, high] [image, similarity, query, reference, auto, computed, database, retrieval, variance, consistent]
Space-Time Tree Ensemble for Action Recognition
Ma, Shugao, Sigal, Leonid, Sclaroff, Stan

cvpr2015papers_第206张图片
[set, human, word, discover, discovered, discovering, mining, mined, latent, appear, search, positive] [discriminative, performance, training, learned, recognition, class, smaller, larger, root, ensemble, frequent, pooling, representation, table, testing, svm] [action, video, highfive, temporal, stss, hierarchical, node, average, edge, sts, eij, movement, segmentation, hug, box, bounding, combination, propose, kiss, recall, wang] [tree, body, model, inference, approach, structure, procedure, single, vertex] [graph, compact, number, method, clustering, exact, note, space, pairwise, cluster, fusion] [small, figure, simple] [spatial, matching, best, partial, large, pair, vocabulary, matched, computed, match]
Fast and Accurate Image Upscaling With Super-Resolution Forests
Schulter, Samuel, Leistner, Christian, Bischof, Horst

cvpr2015papers_第207张图片
[learning, set, work, task, evaluate, domain, vision, include, evaluation, prediction] [training, learn, input, output, standard, feature, factor, table, size, better, investigate] [node, framework, detection, average, based] [single, regression, model, leaf, forest, inference, approach, tree, typically, sisr, depth, operates, yang] [random, linear, data, function, problem, time, objective, sparse, parameter, mapping, space, number, regularization, coding, basis, sample, encoding, proposed, method] [figure, quality, superresolution, direct, constant, computer, simple, upsampling, directly] [image, dictionary, upscaling, splitting, psnr, fast, well, rfl, neighborhood, anr, coupled, compare, locally, srcnn, comparison, bicubic, anchored]
Label Consistent Quadratic Surrogate Model for Visual Saliency Prediction
Luo, Yan, Wong, Yongkang, Zhao, Qi

cvpr2015papers_第208张图片
[learning, visual, dataset, human, knowledge, set, prediction, label, vision, datasets, existing, predict, evaluation, process, natural, processing, leverage, sauc, osie] [training, width, feature, mit, learn, learned, trained, discriminative, map, better, neural, performance] [saliency, online, based, eye, object, itti, attention, tracking, gbvs] [model, prior, scene, ground] [sparse, proposed, algorithm, machine, quadratic, surrogate, lcqs, update, min, convergence, gaussian, arg, analysis, method, problem, sample, journal, function, number, regularization, data, matrix, basis, solution] [ieee, blur, computer, pattern, conference, bias, center, international, conventional, figure] [dictionary, image, consistent, cross, compute, large, computed, leveraging, corresponding, scale]
Early Burst Detection for Memory-Efficient Image Retrieval
Shi, Miaojing, Avrithis, Yannis, Jegou, Herve

cvpr2015papers_第209张图片
[visual, search, set, evaluate, dataset, strategy, representative, concept, focus, evaluation] [feature, size, performance, connected, factor, map, table, representation, larger, gain, conduct] [group, detection, three, object, cost, memory, detect, hierarchical] [initial, model, orientation, approach, position] [kernel, number, spectral, clustering, asymmetric, method, symmetric, function, normalized, product, inverted, matrix, quantization, proposed, hamming, vector] [figure, shift, component] [image, burst, scale, descriptor, large, local, retrieval, query, vlad, vocabulary, similarity, isolated, matching, bursty, burstiness, imbalance, early, oxford, patch, remaining, aggregation, aggregated, quick, match, geometrical, database, compute]
Multi-View Feature Engineering and Learning
Dong, Jingming, Karianakis, Nikolaos, Davis, Damek, Hernandez, Joshua, Balzer, Jonathan, Soatto, Stefano

cvpr2015papers_第210张图片
[vision, test, learning, probability, evaluation, instance, predicted] [training, performance, feature, representation, better, hog, detector, accuracy] [multiple, object, density, group] [scene, reconstruction, point, orientation, surface, single, ground, viewpoint, model, step, joint, shape] [approximation, function, sample, number, distribution, machine, random, complexity, respect, analysis, dependency, approximate, time] [computer, ieee, pattern, conference, excitation, planar, real, requires, synthetic, parallel, translation, intrinsic, plane, minimal] [image, descriptor, spatial, ideal, local, nuisance, computed, daisy, dense, surf, variability, gradient, correspondence, compute, marginalized, compare, marginalization, distance, invariant, sampling, aggregation, scale, area, histogram]
BOLD - Binary Online Learned Descriptor For Efficient Image Matching
Balntas, Vassileios, Tang, Lilian, Mikolajczyk, Krystian

cvpr2015papers_第211张图片
[set, positive, learning, correct, subset, dataset, test] [global, performance, feature, discriminative, table, class, learned, negative, better, training, intra, compared] [online, based, extraction, tracking, object, generated, generate] [error, approach] [binary, optimized, proposed, optimization, hamming, number, selection, random, method, minimize, computational, low, dimensionality, matrix, distribution] [figure, high, simple, robust, intensity, ieee] [descriptor, patch, globally, matching, rate, sift, variance, bold, distance, binboost, image, adapted, surf, large, local, locally, best, orb, false, discriminant, yosemite, demonstrate, dbrief, stable, implementation, masked, true]
Robust Multi-Image Based Blind Face Hallucination
Jin, Yonggang, Bouganis, Christos-Savvas

cvpr2015papers_第212张图片
[work, latent, predicted, set, example, association, learning] [performance, log, previous, generic, max, table, map, training, better, feature, size, improve, global] [multiple, combination, based, viewed, combined, framework, tracking, background] [face, prior, estimation, posterior, model, facial, mixture, single, holistic, approach] [kernel, proposed, covariance, gaussian, method, updating, bound, lower, distribution, note, arg, subspace, number, constraint, optimizing, normalized, special] [estimated, ieee, robust, blind, simulated, real, estimate, paper, figure, observation, observed, deblurring, assumed] [blurring, pca, image, mppca, transformation, hallucination, registration, initialized, database, const, comparison, local, patch, hmrf, est, represents, psnr, applied]
Face Alignment by Coarse-to-Fine Shape Searching
Zhu, Shizhan, Li, Cheng, Change Loy, Chen, Tang, Xiaoou

cvpr2015papers_第213张图片
[set, dataset, test, search, subset, learning] [feature, training, stage, accuracy, learn, zhu, performance, deep, hybrid, coarse, setting] [framework, based, candidate, weight, appearance, detection, including] [shape, searching, cascaded, regression, face, error, alignment, facial, initialisation, probabilistic, approach, initial, pose, challenging, averaged, xij, model, sdm, regressed, landmark, dominant, helen, lbf, cfan, regressors, rcpr, drmf, lfpw, ced] [method, proposed, distribution, space, random, practical, algorithm, time, number, problem, vector, sample] [estimate, robust, center, conventional, speed, accurate, estimated, estimating] [local, sift, comparison, large, early, worth, outperforms, best, subsequent, region]
Interaction Part Mining: A Mid-Level Approach for Fine-Grained Action Recognition
Zhou, Yang, Ni, Bingbing, Hong, Richang, Wang, Meng, Tian, Qi

cvpr2015papers_第214张图片
[human, mining, set, dataset, semantic, novel] [performance, discriminative, recognition, representation, pooling, feature, detector, apply, max, better, achieves, training, top, fisher, extracted, accuracy, achieve, table, improved] [interaction, action, object, video, activity, daily, mpii, detection, candidate, cooking, msr, propose, based, appearance, tak, weight, proposal, segmentation, utilize, generate, frame, score, trajectory, edge, background, skeleton, localized, foreground, seeding, fro] [approach, model, depth, free, represent, formation, holistic] [method, graph, proposed, explicit, number, form, algorithm, vector, sample, parameter] [motion, figure] [dense, large, computed, region, image, compare, matching, consistent]
Traditional Saliency Reloaded: A Good Old Model in New Shape
Frintrop, Simone, Werner, Thomas, Martin Garcia, German

cvpr2015papers_第215张图片
[system, visual, human, evaluation, vision] [feature, map, pyramid, performance, layer, smoothing, global, designed] [saliency, object, salient, contrast, invt, based, detection, surround, attention, twin, measure, segmentation, eye, proposal, itti, three, turned, precise, average, integration, arithmetic, suited, recall] [approach, location, model, structure, prior, current] [method, computation, space, corresponds, fusion, original, data, basic, computational] [color, center, intensity, theory, mobile, figure, simple, elegant, separation, ratio, high] [image, compute, well, large, good, region, outperforms, local, scale, benchmark, weighting, weighted]
Semantics-Preserving Hashing for Cross-View Retrieval
Lin, Zijia, Ding, Guiguang, Hu, Mingqing, Wang, Jianmin

cvpr2015papers_第216张图片
[set, semantic, learnt, supervised, learning, text, probability, predicted, unseen, instance, provided, vision, novel, work, wiki, dataset] [training, performance, size, learn, feature, output, learns, recognition, better, previous, effective] [propose, utilize] [regression, model, probabilistic] [hash, seph, hashing, code, kernel, data, hamming, proposed, logistic, matrix, formula, vector, optimal, binary, distribution, function, objective, minimizing, space, note, bit, mirflickr, largest, cmfh, transforming, derive, approximates] [conference, observed, international, quality, computer, acm, length, ieee, pattern] [query, corresponding, retrieval, sampling, image, view, well, database, preserve, similarity, benchmark, derived, calculated]
SUN RGB-D: A RGB-D Scene Understanding Benchmark Suite
Song, Shuran, Lichtenberg, Samuel P., Xiao, Jianxiong

cvpr2015papers_第217张图片
[evaluate, dataset, understanding, semantic, evaluation, prediction, annotation, task, popular, set, learning, sun, human, study, label, example, tool] [table, recognition, map, training, sliding, raw, performance, svm, effective, size, deep, pascal, context, power, testing] [object, detection, truth, box, iou, segmentation, bounding, based, average, frame, three] [depth, scene, room, kinect, ground, layout, estimation, indoor, point, rgbd, free, xtion, nyu, realsense, orientation, intel, rgb, asus, capture] [space, algorithm, data, convex, number] [figure, color, sensor, total, estimate, camera, angle, computer, captured, inside, quality, accurate, manhattan, result] [image, benchmark, geometric, difference, database, scale]
A Maximum Entropy Feature Descriptor for Age Invariant Face Recognition
Gong, Dihong, Li, Zhifeng, Tao, Dacheng, Liu, Jianzhuang, Li, Xuelong

cvpr2015papers_第218张图片
[set, entropy, probability, dataset, existing, vision] [feature, recognition, age, training, morph, discriminative, identity, table, decision, performance, factor, album, extracted, log, hog, representation, improve, cosine, compared] [framework, based, node] [face, approach, tree, facial, leaf, model, expressive, ifa, estimation, orientation] [vector, maximum, binary, proposed, code, random, analysis, sample, encoding, method, space, number, largest] [pattern, ieee, figure, component, pixel, uniform, frequency, computer] [descriptor, matching, aging, image, local, lbp, sift, gallery, invariant, probe, mefd, scheme, fgnet, aifr, called, lfw, database, tot, sampling, illustrated, splitting, generalizability, representing, distance, histogram, corresponding]
Separating Objects and Clutter in Indoor Scenes
Khan, Salman H., He, Xuming, Bennamoun, Mohammed, Sohel, Ferdous, Togneri, Roberto

cvpr2015papers_第219张图片
[set, learning, task, semantic, obj, labeling, annotated, baseline, evaluate, dataset, understanding, involved] [feature, top, crf, performance, better, training, level, generic] [object, detection, superpixel, cost, based, bounding, three, segmentation, superpixels, incorporate, threshold, appearance, intersection, salient, hypothesis, jaccard] [cuboid, scene, cluttered, indoor, clutter, potential, model, approach, rgbd, unary, milp, layout, estimation, inference, jiang, joint, depth, obstruction, reasoning, room, detected, term, volumetric, huebner] [method, pairwise, number, linear, problem, parameter, function, algorithm, denote, variable, time, random] [color, estimate, figure, geometry] [image, local, geometric, spatial, rate, major, view, geometrical, large]
Learning to Detect Motion Boundaries
Weinzaepfel, Philippe, Revaud, Jerome, Harchaoui, Zaid, Schmid, Cordelia

cvpr2015papers_第220张图片
[dataset, learning, clean, datasets, evaluation, work, annotated, set, predict, prediction, introduce, predicted] [training, structured, train, performance, table, input, trained, decision, feature, map] [boundary, detection, ymb, edge, youtube, temporal, propose, video, based, segmentation, ldof, object, recall, precision, frame, threshold] [approach, estimation, error, occlusion, forest, model, depth, tree, challenging, rgb] [random, method, norm, binary, energy, proposed] [motion, optical, figure, estimated, warping, high, color, middlebury, ieee, flow, estimate, accurately, result, robust] [image, gradient, large, patch, local, compute, compare, computed, fast, deepflow, outperforms, impact, well]
Learning to Rank in Person Re-Identification With Metric Ensembles
Paisitkriangkrai, Sakrapee, Shen, Chunhua, van den Hengel, Anton

cvpr2015papers_第221张图片
[learning, set, visual, existing, introduce, surveillance, test, returned, correct, written, task] [recognition, top, performance, training, ensemble, triplet, learn, cnn, feature, neural, pedestrian, discriminative, deep] [propose, individual, multiple, based, appearance] [approach, represent] [data, number, base, rank, covariance, linear, algorithm, matrix, problem, optimization, proposed, curve, idea, method, regularization, constraint, parameter, analysis, solve] [camera, ieee, color, relative, pattern] [person, distance, metric, cmc, rate, view, image, large, local, best, viper, ilids, gallery, outperforms, benchmark, descriptor, probe, texture, represents, applied, matching, pij, consists, commonly]
Discriminative and Consistent Similarities in Instance-Level Multiple Instance Learning
Rastegari, Mohammad, Hajishirzi, Hannaneh, Farhadi, Ali

cvpr2015papers_第222张图片
[positive, instance, mil, learning, bag, consistency, ranking, set, dataset, migraph, categorization, test, discovered, system, text, discovering, csg, datasets, introduce, rxi, coincidental, discovery, evaluate, include, work, task] [negative, discriminative, training, previous, global, feature, table, svm, standard, learn, accuracy, learned] [multiple, node, based, score, edge, walk, eij, weight, object] [approach, reasoning, step, model] [method, pairwise, algorithm, random, optimization, problem, rank, graph, function, clique, consistently, experimental, time] [high, figure] [similarity, image, consistent, maximal, large, notion, local, outperforms, benchmark, compute, includes]
Prediction of Search Targets From Fixations in Open-World Settings
Sattar, Hosnieh, Muller, Sabine, Fritz, Mario, Bulling, Andreas

cvpr2015papers_第223张图片
[search, visual, book, prediction, set, predict, task, amazon, work, predicting, chance, behaviour, human, learning, evaluation, colour, collage, natural, synthesised, investigated, user, relevance, qtest, vision, qtrain, dataset, test, categorical, focused] [accuracy, setting, level, size, previous, training, performance, train, window, recognition, optimum, higher, learn, svm] [target, eye, three, average, cognitive, movement, based, tracker, appearance, longer] [gaze, potential, model, approach, challenging, control, searching] [data, number, problem, proposed, formulation, minimum, binary, journal, random, method, experiment, implicit] [figure, assume, pattern] [image, query, sampling, well, similarity, retrieval, patch]
Rent3D: Floor-Plan Priors for Monocular Layout Estimation
Liu, Chenxi, Schwing, Alexander G., Kundu, Kaustav, Urtasun, Raquel, Fidler, Sanja

cvpr2015papers_第224张图片
[set, test, dataset, prediction, evaluate, work, predicted, photo, goal, correct, exploit, semantic, refer] [window, train, top, performance, table, accuracy, training, additional, feature] [branch, three, employ] [room, layout, wall, scene, apartment, aspect, plan, inference, approach, indoor, rental, localization, estimation, monocular, provide, model, point, typically, full, challenging, front, face, alignment, pose, single, queue, location] [number, energy, problem, note, random, bound, order, time, data, space, denotes, case, markov] [ratio, camera, figure, total, vanishing, geometry, ray, left, estimating] [well, compute, image, large, contained]
Delving Into Egocentric Actions
Li, Yin, Ye, Zhefan, Rehg, James M.

cvpr2015papers_第225张图片
[set, datasets, understanding, visual, work, dataset, include, encode, surveillance, human, focus, novel, evaluation, rich, vision] [recognition, performance, feature, previous, table, representation, training, discriminative, accuracy, improve] [egocentric, action, gtea, object, video, three, trajectory, idt, propose, key, including, foreground, activity, interest, movement, segmentation, compensation, boost, fathi, lab, combination, recognizing, confusion, attention, combined] [gaze, head, hand, point, manipulation, provide, split, model, modeling] [encoding, method, number, data] [motion, camera, optical, result, figure, captured, directly, color] [local, benchmark, image, best, dense, comparison, large, histogram, demonstrate, descriptor, implementation, radius, systematic]
Sense Discovery via Co-Clustering on Images and Text
Chen, Xinlei, Ritter, Alan, Gupta, Abhinav, Mitchell, Tom

cvpr2015papers_第226张图片
[visual, semantic, text, sense, domain, word, knowledge, google, discovers, create, unsupervised, scoring, disambiguation, automatically, dataset, learning, human, web, discovering, set, perform, associated, discover, noun, discovery, phrase, search, evaluation, apple, work, baseline, correspond] [feature, performance, representation, table, training] [based, multiple, node, likelihood, merge, object, score, intersection, subway, propose, extract, generate] [approach, term, joint, structure, model, represent, split, step, cmu, smoothness] [clustering, data, cluster, algorithm, mapping, problem, space, number, spectral, note, function, experiment, membership, iterative] [figure, estimate, assumption] [image, bow, similarity, histogram, compare]
Supervised Descriptor Learning for Multi-Output Regression
Zhen, Xiantong, Wang, Zhijie, Yu, Mengyang, Li, Shuo

cvpr2015papers_第227张图片
[learning, supervised, existing, dataset, work] [discriminative, feature, performance, representation, learn, achieve, hog, learned, supervision, discrimination, effectiveness, output, training, max, table, achieves, validation, highly] [based, propose, framework] [regression, estimation, pose, head, orientation, sij, term, structure, yaw, volume] [sdl, multivariate, manifold, proposed, compact, reduction, approximation, matrix, dimensionality, function, regularization, objective, algorithm, space, data, smr, kernel, arg, problem, optimization, solved, experimental, solve, vector, simultaneously, great, formulated] [ieee, generalized, accurate, conventional, developed, pitch, computer, camera] [descriptor, image, spatial, gradient, lbp, local, handcrafted, phog, outperforms, cross, large]
Fine-Grained Visual Categorization via Multi-Stage Metric Learning
Qian, Qi, Jin, Rong, Zhu, Shenghuo, Lin, Yuanqing

cvpr2015papers_第228张图片
[learning, address, visual, active, dataset, test, baseline, set] [class, training, learned, triplet, table, performance, size, feature, challenge, auxiliary, learn, better, stage, imagenet, issue, accuracy, euclid, improve, observe, deep] [segmentation, framework, cost, based, bounding] [approach, single] [msml, dimensional, number, dml, proposed, random, method, fgvc, low, matrix, lsvm, problem, data, rank, algorithm, optimization, dual, space, computational, time, dimensionality, linear, projection, original, solving, min, storage, note, approximation, optimal, psd, sparse, function] [high, small, handle, conventional, figure, directly, avoid] [metric, distance, large, image, lmnn, comparison, outperforms, compute]
Mirror, Mirror on the Wall, Tell Me, Is the Error Small?
Yang, Heng, Patras, Ioannis

cvpr2015papers_第229张图片
[human, test, set, consistency, evaluate, feedback, selected, example, interesting, dataset, correct, parse, focus, introduce, representative] [performance, training, table, trained, size, recognition] [correlation, object, detection, art, state, truth, bounding, measure, calculate, based, wang, recall, including] [mirror, error, alignment, face, rcpr, mirrorability, localization, ground, pose, ifa, facial, sdm, model, fmp, body, yang, sem, estimation, coefficient, lsp, ltm, sea, landmark, cascaded, mirrored, shape, ramanan, fact] [method, original, order, restart, proposed, symmetric, low, number, analysis, sample, selection] [result, high, symmetry, figure, second] [image, calculated, difference, large, corresponding, good, invariant, applied, called]
Understanding Classifier Errors by Examining Influential Neighbors
Kabra, Mayank, Robie, Alice, Branson, Kristin

cvpr2015papers_第230张图片
[example, set, learning, labeled, test, weak, label, fastboot, mispredicted, understand, user, active, labeling, unlabeled, prediction, jaaba, selected, positive, novel, understanding, system, interactive, wing, lda, inconsistently, add, mislabeled, work, indicated, insight, predicted, relevant, backup] [training, boosting, version, feature, learn, improve, recognition, imagenet, context, improvement] [behavior, measure, object, weight, propose, bootstrap, performs, score] [current, single, leaf, performed] [data, method, space, machine, algorithm, distribution, number, case, complexity, normalized, consider] [figure, median, opposite, small, estimate, examine] [distance, dissimilarity, closest, large, nearest, best, metric, image, sorted, sampled, well, fast, true]
From Dictionary of Visual Words to Subspaces: Locality-Constrained Affine Subspace Coding
Li, Peihua, Lu, Xiaoxiao, Wang, Qilong

cvpr2015papers_第231张图片
[visual, representative, set, learning, work, probability] [feature, performance, fisher, training, ensemble, voc, size, accuracy, higher, table, map, pascal, better, pyramid, increasing, class, achieves, improved] [measure, based, object, propose, segmentation, indicates, surrounding] [structure, indoor, scene, described, coordinate] [lasc, method, subspace, coding, proximity, llc, vector, linear, manifold, data, number, encoding, space, proposed, sparse, approximation, function, dimension, matrix, note, form, dimensionality, kobayashi, attached, cluster, distribution, solution, principal, problem, reduction] [figure, piecewise, component, geometry, high, simple] [dictionary, local, image, comparison, nearest, large, geometric, statistical, metric, matching, outperforms, compare, weighted]
Automatically Discovering Local Visual Material Attributes
Schwartz, Gabriel, Nishino, Ko

cvpr2015papers_第232张图片
[attribute, visual, discovered, set, human, discover, category, recognize, predicted, plastic, perception, encode, trait, discovery, discovering, foliage, schwartz, sharan, leather, nishino, learning, existing, appear, understanding, predict, labeled, process, embedding] [recognition, accuracy, raw, training, feature, previous, train, input] [perceptual, object, based, propose, framework, recognizing] [fact, represent, model, rely, shape] [matrix, space, method, pairwise, random, binary, distribution, mapping, proposed] [material, figure, water, typical, fabric, smooth, desired, paper, metal, computer, glass] [image, local, similarity, exhibit, distance, expect, compute, patch, texture, consistent, indicate]
Jointly Learning Heterogeneous Features for RGB-D Activity Recognition
Hu, Jian-Fang, Zheng, Wei-Shi, Lai, Jianhuang, Zhang, Jianguo

cvpr2015papers_第233张图片
[learning, human, set, dataset, work, share, test, report] [feature, shared, recognition, tpf, hoi, table, training, performance, compared, hog, extracted, accuracy, reported, jointly, learn, channel, sysu, pyramid] [activity, ing, action, temporal, skeleton, dynamic, video, msr, object, based, appearance, daily, confusion, sequence, wallet, joulesvm, framework, interaction, employed, three] [model, depth, joint, rgb, step, inference, capture] [heterogeneous, method, proposed, matrix, dimensionality, original, number, min, random, algorithm, sample, optimization, china, projection, science, solution, subspace, fusion, function] [figure, signal, fourier, phone, developed, motion, observed] [gradient, best, outperforms, comparison]
Improving Object Proposals With Multi-Thresholding Straddling Expansion
Chen, Xiaozhi, Ma, Huimin, Wang, Xiang, Zhao, Zhichen

cvpr2015papers_第234张图片
[obj, set, existing, diverse, generating, search] [accuracy, achieve, improving, pascal, improve, sliding, effective, achieves, higher, setting, feature, generic] [object, bounding, box, recall, straddling, superpixel, iou, bing, superpixels, threshold, tightness, mtse, mcg, average, proposal, overlap, multiple, cpmc, detection, intersection, generate, union, segmentation, objectness, based, edge, propose, integrated, measure, viewed, numerous, three, background, utilize, truth, generated, strong] [localization, approach, initial, alignment, model, ground] [expansion, method, low, number, distribution, curve, proposed] [figure, ieee, high, bias, quality, accurate, simple, sin] [image, best, area, compute, distinct]
VIP: Finding Important People in Images
Solomon Mathialagan, Clint, Gallagher, Andrew C., Batra, Dhruv

cvpr2015papers_第235张图片
[people, dataset, visual, work, predict, baseline, human, study, annotation, task, annotated, photo, focus, prediction, selected, understanding, pick, category, collected, datasets, test, description, british, set, tend, interface, asked, sentence, identify, predicting, agreement, automatically, question, describe, woman, api] [table, feature, compared, improvement, previous] [saliency, group, salient, bounding, frame, score, three, measure, detection, individual, multiple, average, box] [face, social, pose, model, approach, capture, qualitative] [method, number, order, journal, consider] [center, relative, sharpness, ratio, figure, paper, ieee, international, total] [person, image, pair, distance, difference, weighted, compute, best]
Bilinear Heterogeneous Information Machine for RGB-D Action Recognition
Kong, Yu, Fu, Yun

cvpr2015papers_第236张图片
[learning, visual, dataset, knowledge, work, existing, risk, human] [feature, svm, recognition, learned, shared, training, performance, table, discriminative, accuracy, achieves, learn, representation, compared, better, reduce, generalization, extracted, learns, testing] [action, spatiotemporal, msr, structural, activity, measure, daily, oriented, three] [depth, rgb, model, capture, projected, naturally, term, body, allows] [bhim, bilinear, parameter, data, matrix, linear, modality, method, space, problem, proposed, form, projection, nxyt, rank, missing, effectively, heterogeneous, optimization, algorithm, original, dimensionality, vector, function, min, compress, order, regularizer, regularized, minimization, case, solved, optimized] [noise, motion, figure, solver, total] [outperforms, comparison]
Dynamically Encoded Actions Based on Spacetime Saliency
Feichtenhofer, Christoph, Pinz, Axel, Wildes, Richard P.

cvpr2015papers_第237张图片
[human, novel, dataset, evaluation, work, subset] [recognition, feature, pooling, performance, fisher, training, accuracy, table, svm, alternative, improve, hog, representation, standard] [action, saliency, idt, late, spacetime, spatiotemporal, measure, foreground, video, three, based, contrast, yield, oriented, superpixels, puppet, temporal, truth, employed, sequence, hof, catching, static, frame, mbh] [approach, capture, typically, step, ground, orientation] [energy, encoding, proposed, vector, normalized, fusion, sum, directional, gaussian, coding, number] [motion, figure, second, ball, frequency, presented, plane] [local, spatial, image, weighted, aggregation, dense, variance, applied, weighting, well, comparison, descriptor, large, distance, region, locally]
In Defense of Color-Based Model-Free Tracking
Possegger, Horst, Mauthner, Thomas, Bischof, Horst

cvpr2015papers_第238张图片
[visual, ranking, datasets, exploit, distractors, mil, learning, search, adapt, identify, vote, evaluation, advance] [discriminative, representation, performance, accuracy, table, challenge, achieves, compared, van, top, previous, achieve] [object, tracking, based, online, vot, distracting, kcf, dat, adaptive, appearance, lgt, tracker, likelihood, combined, struck, surrounding, threshold, background, dsst, bounding, segmentation, spot, drifting, frame, nodat, fot, ogt, target, average, distinguish, suppress, correlation, including, adaption, propose] [model, approach, robustness, allows, location, current, detailed, estimation] [proposed, rank, update] [color, figure, robust, high] [benchmark, scale, region, well, best, large, overcome]
Watch-n-Patch: Unsupervised Understanding of Actions and Relations
Wu, Chenxia, Zhang, Jiemi, Savarese, Silvio, Saxena, Ashutosh

cvpr2015papers_第239张图片
[topic, human, unsupervised, set, word, learning, dataset, probability, correlated, interactive, document, evaluate, semantic, indicating] [learned, recognition, training, table, output, performance, learn] [action, temporal, video, activity, forgotten, segment, segmentation, assigned, skeleton, patching, complex, wnd, detection, online, znd, frame, correctly, xkc, sequence, extract, multiple, object, tmnd, hmm, hierarchical, tnd] [model, modeling, rgb, approach, kinect, challenging, structure, prior, normal, capture, recorded, body] [time, distribution, consider, number, order, pairwise, missing, drawn, assignment] [relative, figure, absolute, completely, motion] [well, sampling, local]
Building a Bird Recognition App and Large Scale Dataset With Citizen Scientists: The Fine Print in Fine-Grained Dataset Collection
Van Horn, Grant, Branson, Steve, Farrell, Ryan, Haber, Scott, Barry, Jessie, Ipeirotis, Panos, Perona, Pietro, Belongie, Serge

cvpr2015papers_第240张图片
[citizen, dataset, annotation, vision, bird, corruption, datasets, mturkers, set, label, test, tool, category, mechanical, mturk, throughput, expert, visual, domain, flickr, interface, cub, search, crowdsourcing, help, community, quiz, task, collect, scientist, vote, collecting, knowledge, nabirds, work, labeling, user, turkers, expertise] [training, imagenet, recognition, level, performance, accuracy, larger, size, class, raw] [object, lab, average, cost, based, multiple] [error, provide] [number, time, analysis, data, corrupted] [computer, figure, quality, high, conference, ieee, pattern, total, real, accurate] [image, large, june, false, rate]
How Many Bits Does it Take for a Stimulus to Be Salient?
Hossein Khatoonabadi, Sayed, Vasconcelos, Nuno, Bajic, Ivan V., Shan, Yufeng

cvpr2015papers_第241张图片
[visual, prediction, test, label, processing, vision, set, probability, consistency, entropy, natural, human, amount] [feature, accuracy, map, table, neural, standard, class, higher, global] [saliency, video, measure, obdl, average, stimulus, based, temporal, auc, aws, eye, salient, detection, attention, mam, pmes, mcsdm, surp, compressibility, maxnorm, stsd, pxl, gbvs, sorm, principle, score] [model, control, scene, cmp, volume] [block, compressed, sample, number, energy, coding, mrf, proposed, random, analysis, markov, function, avg, journal, low, data] [ieee, motion, high, computer, accurate, figure, measured, signal] [image, compression, spatial, modern, local, implementation, psnr, statistical]
Class Consistent Multi-Modal Fusion With Binary Features
Shrivastava, Ashish, Rastegari, Mohammad, Shekhar, Sumit, Chellappa, Rama, Davis, Larry S.

cvpr2015papers_第242张图片
[test, vision, dataset, learning, predicted, set, subject, predict, evaluate, multimodal, randomly] [class, feature, recognition, level, training, table, max, performance, decision, svm, input, fusing, better, accuracy] [multiple, based, target, object, score, weight, combined, combination] [depth, continuous] [binary, fusion, algorithm, data, modality, ccmm, solution, greedy, problem, method, kernel, optimization, arg, bit, sparse, iris, linear, machine, denoted, proposed, smbr, vector, gurobi, row, matrix, feasible, min, wvu, sample, ith, perturbation, mip, biometrics, atm, order, mkl, number, solving] [conference, ieee, computer, pattern, international, intensity, figure, color] [consistent, compute, image, computed, large, biometric, mth, change, compare]
Ranking and Retrieval of Image Sequences From Multiple Paragraph Queries
Kim, Gunhee, Moon, Seungwhan, Sigal, Leonid

cvpr2015papers_第243张图片
[text, blog, latent, photo, set, semantic, embedding, work, sentence, associated, ranking, retrieve, language, test, visual, paragraph, natural, disneyland, retrieving, topic, corpus, post, association, disney, ripa, user, relation, evaluation, relevant, learning, perform, dvisor, common, scand, review, representative, lda, evaluate, automatically] [training, svm, feature, input, learn, size, top, performance] [segmentation, sequence, multiple, structural, three, segment, based, groundtruth, key] [approach, model, single] [method, algorithm, problem, data, mapping, vector, number, lsa, space, analysis] [unique, parallel] [image, query, retrieval, large, best, similarity, comparison, consists, visualization, pair, comprehensive]
Motion Part Regularization: Improving Action Recognition via Trajectory Selection
Ni, Bingbing, Moulin, Pierre, Yang, Xiaokang, Yan, Shuicheng

cvpr2015papers_第244张图片
[visual, set, human, latent, select, word, associated, introduce, learning, dataset, selected, ranked] [discriminative, fisher, recognition, training, feature, representation, testing, learned, top, svm, global, improved, better, follow, extracted, class, ifv] [action, video, trajectory, based, group, discriminativeness, temporal, weight, entire, structural, framework, movement, copy, golf, olympic, wang, propose, utilized] [model, body] [vector, method, regularization, proposed, selection, number, form, note, denote, entry, sparse, algorithm, sample, optimization, lasso, problem, respect, parameter, gaussian, solve, objective] [motion, figure, ieee, simple, assume, high] [local, dense, weighted, database, corresponding, scheme, overlapping, best, good, gradient]
Beyond Principal Components: Deep Boltzmann Machines for Face Modeling
Nhan Duong, Chi, Luu, Khoa, Gia Quach, Kha, Bui, Tien D.

cvpr2015papers_第245张图片
[vision, active, domain, set, predict, evaluation, ability] [deep, testing, hidden, age, training, table, layer, better, trained, representation, higher, input, factor, recognition, compared, discriminative, capability] [appearance, based, generate, numerous, three] [facial, shape, model, face, aams, reconstruction, modeling, boltzmann, approach, helen, estimation, lfpw, occlusion, dbm, represent, current, maes, challenging, provide, landmark, term] [method, proposed, order, formulated, arg, original, represented, data, gaussian, problem, algorithm] [computer, figure, reconstructed, ieee, conference, presented, pattern, international, noise, rmse, second, estimated, quality] [texture, image, bicubic, database, aging, pca]
Semi-Supervised Low-Rank Mapping Learning for Multi-Label Classification
Jing, Liping, Yang, Liu, Yu, Jian, Ng, Michael K.

cvpr2015papers_第246张图片
[label, learning, labeled, instance, unlabeled, sun, latent, prediction, msrc, set, datasets, category, dataset, lin, active] [feature, loss, training, performance, accuracy, size, table, learn, testing, class] [based, three, handling, multimedia, auc, indicates, average] [model, structure, capture, term] [data, slrm, space, method, mapping, number, singular, matrix, proposed, function, regularization, linear, missing, algorithm, alternating, vector, machine, complexity, min, faie, nuclear, main, optimization, mediamill, convex, problem, norm, miml, order, denoted, cplst, analysis, mlloc, manifold, close, experimental, computational, ith, remark] [figure, intrinsic, direction, real, square, ieee] [large, image, geometric, corresponding, local]
Can Humans Fly? Action Understanding With Multiple Classes of Actors
Xu, Chenliang, Hsieh, Shao-Hang, Xiong, Caiming, Corso, Jason J.

cvpr2015papers_第247张图片
[semantic, vision, dog, understanding, label, conditional, dataset, bird, set, work, human, eating, european, car, jump, cat, tuple, considers, labeled, labeling, relationship] [recognition, performance, jointly, train, accuracy, training, observe, better, class] [action, actor, segmentation, video, trilayer, adult, bayes, object, bilayer, supervoxel, baby, crawling, walking, three, average, individual, climbing, actoraction, jointps, multiple] [model, joint, modeling, inference, full, graphical, independent, potential] [space, product, consider, problem, roll, journal] [computer, conference, ieee, international, pattern, rolling, figure, ball, motion, directly, paper] [dense, well, eat, large]
Learning Scene-Specific Pedestrian Detectors Without Real Data
Hattori, Hironori, Naresh Boddeti, Vishnu, Kitani, Kris M., Kanade, Takeo

cvpr2015papers_第248张图片
[learning, work, surveillance, domain, dataset, adaptation, evaluation, evaluate, human, labeled, purely, baseline, scenario] [pedestrian, detector, dpm, trained, learn, ssv, training, performance, learned, table, generic, train] [detection, appearance, object, precision, overlap, recall, based, generated, generate, cmusrd, correlation, town, bounding, video, framework, truth, static, three] [scene, approach, location, model, single, smoothness, localization, ground, mellon, pose, cij, carnegie] [data, proposed, method, min, problem, denotes, number, arg, range] [real, synthetic, camera, figure, ratio, geometry, computer, fourier, center, small, perspective, varying, account, simulated, second] [virtual, large, distance, resolution, image, outperforms, geometric, corresponding, spatial, applied]
Visual Recognition by Counting Instances: A Multi-Instance Cardinality Potential Kernel
Hajimirsadeghi, Hossein, Yan, Wang, Vahdat, Arash, Mori, Greg

cvpr2015papers_第249张图片
[cardinality, instance, bag, learning, vision, people, label, latent, positive, work, set, visual, european, dataset, summarization, interesting, yqj, relevant, encode, evaluation, probability, human, labeled, example, hard] [recognition, svm, structured, training, feature, performance, trecvid, standard, global] [video, event, activity, temporal, multiple, segment, collective, based, action, group, framework, counting, propose, likelihood, detection] [model, potential, probabilistic, inference, joint, approach, scene, modeling, single] [kernel, proposed, method, algorithm, number, exp, exact, marginal, machine, parameter, encoding, intelligence] [computer, conference, pattern, international, figure, paper, length] [internet, dense, comparison]
Modeling Deformable Gradient Compositions for Single-Image Super-Resolution
Zhu, Yu, Zhang, Yanning, Bonev, Boyan, Yuille, Alan L.


[set, work, learning, example, woman, bird, exploit, test, select, evaluate, selected] [input, deep, better, network, performance, previous, top, global] [based, edge, mask, baby, path, combination, contrast, propose, integrated, candidate] [deformable, model, deformation, compositional, single, smoothness, prior, structure, position, head, represent, reconstruction, regression, estimation, composition, adjustment] [singular, method, sparse, coding, proposed, function, parameter, linear, base, note, form, literature, constraint, appropriate] [figure, intensity, pixel, sharp, ieee, red, reconstructed] [gradient, patch, image, dictionary, psnr, ridge, ssim, magnitude, corresponding, masked, local, dominating, matching, choose, best, increment, well, neighbor, june]
Learning Multiple Visual Tasks While Discovering Their Structure
Ciliberto, Carlo, Rosasco, Lorenzo, Villa, Silvia

cvpr2015papers_第250张图片
[learning, set, mtl, task, attribute, relevant, dataset, associated, positive, visual, natural, knowledge, report, closed, okl, share] [setting, training, standard, table, class, loss, output, designed, feature] [multiple, detection, framework, object, sequence, generated, propose] [structure, approach, term, model, limb] [kernel, problem, matrix, sparse, skmtl, convex, method, sparsity, form, algorithm, solution, linear, reproducing, optimization, empirical, minimization, function, alternating, respect, proposed, theorem, regularization, hilbert, graph, order, machine, solving, space, stl, solve, operator, convergence, number, mtrl, consider, rkhsvv, objective, selection] [recover, conference, scalar, computer, recovered, figure, color, synthetic, theory] [separable, considered]
Mapping Visual Features to Semantic Profiles for Retrieval in Medical Imaging
Hofmanninger, Johannes, Langs, Georg

cvpr2015papers_第251张图片
[semantic, visual, learning, set, healthy, label, radiology, weak, relevant, reticular, emphysema, mil, honeycombing, bvw, labeling, content, highest, haralick, weakly, vision, bag, occurring, link] [feature, training, class, map, ensemble, table, compared, learn, decision] [based, supervoxels, supervoxel, anomaly, appearance, precision, generated, generate, multiple, partitioning] [volume, performed, leaf, capture] [data, random, space, vector, mapping, distribution, number, binary, analysis, method, machine] [imaging, figure, ieee, pattern, computer, conference, high, relative] [image, retrieval, medical, corresponding, clinical, texture, tissue, lung, local, descriptor, distance, true, randomized, variability, metric, similarity]
On Learning Optimized Reaction Diffusion Processes for Effective Image Restoration
Chen, Yunjin, Yu, Wei, Pock, Thomas

cvpr2015papers_第252张图片
[process, learning, test, work, natural, set, feedback] [training, trained, stage, learned, jpeg, size, table, performance, previous, convolution, train, loss, output, effective, standard, involves, additional, smoothing] [based, propose] [model, term, joint, anisotropic, step, discrete, prior, approach, descent, typically] [diffusion, function, nonlinear, proposed, reaction, penalty, linear, gaussian, time, computation, ugt, greedy, order, algorithm, csf, deblocking, problem, consider, energy, number, method, data, optimized, trdt, kernel, note, lead, critical, minimization, wnnm, optimal] [high, ieee, noise, conventional, parallel, result, supplemental, figure, derivative, material] [image, restoration, denoising, gradient, local, corresponding, variational]
Correlation Filters With Limited Boundaries
Kiani Galoogahi, Hamed, Sim, Terence, Lucey, Simon

cvpr2015papers_第253张图片
[domain, learning, set, versus, vision, selected, visual, predicted] [training, performance, size, output, advantage, log, learn, compared, train, negative, applying] [correlation, tracking, object, cost, boundary, eye, struck, memory, precision, average, frame, threshold, detection, complex, based] [approach, localization, error, full, position, face, fps] [method, number, mosse, objective, proposed, computational, matrix, shifted, fragtrack, vector, vectorized, iterative, denotes, miltrack, operator, algorithm, augmented, solving, admm, expressed, energy, operation, iteration, extremely] [figure, equation, frequency, circular, fourier, shift, signal, desired, length, lagrangian, estimated, paper, robust, synthetic] [spatial, dense, image, rate, well, sampling, dramatically, patch]
Active Learning and Discovery of Object Categories in the Presence of Unnameable Instances
Kading, Christoph, Freytag, Alexander, Rodner, Erik, Bodesheim, Paul, Denzler, Joachim

cvpr2015papers_第254张图片
[active, learning, vision, dataset, unlabeled, expected, emoc, unnameable, label, discovery, example, set, labeled, coco, labeling, erm, work, process, risk, european, annotator, selecting, evaluating, select, pknn, probability, rejected, entropy, unrelated, discover, visual] [class, output, recognition, accuracy, training, loss, previous] [object, density, principle, including] [model, regression, approach, face, valid, current, uncertainty, step, realistic] [data, binary, gaussian, problem, random, machine, number, usps, method, sample, empirical, function, respect, code, case] [conference, computer, international, unknown, technique, high, estimate, pattern, estimated, figure] [change, query, large, image]
A MRF Shape Prior for Facade Parsing With Occlusions
Kozinski, Mateusz, Gadde, Raghudeep, Zagoruyko, Sergey, Obozinski, Guillaume, Marlet, Renaud

cvpr2015papers_第255张图片
[set, dataset, semantic, door, baseline] [class, parsing, performance, grammar, table, window, accuracy, neural, training, vegetation, raw] [segmentation, assigned, hierarchical, structural, allowed, complex] [facade, adjacency, shape, prior, adjacent, ecp, alignment, formalism, anc, neighboring, production, etrims, unary, irregular, split, model, textonboost, structure, artdeco, vertically, architectural] [algorithm, number, column, row, dual, encoding, method, energy, formulation, proposed, denote, linear, lower, optimal, respect, problem, element, random, pairwise, optimization, decomposition, primal] [pixel, pattern, vertical, horizontal, figure, component, simultaneous, result, sky] [image, grid, corresponding, recursive, hierarchy]
Predicting the Future Behavior of a Time-Varying Probability Distribution
Lampert, Christoph H.

cvpr2015papers_第256张图片
[set, probability, learning, prediction, domain, future, predicted, embedding, adaptation, work, extrapolating, predict, visual, vision, source, conditional, predicting, test, extrapolate, embeddings, study, positive] [training, learn, setting, table, feature, input, apply, identity, actual] [target, video, situation, sequence] [step, regression, mixture, model, split] [distribution, data, time, edd, kernel, sample, herding, rkhs, method, machine, hilbert, space, linear, operator, problem, empirical, order, number, predsvm, convex, form, working, reproducing, experiment, practical, inner, vector, regularization] [observed, conference, synthetic, figure, estimate, computer, real, international, second, equation] [true, weighted, distance, consists]
An Approximate Shading Model for Object Relighting
Liao, Zicheng, Karsch, Kevin, Forsyth, David

cvpr2015papers_第257张图片
[user, study, visual, system, dataset, existing, subject, task, set, expert] [coarse, age, layer, representation, standard, passed, table, alternative, better] [object, viewed, target, complex] [shape, model, scene, composition, reconstruction, error, normal, surface, rendered, single, realistic, approach, estimation, allows, realism] [method, approximate, composite, column, lower, number] [shading, detail, barron, illumination, insertion, real, malik, parametric, result, figure, karsch, material, relighting, half, inserted, chose, rendering, synthetic, residual, preferred, acm, light, albedo, simple, color, accurate, smooth, technique, second, subpopulation, high, produce, failed, artist, intrinsic, male, lighting] [image, geometric, view, large, best, well, comparison]
Web Scale Photo Hash Clustering on A Single Machine
Gong, Yunchao, Pawlowski, Marcin, Yang, Fei, Brandy, Louis, Bourdev, Lubomir, Fergus, Rob

cvpr2015papers_第258张图片
[photo, perform, store, purity, work, dataset, set, learning, stream, introduce, amount, web] [table, deep, build, convolutional, neural, standard, size, better] [online, based, truth, propose, object] [step, point, single, ground, represent] [clustering, binary, cluster, number, hash, method, data, algorithm, problem, proposed, hashing, time, trending, compact, speedup, kmeans, lookup, running, spam, huge, complexity, batch, machine, hamming, quantization, endfor, update, faster, kmedoids, optimization, duplicate, easily, extremely, bkmeans] [figure, directly, center, speed, enables, handle] [large, image, nearest, fast, distance, internet, applied]
Hardware Compliant Approximate Image Codes
Kuang, Da, Gittens, Alex, Hamid, Raffay

cvpr2015papers_第259张图片
[vision, set, visual, work, encode] [accuracy, size, table, training, pascal, increase, feature, mit, standard, compared, discriminative, better, class] [average, memory, plot, cost] [error, reconstruction, approach, fact, volume] [encoding, exact, approximate, llc, proposed, approximation, codewords, analysis, linear, coding, note, cluster, matrix, perturbation, codebook, bmt, number, normalized, solving, main, intuition, empirical, assignment, theorem, sparse, computational, function, consider, complexity, multiplication, min, codeword, cholesky] [computer, figure, equation, ieee, hardware, pattern, high] [scheme, nearest, image, dictionary, cvpr, local, counterpart, large, modern, distance]
Multispectral Pedestrian Detection: Benchmark Dataset and Baseline
Hwang, Soonmin, Park, Jaesik, Kim, Namil, Choi, Yukyung, So Kweon, In

cvpr2015papers_第260张图片
[dataset, vision, datasets, human, positive, annotated, introduce, evaluation, baseline, system, existing] [acf, pedestrian, channel, feature, negative, capturing, recognition, additional, compared, trained, better, larger, standard, performance, caltech] [detection, based, object, three, bounding, temporal, average, frame, heavy, video] [occlusion, aligned, height, occluded, visible, shape, kitti] [number, analysis, proposed, machine, extension, algorithm] [thermal, color, multispectral, camera, figure, pattern, computer, ieee, conference, captured, infrared, hardware, day, beam, night, pixel, reasonable, nighttime, international, calibration, braking, splitter, handle, daytime, medium] [image, rate, false, distance, scale, partial, gradient, voting, person]
Semi-Supervised Domain Adaptation With Subspace Learning for Visual Recognition
Yao, Ting, Pan, Yingwei, Ngo, Chong-Wah, Li, Houqiang, Mei, Tao

cvpr2015papers_第261张图片
[domain, source, adaptation, learning, labeled, unlabeled, sdasl, visual, transfer, webcam, concept, test, latent, positive, learnt, risk, work, unsupervised, amazon, dslr, mismatch, set, randomization, dataset, instance, scmv] [training, performance, feature, better, trecvid, learn, imagenet, recognition, svm, jointly, trained, decision] [target, three, video, indicates, structural, object] [structure, predictive, model, represent] [data, subspace, objective, function, proposed, regularization, manifold, distribution, optimization, method, conducted, algorithm, problem, original, mapping, number, space, minimizing, minimize, dimensionality, vector, simultaneously, china] [figure, preservation, shift] [image, distance, similarity, metric, difference, good, outperforms, compute]
Real-Time Part-Based Visual Tracking via Adaptive Correlation Filters
Liu, Ting, Wang, Gang, Yang, Qingxiong

cvpr2015papers_第262张图片
[vision, visual, learning, set] [map, response, recognition, size, learned, table, discriminative, training, structured, performance] [tracking, target, based, correlation, object, kcf, appearance, tracker, adaptive, online, tracked, multiple, structural, state, bounding, frame, weight, struck, updated, framework, likelihood, individual, track, adaptively, average, psr, particle, calculate, entire, video, box, detection] [model, fps, occlusion, joint, inference, location, occluded, layout, challenging] [method, proposed, constraint, updating, problem, maximum, number, complicated, sparse, update, machine, adopted] [computer, ieee, conference, pattern, robust, international, speed, red, observation, smooth, center, small, high, figure] [bayesian, corresponding, scale, image, spatial, partial, rate]
Efficient Sparse-to-Dense Optical Flow Estimation Using a Learned Basis and Layers
Wulff, Jonas, Black, Michael J.

cvpr2015papers_第263张图片
[vision, work, learning, set] [feature, learned, layer, training, accuracy, recognition, learn, performance, combine] [frame, truth, average, propose, assigned] [estimation, kitti, approach, model, ground, volume, current, full, point] [sparse, basis, method, algorithm, linear, principal, note, analysis, machine, runtime, data, number, subspace, journal, computationally, approximate] [optical, motion, computer, ieee, pattern, layered, accurate, figure, estimate, conference, international, robust, interpolation, small, horizontal, traditional, vertical, requires, springer, lecture, simple] [image, dense, compute, pca, matching, spatial, computed, sintel, june, large, resolution, fast, variational, well, good, local, displacement]
Cascaded Hand Pose Regression
Sun, Xiao, Wei, Yichen, Liang, Shuang, Tang, Xiaoou, Sun, Jian

cvpr2015papers_第264张图片
[human, dataset, learning, work] [better, stage, previous, training, global, root, feature, learn, achieve] [hierarchical, object, based, frame, tracking, average, framework, truth, updated, complex] [pose, hand, regression, palm, estimation, holistic, depth, coordinate, articulated, cascaded, viewpoint, approach, body, fps, initial, canonical, error, parameterization, lrf, joint, model, ground, finger, initialization, forest, regressing, offset, regressors, rough, challenging, articulation] [random, algorithm, update, note, data] [pixel, estimate, camera, axis, real, estimated, figure, rotation, angle, second, plane, residual, high, pitch, motion, directly] [image, large, transformation, invariance, normalization, difference, applied, corresponding, good, local]
Joint Photo Stream and Blog Post Summarization and Exploration
Kim, Gunhee, Moon, Seungwhan, Sigal, Leonid

cvpr2015papers_第265张图片
[photo, blog, summarization, text, set, stream, semantic, associated, summary, disney, help, latent, ranking, attraction, subset, include, disneyland, knowledge, titling, visual, work, automatic, representative, task, test, selected, entity, randomly, perform] [svm, size, table, stochastic, jointly, training, feature, better, top, build, achieve] [consecutive, key, based, sequence, generate, online, three, indicates, score] [alignment, joint, approach, localization, initial, adjacency, location] [algorithm, method, number, data, constraint, matrix, vector, optimization] [interpolation, estimated, figure] [image, large, similarity, query, pair, comparison, named, choose, list, best]
Latent Trees for Estimating Intensity of Facial Action Units
Kaltwang, Sebastian, Todorovic, Sinisa, Pantic, Maja

cvpr2015papers_第266张图片
[learning, latent, work, categorical, set, conditional, evaluation, clean, human, existing, emotion] [training, hidden, feature, performance, learn, input, discriminative, better, increase] [average, node, action, likelihood, video, correlation, structural, target, appearance] [model, fau, facial, structure, tree, svr, disfa, blt, corr, svc, faus, shoulderpain, spontaneous, inference, face, landmark, posterior, joint, generative, estimation, continuous, leaf, discrete, marginals, pronounced] [data, number, algorithm, distribution, vector, binary, marginal, iteratively, parameter, random, problem, derive] [intensity, noise, ieee, pattern, figure, estimate] [local, parent, lbp, bayesian]
Material Classification With Thermal Imagery
Saponaro, Philip, Sorensen, Scott, Kolagunda, Abhishek, Kambhamettu, Chandra

cvpr2015papers_第267张图片
[work, system, indicator, correct, plastic, wood, vision, study] [feature, coarse, accuracy, class, extracted, top, fine, refers, map, table, compared, standard] [video, combination, based, combined, long, threshold] [model, described, recorded, aligned, coordinate] [rank, method, function, time, linear, sample, spectral, chosen] [material, water, permeation, color, thermal, fft, spherical, equation, figure, infrared, temperature, ieee, computer, heating, pattern, paper, binned, conference, camera, cooling, wave, international, lamp, lighting, nir, slice, champ, lwir, minute, txk, simple, center, total, varying, setup, bidirectional] [heat, texture, best, image, imagery, patch, rate, invariance, comparison, comparing]
Fixation Bank: Learning to Reweight Fixation Candidates
Zhao, Jiaping, Siagian, Christian, Itti, Laurent

cvpr2015papers_第268张图片
[visual, test, positive, human, selected, set, correlated, green, learning] [feature, map, channel, negative, training, fixation, learn, raw, extracted, representation, build, contextual] [blob, saliency, bank, decomposed, group, auc, weight, diem, dynamic, video, blocked, eye, crcns, making, frame, attention, static, density, judd, decompose, ciom, including, based, reweight] [model, gaze, scene, reconstruction, regression, location, approach] [data, gaussian, block, algorithm, column, lasso, linear, reweighting, method, sum, problem, journal, row, sample, decomposition, number, case] [figure, motion] [dictionary, peak, local, image, spatial, outperforms, weighting, weighted, reference, gap]
Leveraging Stereo Matching With Learning-Based Confidence Measures
Park, Min-Gyu, Yoon, Kuk-Jin

cvpr2015papers_第269张图片
[predicted, prediction, selected, set, learning, datasets, correct, consistency, select] [performance, feature, input, training, improvement, accuracy, map, negative, improves, learned, trained] [cost, measure, detection, based, frame, auc, detecting] [kitti, regression, sgm, challenging, forest, initial, term, occluded, approach, tree, ground, error, step, volume, robustness, depth] [proposed, method, number, data, selection, random, variable, algorithm, vector, order, time, original, permutation, problem, consistently] [stereo, disparity, modulation, figure, pixel, outdoor, quality, robust, left, captured, middlebury] [matching, image, difference, computed, unreliable, modulated, correctness, construct, similarity, local, mismatched, fast, false, peak]
Beyond the Shortest Path : Unsupervised Domain Adaptation by Sampling Subspaces Along the Spline Flow
Caseiro, Rui, Henriques, Joao F., Martins, Pedro, Batista, Jorge

cvpr2015papers_第270张图片
[domain, source, adaptation, amazon, dslr, webcam, ssf, symd, visual, unsupervised, set, learning, paradigm, datasets, work, concept, wrapping, description, task, refer] [intermediate, recognition, caltech, map, training, advantage, fully, trained, representation] [target, object, multiple, complex, integrate, sod] [point, allows, kinematic, model, location] [curve, data, manifold, spline, geodesic, space, grassmann, sgf, tangent, proposed, riemannian, projection, order, matrix, number, rank, symmetric, subspace, problem, algorithm, represented, idea, euclidean, sen, orthogonal, special, son] [rolling, mode, smooth, shift, simple, real, presented, motion] [compute, computed, image, geometric, sampling, deal, large, respective, best, metric]
Discriminant Analysis on Riemannian Manifold of Gaussian Distributions for Face Recognition With Image Sets
Wang, Wen, Wang, Ruiping, Huang, Zhiwu, Shan, Shiguang, Chen, Xilin

cvpr2015papers_第271张图片
[set, vision, learning, positive, embedding] [recognition, discriminative, training, performance, feature, testing, compared, belonging] [based, measure, propose, average, three, video, group, multiple] [face, model, modeling, prior, probabilistic, represent, single, approach, mixture] [gaussian, kernel, manifold, gaussians, analysis, riemannian, method, space, covariance, linear, data, lie, gmm, number, distribution, subspace, euclidean, proposed, ytc, nonlinear, order, derive, cdl, mda, mmd, sparse, cox, gda, dcc, pasc, sample, geda, spd, denoted, conducted, ytf] [computer, conference, ieee, pattern, component, society, international, geometry] [image, distance, discriminant, statistical, local, similarity, compute, weighted, nearest, divergence, gallery, derived]
Reflectance Hashing for Material Recognition
Zhang, Hang, Dana, Kristin, Nishino, Ko

cvpr2015papers_第272张图片
[vision, set, embedding, learner, learning, selected, novel, processing, automotive, paint, wood, visual] [recognition, feature, boosting, training, class, discriminative, performance, neural, size, response] [appearance, recognizing, spot, framework] [surface, approach, mirror, point, single, scene, error, joint, layout, provide] [binary, hashing, method, number, function, machine, hash, space, spectral, compact, analysis] [material, figure, computer, texton, disk, conference, ieee, glossy, pattern, angular, illumination, camera, metal, viewing, paper, parabolic, measurement, light, rubber, international, pixel, angle, high, optical, feather, ceramic, brdf] [image, rate, textons, texture, spatial, nearest, gradient, illustrated, distance, region, characteristic, textured, variation, change, database]
Real-Time 3D Head Pose and Facial Landmark Estimation From Depth Images Using Triangular Surface Patch Features
Papazov, Chavdar, Marks, Tim K., Jones, Michael

cvpr2015papers_第273张图片
[test, set, library, system, side] [training, testing, input, accuracy, trained, map, detector, previous, table, feature] [based, detection, average, online, generated] [head, depth, pose, facial, landmark, triangle, face, point, centroid, tsp, surface, model, estimation, biwi, orientation, triangular, performed, height, kinect, fanelli, single, equilateral, winning, error, nose, cloud, bfm, mesh, detected, localization] [base, data, method, number, sample, parameter, lie, proposed, clustering, time] [rotation, figure, robust, synthetic, estimate, estimated, sensor, accurate, high, ieee] [distance, voting, descriptor, sampled, missed, patch, sampling, compute, local, corresponding, image, consists, rate, nearest]
Making Better Use of Edges via Perceptual Grouping
Qi, Yonggang, Song, Yi-Zhe, Xiang, Tao, Zhang, Honggang, Hospedales, Timothy, Li, Yi, Guo, Jun

cvpr2015papers_第274张图片
[set, human, work, generation, ranking, learning, natural, semantic, visual, dataset, novel, vision, strategy, generating] [contour, map, better, performance, feature, top, table, hog] [edge, grouping, object, gestalt, perceptual, segmentation, objectness, based, box, bounding, candidate, group, sbir, score, generated, multiple, detection, proposal, grouper, framework, generate, key, three, measure, edgeboxes, recall, bing, salient, partition, boundary, mcg] [approach, closure, model, qualitative] [graph, proposed, method, problem, proximity, algorithm, function, energy, grouped, number, rank] [figure, real, continuity, simple, ratio, high] [sketch, image, retrieval, pair, query, scale, distance, formed, compute, local]
Data-Driven Depth Map Refinement via Multi-Scale Sparse Representation
Kwon, HyeokHyen, Tai, Yu-Wing, Lin, Stephen

cvpr2015papers_第275张图片
[learning, dataset, natural, set, human, address, noisy] [map, raw, training, learned, level, performance, higher, learn, size, table, previous, compared] [based, three, truth, park, edge] [depth, rgb, kinect, face, approach, joint, scene, reconstruction, nagdp, single, model, hole, ground, reconstruct] [data, method, sparse, practical, algorithm, missing, fusion, problem, order, min, arg] [figure, upsampling, equation, tof, captured, presented, high, ieee, quality, intensity, middlebury, supplemental, recovered] [dictionary, image, resolution, gradient, large, scale, texture, local, degradation, statistical, uncorrelated, geometric, exhibit, patch, bicubic, variation, illustrated]
Super-Resolution Person Re-Identification With Semi-Coupled Low-Rank Discriminant Dictionary Learning
Jing, Xiao-Yuan, Zhu, Xiaoke, Wu, Fei, You, Xinge, Liu, Qinglong, Yue, Dong, Hu, Ruimin, Xu, Baowen

cvpr2015papers_第276张图片
[learning, set, dataset, test, select, evaluation, randomly] [learned, feature, representation, discriminative, training, compared, table, designed, pedestrian, learn, recognition, learns] [based, updated, appearance] [approach, term, represent, ensure, provide] [mapping, min, function, coding, sparse, ith, formula, objective, matrix, rank, regularization, number, experimental, china, method, updating, cluster, proposed, denote, close, lagrange] [camera, ieee, conference, figure, captured, technique, intrinsic] [person, image, dictionary, gallery, pair, matching, probe, rate, sscdl, coupled, sampling, discriminant, viper, rdc, rplm, well, kissme, prid, distance, patch, converted, metric, characterize, corresponding]
Unsupervised Visual Alignment With Similarity Graphs
Shokrollahi Yancheshmeh, Fatemeh, Chen, Ke, Kamarainen, Joni-Kristian

cvpr2015papers_第277张图片
[visual, unsupervised, work, learning, set, manually, semantic, maximization, select, labeled, annotation] [feature, class, effective, imagenet, accuracy, achieved, table, representation, better] [object, seed, cost, based, measure, detection, bounding] [alignment, face, approach, manual, aligned, term] [graph, assignment, problem, method, algorithm, aij, matrix, approximation, number, function, space, solution, selection, optimization, computation, proposed, computational, form, solved, ransac, minimum, data] [figure, direct, generalized, avoid, distortion] [similarity, image, matching, good, geometric, transformation, congealing, dijkstra, spatial, ideal, centrality, fast, benchmark, large, stepwise, mst, best, lankinen, distance, local, huang, sift, moderately]
Person Re-Identification by Local Maximal Occurrence Representation and Metric Learning
Liao, Shengcai, Hu, Yang, Zhu, Xiangyu, Li, Stan Z.


[learning, occurrence, vision, dataset, set] [feature, performance, table, learn, representation, learned, training, accuracy, effective, applying, recognition, compared] [propose, score] [viewpoint, face, detected] [proposed, method, subspace, algorithm, analysis, rank, time, dimensional, dimension, function, problem, low, eigenvalue] [conference, computer, camera, pattern, ieee, color, robust, illumination, international, figure, horizontal, observed, generalized, second] [person, metric, local, xqda, lomo, discriminant, image, retinex, distance, viper, matching, database, called, maximal, kissme, comparison, applied, best, invariant, cumulative, cvpr, bayesian, well, cmc, derived, texture, grid, itml, histogram, lmnn, large, cuhk, computed]
A Dynamic Programming Approach for Fast and Robust Object Pose Recognition From Range Images
Zach, Christopher, Penate-Sanchez, Adrian, Pham, Minh-Tri

cvpr2015papers_第278张图片
[set, work, predicted, dataset, evaluation, learning, task, consistency] [feature, training, recognition, input, def, level, discriminative, performance, global, highly] [object, detection, based, frame, three, interest] [depth, pose, model, surface, approach, point, coordinate, estimation, shape, message, occupancy, normal, spin, cluttered, current, full, cad, commodity] [data, sample, time, method, range, computation, algorithm, number, respect, pairwise, binary, random, choice] [pixel, minimal, robust, sensor, observed, color, estimating, real] [local, matching, image, descriptor, distance, sampling, putative, respective, compatibility, invariant, inlier, determine, compute, true, hough, query]
Fast and Flexible Convolutional Sparse Coding
Heide, Felix, Heidrich, Wolfgang, Wetzstein, Gordon

cvpr2015papers_第279张图片
[learning, dataset, vision, set, evaluate, datasets, existing, strategy, work, example] [convolutional, learned, feature, better, convolution, table, previous, inversion, standard, training, size] [boundary, cost, propose, framework, supported] [reconstruction, allows, approach, described, term] [sparse, method, proposed, objective, coding, algorithm, data, csc, optimization, solve, formulation, converges, argmin, general, linear, fruit, convergence, sum, number, problem, faster, solution, bristow, original, shrinkage, alternating, solved, solving, gaussian, note, convex, operator, time, proximal, derive] [figure, quality, frequency, computer, signal, compressive, inverse, imaging] [image, incomplete, local, psnr, demonstrate, city, well, sampled, spatial, splitting, compare, good]
Similarity Learning on an Explicit Polynomial Kernel Feature Map for Person Re-Identification
Chen, Dapeng, Yuan, Zejian, Hua, Gang, Zheng, Nanning, Wang, Jingdong

cvpr2015papers_第280张图片
[set, latent, learning, vision, positive, dataset, introduce, discover, correct, task] [feature, triplet, loss, table, negative, map, learn, performance, training, size, recognition, accuracy] [multiple, based, three, weight, utilize, object] [approach, face, mixture, single, term] [function, regularization, explicit, method, vector, kernel, sparsity, binary, linear, rank, problem, number, form, formulation, update, algorithm, data, proposed, objective, machine, respect, min, convex, constraint, parameter, national] [computer, figure, pattern, polynomial, equation, conference, camera, adopt, second, international] [similarity, person, matching, image, gallery, distance, patch, hij, viper, comparison, probe, grid, mahalanobis, metric, matched, descriptor, rate, best, cmc, compare, hik]
24/7 Place Recognition by View Synthesis
Torii, Akihiko, Arandjelovic, Relja, Sivic, Josef, Okutomi, Masatoshi, Pajdla, Tomas

cvpr2015papers_第281张图片
[visual, work, baseline, google, example, dog, associated] [recognition, map, top, compared, performance, representation, feature, fisher, standard, extracted] [appearance, based, generate, recall, structural, individual] [scene, viewpoint, depth, location, structure, position, model, challenging, approach] [method, sparse, original, number, compact, data, vector, note, proposed] [illumination, camera, figure, synthetic, perspective, ratio, captured] [place, image, query, virtual, vlad, matching, dense, synthesized, view, database, sampled, local, descriptor, synth, major, large, invariant, densely, synthesis, inlier, sift, tokyo, change, match, illustrated, imagery, geometrically, well, matched, grid, spatial, rootsift, synthesize, sparsely, sampling, closest]
Tree Quantization for Large-Scale Similarity Search and Classification
Babenko, Artem, Lempitsky, Victor

cvpr2015papers_第282张图片
[set, search, learning, dataset, datasets, evaluate, process, evaluation] [training, accuracy, global, advantage, learned, table, neural, representation, better, fisher] [assigned, recall, average, edge, performs] [tree, error, reconstruction, inference, structure, vertex] [otq, encoding, quantization, codebook, coding, product, codebooks, vector, opq, dimension, compressed, uncompressed, code, codewords, optimal, apq, codeword, number, optimized, euclidean, minimization, quantizer, main, sum, method, additive, complexity, approximate, faster, data, linear, binary, assignment, time, graph, composite, orthogonality, pairwise] [scalar, figure, rotation] [compression, nearest, neighbor, large, scheme, distance, retrieval, well, fast, query, compare, image]
Becoming the Expert - Interactive Multi-Class Machine Teaching
Johns, Edward, Mac Aodha, Oisin, Brostow, Gabriel J.

cvpr2015papers_第283张图片
[teaching, learning, human, student, strategy, active, set, teacher, interactive, visual, datasets, task, work, ability, dataset, labeled, seer, mechanical, incorrect, label, swp, showing, provided, scc, example, sbatch, adapt, correct, future, web, test, amount, curriculum, goal, perform, domain, turk, access, feedback] [class, feature, performance, table, learn, testing, better, compared] [based, teach, average, truth, object, individual] [model, current, ground, challenging, approach, prior, uncertainty] [machine, random, number, matrix, time, method, optimal, problem, data, batch, space, selection, distribution] [figure, real, directly, estimate, presented, computer] [image, similarity, computed]
Encoding Based Saliency Detection for Videos and Images
Mauthner, Thomas, Possegger, Horst, Waltner, Georg, Bischof, Horst

cvpr2015papers_第284张图片
[human, visual, dataset, perform, set, evaluate, supervised, content, weakly] [global, map, training, feature, compared, representation, input, top, detector, applying] [saliency, salient, video, object, ebsg, foreground, based, detection, activity, likelihood, dynamic, contrast, recall, individual, background, ucf, static, appearance, precision, segmentation, action, box, analyzing, ebsl, average, gestalt, border, bounding, combination, propose, asd, lab] [occupied, joint, allows, estimation, described, challenging, approach, cluttered, solely] [encoding, number, distribution, data, proposed, method, maximum, general] [motion, color, figure, camera, pixel, directly, optical, enforce, estimating] [image, local, region, histogram, weighted, magnitude]
Unsupervised Object Discovery and Localization in the Wild: Part-Based Matching With Bottom-Up Region Proposals
Cho, Minsu, Kwak, Suha, Schmid, Cordelia, Ponce, Jean

cvpr2015papers_第285张图片
[unsupervised, discovery, learning, collection, common, supervised, work, dataset, weakly, colocalization, set, source, evaluation, visual, evaluate, example, horse] [class, table, performance, better, discriminative, negative, fully, top, bus, pascal] [object, corloc, corret, foreground, multiple, average, standout, cosegmentation, score, target, localized, box, measure, detection, candidate, based, state, localize, bounding, contrast, viewed, objectness, appearance] [localization, potential, challenging, approach, dominant, offset, probabilistic, single, current] [method, algorithm, proposed, form, number, note, problem] [figure, robust] [region, image, matching, neighbor, retrieval, hough, match, best, metric, correspondence, well]
Semantic Part Segmentation Using Compositional Model Combining Shape and Appearance
Wang, Jianyu, Yuille, Alan L.

cvpr2015papers_第286张图片
[semantic, animal, learning, horse, set, labeling, relation, work, human, dataset, side] [parsing, learn, pascal, voc, performance, fully, combining, top] [segmentation, object, appearance, detection, node, bounding, box, segment, average, hierarchical, boundary, score, iou, edge, dynamic] [compositional, model, mixture, inference, head, leaf, shape, structure, location, neck, pose, represent, leg, torso, pixelwise, body, deformable, challenging, localization, tree, capture, landmark, constrained] [algorithm, method, number, problem, min, symmetric, denote, experiment, energy, lower, function, consider, graph, vector, complexity, linear, approximate] [figure, result, develop] [region, best, image, distance, represents, transform]
Multiclass Semantic Video Segmentation With Object-Level Active Inference
Liu, Buyu, He, Xuming

cvpr2015papers_第287张图片
[semantic, learning, set, active, labeling, consistency, select, relation, introduce, evaluate, entropy, selected, subset, focus, include] [crf, class, performance, accuracy, table, learn, build, multiclass, achieve, improve, parsing, learned, jointly, train, better] [object, video, supervoxel, segmentation, based, policy, subgraph, three, hypothesis, subgraphs, hmn, detection, average, state, generate, singleton, foreground, score, imitation, formulate, multiple, sequence] [inference, model, scene, potential, joint, occlusion, ordering, unary, full, approach, structure, current, forest] [number, method, selection, pairwise, optimal, algorithm, approximate, random, proposed] [figure, informative, pixel, develop, liu, computer] [dense, pair, compute, choose, large]
Global Supervised Descent Method
Xiong, Xuehan, De la Torre, Fernando

cvpr2015papers_第288张图片
[vision, set, strategy, dataset, supervised, labeled, subset, active, work, domain, evaluation, learning, positive] [training, feature, previous, global, generic, performance, learns, recognition, standard, context] [tracking, partition, three, object, based, appearance, track, average, state, policy] [sdm, face, gsdm, descent, facial, shape, model, extrinsic, alignment, pose, error, landmark, recorded, posit, dhd, single, nls, ceh, estimation, xiong] [parameter, optimization, function, data, method, nonlinear, matrix, space, exists, proposed, algorithm, number, minimizing, case, extended, condition, order, iteration, hessian, kernel] [computer, camera, pattern, calibration, figure, rotation, simple] [local, image, gradient, large]
3D All The Way: Semantic Segmentation of Urban Scenes From Start to End in 3D
Martinovic, Andelo, Knopp, Jan, Riemenschneider, Hayko, Van Gool, Luc

cvpr2015papers_第289张图片
[semantic, labeling, weak, work, test, set, learning, evaluation, labeled, knowledge, label, create, novel] [accuracy, crf, parsing, van, performance, class, grammar, connected, table, recognition, pascal, top] [segmentation, propose, individual, object, based, box, generated, group, multiple, detection, iou, three] [facade, point, approach, cloud, architectural, pcl, initial, structure, scene, procedural, shape, reconstruction, urban, split, depth, timing, alignment, modeling, height, projected, sfm, building] [method, optimization, random, time, original, graph, number, data, main] [figure, separation, speed, estimated] [image, best, descriptor, compare, view, city, distance]
Robust Video Segment Proposals With Painless Occlusion Handling
Wu, Zhengyang, Li, Fuxin, Sukthankar, Rahul, Rehg, James M.

cvpr2015papers_第290张图片
[vision, evaluation, annotator, system, set, dataset, report, human, semantic, work, european] [training, size, testing, previous, pool, feature, table, better] [video, segment, object, frame, segmentation, track, tracking, lso, appearance, overlap, truth, score, framework, proposal, target, detection, handling, average, long, propose, lsos, start, tracked, hierarchical, merge, backtracking, moving, updated, preliminary, tracker, generated, entire, temporal] [occlusion, ground, occluded, model, complete, free, approach, step, regression, single, starting] [algorithm, number, matrix, method, time, original, clique, note, order, problem, graph, move] [computer, conference, ieee, pattern, motion, international, handle, figure, assume, tested] [image, metric]
Learning To Look Up: Realtime Monocular Gaze Correction Using Machine Learning
Kononenko, Daniil, Lempitsky, Victor

cvpr2015papers_第291张图片
[system, set, learning, vision, test, dataset, people] [training, input, output, feature, recognition, replacement, decision, learn, version, trained, learned] [eye, based, video, score, box, appearance, bounding, node, threshold] [gaze, tree, redirection, face, single, forest, location, columbia, leaf, head, monocular, approach, point, error, pose, realistic, estimation, split, alignment] [method, number, vector, machine, random, sample, problem, data] [pixel, correction, computer, pattern, figure, conference, camera, coherence, contact, small, result, direction, international, second, motion, vertical, hardware, color] [image, view, randomized, difference, ten, compatibility, corresponding, compare, synthesize, person]
Kernel Fusion for Better Image Deblurring
Mai, Long, Liu, Feng

cvpr2015papers_第292张图片
[vision, set, european, conditional, dataset, existing, evaluate, provided, relation, predict] [better, training, performance, improve, combining, input, trained, train, combine, compared, learn, feature, outperform, testing] [individual, multiple, generate, framework, success, combination, average, generated, measure] [estimation, model, error, neighboring, approach, regression] [kernel, fusion, method, element, gaussian, random, function, denotes, effectively, problem, underlying, fpq, energy] [deblurring, deconvolution, computer, conference, figure, blur, ieee, pattern, quality, result, ratio, blind, blurry, estimated, blurred, motion, sharp, gcrf, averaging, sbdb, produce, recover, classical, cho, camera, develop, international, goldstein] [image, true, best, local, rate, good, large, well, blurring]
DASC: Dense Adaptive Self-Correlation Descriptor for Multi-Modal and Multi-Spectral Correspondence
Kim, Seungryong, Min, Dongbo, Ham, Bumsub, Ryu, Seungchul, Do, Minh N., Sohn, Kwanghoon

cvpr2015papers_第293张图片
[visual, evaluation, set, learning] [pooling, receptive, discriminative, window, performance, training, feature, size, table, max] [adaptive, based, cost, weight, correlation, combination, propose, measure, reliable, average] [error, ground, evaluated, challenging, approach] [support, computational, modality, computation, method, sensitive, complexity, objective] [illumination, pattern, robust, ieee, stereo, conventional, optical, exposure, figure, intensity, multispectral, estimate, pixel, middlebury, mpi, varying, motion, color, acm, disparity] [descriptor, dense, matching, dasc, image, sampling, local, correspondence, fast, sift, similarity, benchmark, randomized, computed, patch, compute, computing, daisy, ancc, inherent, unoccluded, wta, rsncc, registration, dramatically, gradient, densely]
GMMCP Tracker: Globally Optimal Generalized Maximum Multi Clique Problem for Multiple Object Tracking
Dehghan, Afshin, Modiri Assari, Shayan, Shah, Mubarak

cvpr2015papers_第294张图片
[set, association, hard, selected, scenario] [network, layer, input, size, global, table, performance, reported, better] [tracking, dummy, gmmcp, gmcp, appearance, node, tracker, tracklets, edge, track, multiple, cost, object, based, detection, integer, weight, adn, tracklet, propose, formulate, target, lot, parking, mot, town, online, segment] [model, program, occlusion, allows, quantitative] [graph, problem, solution, number, clique, maximum, cluster, method, data, proposed, optimization, batch, constraint, time, complexity, order, function, vector, multi, binary, form, formulation, variable, computational, original] [motion, generalized, real, second, equation, figure, length, small, paper] [implementation, similarity]
A Stable Multi-Scale Kernel for Topological Machine Learning
Reininghaus, Jan, Huber, Stefan, Bauer, Ulrich, Kwitt, Roland

cvpr2015papers_第295张图片
[learning, set, domain, vision, popular, address, study, positive, work] [performance, feature, table, bottleneck, map, class, input, recognition, standard, connected] [segmentation, propose, individual] [shape, point, structure, surface, approach] [persistence, kernel, topological, function, data, space, shrec, diagram, dkl, machine, wasserstein, note, persistent, solution, homology, analysis, hilbert, consider, order, exp, theorem, time, outex, form, denote, vector, parameter, dirac, respect, proposed, sublevel, lead, operator, problem, condition, additive, prove, bijections] [stability, result, figure, high, directly, constant, real, synthetic] [distance, scale, texture, stable, hks, called, metric, image, lipschitz, retrieval, corresponding, partial, large]
Indoor Scene Structure Analysis for Single Image Depth Estimation
Zhuo, Wei, Salzmann, Mathieu, He, Xuming, Liu, Miaomiao

cvpr2015papers_第296张图片
[predicted, vision, encode, predict, dataset, semantic, probability, exploit, test, prediction, retrieve, set, label] [global, training, layer, trained, table, additional, input, ablation, larger] [superpixel, superpixels, three, based, employed, state, boundary, candidate, box, propose, appearance] [depth, scene, estimation, single, model, structure, approach, normal, layout, term, indoor, potential, detailed, inference, rmrc, reasoning, provide, unary, orientation, discrete, represent, semanticdepth, typically, error, relies, rgb] [method, energy, note, analysis, pairwise, data] [pixel, computer, ieee, conference, estimated, pattern, plane, figure, estimate, geometry] [image, local, region, corresponding, query, large]
A Graphical Model Approach for Matching Partial Signatures
Du, Xianzhi, Doermann, David, Abd-Almageed, Wael

cvpr2015papers_第297张图片
[dataset, topic, set, datasets, supervised, work, draw, latent, visual, document, address, probability, predicted, test, refer, existing] [context, accuracy, table, training, build, contour, feature, compared, performance, previous, global] [salient, based, three, hierarchical] [full, shape, model, term, built, point] [number, method, dirichlet, rank, pairwise, lsh, distribution, problem, sample, data, algorithm, journal, space, corresponds, cluster, column, row] [figure, equation, total, observation, estimate, international, estimated] [partial, signature, matching, region, hdp, tobacco, query, vocabulary, author, variational, impact, tth, local, slda, umd, image, nth, reference, histogram, large]
Classifier Based Graph Construction for Video Segmentation
Khoreva, Anna, Galasso, Fabio, Hein, Matthias, Schiele, Bernt

cvpr2015papers_第298张图片
[set, learning, type, learnt, work, subset, common, test, baseline, visual] [performance, validation, training, feature, improves, better, connected, improve, table, improvement, level, learn] [video, segmentation, edge, superpixels, based, topology, partitioning, bpr, superpixel, wij, vpr, boundary, appearance, precision, galasso, measure, frame, temporal, object, propose, ncl, average, hierarchical, partition, mjf] [model, shape, approach, provide, allows, challenging, procedure, split] [graph, proposed, number, consider, spectral, clustering, construction, selection, method, algorithm, exp, optimal, literature, matrix] [figure, calibration, motion, length, optical] [similarity, image, best, distance, large, neighbor, compute]
Pushing the Frontiers of Unconstrained Face Detection and Recognition: IARPA Janus Benchmark A
Klare, Brendan F., Klein, Ben, Taborsky, Emma, Blanton, Austin, Cheney, Jordan, Allen, Kristen, Grother, Patrick, Mah, Alan, Burge, Mark, Jain, Anil K.

cvpr2015papers_第299张图片
[subject, dataset, search, protocol, set, provided, baseline, open, human, manually, datasets, annotation, randomly, operational, collected, automated] [recognition, accuracy, detector, training, performance, testing, reported, table] [detection, bounding, box, multiple, detect, video, key, plot, interest, threshold] [face, template, pose, performed, full, approach, commodity, geographic, ground, facial, allows, single] [number, algorithm, proposed, exists, data, order, scalable] [figure, ieee, captured, paper, computer] [unconstrained, image, probe, imagery, false, gallery, benchmark, rate, person, janus, compare, gots, iarpa, varied, well, overview, match, roc, accept, openbr, database, sampled, corresponding, limited, large, lfw]
Video Summarization by Learning Submodular Mixtures of Objectives
Gygli, Michael, Grabner, Helmut, Van Gool, Luc

cvpr2015papers_第300张图片
[summary, summarization, learning, dataset, set, user, interestingness, select, predict, web, text, representative, representativeness, evaluation, document, existing, maximization, visual, subset, create, automatically, interesting, domain, hard, task, evaluate, prediction, type, created, annotated] [learn, max, training, previous, input, deep, compared, global, follow, jointly, trained, structured, performance, optimizes, version] [video, multiple, score, segment, egocentric, short, temporal, based, frame, cost, event, generated, longer, recall, uniformity, propose] [approach, inference] [submodular, method, objective, order, function, optimize, solution, greedy, data, selection, optimizing, algorithm, problem, approximation, marginal, optimal, optimized, arg, optimization] [uniform, acm, length] [reference, well, good, image, best]
Mining Semantic Affordances of Visual Object Categories
Chao, Yu-Wei, Wang, Zhan, Mihalcea, Rada, Deng, Jia

cvpr2015papers_第301张图片
[verb, semantic, affordances, visual, affordance, knowledge, work, human, plausibility, language, google, mining, collaborative, vision, synset, car, predict, sense, horse, bird, selectional, cat, operate, dog, common, cow, boat, bottle, bicycle, search, airplane, plant, side, sheep, noun, couch, dining, returned, textual, motorcycle, potted, text, learning, dataset, question, predicting, hurt, answer, natural, visualness] [train, table, recognition, chair, feed, bus, pascal, representation, top] [object, action, precision, average, recall, score, based, truth] [ground, prior, performed] [number, matrix, problem, computational, data, lsa] [computer, ieee, international, pattern, figure] [person, image, false]
Fast Action Proposals for Human Action Detection and Search
Yu, Gang, Yuan, Junsong

cvpr2015papers_第302张图片
[human, search, set, dataset, evaluation, positive, green, evaluate, work] [table, compared, max, performance, recognition, feature, reduce, training, larger, highly] [action, detection, path, bounding, based, score, video, proposal, box, msrii, object, actionness, recall, candidate, average, ucf, overlap, trajectory, locate, coverage, segmentation, handclapping, three, precision, generated, boxing, temporal, online, frame, cost, propose] [performed, localization, challenging, detected, approach, potential, cad] [algorithm, random, maximum, number, problem, computational, proposed, method, satisfy, greedy, consider, data, space, linear] [motion, red, figure, high, ratio, presented, blue] [best, applied, well, dense, local, unconstrained, fast, large, good]
A Dataset for Movie Description
Rohrbach, Anna, Rohrbach, Marcus, Tandon, Niket, Schiele, Bernt

cvpr2015papers_第303张图片
[movie, description, semantic, visual, vision, audio, corpus, dataset, sentence, linguistics, smt, association, human, natural, language, script, verbnet, learning, datasets, work, open, automatic, sense, annotated, manually, word, european, evaluation, text, meeting, novel, annual, describing, generating, wsd, verb, select, discussion, understand, source, automatically] [recognition, table, parsing, neural, performance, representation] [video, extract, role, object, multiple] [approach, aligned, alignment, provide, rely, align, scene, manual, model] [computational, data, method, number, proposed, time] [conference, computer, ieee, international, pattern, parallel, translation, total, length] [image, large, nearest, compare, comparing]
Reconstructing the World* in Six Days *(As Captured by the Yahoo 100 Million Image Dataset)
Heinly, Jared, Schonberger, Johannes L., Dunn, Enrique, Frahm, Jan-Michael

cvpr2015papers_第304张图片
[photo, visual, set, processing, dataset, datasets, strategy, add, system, perform, discover, process, leverage, work] [connected, representation, size, feature, table, achieve, recognition, smaller] [overlap, based, propose, detection, memory, merging, candidate, framework] [scene, approach, modeling, model, single, reconstruction, sfm, current] [cluster, data, number, method, proposed, order, scalability, computation, time, minimum] [component, computer, figure, disk, enable] [image, iconic, geometric, streaming, retrieval, frahm, register, sift, query, internet, registration, compute, rate, connectivity, large, match, scale, connecting, vocabulary, nearest, unordered, pair, overlapping, major, matching, registered, limited, leveraging, yahoo]
Pairwise Geometric Matching for Large-Scale Object Retrieval
Li, Xinchao, Larson, Martha, Hanjalic, Alan


[set, system, visual, consistency, evaluate, strategy, work, datasets, ranking, ranked] [performance, feature, map, table, global, bof, improvement, compared, improving, achieved, improve, generic] [object, salient, individual, based, three, multiple, bin, start] [point, initial, model, typically, step, approach] [pairwise, method, proposed, time, scaling, number, case, computational, hamming, experimental, exploited, binary, order, exploiting, original, code] [rotation, figure, high] [image, geometric, spatial, matching, retrieval, correspondence, local, pgm, scale, transformation, query, large, hough, cvpr, hpm, oxford, barcelona, matched, scheme, comparison, applied, weighting, reranking, voting, vocabulary, derived, pair, computed, database, performing, best, similarity, illustrated]
Video Magnification in Presence of Large Motions
Elgharib, Mohamed, Hefeeda, Mohamed, Durand, Fredo, Freeman, William T.

cvpr2015papers_第305张图片
[processing, noisy, reveal, user] [gate, global, larger, input, stage, feature] [sequence, video, frame, generate, youtube, temporal, foreground, temporally, background, generated, mask, parking, interest, entire, eye, handling, nearby, moving] [approach, addition, current, model] [original, time, number, composite] [motion, eulerian, dvmag, lagrangian, generates, examined, stabilization, technique, small, optical, magnify, dashed, white, yellow, figure, estimate, subtle, liu, matting, roi, stabilized, red, slice, integrity, remove, bulb, blue, klt, phase, rectangle, vibration, acm, magnifying, compositing, examine, magnified, gun, estimated] [large, reference, image, ssim, region, remaining, blurring, texture, synthesis, compare]
Face Video Retrieval With Image Query via Hashing Across Euclidean Space and Riemannian Manifold
Li, Yan, Wang, Ruiping, Huang, Zhiwu, Shan, Shiguang, Chen, Xilin

cvpr2015papers_第306张图片
[learning, set, common, character, task, supervised, multimodal, work, natural, shot] [training, svms, map, representation, recognition, train, feature, achieve] [video, multiple, based, frame, precision, big, average, recall] [face, single, supplementary] [hash, space, kernel, riemannian, euclidean, covariance, hashing, hamming, heterogeneous, method, manifold, hilbert, matrix, data, represented, modality, optimization, original, bang, iterative, case, function, update, algorithm, code, linear, optimize, proposed, mlbe, discriminability, application, compact, smh, vampire, ith, reproducing, objective, plmh] [theory, length, traditional, paper] [image, retrieval, query, discriminant, database, large, comparison, similarity, distance, person, local, limited]
Effective Face Frontalization in Unconstrained Images
Hassner, Tal, Harel, Shai, Paz, Eran, Enbar, Roee

cvpr2015papers_第307张图片
[photo, vision, learning, soft, labeled, appear, work, system] [recognition, performance, feature, reported, previous, input, frontal, standard, hybrid, accuracy, effective, better, highly] [based, appearance] [face, facial, visibility, shape, surface, model, alignment, estimation, single, gender, pose, aligned, approach, coordinate, detected, facing, visible, rough] [method, matrix, projection, symmetric, problem, code, approximation, guarantee, order, university] [symmetry, pixel, figure, estimate, pattern, produced, geometry, produce, estimating, result, estimated, computer, varying, rotation] [frontalized, reference, query, frontalization, view, image, lfw, unconstrained, adience, despite, corresponding, similarity, lbp, local, benchmark, well, fplbp, funneled, extreme]
Sparse Representation Classification With Manifold Constraints Transfer
Zhang, Baochang, Perina, Alessandro, Murino, Vittorio, Del Bue, Alessio

cvpr2015papers_第308张图片
[embedding, set, learning, vision, transfer, test, dataset, process, subject] [recognition, representation, training, performance, table, input, testing, previous, compared, feature, size] [based, action, framework, sport, ucf, including] [face, structure, fact, allows, alignment, approach] [manifold, data, method, sparse, optimization, src, problem, linear, proposed, geodesic, solve, variable, objective, solution, sample, original, admm, coding, constraint, function, algorithm, matrix, mct, vector, optimized, lagrange, augmented, solved, linearly, machine, space, initialize, fddl, minimize, cloned, general, alternating, minimization, projection, digit, lie, experimental] [ieee, pattern, computer, paper, robust, direction, illumination, traditional, embedded] [dictionary, scheme, distance, database, best, image, nearest, lfw]
SALICON: Saliency in Context
Jiang, Ming, Huang, Shengsheng, Duan, Juanyong, Zhao, Qi

cvpr2015papers_第309张图片
[mouse, visual, dataset, human, paradigm, coco, natural, osie, datasets, salicon, amt, vision, sauc, collection, category, attentional, collect, crowdsourcing, system, set, understand, work, evaluation, peripheral, collected, task, web, psychophysical, cursor] [map, context, contextual, percent, training, performance, width] [saliency, eye, object, tracking, attention, average, attract, propose, viewed, based, record] [scene, gaze, current, procedure] [data, number, function, sample, low, computational, method, gaussian, general] [figure, center, viewing, computer, relative, blur, high, blending, bias, sensitivity] [image, resolution, large, spatial, practice, distance, rate]
Displets: Resolving Stereo Ambiguities Using Object Knowledge
Guney, Fatma, Geiger, Andreas

cvpr2015papers_第310张图片
[vision, set, semantic, european, evaluation, category, processing] [recognition, map, performance, input, class, training, table, cnn] [object, superpixels, segmentation, superpixel, based, combination] [model, displets, error, kitti, displet, cad, unary, reconstruction, smoothness, scene, hull, mesh, shape, challenging, term, single, surface, depth, current, inference, sgm, joint, plausible] [number, machine, analysis, proposed, method, energy, intelligence, denotes, space, data, problem, order, encourage, random, respect, algorithm] [stereo, computer, ieee, disparity, pattern, international, plane, pixel, figure, simple, robust, inverse] [image, matching, pair, well, large, local, corresponding, dense, illustrated]
Fusion Moves for Correlation Clustering
Beier, Thorsten, Hamprecht, Fred A., Kappes, Jorg H.

cvpr2015papers_第311张图片
[labeling, work, evaluation, set, search, positive, natural] [better, size, improve, network, performance, negative, connected, smaller, larger] [correlation, edge, segmentation, proposal, node, based, multicut, hierarchical, partitioning, weight, wij] [current, social, approach, single] [clustering, fusion, graph, energy, problem, move, cgc, polytope, solution, proposed, runtime, optimal, number, yij, algorithm, original, contracted, anytime, ehc, cut, iteratively, inner, contraction, solving, random, polyhedral, faster] [computer, noise, figure, quality, international, conference, high] [large, best, scale, image, weighted, implementation, fast]
Beyond Mahalanobis Metric: Cayley-Klein Metric Learning
Bi, Yanhong, Fan, Bin, Wu, Fuchao

cvpr2015papers_第312张图片
[learning, set, positive, labeled, datasets] [training, learn, input, table, log, class, performance, effectiveness, learned, better] [based, average, three, group] [point] [data, matrix, euclidean, method, linear, space, objective, pairwise, hyperbolic, symmetric, optimization, function, form, problem, repeat, iteration, riemannian, proposed, proposition, call, machine, constraint] [generalized, geometry, relative, figure, pattern, intersect, balance, small, projective] [metric, mahalanobis, lmnn, distance, mmc, elliptic, osr, dck, pubfig, gradient, transformation, image, nearest, cayleyklein, neighbor, dissimilar, called, similarity, fractional, wine, invariant, query, large, remaining, return, kind, outperforms]
Generalized Deformable Spatial Pyramid: Geometry-Preserving Dense Correspondence Estimation
Hur, Junhwa, Lim, Hwasup, Park, Changsoo, Chul Ahn, Sang

cvpr2015papers_第313张图片
[source, search, dataset, ssf] [pyramid, level, belief, table, previous, layer] [node, target, state, object, cost, considering, complex] [model, deformable, term, deformation, smoothness, viewpoint, child, approach, structure, quantitative, qualitative, offset] [algorithm, regularization, objective, number, pairwise, dimensional, analysis, space, function, method, optimization, original, effectively, optimal, case] [rotation, flow, pixel, generalized, figure, warping, translation, geometry, high, handle, varying, preserving, directly, center] [matching, scale, spatial, dsp, dense, sift, correspondence, image, vij, geometric, dff, parent, grid, gdsp, loopy, calculates, propagation, descriptor, daisy, local, distance, calculating, large, cell, filter, transform]
Discriminative Learning of Iteration-Wise Priors for Blind Deconvolution
Zuo, Wangmeng, Ren, Dongwei, Gu, Shuhang, Lin, Liang, Zhang, Lei

cvpr2015papers_第314张图片
[sun, dataset, learning, set, latent, natural, visually, work] [learned, training, learn, discriminative, map, table, better, compared, designed] [based, three, edge, strong, success, updated, threshold, making] [error, prior, estimation, model, quantitative] [method, kernel, algorithm, min, proposed, updating, alternating, parameter, adopted, minimization, time, regularization, function, shrinkage, sparse, constraint, operator, solution, iteration, running, solve, sparsity, regularizers] [blind, blur, deconvolution, figure, ratio, deblurring, gst, levin, jia, blurry, real, synthetic, cho, ieee, krishnan, competing, lee, adopt, estimated, axiter, total, estimate, directly, small, desired, deblurred] [image, gradient, applied, patch, initialized, comparison, fast]
Active Learning for Structured Probabilistic Models With Histogram Approximation
Sun, Qing, Laddha, Ankit, Batra, Dhruv

cvpr2015papers_第315张图片
[entropy, active, learning, annotation, set, unlabeled, highest, work, goal, probability, prediction, labeling, scoring, label, perform, diverse, probable, dataset, test, focus] [structured, map, output, training, performance, log] [segmentation, bin, object, node, monte, entire, deterministic, video] [model, approach, probabilistic, graphical, typically, mass] [distribution, gibbs, number, approximation, approximate, support, algorithm, space, binary, data, surrogate, random, proposed, inaccurate, graph, problem, machine, optimal, lemma, corresponds, computation, exact, markov, pairwise, literature, function, chain] [accurate, lagrangian, small, estimate, computer, estimated, high, unstructured] [image, histogram, compute, large, sampling, variational, outperforms, geometric, computing, partial, compare, true]
SOLD: Sub-Optimal Low-rank Decomposition for Efficient Video Segmentation
Li, Chenglong, Lin, Liang, Zuo, Wangmeng, Yan, Shuicheng, Tang, Jin

cvpr2015papers_第316张图片
[set, unsupervised, dataset, semantic, natural, label, work, evaluation, consistency] [discriminative, representation, feature, previous, global, performance, table, better, window] [video, segmentation, temporal, supervoxel, supervoxels, framework, hierarchical, replication, sold, based, bpr, indicates, three, shgb, recall, hgb, precision, appearance, average] [approach, prior, model, neighboring, current] [matrix, rank, algorithm, method, number, decomposition, low, proposed, denotes, graph, constraint, computational, formulated, optimal, superior, optimization, norm, solution, solved, spectral, science, linear, tuning, denote, min, analysis] [ieee, pattern, figure, motion, robust] [consistent, streaming, image, large, comparison, local, region]
Probability Occupancy Maps for Occluded Depth Images
Bagautdinov, Timur, Fleuret, Francois, Fua, Pascal

cvpr2015papers_第317张图片
[people, report, set, introduce, probability, vision, learning, datasets, human, dataset, type, work] [acf, performance, detector, hidden, log, hog, rectangular, trained, map] [detection, object, threshold, background, overlap, precision, recall, tracking, target, truth, detecting, multiple, evidence, bounding, based, segmentation] [depth, ground, dpom, model, moda, generative, approach, rgb, location, kinect, single, occupancy, silhouette, point, occluding, occluded, unihall, scene, posterior, inference] [distribution, algorithm, euclidean, data] [pixel, presence, camera, plane, conference, estimate, observed, computer, figure, pattern, assume, small, observing, ieee] [image, corresponding, computing, large, compare]
Articulated Motion Discovery Using Pairs of Trajectories
Del Pero, Luca, Ricco, Susanna, Sukthankar, Rahul, Ferrari, Vittorio

cvpr2015papers_第318张图片
[set, human, dataset, dog, selected, unsupervised, label, discovery, perform, discovering, work, purity] [recognition, better, hog, channel, class, discriminative, feature, top] [object, pot, tiger, action, foreground, video, trajectory, appearance, frame, interval, idtfs, segmentation, swing, temporal, based, individual, turning, uniformity, idtf, average, partition, walking, behavior, moving, hierarchical] [articulated, model, single, head, ordered] [method, clustering, number, cluster, pairwise, analysis, consider, move, close] [motion, relative, velocity, median, pattern, optical, periodic, ieee, second, total, figure] [anchor, descriptor, displacement, histogram, large, compare, pair, local, dense]
Just Noticeable Defocus Blur Detection and Estimation
Shi, Jianping, Xu, Li, Jia, Jiaya

cvpr2015papers_第319张图片
[natural, vision, set, work, processing, existing] [feature, map, input, representation, recognition, size, standard, learn] [based, detection, edge, mask, including, strength, strong, recall, propose, generate, precision] [estimation, depth, represent, single, slight] [method, sparsity, sparse, gaussian, number, analysis, data, machine, intelligence, kernel, proposed] [blur, clear, jnb, blurred, small, ieee, conference, figure, defocus, estimate, pattern, blurriness, computer, result, deblurring, sharp, motion, noticeable, frequency, camera, signal, blurry, refocus, handle, directly, generally, spatially, varying, produce, blind, traditional, shi] [image, dictionary, local, patch, region, corresponding, gradient, chakrabarti, well, comparison, difference, commonly, dxi]
Propagated Image Filtering
Rick Chang, Jen-Hao, Frank Wang, Yu-Chiang

cvpr2015papers_第320张图片
[process, highest, example, probability, relationship, processing, existing, expected, goal] [smoothing, context, input, output, apply, better, window, belief, standard, highly, intermediate, performance] [path, weight, based, viewed, calculate, utilize, propose] [adjacent, neighboring] [geodesic, note, algorithm, consider, computation, parameter, explicit, proposed, base, fusion, problem, solving, speedup] [pixel, figure, photometric, acm, result, produced, noise, preserving, pattern, small, color, verify] [image, propagation, bilateral, guided, mixing, psnr, ssim, spatial, denoising, dpf, filtering, textural, recursive, filter, distance, connecting, applied, performing, large, compare, choose, preserve, dgf, determine, calculates, calculated, desirable, difference, gradient, undesirable]
Exploiting Uncertainty in Regression Forests for Accurate Camera Relocalization
Valentin, Julien, Niessner, Matthias, Shotton, Jamie, Fitzgibbon, Andrew, Izadi, Shahram, Torr, Philip H. S.

cvpr2015papers_第321张图片
[set, predicted, entropy, predict, work, probability, randomly, vision, test] [training, table, train, top, accuracy, feature, better] [hypothesis, state, candidate, art, node, propose] [pose, regression, scene, relocalization, isotropic, tree, anisotropic, error, forest, mixture, model, continuous, approach, kabsch, uncertainty, leaf, point, described, coordinate, percentage, predicts, current, body, rigid] [proposed, energy, optimization, method, gaussian, note, distribution, space, update, random, empirical, function, ransac, main, convergence, algorithm, mapping, machine, gaussians] [camera, pixel, computer, pattern, robust, observed, estimate, slam] [sampling, sampled, good, corresponding, image, difference, comparison, major, distance, metric]
Visual Recognition by Learning From Web Data: A Weakly Supervised Domain Generalization Approach
Niu, Li, Li, Wen, Xu, Dong


[domain, latent, learning, web, wsdg, bag, test, visual, mil, set, source, dataset, label, weakly, supervised, learnt, privileged, adaptation, unseen, textual, discovering, indicator, vision, work, proceedsing, existing, select, instance, introduce, kodak, european, ccv] [training, generalization, learn, class, recognition, svm, better, feature, max, effectiveness, testing, inspired, setting] [target, multiple, video, event, based, three, propose, integrated] [approach] [problem, method, data, machine, optimization, constraint, vector, dual, solve, kernel, form, sample, order, exploiting, proposed, function, denote, number, objective, cluster, formulation, optimal, min] [conference, computer, ieee, international, robust, pattern, noise] [image, corresponding, best, well]
PatchCut: Data-Driven Object Segmentation via Local Shape Transfer
Yang, Jimei, Price, Brian, Cohen, Scott, Lin, Zhe, Yang, Ming-Hsuan

cvpr2015papers_第322张图片
[label, test, transfer, example, soft, horse, dataset, selected, work, set, semantic, labeling, interactive, learning, transferring, novel] [structured, window, table, transferred, performance, better, pascal, training, cascade, size, achieve] [segmentation, object, patchcut, mask, gbvs, jaccard, based, cpmc, saliency, grabcut, salient, thres, candidate, zkm, foreground, cost, performs, appearance, truth, score, boundary, three, weizmann] [shape, prior, term, ground, qualitative] [algorithm, energy, mrf, space, note, binary, function, upper, bound, method, problem, graph, solution, solved, thresholding, number, random] [figure, color, estimated, high, estimate, quality] [image, local, patch, matching, scale, constitute, large, compare, best, database]
Flying Objects Detection From a Single Moving Camera
Rozantsev, Artem, Lepetit, Vincent, Fua, Pascal

cvpr2015papers_第323张图片
[dataset, vision, correct, weak, learning, datasets, variety, set] [size, training, performance, achieve, train, hog, effective, detector, hybrid, smaller, adaboost] [aircraft, object, detection, uav, compensation, background, appearance, moving, average, drone, precision, subtraction, target, based, detect, video, complex, apparent, frame, uavs, compensate, course, illustrates, detecting, consecutive, three] [approach, collision, single, allows, rely, challenging, regressors, detected, position] [method, algorithm, machine, number, random, case, problem, close] [motion, optical, center, computer, conference, small, camera, figure, pattern, shift, stabilization, international, ieee, inside, second, horizontal, vertical] [image, large, fast, patch, aerial, comparison, compare]
Video Anomaly Detection and Localization Using Hierarchical Feature Representation and Gaussian Process Regression
Cheng, Kai-Wen, Chen, Yie-Tarng, Fang, Wen-Hsien

cvpr2015papers_第324张图片
[test, process, learning, visual, dataset, set, semantic, unusual, datasets] [ensemble, global, training, log, feature, representation, learned, knn] [gpr, detection, anomaly, stc, based, interaction, stip, event, abnormal, interest, stips, ibc, nearby, video, likelihood, detect, junction, hierarchical, subway, structural, multiple, three, behavior, indicates, frame] [model, normal, regression, inference, template, detected, point] [sparse, method, gaussian, proposed, time, data, codebook, algorithm, matrix, function, computation, space, note, respect, computational, vector] [pattern, figure, ieee, quality, noise, handle, observed, motion] [local, dense, qmul, geometric, similarity, june, spatial, rate, integral, roc]
FAemb: A Function Approximation-Based Embedding Method for Image Retrieval
Do, Thanh-Toan, Tran, Quang D., Cheung, Ngai-Man

cvpr2015papers_第325张图片
[embedding, learning, set, dataset, visual, relationship, work] [map, table, higher, compared, fisher, version, designed, improving, representation, generalization, improvement, gain, applying, learned] [framework] [coordinate, single, ensure, step] [function, coding, method, vector, approximation, dimension, problem, nonlinear, approximated, linear, objective, sum, dimensional, number, order, whitening, hessian, matrix, linearly, main, motivation, lemma, optimization, sparse, sample, nonnegative] [embedded, high, paper, rotation, assumption, second, figure] [image, faemb, vlat, tlcc, vlad, retrieval, local, anchor, temb, aggregated, applied, democratic, comparison, aggregating, vlatimproved, corresponding, aggregation, good, sift, descriptor, lipschitz, large, comparing, rsift, fair, normalization, similarity, appendix, compare, gradient]
Optimal Graph Learning With Partial Tags and Multiple Features for Image and Video Annotation
Gao, Lianli, Song, Jingkuan, Nie, Feiping, Yan, Yan, Sebe, Nicu, Tao Shen, Heng

cvpr2015papers_第326张图片
[learning, label, annotation, set, datasets, ssl, address, noisy, visual, existing, labeled, unlabeled, dataset, work] [performance, training, feature, map, representation, better, learn, adding, achieve, improved] [multiple, based, precision, video, recall, propose] [model, sij, approach, term, reconstruction, point] [graph, data, ogl, optimal, function, min, objective, tagged, construction, method, lgc, sparse, subspace, linear, number, update, manifold, proposed, problem, optimizing, regularization, university, iterative, rank, pairwise, fasttag, harmonic, yan, solve, rest] [ieee, robust, figure, high, color] [image, partial, similarity, lle, ixmas, compare, large, construct, local, variance]
FaLRR: A Fast Low Rank Representation Solver
Xiao, Shijie, Li, Wen, Xu, Dong, Tao, Dacheng

cvpr2015papers_第327张图片
[dataset, datasets, set, existing, learning, positive, study] [representation, table, achieves, global, feature, performance, recognition] [based, framework, three, calculate, generate] [face, term] [lrr, problem, time, algorithm, falrr, running, clustering, solving, data, rank, optimal, solve, svd, parameter, solution, matrix, objective, number, proposed, singular, complexity, theoretical, subspace, method, updating, skinny, theorem, iteration, min, resultant, denote, optimization, original, norm, denotes, harus, truncated, spectral, exactly, formulation, subproblems, note, lemma, nuclear, subproblem, column, low, alternating, solved, vector, speedup, convergence] [figure, solver, distributed, synthetic, total, direction] [lfw, large, corresponding, fast, called, compare]
Unifying Holistic and Parts-Based Deformable Model Fitting
Alabort-i-Medina, Joan, Zafeiriou, Stefanos


[vision, active, novel, search, set, work, probability, learning, popular, provided, associated] [previous, recognition, global, max, reported, learn, cascade, training, trained, discriminative, response, accuracy, combining] [appearance, object, interpretation, combination] [shape, model, deformable, landmark, face, probabilistic, approach, holistic, clms, aams, regression, alignment, constrained, term, prior, misalignment, generative, point, error, hdms, aic, pbdms, cootes, fitting, position, rlms, current, procedure, descent, independent, compositional] [arg, proposed, optimal, min, problem, formulation, algorithm, linear, method, experiment, note, regularized, function, solution, data, optimization, eigenvectors, normalized, distribution, original] [computer, conference, pattern, international, figure, inverse, accurate] [texture, local, image, similarity]
Joint SFM and Detection Cues for Monocular 3D Localization in Road Scenes
Song, Shiyu, Chandraker, Manmohan

cvpr2015papers_第328张图片
[system, set, understanding, work, green] [feature, accuracy, output, raw, detector, achieve, input, size, table, better] [object, bounding, detection, tracking, box, framework, tracked, based, truth, frame, cost, incorporate, moving, background, adaptive, incorporates, mechanism, complementary] [ground, sfm, localization, pose, joint, monocular, scene, driving, estimation, kitti, point, autonomous, model, prior, alignment, error, coordinate, adjustment, multibody, approach, initial, challenging, height, position] [optimization, sparse, note, method, energy, denoted] [plane, optical, estimated, camera, figure, estimate, high, accurate, epipolar, motion, intensity, quality, robust] [dense, consistent, image, demonstrate, well, fast, distance, computed, corresponding]
Subgraph Decomposition for Multi-Target Tracking
Tang, Siyu, Andres, Bjoern, Andriluka, Miykhaylo, Schiele, Bernt

cvpr2015papers_第329张图片
[set, probability, people, association, selected, subset, par] [performance, jointly, feature, connected, achieves, compared, learned] [subgraph, multicut, tracking, disjoint, detection, tracklet, cost, mota, frame, target, multiple, ilp, yvw, object, hypothesis, undirected, track, based, state, tracklets, edge, suppression, video, path] [model, unary, approach, initial, procedure, full] [problem, minimum, graph, number, respect, pairwise, solution, feasible, time, method, space, objective, optimization, proposed, greedy, optimal, data, linear, decomposition, function, formulation, solving, variable] [solver, figure, component] [characteristic, comparison, cycle, good, person, overlapping, spatial]
Supervised Discrete Hashing
Shen, Fumin, Shen, Chunhua, Liu, Wei, Tao Shen, Heng

cvpr2015papers_第330张图片
[supervised, learning, set, label, test, dataset, embedding] [training, loss, table, learned, map, compared, achieve, learn, performance, achieves, better, feature] [precision, propose, cyclic, based, generate] [discrete, approach, joint, coordinate] [hashing, binary, hash, sdh, code, problem, method, optimization, linear, data, proposed, fasthash, ksh, hamming, solve, min, ssh, bit, solution, optimal, bre, hinge, algorithm, nonlinear, kernel, mlh, function, original, regularization, matrix, time, number, mnist, comparative, form, maximum, formulation, graph, vector, iteratively, imh] [length, high, figure, ieee] [large, retrieval, image, distance, anchor, similarity, best, query]
Face Alignment Using Cascade Gaussian Process Regression Trees
Lee, Donghoon, Park, Hyunsin, Yoo, Chang D.

cvpr2015papers_第331张图片
[dog, learning, prediction, set, dataset, process, vision, test, randomly] [training, cascade, recognition, learned, log, smoothing, feature, input, extracted, compared, better, distant] [based, framework, nearby, extraction] [shape, regression, cgprt, face, split, gprt, predictive, facial, helen, tree, lfpw, landmark, crt, alignment, leaf, depicted, performed, described, gprts, fall, estimation, lbf, point, error] [gaussian, number, method, function, data, proposed, kernel, regularization, greedy, computation, note, variable, random, binary] [computer, figure, ieee, conference, estimate, pattern, pixel, equation, second, international] [sampling, local, comparison, retinal, computed, difference, similarity, gradient, consists, distance]
Hierarchical Sparse Coding With Geometric Prior For Visual Geo-Location
Gopalan, Raghuraman

cvpr2015papers_第332张图片
[learning, test, dataset, set, vision, visual, perform, latent, knowledge, utility, novel, collection, european, transferring, datasets, existing] [training, feature, recognition, svm, discriminative, representation, accuracy, learn, level] [appearance, hierarchical, average, recall, based, cost, truth, three, video] [location, approach, prior, ground, structure, inference, landmark, single] [data, sparse, space, heterogeneous, coding, function, vector, problem, number, code, method, matrix, manifold, machine, geodesic, random, form, algorithm, case, cluster, analysis, tangent] [conference, computer, ieee, parallel, pattern, figure, transport, mobile, uniform] [geometric, image, san, francisco, similarity, corresponding, contained, pertaining, matching, best, obtaining, compute, large]
Social Saliency Prediction
Soo Park, Hyun, Shi, Jianbo

cvpr2015papers_第333张图片
[predictor, predict, relationship, set, prediction, positive, people, predicted] [feature, representation, learn, top, context, apply, accuracy] [attention, saliency, group, interaction, detection, basketball, based, truth, multiple, team, detect, video] [social, joint, formation, gaze, dipole, ground, scene, detected, member, busker, orientation, model, mass, location, point, circumcircle, represent] [method, space, number, data, third, distribution, respect, analysis, note, computational, form] [center, figure, moment, captured, motion, direction, camera, circular, equation] [person, geometric, scale, spatial, rate, magnitude, view, image, compare, distance, false, true, large]
Reliable Patch Trackers: Robust Visual Tracking by Exploiting Reliable Patches
Li, Yang, Zhu, Jianke, Hoi, Steven C.H.

cvpr2015papers_第334张图片
[visual, positive, set, process, probability, ranking, label] [performance, negative, response, top, feature, map] [tracking, reliable, object, kcf, rpt, tracker, tracked, scm, tgpr, target, precision, struck, success, particle, threshold, tld, cxt, ope, asla, score, resample, state, background, bounding, box, trackable, track, employed, overlap, cel, employ, monte, carlo, correlation, framework, likelihood, vor, video, based, measure, foreground, plot, group, online, kwon] [location, error, approach, structure, represent, challenging, sequential, position, model, template] [proposed, method, base, function, denotes, experimental, algorithm] [motion, figure, estimate, robust, rotation, observation, illumination] [patch, scale, large, rate, image, fast, benchmark, scheme, implementation, hough, comparing]
Complexity-Adaptive Distance Metric for Object Proposals Generation
Xiao, Yao, Lu, Cewu, Tsougenis, Efstratios, Lu, Yongyi, Tang, Chi-Keung

cvpr2015papers_第335张图片
[set, search, evaluation, vision, work, process, novel, example] [performance, level, achieve, selective, output, larger, better, table, recognition, pascal] [object, superpixel, superpixels, edge, recall, iou, grouping, bounding, based, proposal, box, combination, cost, segmentation, mcg, threshold, dmin, candidate, mabo, overlap, auc, spa, detection, transition, merging, average, score, hierarchical, merged, multiple] [searching] [number, complexity, low, method, algorithm, time, graph, space] [color, high, figure, computer, ieee, conference, pattern, small, tested, international, total, produced] [distance, metric, image, dmax, area, spatial, large, best, texture, local]
Multihypothesis Trajectory Analysis for Robust Visual Tracking
Lee, Dae-Youn, Sim, Jae-Young, Kim, Chang-Su

cvpr2015papers_第336张图片
[visual, select, set, selected, test, search] [feature, table, better, previous, achieve] [tracker, forward, tracking, trajectory, backward, frame, object, target, appearance, mta, bounding, multiple, cyclic, three, based, struck, box, interval, multihypothesis, meem, consecutive, yield, updated, average, success, failure, weight, online, precision, employ, kcf, backwardly, employing, score, analyzing, measure, detect, tld] [position, robustness, single, current] [proposed, algorithm, vector, analysis, time, optimal, support, selection, denote, note, experimental] [component, figure, robust, conventional, estimated, illumination, result, motion, color, red, length, korea] [geometric, similarity, best, pair, benchmark, corresponding, image, discriminant, invariant]
Completing 3D Object Shape From One Depth Image
Rock, Jason, Gupta, Tanmay, Thorsen, Justin, Gwak, JunYoung, Shin, Daeyun, Hoiem, Derek


[novel, set, entropy, dataset, predict, retrieve, category, include, work] [exemplar, training, better, class, table, input, recognition] [object, based, detection, propose, target] [depth, mesh, model, shape, reconstruction, voxels, deformation, voxel, deformed, surface, point, complete, tps, approach, occupied, reconstruct, alignment, volume, full, single, aligned, forest, tree, silhouette, constrained, cluttered, control] [random, method, approximation, function, machine, linear, number, analysis, consider] [symmetry, observed, median, computer, figure, acm, ieee, plane, conference, international, technique, camera] [query, retrieved, image, similarity, matching, distance, view, retrieval, match, correspondence, matched, partial, corresponding, compute]
PAIGE: PAirwise Image Geometry Encoding for Improved Efficiency in Structure-From-Motion
Schonberger, Johannes L., Berg, Alexander C., Frahm, Jan-Michael

cvpr2015papers_第337张图片
[visual, evaluation, predict, incremental, positive, strategy, describe] [feature, stage, reduce, training, performance] [overlap, based, eye, propose, multiple] [location, orientation, approach, scene, sfm, volume, structure, reconstruction, model] [number, pairwise, method, approximate, time, computational, proposed, encoding, random, quadratic, data, matrix, scalable, low, ransac] [geometry, camera, figure, square, produce, small, rotation, translation, motion, relative, robust] [image, matching, paige, change, geometric, retrieval, rome, vocabulary, large, pair, match, exhaustive, trafalgar, london, tate, san, histogram, overhead, modern, marco, true, preemptive, false, scale, sift, overlapping, compute, similarity, descriptor, putative, fast, corresponding]
Single Target Tracking Using Adaptive Clustered Decision Trees and Dynamic Multi-Level Appearance Models
Xiao, Jingjing, Stolkin, Rustam, Leonardis, Ales

cvpr2015papers_第338张图片
[visual, set, work, test, select] [feature, level, decision, top, accuracy, performance, highly, discriminative, previous] [target, tracking, adaptive, middle, tracker, superpixels, bounding, appearance, background, candidate, box, object, overlap, online, adaptively, vot, superpixel, relearning, foreground, threshold, propose, success, frame, favour, three, combination, successive, based, scm, video, ivt, csk] [model, tree, robustness, current, continuously, approach, occlusion, represent] [method, algorithm, proposed, update, updating, number, minimum, data, ith] [clustered, robust, ratio, figure, high, simple, enable, illumination, pixel] [matching, image, best, local, region, benchmark, histogram, discriminating, computed, similarity, mth, unmatched, change]
CIDEr: Consensus-Based Image Description Evaluation
Vedantam, Ramakrishna, Lawrence Zitnick, C., Parikh, Devi

cvpr2015papers_第339张图片
[sentence, human, evaluation, description, cider, dataset, automatic, meteor, evaluate, bleu, automated, datasets, rouge, set, popular, existing, visual, language, evaluating, describing, uiuc, work, association, annotation, generation, natural, generating, protocol, coco, collect, describe, include, vision, learning, perform, judgment, abstract] [performance, accuracy, triplet, pascal, higher, better, recognition, version] [candidate, based, generated, correlation, propose, measure, score, truth, precision] [provide, ground] [machine, number, computational, proposed, method, form] [translation, computer, quality, high, relative, conference] [image, metric, reference, consensus, well, best, similarity, computed, commonly]
Category-Specific Object Reconstruction From a Single Image
Kar, Abhishek, Tulsiani, Shubham, Carreira, Joao, Malik, Jitendra

cvpr2015papers_第340张图片
[instance, automatic, learnt, vision, predicted, noisy, set, car, visual, work, dataset, test, predict] [pascal, training, voc, learn, fully, table, map, learned, recognition, subcategory, pipeline] [object, truth, bounding, segmentation, mask, based, detection] [shape, model, reconstruction, ground, silhouette, nrsfm, deformation, depth, single, rigid, viewpoint, point, approach, pose, deformable, mesh, morphable, capture, chamfer, projected] [method, data, projection, note, algorithm, energy, basis, optimization, proposed, analysis, machine] [computer, figure, conference, camera, high, ieee, estimate, quality, pattern, illumination, frequency] [image, distance, well, keypoints, scale, keypoint, local, large]
Book2Movie: Aligning Video Scenes With Book Chapters
Tapaswi, Makarand, Bauml, Martin, Stiefelhagen, Rainer

cvpr2015papers_第341张图片
[book, novel, text, chapter, set, dialog, perform, character, story, adaptation, source, textual, rich, game, harry, work, language, shot, evaluation, automatic, dlgs, semantic, potter, finding, learning, create] [performance, accuracy, table, global, max, neural] [video, series, node, shortest, path, truth, based, detection, edge, plot, multiple, three, score, assigned, indicates, complex, track] [alignment, scene, prior, face, ground, approach, allows, term, aligning, align, model] [data, graph, method, number, problem, analysis, corresponds, proposed, column] [figure] [matching, similarity, distance, large, best, list, corresponding, person]
Active Pictorial Structures
Antonakos, Epameinondas, Alabort-i-Medina, Joan, Zafeiriou, Stefanos

cvpr2015papers_第342张图片
[vision, active, set, existing, correspond] [recognition, training, trained, connected, table, applying, global] [appearance, object, cost, precision, employ] [shape, model, aps, landmark, deformation, prior, gmrf, normal, compositional, deformable, structure, face, pictorial, alignment, aams, jacobian, single, error, aam, generative, tree, location, facial] [graph, proposed, matrix, vector, distribution, optimization, random, method, covariance, algorithm, case, order, function, note, close, respect, analysis, multivariate, denotes, equivalent, form, denote, markov] [computer, ieee, inverse, conference, pattern, international, robust, accurate, high, figure] [pca, image, comparison, local, patch, statistical, best]
Project-Out Cascaded Regression With an Application to Face Alignment
Tzimiropoulos, Georgios

cvpr2015papers_第343张图片
[learning, set, work, system, active, process, vision, strategy] [training, performance, learn, learned, learns, level] [appearance, sequence, propose, employed, based, detection, publicly, bounding] [regression, shape, cascaded, facial, face, model, averaged, alignment, descent, deformable, ibug, jacobians, structure, jacobian, approach, sdm, afw, pose, prior, analytic, chehra, challenging, generative, current, lfpw, constrained, landmark, step] [problem, data, proposed, optimization, iteration, method, solution, hessian, vector, number, subspace, orthogonal, idea, note, algorithm, computational, update] [parametric, inverse, estimate, computer, accurate, noise] [image, large, variation, gradient, compute, local, computed, difference]
Target Identity-Aware Network Flow for Online Multiple Target Tracking
Dehghan, Afshin, Tian, Yicong, Torr, Philip H. S., Shah, Mubarak

cvpr2015papers_第344张图片
[association, learning, set, human, introduce, annotation] [network, global, structured, performance, detector, discriminative, highly, better, identity] [tracking, object, target, candidate, multiple, cost, based, detection, online, track, node, frame, fij, bounding, appearance, edge, video, overlap, tracker, lot, temporal, transition, segment, score, consecutive, assigned, struck, sequence, propose, gog, spot, three] [model, inference, location, manual, allows, penalizes] [method, data, constraint, solution, proposed, number, lagrange, problem, graph, relaxation, optimization, algorithm, function, vector, minimum, linear, greedy, solving, main, relaxing] [observation, lagrangian, figure, color, traditional] [spatial, best, pair]
Single Image Super-Resolution From Transformed Self-Exemplars
Huang, Jia-Bin, Singh, Abhishek, Ahuja, Narendra

cvpr2015papers_第345张图片
[natural, source, search, learning, visual, exploit, dataset] [training, better, input, performance, version, achieve, structured] [target, based, cost, propose, indicates, appearance] [urban, deformation, quantitative, single, scene, regression, model, shape, estimation, searching, detected, initialization, driven] [algorithm, matrix, method, space, proposed, parameter, mapping, number, function, simultaneously] [perspective, figure, planar, plane, estimate, ieee, geometry, second, acm] [image, patch, transformation, neighbor, internal, external, nearest, matching, geometric, scale, best, srcnn, dictionary, glasner, kim, scsr, database, patchmatch, transformed, downsampled, large, regular, bsd, comparison, psnr, indicate, local, fast, similarity]
A Statistical Model of Riemannian Metric Variation for Deformable Shape Analysis
Gasparetto, Andrea, Torsello, Andrea

cvpr2015papers_第346张图片
[set, dataset, learning, human, process, supervised, test, datasets, associated, example] [representation, training, performance, class, log, table, build, pyramid, global, belonging, follow] [based, object, density] [shape, mesh, centroid, model, approach, represent, deformable, lack, eigendecomposition, introduced] [matrix, eigenvector, spectral, geodesic, laplacian, eigenvectors, proposed, order, method, decomposition, number, eigenvalue, problem, orthogonal, respect, distribution, algorithm, computation, parameter, main, riemannian, analysis, manifold, permutation] [rotation, computer, ieee, conference, second, international, pattern, acm, assume, component, geometry, intrinsic, synthetic] [retrieval, invariant, distance, compute, computed, transformation, applied, descriptor, statistical, metric, spatial, june, eurographics, matching]
Representing 3D Texture on Mesh Manifolds for Retrieval and Recognition Applications
Werghi, Naoufel, Tortorici, Claudio, Berretti, Stefano, Del Bimbo, Alberto

cvpr2015papers_第347张图片
[set, work, vision] [recognition, representation, performance, compared, global, unit, reported, feature, extracted, table] [three, based, framework, appearance, action, score, pot, combination, object, sequence, middle] [face, shape, mesh, surface, facial, approach, facet, depth, ordered, expression, owl, full, triangular, single, point, procedure, variant, potential, normal, azimuthal] [fusion, analysis, binary, proposed, machine, method, random, function, construction, case, lower] [photometric, pattern, ieee, central, figure, geometry, computer, rotation] [texture, geometric, local, image, lbp, retrieval, eat, best, distance, histogram, grid, corresponding, comparison, descriptor, database, probe, ring, computed, representing, gap, meshlbp]
FlowWeb: Joint Image Set Alignment by Weaving Consistent, Pixel-Wise Correspondences
Zhou, Tinghui, Jae Lee, Yong, Yu, Stella X., Efros, Alyosha A.

cvpr2015papers_第348张图片
[set, consistency, collection, visual, unsupervised, work, source, vision, human, discovery, example, provided, dataset, car] [representation, global, size, accuracy, table, pascal, jointly, feature, max, better] [object, score, appearance, target, boundary, measure, three, propose] [alignment, tij, joint, model, shape, initial, sij, term] [pairwise, graph, update, method, number, optimization, lower, algorithm, iteration, subspace, denote, sample, regularization] [pixel, figure, optical, pattern, flow, accurate, phase, ieee] [image, correspondence, matching, cycle, consistent, keypoint, priority, dsp, flowweb, tkj, transitive, sift, pair, tik, compute, congealing, good, rasl, local, region, mobahi, dense, spatial, similarity]
Heteroscedastic Max-Min Distance Analysis
Su, Bing, Ding, Xiaoqing, Liu, Changsong, Wu, Ying

cvpr2015papers_第349张图片
[lda, dataset, set, datasets, separability, learning, latent, perform] [class, training, max, better, margin, log, fisher, standard] [three, average] [face, error, projected] [hmmda, whmmda, dimensionality, pairwise, ohmmda, scatter, proposed, original, mmda, hlda, apac, problem, analysis, reduced, orthogonal, heteroscedastic, data, linear, dimension, chernoff, reduction, gaussian, solution, yale, optimal, feasible, lower, qdf, min, marohmmda, whitening, subspace, general, bound, distribution, yaleface, matrix, objective, trace, quotient, criterion, space, supohmmda, number, relaxation, uci, whitened, homoscedastic] [separation] [distance, discriminant, difference, best, database, nearest, rate, transformation, weighted, transformed, corresponding]
Structural Sparse Tracking
Zhang, Tianzhu, Liu, Si, Xu, Changsheng, Yan, Shuicheng, Ghanem, Bernard, Ahuja, Narendra, Yang, Ming-Hsuan

cvpr2015papers_第350张图片
[visual, learning, set, relationship, existing, human] [representation, learn, global, learned, discriminative, performance, achieves] [tracking, target, appearance, object, based, structural, sst, particle, frame, candidate, track, tracker, online, propose, three, performs, lgt, mixed, rtct, mtt, abrupt, combination, dft, multiple] [model, represent, joint, structure, layout, drift, pose, template, error, occlusion, reconstruction, challenging, holistic, location] [sparse, proposed, algorithm, represented, method, note, matrix, number, sparsity, formulation, time] [figure, inside, robust, intrinsic, illumination, center, observation, second] [local, image, dictionary, corresponding, spatial, patch, sampled, partial, best, well, sampling, region]
Person Count Localization in Videos From Noisy Foreground and Detections
Chen, Sheng, Fern, Alan, Todorovic, Sinisa

cvpr2015papers_第351张图片
[people, noisy, work, add, correct, set, introduce, dataset, address, human] [pedestrian, accuracy, input, detector, previous, output, training, performance, smaller, train, hyper, adding] [count, foreground, detection, ilp, football, egr, counting, based, frame, truth, tracking, assigned, cla, american, video, edge, integer, localized, segmentation, crowded, object, score, tracker, dei, framework, detecting, assign, crowd, measure, activity] [localization, approach, ground, initial, forest, provide, prior, occlusion, single, program, error] [graph, problem, solution, number, missing, random, note, iterative, optimization, constraint, linear, order, iteration] [integrity, figure, small] [person, metric, large, false, isolated, corresponding, local, severe]
Transport-Based Single Frame Super Resolution of Very Low Resolution Face Images
Kolouri, Soheil, Rohde, Gustavo K.

cvpr2015papers_第352张图片
[dataset, set, test, learning, subject, amount, work, process, written, vision, instance] [training, learned, size, standard, testing, learn, performance] [based, combination, appearance, framework, likelihood, average] [face, model, modeling, introduced, described, reconstruct, approach, facial, prior, reconstruction] [method, optimal, linear, space, problem, proposed, low, function, principal, subspace, yaleb, note, original, nonlinear, data, eigenvectors, sparse, basis, spline, mapping, matrix, corresponds, optimization, journal] [figure, ieee, transport, reconstructed, computer, high, technique, lagrangian, pattern, international, intensity, interpolation] [image, sfsr, displacement, resolution, corresponding, degraded, cubic, ssim, gradient, grid, morphing, reference, nearest, compare, best]
Modeling Object Appearance Using Context-Conditioned Component Analysis
Turmukhambetov, Daniyar, Campbell, Neill D.F., Prince, Simon J.D., Kautz, Jan

cvpr2015papers_第353张图片
[visual, set, dataset, vision, learning, example, test, datasets, active, semantic, address] [context, training, recognition, hidden, learn, table, additional, log] [appearance, object, segmentation] [model, xij, face, reconstruction, cij, approach, error, hif, argmax, shape, term, generative, modeling, regression, inpainted, structure, allows] [vector, subspace, function, form, inpainting, number, analysis, solution, matrix, parameter, linear, general, space, case, gaussian, yij, principal, problem, proposed] [computer, equation, pixel, figure, conference, international, pattern, rotation, estimate, color, component, varying, parametric, translation, noise] [image, consists, corresponding, dense, sift, spatial]
Feature-Independent Context Estimation for Automatic Image Annotation
Tariq, Amara, Foroosh, Hassan


[annotation, visual, textual, test, process, automatic, tucker, set, strategy, description, probability, system, relevance, association, esp, word, evaluation, semantic, text, vision, iapr, learning, work, visually, task, popular, conditional, processing, dataset, keywords, pick, tag] [context, training, representation, performance, size, auxiliary] [group, based, recall, employed, video, three, evidence] [estimation, model, formation, represent, joint, distinctive, height, prior] [tensor, decomposition, vector, proposed, form, number, analysis, core, computational, problem, distribution, data, gaussian, denoted, computationally, represented, suitable] [estimate, computer, estimated, ieee, conference, international, figure, assume, high, inserted, assumed] [image, vocabulary, similarity, signature, grid]
Understanding Tools: Task-Oriented Object Modeling, Learning and Recognition
Zhu, Yixin, Zhao, Yibiao, Chun Zhu, Song

cvpr2015papers_第354张图片
[human, tool, physical, affordance, learning, ranking, vision, imagined, parse, concept, demonstration, task, essential, understanding, work, set, stone, rational, cracking, hammer, causal, shovel, household, animal, imagining, automation, paint, nut, ability, pgs, selected, visual, natural, annotated] [recognition, representation, learned, top] [object, action, three, target, based, tracking, cognitive, detection, sequence, video, including, temporal, framework] [hand, pose, robotics, scene, represent, inference, inferred, reasoning, force] [algorithm, functional, basis, graph, choice, consider, number, space, science, sample, optimal] [computer, conference, ieee, international, figure, typical, red, assume, blue, produced] [best, image, area, spatial, represents]
Temporally Coherent Interpretations for Long Videos Using Pattern Theory
Souza, Fillipe, Sarkar, Sudeep, Srivastava, Anuj, Su, Jingyong

cvpr2015papers_第355张图片
[vision, concept, semantic, correct, set, describing, generating, pick, knowledge, search, human, closed, learning, identifying] [performance, recognition, feature, window, representation, level] [temporal, video, bond, interpretation, complex, generator, based, multiple, ontological, action, segment, consecutive, structural, theoretic, framework, generated, individual, coherent, sequence, allowed, combinatorial, interaction, correctly, activity, average, depicts] [approach, represent, structure, inference, single, provide] [modality, number, space, form, data, time, graph, proposed, function, machine, main] [pattern, figure, conference, computer, ieee, quality, international] [indicate, best, illustrated, consists, local]
Robust Saliency Detection via Regularized Random Walks Ranking
Li, Changyang, Yuan, Yuchen, Cai, Weidong, Xia, Yong, Dagan Feng, David

cvpr2015papers_第356张图片
[ranking, vision, visual, dataset, set, evaluation, process] [input, map, recognition, dir, performance] [saliency, boundary, background, detection, foreground, salient, erroneous, superpixel, recall, final, based, precision, object, seed, contrast, three, controlling, weight, calculate, pbo, wij, propose, sstep, superpixels, attention, seg] [estimation, model, step, approach, prior, fitting, ground, adjacent] [random, proposed, regularized, algorithm, method, matrix, graph, manifold, vector, element, analysis, defined] [figure, pattern, ieee, computer, conference, removal, color, second, blue, left, pixel] [image, comparison, pca, outperforms, major, remaining, region]
Semantic Object Segmentation via Detection in Weakly Labeled Video
Zhang, Yu, Chen, Xiaowu, Li, Jia, Wang, Chen, Xia, Changqun

cvpr2015papers_第357张图片
[semantic, weakly, set, labeled, dataset, supervised, consistency, learning, noisy, selected, unsupervised, bird, label, perform, car, address] [performance, training, train, accuracy, table] [object, segmentation, video, detection, track, segment, appearance, framework, foreground, temporal, generate, background, segtrack, based, bounding, cost, tracking, spatiotemporal, threshold, frame, overlap, likelihood] [shape, approach, initial, step, term, joint, procedure, single, represent, provide] [proposed, method, problem, solving, objective, algorithm, note] [figure, estimated, preserving, high, pixel, paper, motion, relative] [region, area, consists, initialized, compare, consistent, statistical, outperforms]
Saturation-Preserving Specular Reflection Separation
Liu, Yuanliu, Yuan, Zejian, Zheng, Nanning, Wu, Yang

cvpr2015papers_第358张图片
[set, natural, vision, preference] [input, stage, class, global, transferred] [based] [surface, model, smoothness, term, fall] [method, problem, note, bound, upper, maximum, lower, space, linear, solve, sparsity, original, main, number, analysis, proposed, sparse, function, row] [specular, diffuse, chromaticity, color, saturation, hue, achromatic, highlighted, separation, computer, ratio, recovered, pure, chromatic, smooth, constancy, highlight, figure, pixel, piecewise, component, produced, ieee, recover, result, conference, removal, blue, suv, separating, red, intensity, removed, pattern, specularfree, tan, recovering, paper, separate, yellow, small] [image, anchor, local, propagation, spatial, isolated, well, partial, difference]
Unconstrained Realtime Facial Performance Capture
Hsieh, Pei-Lun, Ma, Chongyang, Yu, Jihun, Li, Hao

cvpr2015papers_第359张图片
[system, user] [input, map, identity, previous, performance, feature, designed, effective, training] [tracking, segmentation, video, frame, appearance, tracked, based, reliable, superpixel, dynamic, framework, superpixels, detection] [facial, face, model, depth, blendshape, capture, occlusion, realtime, shape, personalization, expression, current, occluded, deformation, rigid, animation, landmark, rgb, mesh, term, approach, uncontrolled, visibility, personalized, completion, modeling, vertex, penetration, ensure, prior, introduced, uninterrupted] [method, linear, data, laplacian, exponential, solve, sparse, binary, algorithm, spectral, exponentially, number] [figure, motion, acm, color, illumination, robust, accurate, lighting, handle, captured, caused, siggraph] [texture, large]
Automatic Construction Of Robust Spherical Harmonic Subspaces
Snape, Patrick, Panagakis, Yannis, Zafeiriou, Stefanos

cvpr2015papers_第360张图片
[set, provided, existing, automatically, perform, active, example] [training, class, input, representation, build, involves] [appearance, object, propose] [shape, facial, alignment, model, single, surface, provide, face, performed, helen, expression, morphable, aam, normal, allows] [order, subspace, basis, problem, matrix, decomposition, rank, method, low, algorithm, linear, number, harmonic, proposed, time, sparse, data, tensor, svd, solution, denotes, case, argmin, update, solved, solving, seek] [recover, photometric, recovered, lighting, robust, recovery, stereo, uncalibrated, ieee, spherical, accurate, illumination, figure, recovering, tom, component, frequency, directly, lambertian, light, unknown, optical, lee, technique, pixel, warped] [image, dense, correspondence, reference, large]
KL Divergence Based Agglomerative Clustering for Automated Vitiligo Grading
Das Gupta, Mithun, Srinivasa, Srinidhi, J., Madhukara, Antony, Meryl

cvpr2015papers_第361张图片
[label, expert, evaluation, set, task, work, generation] [feature, reported, size] [based, segmentation, hierarchical, score, generate, cost, merge, generated, segmenting, propose, hypothesis, segment, superpixels, swa, boundary, partially, truth, merging] [model, ground, represent, mixture] [clustering, proposed, method, algorithm, symmetric, number, note, analysis, data, mrf, cluster, graph, gaussian, denotes, random, markov, matrix, function, respect, exponential, convex, decomposition, machine] [pixel, skin, albedo, color, completely, shading, technique, figure, ieee, simple] [image, divergence, vitiligo, mncut, depigmented, disease, agglomerative, distance, region, well, bregman, missed, compare, depigmentation, patch, partial, area, similarity, neighborhood, local, comparison, weighted, clinical]
Real-Time Joint Estimation of Camera Orientation and Vanishing Points
Lee, Jeong-Kyun, Yoon, Kuk-Jin

cvpr2015papers_第362张图片
[noisy, set, system, datasets, existing, vision, mutually, evaluation] [performance, feature, detector, accuracy, jointly, table, unit] [frame, segment, based, state, three, tracking, propose, detection] [orientation, estimation, point, scene, joint, normal, model, approach, robustness, current, solid, detected] [method, proposed, orthogonal, vector, great, constraint, consider, singular, nonlinear, gaussian, matrix, space, covariance, projection, number] [camera, vps, rotation, manhattan, parallel, vanishing, spurious, accurate, robust, estimate, figure, measurement, estimated, synthetic, real, noise, circle, management, direction, center, rotational, sphere, plane, lbd, blue, estimating, observed, pynew] [image, matching, nearest, bayesian, descriptor, gradient]
A Low-Dimensional Step Pattern Analysis Algorithm With Application to Multimodal Retinal Image Registration
Addison Lee, Jimmy, Cheng, Jun, Hai Lee, Beng, Ping Ong, Ee, Xu, Guozhen, Wing Kee Wong, Damon, Liu, Jiang, Laude, Augustinus, Han Lim, Tock

cvpr2015papers_第363张图片
[multimodal, dataset, test, datasets, describe] [feature, higher, window, factor, achieve] [average, success, based, edge, detection, extraction, bin, contrast, eye] [step, mae, point, volume, error, robustness] [algorithm, method, number, proposed, low, form] [intensity, rotation, robust, color, figure, pattern, distributed, high, uniformly, angle] [image, retinal, registration, scale, corner, sift, severe, unhealthy, rate, fundus, vascular, geometric, lospa, descriptor, invariant, corresponding, transformation, local, register, matched, gradient, bifurcation, applied, harris, closest, mutual, change, repeatability, match, matching, connecting, difference, mee]
Video Co-Summarization: Video Summarization by Visual Co-Occurrence
Chu, Wen-Sheng, Song, Yale, Jaimes, Alejandro

cvpr2015papers_第364张图片
[mbf, visual, shot, set, summarization, summary, biclique, aca, unsupervised, jump, collection, subject, coc, learning, dataset, appear, selected, labeled, knowledge, relevant, human, discovering, novel, eiffel, content, discovered, visually, identify, work, example] [feature, top, compared, effectiveness, better, standard] [video, temporal, multiple, recall, truth, precision, score, three, generate, online, segmentation, walk, generated] [approach, model, ground, performed, capture, cij, single, represent, averaged] [number, algorithm, matrix, graph, method, cluster, clustering, suppose, sparsity, problem, solution, vector, analysis] [quality, figure, high, total] [maximal, query, bipartite, compute, pair, large, retrieved, computed, image]
Toward User-Specific Tracking by Detection of Human Shapes in Multi-Cameras
Huang, Chun-Hao, Boyer, Edmond, do Canto Angonese, Bibiana, Navab, Nassir, Ilic, Slobodan

cvpr2015papers_第365张图片
[human, visual, strategy, set, association, subject] [input, better, apply, feature, previous, discriminative, training, testing, standard] [tracking, overlap, framework, frame] [surface, error, pose, shape, icp, forest, joint, regression, vertex, voxel, approach, mesh, normal, model, generative, vsuf, deformation, meshhog, voxels, patchedol, ground, averaged, silhouette, capture, volumetric, skeletal, bpsvm, current, straka, cagniart, split, coordinate, offset, depth] [method, data, vector, time, outlier, random, proposed, denoted, close] [figure, estimated, motion, black] [local, reference, correspondence, fail, image, large, distance, matching, compare, medical, metric]
Unsupervised Learning of Complex Articulated Kinematic Structures Combining Motion and Skeleton Information
Jin Chang, Hyung, Demiris, Yiannis

cvpr2015papers_第366张图片
[learning, generation, human, description, set, novel, dataset, entropy] [feature, highly, smoothing, margin, performance, input, recognition] [skeleton, object, boundary, segmentation, complex, based, segment, cost, generated, target, merging, framework, generate, centre, factorisation, robot, adaptive, detection] [articulated, structure, kinematic, estimation, body, icub, point, error, skeletal, silhouette, approach, hand, shape, prior, rigid, ground, single] [method, number, function, proposed, kernel, data, iterative, parameter, sparse, sample, algorithm, minimum, vector, order, support, geodesic, analysis, journal, baxter, topological, iteration, yan, ransac] [motion, figure, estimated, estimate, computer, pattern, international, ieee, small] [distance, arm, good, randomized, voting, image]
Joint Tracking and Segmentation of Multiple Targets
Milan, Anton, Leal-Taixe, Laura, Schindler, Konrad, Reid, Ian

cvpr2015papers_第367张图片
[set, label, evaluation, work, positive, task, association, learning, colour] [representation, crf, detector, global, table, standard, negative] [tracking, detection, object, segmentation, target, trajectory, multiple, superpixel, video, superpixels, temporal, based, cost, bounding, background, tracker, likelihood, track, hypothesis, evidence, crowded, state, foreground, three, edge, weight, entire, propose, online, frame] [model, prior, approach, ground, occluded, location, volume, shape, challenging, scene, continuous, joint, quantitative] [number, method, energy, problem, space, optimization, time, data, pairwise, note, minimum, linear, random] [ieee, motion, velocity, pattern] [image, cvpr, eccv, partial, consistent]
Constrained Planar Cuts - Object Partitioning for Point Clouds
Schoeler, Markus, Papon, Jeremie, Worgotter, Florentin


[human, set, vision, work, supervised, unsupervised, consistency, scoring, introduce, evaluation, annotated, label, evaluate] [recognition, performance, level, training, class, achieve, reported, shoulder] [object, segmentation, partitioning, edge, supervoxels, rand, segment, score, strong, supervoxel, leading, boundary] [point, adjacency, constrained, model, concavity, princeton, cloud, shape, mesh, single, step, full, ground, approach, error, pose, smin] [concave, cut, algorithm, method, convex, graph, euclidean, ransac, directional, proposed, support, number, data, clustering] [figure, computer, ieee, pattern, conference, planar, noise, acm, separate] [local, cutting, locally, benchmark, well, weighted, gap, randomized, distance]
Light Field Layer Matting
Fiss, Juliet, Curless, Brian, Szeliski, Rick

cvpr2015papers_第368张图片
[novel, work, system, content, focus, natural, common, create, vision] [layer, input, map, representation, window] [foreground, background, multiple, threshold, video, propose, weight] [depth, scene, occluded, estimation, single, reconstruction, typically, model, ground] [method, composite, range, operation, problem, algorithm, original, linear, iteration] [light, matting, rendering, layered, color, estimate, ray, alpha, lytro, figure, plane, siggraph, conference, acm, computer, shift, ieee, refocused, matte, perspective, occluders, synthetic, technique, thin, estimating, pattern, spatially, defocus, varying, camera, multicolor, trimap, focal, separate, small, plenoptic, removing, reconstructing, refocusing] [image, compute, view, compare, well, limited, bayesian, computing]
Understanding Pedestrian Behaviors From Stationary Crowd Groups
Yi, Shuai, Li, Hongsheng, Wang, Xiaogang

cvpr2015papers_第369张图片
[dataset, prediction, example, learning, probability, annotated, source, existing, topic] [pedestrian, map, factor, table, smaller, larger, learned] [crowd, stationary, walking, group, moving, path, destination, personality, based, behavior, abnormal, video, route, sgi, average, walk, frame, including, weight, key, three, modeled, detection, obstacle, density, dynamic, crowded, regarded, event, activity] [scene, model, modeling, personalized, layout, location, social, current, estimation, force] [energy, proposed, analysis, general, optimal, parameter, machine, number, close, time, optimized, method] [figure, observed, observation, ieee, motion, pattern, estimated, observing] [calculated, distance, region, large, change, represents]
Holistic 3D Scene Understanding From a Single Geo-Tagged Image
Wang, Shenlong, Fidler, Sanja, Urtasun, Raquel

cvpr2015papers_第370张图片
[semantic, labeling, car, visual, help, dataset, set, freely] [map, performance, context, jointly, input, road, trained] [object, detection, segmentation, based, generate, bounding, employ, box, multiple] [depth, geographic, scene, holistic, pose, single, model, reconstruction, inference, approach, estimation, localization, cad, error, kitti, rad, shape, rendered, structure, osm, reasoning, edo, layout, challenging, fidler, render, initial, reason, prior, edep, point, provide, eseg, reader, potential, ground, current] [method, proposed, min, energy, encoding, solve, note, random, gaussian, algorithm, solving, minimizing] [median, figure, camera, paper, estimated, pixel, computer, rendering, relative] [image, well, geometric, local, dense, distance]
Image Segmentation in Twenty Questions
Rupprecht, Christian, Peter, Loic, Navab, Nassir

cvpr2015papers_第371张图片
[question, probability, set, human, dataset, user, answer, interactive, visual, ibsr, vision, label, active, datasets, learning, strategy, work, natural, twenty] [stanford, hidden, training, size, design, performance] [segmentation, object, seed, background, guess, state, foreground, score, proposal, berkeley, based, three, dice, allowed, framework, truth, segment, gop, series] [mcmc, location, performed, ground, posterior, picked, current, provide, approach, uncertainty, described, volume, rely, fact] [method, number, geodesic, distribution, oracle, markov, algorithm, machine, binary, space, chain, selection, proposed, random, case, solution] [computer, pixel, figure, small, ieee, inside, international, pattern, informative, conference] [image, sampling, distance, brain, sampled, consists, large]
Towards Force Sensing From Vision: Observing Hand-Object Interactions to Infer Manipulation Forces
Pham, Tu-Hoa, Kheddar, Abderrahmane, Qammaz, Ammar, Argyros, Antonis A.

cvpr2015papers_第372张图片
[human, vision, visual, learning, work, process, physical] [net, training, apply, learn, performance] [tracking, object, based, interaction, framework, individual] [force, hand, estimation, kinematics, manipulation, tactile, pose, grasp, mass, nominal, socp, single, approach, normal, manipulating, capture, differentiation, grasping, fsv, full, friction, anns, step, performed, fact, model, environment, body, control, challenging] [method, distribution, time, experimental, problem, consider, optimization, sum, data, gaussian] [contact, motion, ieee, estimate, observed, numerical, algebraic, computer, sensing, estimated, relative, acceleration, minimal, figure, cone, acm, camera, rotational, velocity] [internal, applied, compute, brain]
Dual Domain Filters Based Texture and Structure Preserving Image Non-Blind Deconvolution
Yang, Hang, Zhu, Ming, Niu, Yan, Guan, Yujing, Zhang, Zhongbo

cvpr2015papers_第373张图片
[domain, scenario, noisy, visual, process] [window, stp, top, table, observe, smaller, representation, performance] [based, propose, forward, adaptive] [structure, approach, joint, volume] [method, proposed, sparse, algorithm, gaussian, dual, shrinkage, parameter, kernel, original, iteration, energy, iterative, regularized, low, house, matrix] [deconvolution, rolling, fourier, ieee, deblurring, noise, figure, preserving, stk, blurred, result, small, blur, residual, simple, quality, left, remove, ringing, signal, detail, produce, frequency] [image, guidance, denoising, transform, bilateral, isnr, variance, texture, outperforms, barbara, uorig, denoised, restoration, large, good, nitr, gray, scale, pieias, spatial, stft, local]
Robust Image Filtering Using Joint Static and Dynamic Guidance
Ham, Bumsub, Cho, Minsu, Ponce, Jean

cvpr2015papers_第374张图片
[set] [input, map, output, table, global, coarse] [static, dynamic, structural, average, framework, contrast, truth] [depth, model, joint, structure, ground, term, localization, approach, initialization, typically] [regularization, function, parameter, data, objective, solution, nonconvex, regularizer, surrogate, regularized, gaussian, implicit, converges, time, method, iteratively, explicit, random, proposed, convergence, minimum, note, bilinear] [color, figure, constant, intensity, acm, solver, varying, robust, high, ieee, upsampling, noise, middlebury, handle] [guidance, image, rgf, wls, scale, gradient, texture, upsampled, local, applied, oall, wmf, weighted, bmes, ois, comparison, difference, correspondence, outperforms]
Efficient Illuminant Estimation for Color Constancy Using Grey Pixels
Yang, Kai-Fu, Gao, Shao-Bing, Li, Yong-Jie


[dataset, datasets, natural, set, work, source, vision] [performance, table, better, standard, input, van] [based, three, edge, multiple, hypothesis, contrast, measure, utilized] [error, estimation, indoor, scene, model, bright, detected, percentage, evaluated, volume] [method, proposed, journal, analysis, note, science, machine, low, order, number, parameter, lower] [grey, color, illuminant, angular, constancy, sfu, figure, ieee, median, pixel, iim, computer, society, imaging, pattern, light, ball, detectable, simple, checker, logarithmic, equation, conference, estimated, white, optical, small, second, estimate, america] [image, local, computed, benchmark, patch, best, equal]
Computationally Bounded Retrieval
Rastegari, Mohammad, Keskin, Cem, Kohli, Pushmeet, Izadi, Shahram

cvpr2015papers_第375张图片
[set, vision, learning, search, sun, dataset, noisy, prediction, embedding, task] [accuracy, feature, imagenet, training] [recall, based, auc, cost] [joint, error] [binary, method, optimization, itq, number, projection, data, function, matrix, sparsity, bit, computational, quantization, hashing, code, orthogonal, sparse, update, time, optimal, mapping, complexity, dimensional, case, linear, iterative, proposed, solution, hash, solve, problem, machine, nop, space, curve, solving, coding, cbe, circulant, objective, constraint, column, faster, original, consider, computation] [figure, equation, high, computer, ieee, conference, pattern, naive] [retrieval, nearest, image, rate, neighbor, fast, query, comparison, large]
Geodesic Exponential Kernels: When Curvature and Linearity Conflict
Feragen, Aasa, Lauze, Francois, Hauberg, Soren

cvpr2015papers_第376张图片
[positive, vision, learning, popular, example, sense, cat, embedding, processing] [negative, fisher, recognition, neural, table] [path, shortest, based, framework] [curvature, triangle, point, shape, continuous, model, structure] [geodesic, kernel, riemannian, space, gaussian, data, theorem, euclidean, laplacian, manifold, cnd, machine, exponential, sectional, hilbert, general, main, linear, respect, analysis, consider, isometric, journal, proof, prove, hyperbolic, mathematical, isometrically, inner, grassmann, laplace, vector, tangent, mapping, diffusion, form, implies, graph, spectral] [computer, conference, ieee, chordal, intrinsic, pattern, geometry, embedded, international, classical, radial, result, straight, generally, assume, figure] [metric, distance, corresponding, image, comparison, statistical, geometric]
Efficient SDP Inference for Fully-Connected CRFs Based on Low-Rank Decomposition
Wang, Peng, Shen, Chunhua, van den Hengel, Anton

cvpr2015papers_第377张图片
[positive, set, label, work] [map, feature, table, crf, input, standard, neural, achieves, convolution] [segmentation, based, memory, cost] [approach, inference, estimation, evaluated, unary, performed] [method, approximation, matrix, energy, product, sdp, problem, kernel, sdcut, proposed, function, pairwise, computational, crfs, algorithm, gaussian, linear, number, random, sdlr, expressed, rank, dual, optimal, complexity, note, solution, vector, relaxation, objective, iteration, original, binary, denotes, rounding, exp, requirement, symmetric, lower, minimization, approximate, faster, lanczos, solve, column, quadratic, general, computation, solves] [ieee, speed, pixel, pattern, accurate] [image, corresponding, compatibility, applied, fast, local, best]
Diversity-Induced Multi-View Subspace Clustering
Cao, Xiaochun, Zhang, Changqing, Fu, Huazhu, Liu, Si, Zhang, Hua

cvpr2015papers_第378张图片
[diversity, learning, work, dataset, diverse, novel, existing] [representation, table, standard, performance, learned, feature, explore, jointly] [complementary, based, multiple, video, employ, recall, performs, precision] [face, approach, multiview, term, ensure, joint] [clustering, subspace, method, data, matrix, spectral, dimsc, namsc, objective, function, proposed, kernel, dependence, hsic, complementarity, extended, algorithm, product, independence, independently, low, row, nmi, number, singlebest, problem, inner, constructed, consider, graph, china, denotes, minimizing, underlying, denote, space, alternating, featconcate, solution, nonlinear] [figure, spc, result, smooth, enforce, equation, second] [similarity, view, corresponding, outperforms, best, well, image]
Simulating Makeup Through Physics-Based Manipulation of Intrinsic Image Layers
Li, Chen, Zhou, Kun, Lin, Stephen

cvpr2015papers_第379张图片
[transfer, personal, work, example, style] [layer, transferred, additional] [target, appearance, eye, based, three, including] [face, model, surface, facial, shadow, approach, term] [method, row, decomposition, approximation, energy, proposed, sparse, approximate, linear, expressed, distribution, data] [makeup, skin, specular, cosmetic, diffuse, shading, intrinsic, foundation, albedo, color, highlight, lighting, lipstick, scattering, figure, light, simulated, subsurface, refractive, thickness, computer, real, rendering, captured, second, roughness, optical, acm, ratio, geometry, blush, bst, simulating, microfacet, beckmann, yellow, opaque, barefaced] [image, reference, change, gradient, illustrated, applied, synthesized]
Riemannian Coding and Dictionary Learning: Kernels to the Rescue
Harandi, Mehrtash, Salzmann, Mathieu

cvpr2015papers_第380张图片
[learning, vision, supervised, set, positive, perform, example, unsupervised, introduce, visual, dataset, embedding] [recognition, training, achieved, table, discriminative, representation, learn, previous, version, learned, better, accuracy] [] [approach, term, reconstruction, shape, performed, point, introduced, prior, fact] [coding, riemannian, sparse, kernel, general, space, manifold, formulation, linear, euclidean, min, note, case, machine, tangent, vector, analysis, spd, special, form, solution, derive, problem, solving, hilbert, function, method, ksc, logarithm, data, gaussian, matrix, approximate, nonlinear, logx, geodesic, kslcc, objective, logp, covariance] [conference, computer, ieee, intrinsic, pattern] [dictionary, image, local, query, gradient, well]
Shape-Based Automatic Detection of a Large Number of 3D Facial Landmarks
Zulqarnain Gilani, Syed, Shafait, Faisal, Mian, Ajmal

cvpr2015papers_第381张图片
[set, automatic, manually, source, annotated, automatically, dataset, select, transfer, existing] [level, table, reported, feature, recognition, coarse] [detect, detection, truth, target, adaptive, seed, based] [face, facial, model, point, surface, shape, landmark, localization, neutral, error, expression, curvature, ground, anthropometric, approach, detected, bending, morphable, landmarking, autism, nose, pose, craniofacial] [number, algorithm, proposed, row, method, analysis, journal, energy, function, note, distribution, order] [figure, speed, ieee, computer, paper, subtle, intensity, international, presented] [dense, correspondence, corresponding, large, matching, statistical, query, database, sampled, difference, medical, extreme, well, comparison, image, mild]
Adopting an Unconstrained Ray Model in Light-Field Cameras for 3D Shape Reconstruction
Bergamasco, Filippo, Albarelli, Andrea, Cosmo, Luca, Torsello, Andrea, Rodola, Emanuele, Cremers, Daniel

cvpr2015papers_第382张图片
[vision, set, process, work, common] [level, recognition, feature, standard, accuracy, global, capturing, fully] [target, behavior] [model, point, error, reconstruction, surface, estimation, rigid, performed, single, full, procedure] [code, coding, method, number, optimization, problem, proposed, distribution, order, main, adopted, ransac, space] [calibration, camera, ray, imaging, computer, figure, light, interpolation, ieee, plenoptic, zoom, conference, observed, sensor, pattern, microlens, parametric, paper, motion, plane, lytro, triangulation, maxzoom, pinhole, monitor, phase, calibrated, accurate, detail, horizontal, pixel, rotation, rms, adopt, result, captured, acm, international, high, axis, recovery, squared, white, translation] [image, dense, large, well, view, applied]
Human Action Segmentation With Hierarchical Supervoxel Consistency
Lu, Jiasen, Xu, ran, Corso, Jason J.

cvpr2015papers_第383张图片
[human, set, evaluate, vision, understanding, automatically, consistency, european, label] [better, recognition, map, feature, higher, level, representation, table] [action, segmentation, supervoxel, saliency, supervoxels, video, detection, hierarchical, actionness, foreground, bounding, truth, mask, jhmdb, background, trajectory, iou, box, based, frame, penn, strong, puppet, golf, score, precision, average] [model, unary, potential, ground, localization, full, body, prior, detailed, term, estimation, shape, pose, joint, approach] [pairwise, data, order, graph, method, mrf, analysis, energy, parameter, denotes, number, machine] [motion, computer, ieee, conference, pattern, international, camera, optical, figure, high, color, robust] [hierarchy, large, person]
Non-Rigid Registration of Images With Geometric and Photometric Deformation by Using Local Affine Fourier-Moment Matching
Su, Hong-Ren, Lai, Shang-Hong

cvpr2015papers_第384张图片
[dataset, relationship, associated, lin, example] [global, accuracy, feature, table, max, sliding, layer, map, increase] [target, edge, based, lab, cost, framework, hierarchical, propose, average, three, partitioning] [estimation, model, point, initial, smoothness, term, approach, step, depicted, challenging] [block, proposed, algorithm, method, minimizing, vector, matrix, function, problem, solved, order, energy, derive, parameter, denoted] [photometric, color, fourier, moment, estimate, translation, figure, estimated, lighting, spectrum, robust, real, second, illumination, camera, estimating] [image, registration, local, geometric, transform, fmm, matching, corresponding, reference, transforms, weighting, hierarchal, overlapping, voting, shen, similarity, large, zaragoza, deal, transformation, registering, transformed, histogram, sift, registered]
Hierarchically-Constrained Optical Flow
Kennedy, Ryan, Taylor, Camillo J.

cvpr2015papers_第385张图片
[set, dataset, label, annotated, store, associated] [size, global, table, training, feature] [cost, segmentation, hierarchical, node, object, background, frame, based, groundtruth, edge, superpixels, final, multiple, ldof, weight, entire] [error, term, offset, model, vertex, smoothness, tree, location, estimation, discrete, continuous, approach, procedure, full] [optimization, function, method, matrix, random, data, denoted, algorithm, graph, diagram, note, base, time, problem] [optical, motion, figure, small, pixel, computer, color, ieee, conference, synthetic, international] [image, large, voronoi, hcof, computed, matching, deepflow, epe, endpoint, distance, local, compute, computing, similarity, magnitude, patch, region, area, displacement]
Bayesian Sparse Representation for Hyperspectral Image Super Resolution
Akhtar, Naveed, Shafait, Faisal, Mian, Ajmal

cvpr2015papers_第386张图片
[process, indicator, set, learning, existing, probability, knowledge] [representation, table, performance, public, learned, reported] [based, truth, propose, state, weight] [approach, ground, scene, prior, posterior] [sparse, proposed, spectral, matrix, support, coding, method, sample, sik, parameter, code, vector, denotes, low, range, distribution, data, factorization, gibbs, selection, literature] [high, ieee, imaging, bernoulli, component, estimate, rmse, reconstructed, sensing, multispectral, figure, material, directly] [image, hyperspectral, resolution, dictionary, bayesian, super, remote, spatial, sampling, inferencing, database, zik, gsomp, beta, computed, harvard, cave, acquired, transformed, june, transform, compute, initialized, distinct, exhaustive, construct]
Revisiting Kernelized Locality-Sensitive Hashing for Improved Large-Scale Image Retrieval
Jiang, Ke, Que, Qichao, Kulis, Brian

cvpr2015papers_第387张图片
[search, vision, work, set, probability, flickr] [performance, standard, improvement, feature, applying, increasing, representation, better] [viewed, based, intersection] [projected, provide] [kernel, klsh, rank, random, lsh, analysis, gaussian, hashing, number, data, function, hash, matrix, choice, method, projection, space, note, covariance, empirical, original, vector, approximation, bound, approximate, theoretical, principal, limit, kernelized, theorem, sample, explicit, subspace, hilbert, consider, sum, product, main, kulis, distribution, decay] [international, computer, figure, simple, central, conference, ieee, pattern] [retrieval, nearest, scale, neighbor, image, database, similarity, query, difference, transformation, large, corresponding, tradeoff, best, pca]
Low-Level Vision by Consensus in a Spatial Hierarchy of Regions
Chakrabarti, Ayan, Xiong, Ying, Gortler, Steven J., Zickler, Todd

cvpr2015papers_第388张图片
[set, consistency, visual, subset, vision, physical] [map, larger, global, table, representation] [framework, cost, hierarchical, multiple, organized, truth, measure, three] [scene, model, error, estimation, kitti, approach, term, ground, occluded, binocular, initial, slanted, noc] [data, optimization, iteration, minimization, computational, computation, objective, algorithm, alternating, number, note, lower, core, greater, lowest, explicitly, proposed, support] [stereo, ieee, degree, disparity, piecewise, second, figure, small, parallel, pixel, estimate, inliers, requires] [region, consensus, local, image, inlying, matching, scale, variational, overlapping, aggregation, spatial, benchmark, computed, well, includes, large, consistent]
Clustering of Static-Adaptive Correspondences for Deformable Object Tracking
Nebehay, Georg, Pflugfelder, Roman

cvpr2015papers_第389张图片
[evaluation, dataset, set, identify] [output, performance, standard, global] [object, tracking, frame, employ, measure, success, based, static, bounding, tracker, adaptive, individual, interest, overlap, box, plot, recall, threshold, state, average, scm, multiple, art, candidate, cutoff, allowing, truth, tld] [model, approach, initial, current, ground] [clustering, method, order, algorithm, number, proposed, parameter, idea, pairwise, respect, computational, time, ransac, denotes] [figure, robust, estimating, estimate, motion, second, str, simple] [keypoints, matching, vojir, dissimilarity, keypoint, scale, geometric, match, transform, compute, correspondence, distance, rate, algorithmic, similarity, hough, computed, well, comparison, image, sgt, austrian, agglomerative, local]
Efficient ConvNet-Based Marker-Less Motion Capture in General Scenes With a Low Number of Cameras
Elhayek, Ahmed, de Aguiar, Edilson, Jain, Arjun, Tompson, Jonathan, Pishchulin, Leonid, Andriluka, Micha, Bregler, Chris, Schiele, Bernt, Theobalt, Christian

cvpr2015papers_第390张图片
[human, people] [input, convnet, accuracy, network, convolutional, table, additional, discriminative, combining] [tracking, detection, sequence, frame, multiple, based, combined, moving, average, video, overlap, likelihood, tracked, truth, individual, track, three] [pose, joint, articulated, body, approach, capture, model, term, estimation, position, recorded, pictorial, full, posterior, challenging, location, generative, sog, single, ground, error, arker, projected, scene, depth, disc] [method, number, energy, optimization, algorithm, gaussian, time, note, gaussians, low] [motion, camera, outdoor, figure, color, high, estimate] [image, computed, similarity, view, resolution, weighted, local, sampling]
Semantic Alignment of LiDAR Data at City Scale
Yu, Fisher, Xiao, Jianxiong, Funkhouser, Thomas

cvpr2015papers_第391张图片
[semantic, work, car, set, google, correct, vision, associated, provided, system, test, collected, add, mutually] [feature, global, better, recognition] [based, detect, object, detection, extract] [lidar, alignment, point, icp, urban, initial, provide, detected, scanner, control, building, pose, align, shape, represent, percentage, york, distinctive, approach, valuable, reconstruction, scan, ground] [data, algorithm, method, optimization, range, mapping, spline, solution] [computer, figure, international, conference, ieee, pattern, laser, small, simultaneous, paper, rotation, translation] [large, street, view, registration, city, closest, distance, image, matched, correspondence, transformation, scale, dense, remote]
A Weighted Sparse Coding Framework for Saliency Detection
Li, Nianyi, Sun, Bilin, Yu, Jingyi

cvpr2015papers_第392张图片
[set, human, visual, vision, focus, existing] [feature, performance, input, map, combine, compared, highly] [saliency, superpixels, detection, focusness, salient, precision, recall, dsr, foreground, lrmr, based, superpixel, background, gbmr, framework, hdct, orbd, sod, contrast, measure, prc, object, handling, lfsd, generate, weight] [depth, approach, scene, template, rgb, reconstruction, represent, error] [algorithm, data, sparse, matrix, solution, vector, coding, notice, iterative, function, number] [color, light, high, computer, focal, figure, stereo, stack, technique, disparity, conference, ieee] [image, dictionary, texture, weighted, histogram, reference, region, database, scheme, compute, similarity, connectivity, comprehensive, computed]
Bayesian Adaptive Matrix Factorization With Automatic Model Selection
Chen, Peixian, Wang, Naiyan, Zhang, Nevin L., Yeung, Dit-Yan


[set, process, draw, automatic, text, work] [input, factor, performance, table, better, advantage, comparable, setting] [based, background, three, adaptive] [model, face, error, initial, mixture, shadow, prior, pcp, full, inference, approach, modeling, probabilistic] [matrix, amf, factorization, rank, gaussian, distribution, number, data, method, selection, dirichlet, missing, proposed, brpca, vblr, norm, prmf, gmm, original, gaussians, parameter, denotes, problem, analysis, mbrmf, row, mrpca, cwm, principal, ard, optimization, approximate, construction, journal, update] [noise, robust, removal, component, relative, recovered, figure, unknown, computer, proper] [bayesian, variational, image, variance, corresponding, large, compare, mixing, true]
Robust Regression on Image Manifolds for Ordered Label Denoising
Wu, Hui, Souvenir, Richard

cvpr2015papers_第393张图片
[label, set, predicted, vision, mislabeled, associated, corruption, natural, work, learning, visual, processing] [input, global, learn, neural, performance, loss] [truth, average, measure] [regression, svr, ordered, ground, model, pose, face, percentage, approach, error] [manifold, data, method, function, hessian, ransac, problem, kspca, rbfn, sparse, algorithm, cloudy, space, original, digit, regularization, swiss, row, dimensionality, vector, cloudiness, represented, corrupted, outlier, matrix, linear, number, nonlinear, weather, roll, kernel, statue, underlying] [robust, figure, computer, conference, noise, ieee, rotation, rmse, pattern, international, clear, high, equation, paper, elevation, estimate, varying, red, intrinsic] [image, denoising, local, applied, neighborhood, indicate]
Visual Vibrometry: Estimating Material Properties From Small Motion in Video
Davis, Abe, Bouman, Katherine L., Chen, Justin G., Rubinstein, Michael, Durand, Fredo, Freeman, William T.

cvpr2015papers_第394张图片
[work, physical, engineering, vision, tend] [testing, training, power, trained, learn, extracted, reported] [video, object, weight, truth] [regression, model, approach, predictive, point, stiffness, ground, single, force, surface, elastic] [analysis, method, data, journal, linear] [material, motion, mode, resonant, fabric, figure, vibration, estimate, small, recovered, frequency, sound, laser, ambient, clamped, geometry, high, estimating, computer, rod, camera, theory, modulus, sensitivity, excitation, estimated, equation, acoustic, spectrum, hanging, technique, measured, aluminum, simple, captured, paper, plsr, setup, pixel, tested] [local, area, modal, corresponding, image]
Reweighted Laplace Prior Based Hyperspectral Compressive Sensing for Unknown Sparsity
Zhang, Lei, Wei, Wei, Zhang, Yanning, Tian, Chunna, Li, Fei

cvpr2015papers_第395张图片
[datasets, dataset, learning, latent] [log, learned, table, max, reduce] [three, based, bayes, hierarchical, adaptive, likelihood] [prior, reconstruction, urban, model, structure, estimation, capture] [rlphcs, sparse, data, lasso, rcs, stomp, omp, sparsity, proposed, laplace, reweighted, matrix, hsi, paviau, variable, method, indiana, sam, function, linear, snr, hcs, min, optimization, underlying, norm, orthogonal, bar, random, basis, covariance, diag, optimizing, denotes, distribution, ady, update, row, solution, regularization, gaussian, undemocratic, diagonal, adopted, penalization, exp, experimental, unmixing, proved] [signal, noise, figure, ieee, unknown, compressive, sensing, recover, traditional] [sampling, rate, hyperspectral, image, ssim, psnr, band, comparison, wavelet, bayesian, matching, transform]
Online Sketching Hashing
Leng, Cong, Wu, Jiaxiang, Cheng, Jian, Bai, Xiao, Lu, Hanqing

cvpr2015papers_第396张图片
[learning, dataset, stream, search, work, vision, set, existing, datasets, annual] [training, size, learned, smaller, learn, performance, top, loss, version, map] [online, based, average, precision, extensive, propose] [step, approach, symposium] [data, matrix, hash, hashing, algorithm, round, method, osh, proposed, function, sketching, okh, time, complexity, space, lsh, itq, chunk, random, binary, accumulated, problem, covariance, machine, objective, ytt, batch, column, code, approximation, number, hamming, isoh, decomposition, denote, original, projection, approximate, optimized, scalability, eigenvectors] [computer, conference, ieee, rotation, pattern, second, international, small] [sketch, streaming, large, scale, pair, nearest]
Saliency Propagation From Simple to Difficult
Gong, Chen, Tao, Dacheng, Liu, Wei, Maybank, Stephen J., Fang, Meng, Fu, Keren, Yang, Jie


[teacher, set, labeled, vision, learning, curriculum, learner, unlabeled, existing, processing, knowledge, evaluate, example, process] [map, recognition, input, learned, land, stage, coarse, better, connected] [saliency, superpixels, superpixel, background, detection, salient, based, foreground, individuality, precisionw, tllt, sequence, ihm, object, target, msra, boundary, key, ecssd, recallw, informativity, generated, updated, generate] [hull, adjacent, approach, prior] [convex, matlab, proposed, matrix, method, iteration, graph, algorithm, vector, inhomogeneity, machine, geodesic, gaussian, adopted, diffusion, random] [conference, ieee, pattern, simple, computer, figure, small, reasonable, inhomogeneous, result, decide, traditional] [propagation, image, weighted, connectivity, region]
Long-Term Correlation Tracking
Ma, Chao, Yang, Xiaokang, Zhang, Chongyang, Yang, Ming-Hsuan

cvpr2015papers_第397张图片
[visual, mil, learning, vision, set, european, evaluation] [context, detector, feature, hog, train, effective, window] [tracking, target, correlation, success, overlap, kcf, lct, frame, tracker, performs, stc, online, tld, threshold, struck, temporal, object, appearance, meem, scm, csk, tgpr, asla, based, precision, adaptive, favorably, heavy, background] [model, regression, occlusion, estimation, location, approach, challenging, error, deformation] [method, algorithm, proposed, number, random, case, update, problem, effectively, machine, gaussian, kernel] [figure, computer, conference, ieee, translation, pattern, estimated, robust, motion] [scale, rate, well, fern, distance, fast, change, benchmark, regressor]
Learning Similarity Metrics for Dynamic Scene Segmentation
Teney, Damien, Brown, Matthew, Kit, Dmitry, Hall, Peter

cvpr2015papers_第398张图片
[learning, unsupervised, supervised, dataset, semantic, existing, work, set] [learned, training, learn, size, performance, larger, input, additional, smaller, better] [segmentation, dynamic, video, segment, appearance, hierarchical, truth, spatiotemporal, bank, merged, memory, object, edge, segmenting, framework, rand, updated, boundary, complex, static] [rigid, ground, capture, allows, adjacent, approach, occlusion, manual] [algorithm, dimensionality, proposed, reduction, number, data, method, function, optimize, linear, yij, objective, pairwise, note, general, graph] [motion, color, small, figure, optical, pattern, ieee] [metric, distance, image, texture, large, compare, mahalanobis, consistent, pca]
Causal Video Object Segmentation From Persistence of Occlusions
Taylor, Brian, Karasev, Vasiliy, Soatto, Stefano

cvpr2015papers_第399张图片
[set, domain, causal, work, evaluation, perform, evaluate, knowledge, report] [layer, max, previous, training, extracted] [video, object, segmentation, occluder, foreground, precision, recall, temporal, partition, appearance, docc, based, cost, detection, moving, tracking, entire, violating, framework, truth, frame] [occlusion, occluded, depth, prior, model, full, ordering, estimation, term, ground, occluding, current, approach, described, visible, front] [optimization, row, convex, time, min, problem, method, constraint, algorithm, note, binary, function, regularization, sample, number] [motion, optical, figure, requires, result, small, pixel, determining] [image, region, scheme, computed, large, local, comparison]
3D Model-Based Continuous Emotion Recognition
Chen, Hui, Li, Jiangdong, Zhang, Fengjun, Li, Yang, Wang, Hongan

cvpr2015papers_第400张图片
[emotion, set, natural, system, test, vision, work, human, recognize, dataset, multimodal, selected] [training, recognition, table, input, achieve] [tracking, correlation, video, based, appearance, dynamic, node] [facial, continuous, shape, affective, regression, current, presentation, expression, model, estimation, sij, forest, head, mij, picked, wid, avec, coordinate, arousal, face, step, valence, blendshapes, landmark, represent] [random, method, algorithm, time, matrix, dimensional, fusion, number, proposed, space, energy, binary, journal, vector, dimension, original, covariance] [result, international, figure, ieee, conference, computer, robust, recovery, estimated] [image, large, lbp, affect, difference]
New Insights Into Laplacian Similarity Search
Wu, Xiao-Ming, Li, Zhenguo, Chang, Shih-Fu

cvpr2015papers_第401张图片
[ranking, set, positive, learning, dataset, test, interesting, select] [better, larger, map, top, design] [density, average, propose, performs, entire, surrounding, wij, complementary] [vertex, structure, personalized, robustness] [cluster, graph, sparse, regularizer, lim, parw, theorem, column, vector, behaves, lemma, laplacian, matrix, usps, analysis, pagerank, denote, random, hitting, corollary, respect, absorption, pii, consider, normalized, machine, gaussians, note, sparsest, suppose, proposed, densest, limit, digit, order] [degree, ratio, paper, robust, assume, balancing, figure, computer, prefers, ieee, constant, small] [local, dense, image, query, similarity, divergence, gap, weighted, compare, impact, essentially, large, retrieval, metric, desirable, well]
Active Sample Selection and Correction Propagation on a Gradually-Augmented Graph
Su, Hang, Yin, Zhaozheng, Kanade, Takeo, Huh, Seungil

cvpr2015papers_第402张图片
[active, human, label, learning, unlabeled, annotation, selected, transductive, select, expected, labeled, prediction, annotated, soft, set, dataset, hard] [training, testing, performance, improved, feature, accuracy, achieve, implemented, kth] [segmentation, based, propose, frame] [initial, error, sequential, triangle] [sample, selection, data, rademacher, matrix, luu, minimizing, algorithm, complexity, wul, graph, number, actively, random, method, vector, theorem, column, bound, upper, zsu, proposed, augmented, machine, criterion, min, note, arg, order, ith, gaussian, drawn, harmonic, laplacian, time, solution] [correction, informative, conference, ieee, international, corrected, computer, pattern, gradually, inverse] [propagation, cell, corresponding, subsequent, large, marked, propagated, change]
Continuous Visibility Feature
Liu, Guilin, Gingold, Yotam, Lien, Jyh-Ming

cvpr2015papers_第403张图片
[weak, example, human, semantic, existing, positive, set, evaluation, address, search, visual, airplane, meaningful] [feature, better, achieved, table, designed, reduce] [segmentation, path, based, strong, propose, shortest, intersection, generate] [cvf, shape, visibility, visible, continuous, continuously, vertex, cvfavg, approach, point, mesh, model, diameter, volume, surface, princeton, invisible, provide, entirely, wiley] [geodesic, running, function, time, decomposition, machine, idea, lower, analysis, cut, distribution, graph, vector] [figure, sdf, computer, inside, thickness, component, pattern, acm, second, ieee, length, ray] [region, retrieval, connecting, benchmark, matching, comparison, image, true, area, consistent, change]
Best-Buddies Similarity for Robust Template Matching
Dekel, Tali, Oron, Shaul, Rubinstein, Michael, Avidan, Shai, Freeman, William T.

cvpr2015papers_第404张图片
[set, visual, evaluation, sad] [top, accuracy, window, performance] [target, measure, object, background, foreground, overlap, success, bounding, key, interest, likelihood, based, detection, box, threshold, appearance, complex, candidate] [template, point, nonrigid, challenging, model, deformation] [drawn, case, number, method, distribution, analysis, underlying, data, gaussian, note, space, range] [robust, pattern, figure, ieee, high, real, signal] [image, matching, similarity, distance, expectation, marked, pair, ebbp, distinct, patch, bbp, bbps, ssd, emd, best, nearest, ncc, computing, cpk, integral, considered, geometric, cql, match, commonly, sharply, well, deal, demonstrated, neighbor, fast]
Pose-Conditioned Joint Angle Limits for 3D Human Pose Reconstruction
Akhter, Ijaz, Black, Michael J.

cvpr2015papers_第405张图片
[human, dataset, vision, report, evaluate] [previous, representation, learn, training, learned, network] [propose] [pose, joint, prior, estimation, valid, invalid, bone, body, cmu, reconstruction, error, coordinate, model, kinematic, manual, torso, single, fan, ramakrishna, elbow, represent, monocular, full, wrist, modeling] [method, basis, proposed, algorithm, vector, matrix, data, denotes, number, problem, procrustes, consider, space, min, optimization, sparse, range, distribution, gaussian, extended, linear, arbitrary, exploiting, function, general] [estimate, computer, estimated, camera, figure, angle, equation, international, rotation, relative, conference, estimating, motion, length, noise, plane] [local, image, dictionary, corresponding, compare, mocap, bayesian, parent]
EpicFlow: Edge-Preserving Interpolation of Correspondences for Optical Flow
Revaud, Jerome, Weinzaepfel, Philippe, Harchaoui, Zaid, Schmid, Cordelia

cvpr2015papers_第406张图片
[set, work, perform, evaluate, dataset, incorrect, test] [table, contour, training, better, input, level, performance] [based, propose, edge, performs, cost, density, state] [kitti, estimation, approach, error, occluded, initialization, term, model] [geodesic, energy, minimization, euclidean, method, approximation, order, sparse, computation, note, algorithm, time] [interpolation, optical, motion, pixel, figure, estimate, middlebury, small, ieee, flow, robust, second, interpolated, quality, result] [distance, epicflow, matching, image, dense, large, variational, aee, correspondence, scheme, local, match, sed, compute, impact, kpm, comparison, fast, deepmatching, deepflow, well, kdg, displacement, compare, neighbor, outperforms, computed, region]
Projection Metric Learning on Grassmann Manifold With Application to Video Based Face Recognition
Huang, Zhiwu, Wang, Ruiping, Shan, Shiguang, Chen, Xilin

cvpr2015papers_第407张图片
[learning, set, work, embedding, positive] [recognition, discriminative, learn, feature, size, learned, challenge] [video, based, target, youtube] [face, canonical, approach, represent, constrained] [grassmann, manifold, matrix, projection, method, linear, proposed, subspace, dimensionality, data, analysis, space, kernel, ytf, algorithm, original, riemannian, psd, euclidean, gda, function, pml, reduction, hilbert, number, orthonormal, dcc, mapping, grassmannian, optimization, ggda, pasc, academy, basis, cmsm, time, objective, dimensional, convergence, problem, aij, solution] [geometry, figure, high, developed, directly] [metric, discriminant, image, distance, gradient, chinese, transformed, transformation, well, rate]
Multiple Random Walkers and Their Application to Image Cosegmentation
Lee, Chulwoo, Jang, Won-Dong, Sim, Jae-Young, Kim, Chang-Su

cvpr2015papers_第408张图片
[probability, rule, process, system, set, interactive] [input, recursion, achieve, accuracy, neural, fully, top, performance, connected] [foreground, cosegmentation, segmentation, node, stationary, background, multiple, based, agent, object, transition, wij, employing, walk, ncut, simulate, edge, eij, propose, interact, segment, icoseg] [initial, single, posterior, prior, dominant, point] [mrw, restart, random, proposed, clustering, distribution, algorithm, walker, graph, repulsive, data, concurrence, rwr, matrix, auv, spectral, computation, time, element, number, normalizing, note, experimental, taj, mahal] [figure, ieee, conventional, center, desired, red, result, developed, korea] [image, represents, distance, compute, computed]
More About VLAD: A Leap From Euclidean to Riemannian Manifolds
Faraki, Masoud, Harandi, Mehrtash T., Porikli, Fatih

cvpr2015papers_第409张图片
[vision, set, dataset, learning, visual, example] [recognition, table, extracted, size, better, standard, previous, power, training, accuracy, deep, reported] [video, dynamic, action, framework, cost] [scene, point, face, provide] [riemannian, manifold, grassmannian, vector, spd, geodesic, machine, coding, matrix, projection, linear, euclidean, analysis, form, space, function, codebook, intelligence, chet, covariance, bowg, ldv, computational, algorithm, method, logarithm, data, symmetric, problem, jeffrey, proposed, general, studied, tangent, block, note, stein, journal, upenn, formulation, complexity] [computer, conference, pattern, ieee, conventional, second] [vlad, local, image, metric, gradient, distance, computing, descriptor, bowle, best, texture]
Bilinear Random Projections for Locality-Sensitive Binary Codes
Kim, Saehoon, Choi, Seungjin

cvpr2015papers_第410张图片
[expected, vision, natural, search] [compared, performance, feature, size, reduce, recognition] [precision, recall, correlation, memory, generate] [single, represent] [bilinear, binary, projection, upper, kernel, gaussian, random, bound, lower, data, matrix, time, hamming, hashing, hash, covariance, computational, linear, code, space, analysis, theorem, case, function, lemma, number, respect, rdw, sgn, consumption, theoretical, lsh, induced, drawn, easily, algorithm, vector, mnist, method, empirical, approximate, plotted] [figure, high, international, fourier, computer, conference, ieee, pattern, property] [distance, large, similarity, randomized, nearest, computed, neighbor, image]
Coarse-To-Fine Region Selection and Matching
Yang, Yanchao, Lu, Zhaojin, Sundaramoorthi, Ganesh

cvpr2015papers_第411张图片
[search, test, subset, set, baseline, dataset, vision, existing, correct, highest] [size, level, top, performance, achieve] [detection, interest, hierarchical, threshold, group, salient, recall, precision, entire, cost] [viewpoint, point, procedure, approach] [method, number, algorithm, denote, computational, journal, sample, chosen, element, decomposition, block, time] [figure, computer, pixel, ieee, pattern, second, conference, international, piecewise, robust] [region, matching, hierarchy, image, invariant, large, rate, match, descriptor, correspondence, matched, corresponding, wide, change, false, turntable, maximal, ncc, invariance, alarm, orbit, best, transforms, sift, asift, roc, comparing, local, bij, scale, compare, indicate, patch, neighborhood, haraff]
Elastic Functional Coding of Human Actions: From Vector-Fields to Latent Variables
Anirudh, Rushil, Turaga, Pavan, Su, Jingyong, Srivastava, Anuj

cvpr2015papers_第412张图片
[human, dataset, latent, set, embedding, work, learning, diverse, perform, datasets, search] [recognition, representation, feature, learn, identity] [action, sequence, group, trajectory, skeleton] [shape, skeletal, point, joint, reconstruction, performed, provide] [space, dimensional, manifold, mfpca, riemannian, lower, analysis, product, tangent, data, lie, tsrvf, principal, grassmann, original, proposed, low, matrix, geodesic, functional, time, vector, variable, dimensionality, utkinect, exponential, euclidean, dimension, decay, linear, reduction, function, nonlinear, element, coding, eigenvalue, representational, orthogonal] [high, relative, warping, component, ieee, transport, figure, velocity, pattern] [invariant, distance, sampling, pca, metric, applied, demonstrate]
Reflection Removal Using Ghosting Cues
Shih, YiChang, Krishnan, Dilip, Durand, Fredo, Freeman, William T.

cvpr2015papers_第413张图片
[vision, work, example, natural, side, user, automatic, set, include] [layer, input, recognition, factor, convolution, map] [truth, multiple, based, generated, cost] [model, ground, prior, volume, single, mixture, scene] [gmm, method, kernel, algorithm, shifted, optimization, analysis, solve, data, energy, gaussian, sparsity, decomposition, problem, product, machine] [ghosting, transmission, glass, figure, recovered, ieee, separation, computer, camera, conference, pattern, attenuation, light, separating, relative, ghosted, synthetic, shift, angle, incident, requires, removal, pane, transparent, estimate, attenuated, autocorrelation, assume, equation, international, ratio, supplemental, sin, pixel, imaging] [image, patch, local, spatial, gradient, distance, pair, corner]
Unsupervised Simultaneous Orthogonal Basis Clustering Feature Selection
Han, Dongyoon, Kim, Junmo

cvpr2015papers_第414张图片
[selected, unsupervised, latent, select, learning, set, datasets, type, unlabeled, dataset] [feature, discriminative, better, performance] [target, propose, based] [projected, structure, single, face, regression, term] [socfs, selection, matrix, clustering, data, orthogonal, nmi, acc, ndfs, nonnegative, proposed, udfs, mcfs, number, cluster, constraint, optimization, basis, algorithm, min, regularized, formulation, objective, selects, problem, update, projection, method, bet, analysis, umist, experimental, subproblem, row, spectral, usps, practical, function, solution, orthogonality, minimized, factorization, yaleb, tune, variable, parameter, relates, diagonal, encoding] [figure, directly, simultaneous] [local, corresponding, computed, lung, best]
Saliency Detection via Cellular Automata
Qin, Yao, Lu, Huchuan, Xu, Yiqun, Wang, He

cvpr2015papers_第415张图片
[vision, set, visual, existing, eqn] [map, recognition, global, factor, improve, input, level, apply, effective, better, achieve] [saliency, cellular, salient, object, precision, boundary, recall, bsca, detection, mca, based, background, threshold, propose, superpixels, integration, superpixel, gcd, foreground, wco, generated, contrast, dsr, integrated, integrate, ufo, asd, state, segmentation, evolution] [ground, prior] [matrix, algorithm, optimized, proposed, method, time, updating, row, order, number, denoted, original] [ieee, computer, conference, figure, pattern, color, international, pixel, coherence, intrinsic, measured] [image, distance, bayesian, construct, impact, cell, represents, local, region, spatial]
Random Tree Walk Toward Instantaneous 3D Human Pose Estimation
Yub Jung, Ho, Lee, Soochahn, Seok Heo, Yong, Dong Yun, Il

cvpr2015papers_第416张图片
[set, human, vision, randomly, test, evaluation, probability, process, operate] [trainin

你可能感兴趣的:(计算机视觉)