3D Human Pose Estimation on Human3.6M H36MA
39.6MPJPESnapPose3D
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SnapPose3DH=20, Selection=Best2026.04 | 39.6 | 31.6 | — | — | |
| SnapPose3DH=20, Aggregation=Median2026.04 | 45.8 | 36.2 | — | — | |
| SnapPose3DH=20, Aggregation=Average2026.04 | 46.2 | 36.5 | — | — | |
| SnapPose3DH=1, Selection=Random2026.04 | 48.8 | 39.8 | — | — | |
| DiffPoseH=200, Selection=Best2026.04 | 63.1 | 46.7 | — | — | |
| Normalizing Flow based 3D Human Pose Estimation2021.07 | 71 | 54.2 | 93.4 | 171 | |
| Wehrbein et al.N=200, # Params=2,157,1762022.10 | 71 | — | — | — | |
| Wehrbein et al.2026.04 | 71 | 54.2 | — | — | |
| cGNF xlarge w LsampleN=200, # Params=8,318,7412022.10 | 72.6 | — | — | — | |
| Sharma et al.2021.07 | 78.3 | 61.1 | 88.5 | 136.4 | |
| Sharma et al.N=200, # Params=9,100,0802022.10 | 78.3 | — | — | — | |
| Sharma et al.2026.04 | 78.3 | 61.1 | — | — | |
| cGNF w LsampleN=200, # Params=852,5462022.10 | 79.3 | — | — | — | |
| Li et al.2021.07 | 81.1 | 66 | 85.7 | 119.9 | |
| Li & LeeN=5, # Params=4,498,6822022.10 | 81.1 | — | — | — | |
| Li et al.2026.04 | 81.1 | 66 | — | — | |
| cGNFN=200, # Params=852,5462022.10 | 87.3 | — | — | — |