3D Human Pose Prediction on Human3.6M
13.2Avg 3D Error (160ms)MotionMixer
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MotionMixerEvaluation protocol=average error over all frames2022.07 | 13.2 | 6.9 | 26.9 | 33.6 | 46.1 | 56.5 | 65.7 | 71.6 | |
| GAGCNEvaluation protocol=average error over all frames2022.07 | 16.9 | 10.1 | 32.5 | 38.5 | 50 | — | 72.9 | — | |
| STSGCNEvaluation protocol=average error over all frames2022.07 | 17.1 | 10.1 | 33.1 | 38.3 | 50.8 | 60.1 | 68.9 | 75.6 | |
| siMLPe2023.04 | 21.7 | — | — | 57.3 | 75.7 | — | — | 109.4 | |
| RNN-GCNEvaluation protocol=particular frame2022.07 | 22.6 | 10.4 | 47.1 | 58.3 | 77.3 | 91.8 | 104.1 | 112.1 | |
| HisRep2023.04 | 22.6 | — | — | 58.3 | 77.3 | — | — | 112.1 | |
| LTD-10-25Evaluation protocol=particular frame2022.07 | 23.4 | 11.2 | 47.9 | 58.9 | 78.3 | 93.3 | 106 | 114 | |
| LTD2023.04 | 23.4 | — | — | 58.9 | 78.3 | — | — | 114 | |
| Motion-AttentionEvaluation protocol=particular frame2022.07 | 23.6 | 11 | 49.1 | 60 | 75.9 | 90.4 | 102.5 | 110.1 | |
| MotionMixerEvaluation protocol=particular frame2022.07 | 23.6 | 11 | 47.8 | 59.3 | 77.8 | 91.4 | 106 | 111 | |
| MSR-GCN2023.04 | 25.5 | — | — | 63.3 | 81.1 | — | — | 114.1 | |
| 2-Body Pose Forecasting Best Practicesinitialization=true2023.04 | 26.8 | — | — | 63.1 | 81.1 | — | — | 113.2 | |
| 2-Body Pose Forecasting Best Practicesinitialization=false2023.04 | 27.3 | — | — | 64.6 | 83.1 | — | — | 116.3 | |
| SeS-GCN2023.04 | 29 | — | — | 64 | 84.4 | — | — | 113.9 | |
| convSeq2SeqEvaluation protocol=particular frame2022.07 | 33.3 | 16.6 | 77 | 72.7 | 90.7 | 104.7 | 116.7 | 124.2 | |
| Res. sup.Evaluation protocol=particular frame2022.07 | 46.2 | 25 | 61.4 | — | 88.3 | 119.4 | 130 | 136.6 |