Loading...
Please wait, while we are loading the content...
Similar Documents
Compositional Human Pose Regression
| Content Provider | arXiv |
|---|---|
| Author | Sun, Xiao Shang, Jiaxiang Liang, Shuang Wei, Yichen |
| Date of Submission | 2017-08-01 |
| Abstract | Regression based methods are not performing as well as detection based methods for human pose estimation. A central problem is that the structural information in the pose is not well exploited in the previous regression methods. In this work, we propose a structure-aware regression approach. It adopts a reparameterized pose representation using bones instead of joints. It exploits the joint connection structure to define a compositional loss function that encodes the long range interactions in the pose. It is simple, effective, and general for both 2D and 3D pose estimation in a unified setting. Comprehensive evaluation validates the effectiveness of our approach. It significantly advances the state-of-the-art on Human3.6M and is competitive with state-of-the-art results on MPII. |
| Related Links | https://arxiv.org/pdf/1704.00159.pdf |
| arXiv | 1704.00159 |
| Language | English |
| Access Restriction | Open |
| Subject Keyword | Computer Science - Computer Vision and Pattern Recognition Computer Science |
| Content Type | Text |
| Resource Type | Article |
| Subject | Computer Science |
