Multiple View Geometry Transformers for 3D Human Pose Estimation
About
In this work, we aim to improve the 3D reasoning ability of Transformers in multi-view 3D human pose estimation. Recent works have focused on end-to-end learning-based transformer designs, which struggle to resolve geometric information accurately, particularly during occlusion. Instead, we propose a novel hybrid model, MVGFormer, which has a series of geometric and appearance modules organized in an iterative manner. The geometry modules are learning-free and handle all viewpoint-dependent 3D tasks geometrically which notably improves the model's generalization ability. The appearance modules are learnable and are dedicated to estimating 2D poses from image signals end-to-end which enables them to achieve accurate estimates even when occlusion occurs, leading to a model that is both accurate and generalizable to new cameras and geometries. We evaluate our approach for both in-domain and out-of-domain settings, where our model consistently outperforms state-of-the-art methods, and especially does so by a significant margin in the out-of-domain setting. We will release the code and models: https://github.com/XunshanMan/MVGFormer.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| 3D Human Pose Estimation | Human3.6M (test) | -- | 547 | |
| 3D Human Pose Estimation | Campus (test) | Actor 1 Score40.2 | 66 | |
| 3D Human Pose Estimation | Shelf (test) | -- | 27 | |
| 3D Human Pose Estimation | CMU Panoptic JLT+15 (test) | MPJPE15.99 | 14 | |
| 3D Multi-person Pose Estimation (In-domain) | Shelf 2 (test) | PCP98 | 12 | |
| 3D Multi-person Pose Estimation (In-domain) | Campus 2 (test) | PCP96.7 | 11 | |
| 3D Human Pose Estimation | CMU Panoptic Average K=1-7 CMU0 (test) | AP@2583.3 | 10 | |
| 3D Multi-person Pose Estimation (In-domain) | CMU Panoptic 17 (test) | MPJPE16 | 7 | |
| 3D Keypoints Estimation | CMU Panoptic conventional (test) | MPJPE (mm)16 | 7 | |
| 3D Human Pose Estimation | CMU Panoptic 6 cameras CMU0 (test) | AP@0.259.47e+3 | 5 |