Chenxin Xu, Siheng Chen, Maosen Li, Ya zhang
We propose a novel method based on teacher-student learning framework for 3D human pose estimation without any 3D annotation or side information. To solve this unsupervised-learning problem, the teacher network adopts pose-dictionary-based modeling for regularization to estimate a physically plausible 3D pose. To handle the decomposition ambiguity in the teacher network, we propose a cycle-consistent architecture promoting a 3D rotation-invariant property to train the teacher network. To further improve the estimation accuracy, the student network adopts a novel graph convolution network for flexibility to directly estimate the 3D coordinates. Another cycle-consistent architecture promoting 3D rotation-equivariant property is adopted to exploit geometry consistency, together with knowledge distillation from the teacher network to improve the pose estimation performance. We conduct extensive experiments on Human3.6M and MPI-INF-3DHP. Our method reduces the 3D joint prediction error by 11.4% compared to state-of-the-art unsupervised methods and also outperforms many weakly-supervised methods that use side information on Human3.6M. Code will be available at https://github.com/sjtuxcx/ITES.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| 3D Reconstruction | MPI-INF-3DHP | AUC | 35.2 | ITES-TS |
| 3D Reconstruction | MPI-INF-3DHP | PCK | 68.2 | ITES-TS |
| 3D Reconstruction | Human3.6M | MPJPE | 85.3 | ITES-TS |
| 3D Reconstruction | Human3.6M | P-MPJPE | 59.8 | ITES-TS |
| 3D | MPI-INF-3DHP | AUC | 35.2 | ITES-TS |
| 3D | MPI-INF-3DHP | PCK | 68.2 | ITES-TS |
| 3D | Human3.6M | MPJPE | 85.3 | ITES-TS |
| 3D | Human3.6M | P-MPJPE | 59.8 | ITES-TS |