Action Recognition on NTU RGB+D
91.7AccuracyGATS
Evaluation Results
| Method | Links | |
|---|---|---|
| GATSInput=point2026.03 | 91.7 | |
| PST-TransformerInput=point2026.03 | 91 | |
| MaST-PreInput=point2026.03 | 90.8 | |
| PSTNetInput=point2026.03 | 90.5 | |
| SequentialPointNetInput=point2026.03 | 90.3 | |
| P4TransformerInput=point2026.03 | 90.2 | |
| DGNNInput=skeleton2026.03 | 89.9 | |
| VA-fusionInput=skeleton2026.03 | 89.4 | |
| AGC-LSTMInput=skeleton2026.03 | 89.2 | |
| PvNeXtInput=point2026.03 | 89.2 | |
| 3DV-PointNet++Input=voxel+point2026.03 | 88.8 | |
| 2s-AGCNInput=skeleton2026.03 | 88.5 | |
| 3s-TranCLRStream=Joint+Motion+Bone, Evaluation Protocol=Linear Evaluation2026.04 | 88.5 | |
| Wang et al.Input=depth2026.03 | 87.1 | |
| AS-GCNInput=skeleton2026.03 | 86.8 | |
| 3s-ActCLRStream=Joint+Motion+Bone, Evaluation Protocol=Linear Evaluation2026.04 | 86.6 | |
| TranCLRStream=Joint, Evaluation Protocol=Linear Evaluation2026.04 | 85.9 | |
| TranCLRStream=Bone, Evaluation Protocol=Linear Evaluation2026.04 | 85.2 | |
| MVDIInput=depth2026.03 | 84.6 | |
| 3DVInput=voxel2026.03 | 84.5 | |
| ActCLRStream=Joint, Evaluation Protocol=Linear Evaluation2026.04 | 83.8 | |
| TranCLRStream=Motion, Evaluation Protocol=Linear Evaluation2026.04 | 83 | |
| ActCLRStream=Bone, Evaluation Protocol=Linear Evaluation2026.04 | 82.6 | |
| Geometric Features2017.05 | 82.39 | |
| Proposed Method2017.05 | 82.31 | |
| ActCLRStream=Motion, Evaluation Protocol=Linear Evaluation2026.04 | 81.5 | |
| 3s-AimCLRStream=Joint+Motion+Bone, Evaluation Protocol=Linear Evaluation2026.04 | 81.4 | |
| STA-LSTM2017.05 | 81.2 | |
| PointNet++Input=point2026.03 | 80.1 | |
| ST-LSTM+Trust Gate2017.05 | 77.7 | |
| 3s-SkeletonCLRStream=Joint+Motion+Bone, Evaluation Protocol=Linear Evaluation2026.04 | 77.4 | |
| AttentionLSTMInput=skeleton2026.03 | 77.1 | |
| AimCLRStream=Joint, Evaluation Protocol=Linear Evaluation2026.04 | 77 | |
| JTM2017.05 | 75.2 | |
| AimCLRStream=Bone, Evaluation Protocol=Linear Evaluation2026.04 | 75.1 | |
| GCA-LSTMInput=skeleton2026.03 | 74.4 | |
| SkeletonCLRStream=Joint, Evaluation Protocol=Linear Evaluation2026.04 | 72.4 | |
| Part-aware LSTM2017.05 | 70.27 | |
| SkeleMotionInput=skeleton2026.03 | 69.6 | |
| AimCLRStream=Motion, Evaluation Protocol=Linear Evaluation2026.04 | 68.7 | |
| SkeletonCLRStream=Bone, Evaluation Protocol=Linear Evaluation2026.04 | 68.4 | |
| Li et al.Input=depth2026.03 | 68.1 | |
| Dynamic Skeletons2017.05 | 65.22 | |
| Deep RNN2017.05 | 64.09 | |
| HBRNN2017.05 | 63.97 | |
| Lie Group2017.05 | 52.76 | |
| SkeletonCLRStream=Motion, Evaluation Protocol=Linear Evaluation2026.04 | 52.1 | |
| HOG2Input=depth2026.03 | 32.2 | |
| SNVInput=depth2026.03 | 31.8 | |
| HON4DInput=depth2026.03 | 30.6 |