Skeleton-based Action Recognition on NTU-60 Scenario (A) Without Occlusion
74.19AccuracyTrans4SOAR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Trans4SOARScale=Base2022.02 | 74.19 | 74.34 | 75.91 | 74.2 | |
| Trans4SOARScale=Small2022.02 | 69.74 | 70.52 | 72.45 | 69.82 | |
| SL-DML (LeViT) + LSCEncoder=LeViT, LSC=true2022.02 | 67.67 | 67.87 | 68.74 | 67.67 | |
| SL-DML (LeViT)Encoder=LeViT2022.02 | 64.45 | 64.17 | 66.35 | 64.47 | |
| SL-DML (Swin)Encoder=Swin2022.02 | 56.99 | 56.24 | 58.67 | 56.99 | |
| Skeleton-DML2022.02 | 55.54 | 50.88 | 53.13 | 51.24 | |
| SL-DML2022.02 | 54.82 | 54.31 | 56.72 | 54.65 |