Action Recognition on N-UCLA cross-view 1.0 (V1,2-3)
93.5AccuracyGlobal Model
Evaluation results
| Method | Links | |
|---|---|---|
| Global ModelData=RGB+Pose, Attention mechanism=true2021.05 | 93.5 | |
| VPNData=RGB+Pose, Attention mechanism=true2021.05 | 93.5 | |
| VPN++ + 3D PosesData=RGB+Pose, Attention mechanism=true2021.05 | 93.5 | |
| P-I3DData=RGB+Pose, Attention mechanism=true2021.05 | 93.1 | |
| SGNData=Pose, Attention mechanism=false2021.05 | 92.5 | |
| Separable STAData=RGB+Pose, Attention mechanism=true2021.05 | 92.4 | |
| HPM+TMData=Depth, Attention mechanism=false2021.05 | 91.9 | |
| VPN++Data=RGB+Pose, Attention mechanism=true2021.05 | 91.9 | |
| Glimpse CloudData=RGB+Pose, Attention mechanism=true2021.05 | 90.1 | |
| Ensemble TS-LSTMData=Pose, Attention mechanism=false2021.05 | 89.2 | |
| I3D*Data=RGB, Attention mechanism=false2021.05 | 86 | |
| NKTMData=RGB, Attention mechanism=false2021.05 | 85.6 |