Skeleton Based Action Recognition On J Hmdb
評価指標
Accuracy (RGB+pose)
Accuracy (pose)
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | Accuracy (RGB+pose) | Accuracy (pose) | Paper Title | Repository |
|---|---|---|---|---|
| DD-Net | - | 77.2 | Make Skeleton-based Action Recognition Model Smaller, Faster and Better | - |
| Potion | 90.4 | 67.9 | PoTion: Pose MoTion Representation for Action Recognition | - |
| MR Two-Sream R-CNN | 71.1 | - | Multi-region two-stream R-CNN for action detection | - |
| Action Tubes | 62.5 | - | Finding Action Tubes | - |
| RPAN | 83.9 | - | RPAN: An End-to-End Recurrent Pose-Attention Network for Action Recognition in Videos | - |
| EHPI | - | 65.5 | Simple yet efficient real-time pose-based action recognition | - |
| I3D + Potion | 85.5 | - | PoTion: Pose MoTion Representation for Action Recognition | - |
| I3D | 84.1 | - | Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset | - |
| PA3D | 69.5 | - | PA3D: Pose-Action 3D Machine for Video Recognition | - |
| STAR-Net | 64.3 | - | STAR-Net: Action Recognition using Spatio-Temporal Activation Reprojection | - |
| PA3D+RPAN | 86.1 | - | PA3D: Pose-Action 3D Machine for Video Recognition | - |
| Chained (RGB+Flow +Pose) | 76.1 | 56.8 | Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection | - |
0 of 12 row(s) selected.