人体姿态估计资料整理
目录
1. 基础知识
1.1相关术语
2. 资料博客
2.1 OpenPose相关
3. 论文
3.1 Convolutional Pose Machines(CPM)
3.2 Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
3.3 Stacked Hourglass Networks for Human Pose Estimation
3.4 RMPE: Regional Multi-person Pose Estimation
3.5 Hand Keypoint Detection in Single Images using Multiview Bootstrapping
4. 数据集
4.1 LSP
4.2 FLIC
4.3 MPII
1. 基础知识
1.1 相关术语
- articulated pose estimation: 关节姿态估计
- bottom-up & top-down: 多人姿态估计的两种方式,前者是先检测图中关节,再组合成人;后者是先检测出人,再分别进行单人姿态估计。
- Bipartite graph: 二分图(出自论文3.2)
- Hungarian algorithm: 匈牙利算法(论文3.2中做关节最佳匹配用到的)
2. 资料博客
人体姿态估计综述(这是个比较全的论文整理,博主对每篇相关论文也有自己的总结和分析)
2.1 OpenPose相关
- 如何评价卡内基梅隆大学的开源项目 OpenPose?
- Openpose开源代码
- Openpose训练代码分析(1)(2)
3. 论文
- 1. Convolutional Pose Machines(CPM)
- CVPR-2016
- 单人姿态估计
- The Robotics Institute, Carnegie Mellon University
- Paper
- 论文翻译
- 论文阅读
- 论文阅读2
- Code(Caffe)
- Code(Tensorflow)
- 2. Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
- CVPR-2017
- 多人姿态估计(bottom-up)(CPM + PAF)
- The Robotics Institute, Carnegie Mellon University
- Paper
- 论文翻译
- 论文阅读
- Code(README中有Tensorflow,pytorch等版本的链接)
- Code(Pytorch)
- Code(Pytorch)代码阅读
- 3. Stacked Hourglass Networks for Human Pose Estimation
- ECCV-2016
- 单人姿态估计
- Paper
- 论文阅读
- 算法详解
- Code(torch)
- Code(Tensorflow)
- Code(Pytorch)
- 代码实践
- RMPE: Regional Multi-person Pose Estimation
- ICCV-2017
- 多人姿态估计(top-down)(SSD + stacked hourglass)
- SJTU,Tencent Youtu
- Paper
- 论文阅读
- Code(Caffe)
- Code(Torch)
- 5. Hand Keypoint Detection in Single Images using Multiview Bootstrapping
- CVPR-2017
- Carnegie Mellon University
- Paper