Skeleton-based One-shot Action Recognition on NTU-120 (RE)
52.35AccuracyTrans4SOAR (Base)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Trans4SOAR (Base)Model Size=Base2022.02 | 52.35 | 48.79 | 52.87 | 52.43 | |
| Trans4SOARModel Scale=Base2022.02 | 52.35 | 48.79 | 52.87 | 52.43 | |
| Trans4SOAR (Small)Model Size=Small2022.02 | 51.64 | 50.47 | 52.36 | 51.7 | |
| Trans4SOARModel Scale=Small2022.02 | 51.64 | 50.47 | 52.36 | 51.7 | |
| Skeleton-DML2022.02 | 49.21 | 46.82 | 48.1 | 49.18 | |
| Skeleton-DMLEncoder=Skeleton-DML2022.02 | 49.21 | 46.82 | 48.1 | 49.18 | |
| SL-DML + LSCBackbone=LeViT, Feature Enhancement=LSC2022.02 | 48.28 | 46.03 | 47.58 | 48.31 | |
| SL-DML (LeViT) + LSCEncoder=LeViT, Enhanced=LSC loss2022.02 | 48.28 | 46.03 | 47.58 | 48.31 | |
| SL-DMLBackbone=Swin2022.02 | 47.19 | 45.64 | 46.78 | 47.29 | |
| SL-DMLEncoder=Swin2022.02 | 47.19 | 45.64 | 46.78 | 47.29 | |
| SL-DMLBackbone=LeViT2022.02 | 44.22 | 42.29 | 44.2 | 44.31 | |
| SL-DMLEncoder=LeViT2022.02 | 44.22 | 42.29 | 44.2 | 44.31 | |
| SL-DML2022.02 | 39.82 | 37.85 | 39.32 | 39.86 | |
| SL-DMLEncoder=Standard SL-DML2022.02 | 39.82 | 37.85 | 39.32 | 39.86 |