Spatio-temporal Action Detection on UCF101 24 (V@0.5, F@0.5)
83.2F@0.5TubeR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TubeRStream=RGB, Backbone=CSN-1522026.07 | 83.2 | 58.4 | |
| TubeLiteStream=RGB, Backbone=ConvNeXt-T2026.07 | 82.1 | 71.7 | |
| 2in1Stream=Two-stream, Backbone=VGG2026.07 | 78.5 | 50.3 | |
| I3D Det.Stream=Two-stream, Backbone=I3D2026.07 | 76.3 | 59.9 | |
| STEPStream=Two-stream, Backbone=I3D2026.07 | 75 | — | |
| CFADStream=Two-stream, Backbone=I3D2026.07 | 72.5 | 64.6 | |
| TacNetStream=Two-stream, Backbone=VGG2026.07 | 72.1 | 52.9 | |
| MOCStream=RGB, Backbone=DLA342026.07 | 72.1 | 50.7 | |
| ACTStream=Two-stream, Backbone=VGG2026.07 | 67.1 | 51.4 | |
| T-CNNStream=RGB, Backbone=C3D2026.07 | 41.4 | — |