Action Recognition Suite (HMDB, SSV2, Diving, UCF, RareAct)
77.82HMDB AccuracyTAMT*
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| TAMT*Source Dataset=K-400, Backbone=ViT-S, Resolution=224x224, Protocol=5-way 5-shot2024.11 | 77.82 | 64.2 | 49.16 | 97.08 | 73.31 | 72.31 | |
| TAMTSource Dataset=K-400, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 74.14 | 59.18 | 45.18 | 95.92 | 67.44 | 68.37 | |
| TAMT*Source Dataset=K-100, Backbone=ViT-S, Resolution=224x224, Protocol=5-way 5-shot2024.11 | 70.87 | 58.38 | 44.15 | 93.69 | 61.67 | 65.75 | |
| TAMTSource Dataset=SSV2, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 63.66 | — | 38.75 | 83.45 | 42.23 | 57.02 | |
| TAMTSource Dataset=K-100, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 61.76 | 48.9 | 38.33 | 87.76 | 52.81 | 57.91 | |
| DMSDSource Dataset=K-100, Backbone=ResNet-18, Resolution=224x224, Protocol=5-way 5-shot2024.11 | 54.9 | 32.1 | 42.28 | 81.9 | 53.3 | 52.9 | |
| CDFSL-VSource Dataset=K-400, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 53.23 | 49.92 | 17.84 | 65.42 | 49.8 | 47.24 | |
| SEENSource Dataset=K-100, Backbone=ResNet-18, Resolution=224x224, Protocol=5-way 5-shot2024.11 | 52.8 | 31.2 | 40.9 | 79.6 | 50.2 | 50.94 | |
| DD++Source Dataset=K-400, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 48.04 | 44.5 | 16.23 | 63.26 | 47.01 | 43.81 | |
| TAMTSource Dataset=UCF, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 45.34 | — | 33.38 | — | 41.08 | 39.93 | |
| TAMTSource Dataset=Diving, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 45.18 | 38.09 | — | 63.52 | 36.88 | 45.92 | |
| STARTUP++Source Dataset=K-400, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 44.71 | 39.6 | 14.92 | 60.82 | 45.22 | 41.05 | |
| CDFSL-VSource Dataset=SSV2, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 29.86 | — | 17.6 | 33.3 | 35.25 | 29 | |
| HYRSMSource Dataset=K-400, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 29.81 | 40.09 | 17.57 | 45.65 | 44.27 | 35.49 | |
| CDFSL-VSource Dataset=K-100, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 29.8 | 17.21 | 16.37 | 36.53 | 33.91 | 26.76 | |
| CDFSL-VSource Dataset=UCF, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 28.86 | — | 16.07 | — | 33.91 | 26.82 | |
| CDFSL-VSource Dataset=Diving, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 28.45 | 17.46 | — | 31.98 | 34.11 | 28 | |
| DD++Source Dataset=K-100, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 25.99 | 16 | 16.24 | 34.1 | 31.2 | 24.71 | |
| STRMSource Dataset=K-400, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 24.98 | 35.01 | 16.69 | 42.33 | 39.01 | 31.6 | |
| STARTUP++Source Dataset=K-100, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 24.97 | 15.16 | 14.55 | 32.2 | 31.77 | 23.73 | |
| DD++Source Dataset=UCF, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 24.06 | — | 16.15 | — | 32 | 24.07 | |
| STARTUP++Source Dataset=UCF, Backbone=ViT-S, Resolution=112x112, Protocol=5-way 5-shot2024.11 | 23.56 | — | 14.84 | — | 31.31 | 23.24 |