Action Localization on BABEL Subset-3 v1.0 (test)
42mAP@0.1RVTCLR + Ours
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RVTCLR + OursEvaluation Protocol=KNN evaluation, Feature Enhancement=Skeleton-Snippet Contrastive Learning2025.12 | 42 | 34.3 | 29.8 | 25.6 | 24.2 | 31.2 | |
| AimCLR + OursEvaluation Protocol=KNN evaluation, Feature Enhancement=Skeleton-Snippet Contrastive Learning2025.12 | 36.5 | 32.3 | 27.4 | 26.4 | 20 | 28.5 | |
| AimCLREvaluation Protocol=KNN evaluation2025.12 | 34.2 | 29.5 | 25.4 | 22.4 | 18.3 | 25.9 | |
| HiCLR + OursEvaluation Protocol=KNN evaluation, Feature Enhancement=Skeleton-Snippet Contrastive Learning2025.12 | 31.9 | 27.7 | 24.8 | 21.3 | 19.2 | 25 | |
| HiCLREvaluation Protocol=KNN evaluation2025.12 | 31.5 | 27.2 | 21.4 | 17.4 | 13.7 | 22.2 | |
| RVTCLREvaluation Protocol=KNN evaluation2025.12 | 29.2 | 24.4 | 21.6 | 20.2 | 18 | 22.7 |