CVPR/AAAI/ECCV顶会论文/代码

一，CVPR/AAAI/ECCV顶会论文/代码

本页面是自己学习时候，觉得有用的一些文章，暂时MARK一下，用得着的地方就再细看。目标检测、 图像分割、人脸识别、 目标跟踪、 三维点云、 图像处理、 图像分类、 姿态估计、 视频分析、 OCR
、GAN、小样本&零样本、 弱监督&无监督、神经网络、 模型压缩、NAS、 视觉常识

1.cvpr20200

1. 压缩方面

• Deep compression(song han):权重共享，梯度更新
  论文解读：https://blog.csdn.net/weixin_36474809/article/details/80643784#%E6%84%8F%E4%B9%89

• PU-Accelerated Mobile Multi-view Style Transfer
  论文地址：https://arxiv.org/abs/2003.00706

• GPU-Accelerated Mobile Multi-view Style Transfer
  论文地址：https://arxiv.org/abs/2003.00706

• Filter Grafting for Deep Neural Network
  论文地址：https://arxiv.org/pdf/2001.05868.pdf
  论文解读：https://blog.csdn.net/m0_37400316/article/details/105298538

• HRank: Filter Pruning using High-Rank Feature Map
  github:https://github.com/lmbxmu/HRank
  paper:https://128.84.21.199/pdf/2002.10179.pdf
  论文讲解：https://blog.csdn.net/m0_37400316/article/details/105488195

• Out-of-the-box Channel Pruned Networks（基于RL）
  论文地址：https://arxiv.org/pdf/2004.14584.pdf

• Training with Quantization Noise for Extreme Model Compression（量化）
  论文地址： https://arxiv.org/pdf/2004.07320.pdf
  论文分析：https://blog.csdn.net/m0_37400316/article/details/105990601

2. NAS方面

• A Semi-Supervised Assessor of Neural Architectures

• CARS: Contunuous Evolution for Efficient Neural Architecture Search
  https://arxiv.org/pdf/1909.04977.pdf
  开源代码：https://github.com/huawei-noah/CARS
  讲解1:https://mp.weixin.qq.com/s/EW09lqADAoyIlvuyuhLY1w

• Rethinking Performance Estimation in Neural Architecture Search（NAS）
  由于block wise neural architecture search中真正消耗时间的是performance estimation部分，本文针对 block wise的NAS找到了最优参数，速度更快，且相关度更高
  论文：
  代码： https://github.com/zhengxiawu/rethinking_performance_estimation_in_NAS

• NAS 目标检测 Hit-Detector: Hierarchical Trinity Architecture Search for Object Detection
  github地址：https://www.ctolib.com/https://github.com/ggjy/HitDet.pytorch
  讲解地址：https://blog.csdn.net/weixin_43383972/article/details/105646189
  备注：git代码未提供搜索过程

• NAS 目标检测 NAS-FCOS: Fast Neural Architecture Search for Object Detection
  文章地址：https://arxiv.org/pdf/1906.04423.pdf

3. 目标检测
排行榜：目标检测排行榜

• Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection
  论文地址：https://arxiv.org/abs/1912.02424
  代码：https://github.com/sfzhang15/ATSS

• Few-Shot Object Detection with Attention-RPN and Multi-Relation Detector
  论文地址：https://arxiv.org/abs/1908.01998、

• CentripetalNet: Pursuing High-quality Keypoint Pairs for Object Detection
  github：https://github.com/KiveeDong/CentripetalNet

4. 人脸识别

• Towards Universal Representation Learning for Deep Face Recognition
  论文地址：https://arxiv.org/abs/2002.11841
  论文分析：https://blog.csdn.net/m0_37400316/article/details/104900025
• Suppressing Uncertainties for Large-Scale Facial Expression Recognition
  论文地址：https://arxiv.org/abs/2002.10392
  代码：https://github.com/kaiwang960112/Self-Cure-Network
• Face X-ray for More General Face Forgery Detection
  论文地址：https://arxiv.org/pdf/1912.13458.pdf
• 基于元学习的泛化人脸识别（CVPR2020 oral）
  Learning Meta Face Recognition in Unseen Domains
  论文地址https://arxiv.org/pdf/2003.07733.pdf

5. 无监督学习

• A Simple Framework for Contrastive Learning of Visual Representation（）
  论文地址：https://arxiv.org/pdf/2002.05709.pdf
  代码：https://github.com/google-research/simclr
• Kaiming He的Momentum Contrast for Unsupervised?何凯明的无监督方面2篇
  a.Momentum Contrast for Unsupervised Visual Representation Learning
  代码地址： https://github.com/facebookresearch/moco
  b.Improved Baselines with Momentum Contrastive Learning

6. OCR识别

• ABCNet: Real-time Scene Text Spotting with Adaptive Bezier-Curve Network
  论文地址：https://arxiv.org/pdf/2002.10200.pdf
  代码地址：https://github.com/aim-uofa/AdelaiDet

7. 图像和谐化

• “DoveNet: Deep Image Harmonization via Domain Verification”
  文章地址：https://arxiv.org/pdf/1911.13239.pdf
  代码地址：https://github.com/bcmi

8.迁移

• Towards Discriminability and Diversity: Batch Nuclear-norm Maximization under Label Insufficient Situations（CVPR oral）
  论文地址：https://arxiv.org/abs/2003.12237
  代码地址：https://github.com/cuishuhao/BNM

9.视觉常识

• Visual Commonsense R-CN
  论文地址：https://arxiv.org/pdf/2002.12204.pdf
  代码地址：

10.Gradient 优化

• Gradient Centralization: A New Optimization Technique for Deep Neural Networks
  论文地址：https://arxiv.org/pdf/2004.01461.pdf
  代码地址：https://github.com/Yonghongwei/Gradient-Centralization
  论文解读：https://blog.csdn.net/m0_37400316/article/details/105955193

Reference

• http://bbs.cvmart.net/articles/1538#15

• https://github.com/Sophia-11/Awesome-CVPR-Paper

• https://blog.csdn.net/mrjkzhangma/article/details/104579774

• cvpr论文合集https://zhuanlan.zhihu.com/p/112337176

• https://zhuanlan.zhihu.com/p/121507249

• https://github.com/extreme-assistant/CVPR2020-Paper-Code-Interpretation/blob/master/CVPR2020.md

• https://www.ctolib.com/amp/amusi-CVPR2020-Code.html

• https://std.xmu.edu.cn/2020/0303/c4739a395897/page.htm

• 检测类代码

• 检测 https://blog.csdn.net/weixin_40520963/article/details/105224141

• 各类Net代码实现https://github.com/CeLuigi/models-comparison.pytorch

• 物体检测系列代码https://github.com/hoya012/deep_learning_object_detection

• 检测类论文系列https://github.com/amusi/awesome-object-detection