3D Joint Position Prediction on CMU-Mocap
14Basketball MPJPEDCT-based Graph Convolutional Network 3D
Evaluation results
| Method | Links | ||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DCT-based Graph Convolutional Network 3Dmilliseconds=802019.08 | 14 | 3.5 | 7.4 | 16.9 | 25.5 | 11.3 | 7.7 | 5.9 | 11.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| convSeq2Seq 3Dmilliseconds=802019.08 | 16.7 | 8.4 | 10.6 | 22.4 | 14.3 | 12.1 | 7.6 | 8.2 | 12.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Residual sup. 3Dmilliseconds=802019.08 | 18.4 | 12.7 | 15.2 | 36 | 15.6 | 20.3 | 8.2 | 8.4 | 16.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DCT-based Graph Convolutional Network 3Dmilliseconds=1602019.08 | 25.4 | 6.1 | 15.1 | 34.4 | 36.7 | 21.5 | 11.8 | 11.9 | 20.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| convSeq2Seq 3Dmilliseconds=1602019.08 | 30.5 | 16.2 | 20.3 | 44 | 16.3 | 21.8 | 12.5 | 15.9 | 22.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Residual sup. 3Dmilliseconds=1602019.08 | 33.8 | 23.8 | 29.6 | 68.7 | 19.4 | 39.5 | 13.7 | 15.8 | 30.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DCT-based Graph Convolutional Network 3Dmilliseconds=3202019.08 | 49.6 | 11.7 | 31.7 | 76.3 | 39.3 | 44.2 | 19.4 | 30.3 | 37.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| convSeq2Seq 3Dmilliseconds=3202019.08 | 53.8 | 30.8 | 38.7 | 87.5 | 18 | 41.9 | 23 | 32.1 | 40.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Residual sup. 3Dmilliseconds=3202019.08 | 59.5 | 40.3 | 55.1 | 125 | 31.2 | 71.3 | 21.9 | 29.3 | 54.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DCT-based Graph Convolutional Network 3Dmilliseconds=4002019.08 | 61.4 | 15.2 | 42.2 | 96.8 | 39.9 | 55.8 | 23.1 | 40 | 46.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| convSeq2Seq 3Dmilliseconds=4002019.08 | 64.3 | 37.8 | 48.4 | 106.3 | 20.2 | 52.9 | 27.5 | 39.9 | 49.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Residual sup. 3Dmilliseconds=4002019.08 | 70.5 | 46.7 | 66.1 | 145.5 | 36.2 | 84 | 24.5 | 35.4 | 63.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| convSeq2Seq 3Dmilliseconds=10002019.08 | 91.5 | 76.5 | 115.5 | 162.6 | 27.5 | 94.6 | 49.8 | 58.9 | 84.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DCT-based Graph Convolutional Network 3Dmilliseconds=10002019.08 | 106.1 | 53.9 | 152.4 | 164.6 | 58.2 | 117.5 | 40.2 | 79.3 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Residual sup. 3Dmilliseconds=10002019.08 | 106.7 | 77.5 | 127.1 | 195.5 | 43.3 | 129.6 | 32.2 | 61.1 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DMGNNShort-term prediction protocol=Standard MPJPE2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 15.57 | 28.72 | 59.01 | 73.05 | 5.03 | 9.28 | 20.21 | 26.23 | 10.21 | 20.9 | 41.55 | 52.28 | 31.97 | 54.32 | 96.66 | 119.92 | 17.42 | 26.82 | 38.27 | 40.08 | 14.86 | 25.29 | 52.21 | 26.23 | 9.57 | 15.53 | 26.03 | 30.37 | 7.93 | 14.68 | 33.34 | 44.24 | — | — | — | — | — | |
| DMGNN2023.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.03 | 9.28 | 20.21 | 26.23 | 52.04 | |
| DPnetShort-term prediction protocol=Standard MPJPE2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 10.7 | 17.8 | 38.4 | 49.5 | 2.6 | 4.4 | 10 | 13.4 | 5.9 | 11.8 | 26.6 | 33.5 | 12.4 | 23.3 | 70.2 | 89.2 | 16.7 | 18.4 | 19.6 | 25.1 | 9 | 17.1 | 35.8 | 48.7 | 5.8 | 9 | 17.2 | 21.4 | 4.5 | 9.8 | 27.3 | 36.7 | — | — | — | — | — | |
| FMS-AMShort-term prediction protocol=Standard MPJPE, Number of Parameters=30.82M2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 6.2 | 14.6 | 26.2 | 34.1 | 2.4 | 3.2 | 6.8 | 10.6 | 5.8 | 8.4 | 19.8 | 26.4 | 9.3 | 16.3 | 56.2 | 79.2 | 10.1 | 12.6 | 15.3 | 18.8 | 6.7 | 11.2 | 26.8 | 31.4 | 3.9 | 6.1 | 11.8 | 14.9 | 4.4 | 10 | 18.1 | 21.7 | — | — | — | — | — | |
| GA-MINShort-term prediction protocol=Standard MPJPE2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 10.3 | 19.8 | 40.3 | 51.8 | 2.5 | 4.6 | 10.5 | 15.3 | 5.7 | 10.8 | 27.2 | 33.4 | 14.2 | 28.2 | 71.8 | 91.1 | 17.5 | 22.3 | 22.1 | 26.1 | 9.8 | 18.3 | 39 | 49.4 | 5.2 | 8.9 | 16.2 | 18.2 | 4.5 | 9.9 | 27.8 | 35.2 | — | — | — | — | — | |
| HRI2021.11 | — | — | — | — | — | — | — | — | — | 0.96 | 2.06 | 3.11 | 1.05 | 1.37 | 1.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LTD2021.11 | — | — | — | — | — | — | — | — | — | 0.97 | 1.73 | 2.62 | 0.98 | 1.21 | 1.37 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MSR-GCNShort-term prediction protocol=Standard MPJPE2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 10.28 | 18.94 | 37.68 | 47.03 | 3.03 | 5.68 | 12.35 | 16.26 | 5.92 | 12.09 | 28.36 | 38.04 | 14.99 | 28.66 | 55.86 | 62.93 | 12.84 | 20.42 | 30.58 | 34.42 | 3.03 | 5.68 | 12.35 | 16.26 | 6.31 | 10.3 | 17.64 | 21.12 | 5.49 | 11.07 | 25.05 | 32.51 | — | — | — | — | — | |
| MSR-GCN2023.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 3.03 | 5.68 | 12.35 | 16.26 | 47.91 | |
| Multi-Range Transformers2021.11 | — | — | — | — | — | — | — | — | — | 0.6 | 1.12 | 1.71 | 0.79 | 1.05 | 1.22 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ResChunk2023.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 2.85 | 5.54 | 9.27 | 15.55 | 45.23 | |
| Residual SupShort-term prediction protocol=Standard MPJPE2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 15.45 | 26.88 | 43.51 | 49.23 | 20.17 | 32.98 | 42.75 | 44.65 | 20.52 | 40.58 | 75.38 | 90.36 | 26.85 | 48.07 | 93.5 | 108.9 | 25.76 | 48.91 | 88.19 | 100.8 | 17.75 | 31.3 | 52.55 | 61.4 | 44.35 | 76.66 | 216.83 | 151.43 | 22.84 | 44.71 | 86.78 | 104.68 | — | — | — | — | — | |
| Seq2seq2023.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 20.17 | 32.98 | 42.75 | 44.65 | 60.57 | |
| SGSN2023.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 2.2 | 4.1 | 9.7 | 14.7 | 51.5 | |
| SocialPool2021.11 | — | — | — | — | — | — | — | — | — | 0.96 | 2.01 | 2.96 | 1.03 | 1.41 | 1.71 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Traj-GCNShort-term prediction protocol=Standard MPJPE2023.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 11.68 | 21.26 | 40.99 | 50.78 | 3.33 | 6.25 | 13.58 | 17.98 | 6.92 | 13.69 | 30.3 | 39.97 | 17.18 | 32.37 | 60.12 | 72.55 | 14.53 | 24.2 | 37.44 | 41.1 | 3.33 | 6.25 | 13.58 | 17.98 | 6.62 | 10.74 | 17.4 | 20.35 | 5.96 | 11.62 | 24.77 | 31.63 | — | — | — | — | — | |
| TrajDep2023.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 3.33 | 6.25 | 13.58 | 17.98 | 54 |