CVPR 2022 3月7日论文速递（17 篇打包下载）涵盖 3D 目标检测、医学影像、图像去模糊、车道线检测等方向

CVPR2022论文速递系列：

CVPR 2022 3月3日论文速递（22 篇打包下载）涵盖网络架构设计、姿态估计、三维视觉、动作检测、语义分割等方向
CVPR 2022 3月4日论文速递（29 篇打包下载）涵盖目标检测、全景分割、异常检测、度量学习、对比学习、目标跟踪等方向

全部论文汇总：

CVPR 2022 最全整理：论文分方向汇总 / 代码 / 解读 / 直播 / 项目（更新中）【计算机视觉】

以下是今日更新的 CVPR 2022 论文，包括的研究方向有：风格迁移、医学影像、图像去模糊、图像生成/合成、3D目标检测、深度估计、超分辨率、车道线检测、人脸反欺诈、半监督学习和图像重建。打包合集：下载地址

风格迁移

[1] CLIPstyler: Image Style Transfer with a Single Text Condition(具有单一文本条件的图像风格转移)

关键词：Style Transfer, Text-guided synthesis, Language-Image Pre-Training (CLIP)

论文：https://arxiv.org/abs/2112.00374

医学影像

[1] Temporal Context Matters: Enhancing Single Image Prediction with Disease Progression Representations(时间上下文很重要：使用疾病进展表示增强单图像预测)

关键词：Self-supervised Transformer, Temporal modeling of disease progression

论文：https://arxiv.org/abs/2203.01933

图像去模糊

[1] E-CIR: Event-Enhanced Continuous Intensity Recovery(事件增强的连续强度恢复)

论文：https://arxiv.org/abs/2203.01935)

代码：https://github.com/chensong1995/E-CIR

图像生成/图像合成

[4] 3D Shape Variational Autoencoder Latent Disentanglement via Mini-Batch Feature Swapping for Bodies and Faces(基于小批量特征交换的三维形状变化自动编码器潜在解纠缠

论文：https://arxiv.org/abs/2111.12448

代码：https://github.com/simofoti/3DVAE-SwapDisentangled

[3] Interactive Image Synthesis with Panoptic Layout Generation(具有全景布局生成的交互式图像合成)

论文：https://arxiv.org/abs/2203.02104

[2] Polarity Sampling: Quality and Diversity Control of Pre-Trained Generative Networks via Singular Values(极性采样：通过奇异值对预训练生成网络的质量和多样性控制)

论文：https://arxiv.org/abs/2203.01993

demo：http://bit.ly/polarity-demo-colab

[1] Autoregressive Image Generation using Residual Quantization(使用残差量化的自回归图像生成

论文：https://arxiv.org/abs/2203.01941)

代码：https://github.com/kakaobrain/rq-vae-transformer

3D目标检测

[2] A Versatile Multi-View Framework for LiDAR-based 3D Object Detection with Guidance from Panoptic Segmentation(在全景分割的指导下，用于基于 LiDAR 的 3D 对象检测的多功能多视图框架

关键词：3D Object Detection with Point-based Methods, 3D Object Detection with Grid-based Methods, Cluster-free 3D Panoptic Segmentation, CenterPoint 3D Object Detection

论文：https://arxiv.org/abs/2203.02133

[1] Pseudo-Stereo for Monocular 3D Object Detection in Autonomous Driving(自动驾驶中用于单目 3D 目标检测的伪立体)

关键词：Autonomous Driving, Monocular 3D Object Detection

论文：https://arxiv.org/abs/2203.02112

代码：https://github.com/revisitq/Pseudo-Stereo-3D

深度估计

[5] ITSA: An Information-Theoretic Approach to Automatic Shortcut Avoidance and Domain Generalization in Stereo Matching Networks(立体匹配网络中自动避免捷径和域泛化的信息论方法)

关键词：Learning-based Stereo Matching Networks, Single Domain Generalization, Shortcut Learning

论文：https://arxiv.org/pdf/2201.02263.pdf

ACVNet: Attention Concatenation Volume for Accurate and Efficient Stereo Matching(用于精确和高效立体匹配的注意力连接体积)

关键词：Stereo Matching, cost volume construction, cost aggregation

论文：https://arxiv.org/pdf/2203.02146.pdf

代码：https://github.com/gangweiX/ACVNet

超分辨率

[1] HDNet: High-resolution Dual-domain Learning for Spectral Compressive Imaging(光谱压缩成像的高分辨率双域学习)

关键词：HSI Reconstruction, Self-Attention Mechanism, Image Frequency Spectrum Analysis

论文：https://arxiv.org/pdf/2203.02149.pdf

车道线检测

[1] Rethinking Efficient Lane Detection via Curve Modeling(通过曲线建模重新思考高效车道检测)

关键词：Segmentation-based Lane Detection, Point Detection-based Lane Detection, Curve-based Lane Detection, autonomous driving

论文：https://arxiv.org/abs/2203.02431)

代码：https://github.com/voldemortX/pytorch-auto-drive

人脸反欺诈

[2] Voice-Face Homogeneity Tells Deepfake

论文：https://arxiv.org/abs/2203.02195

代码：https://github.com/xaCheng1996/VFD

半监督学习

[2] Class-Aware Contrastive Semi-Supervised Learning(类感知对比半监督学习

关键词：Semi-Supervised Learning, Self-Supervised Learning, Real-World Unlabeled Data Learning

论文：https://arxiv.org/abs/2203.02261

图像重建

[1] Event-based Video Reconstruction via Potential-assisted Spiking Neural Network(通过电位辅助尖峰神经网络进行基于事件的视频重建

论文：https://arxiv.org/abs/2201.10943

暂无分类

[2] Do Explanations Explain? Model Knows Best(解释解释吗？ 模型最清楚

论文：https://arxiv.org/abs/2203.02269