Invariant Teacher and Equivariant Student for Unsupervised 3D Human Pose Estimation

17 Dec 2020  ·  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.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Unsupervised 3D Human Pose Estimation Human3.6M ITES-TS MPJPE 85.3 # 4
P-MPJPE 59.8 # 1
Unsupervised 3D Human Pose Estimation MPI-INF-3DHP ITES-TS PCK 68.2 # 2
AUC 35.2 # 2

Methods