Temporal Action Detection on THUMOS14
66.66mAPRaw Videos
Evaluation Results
| Method | Links | |
|---|---|---|
| Raw VideosNetwork=VJEPA-H2025.11 | 66.66 | |
| SPLAVUNetwork=VJEPA-H2025.11 | 66.3 | |
| Raw VideosNetwork=VideoMAE-B2025.11 | 60.82 | |
| SPLAVUNetwork=VideoMAE-B2025.11 | 60.5 | |
| OursEnd-to-End=true, Uses Optical Flow=false, FLOPs=475G, Latency=587ms2022.04 | 54.2 | |
| MUSESEnd-to-End=false, Uses Optical Flow=false, FLOPs=17.5T, Latency=72s*+2.1s2022.04 | 53.4 | |
| AFSDEnd-to-End=true, Uses Optical Flow=true, FLOPs=2780G, Latency=2472ms2022.04 | 52 | |
| Raw VideosNetwork=I3D2025.11 | 25.29 | |
| SPLAVUNetwork=I3D2025.11 | 24.92 | |
| BlurringNetwork=I3D2025.11 | 17.65 | |
| TeD-SPADNetwork=I3D2025.11 | 17.27 | |
| Downsample-2xNetwork=I3D2025.11 | 16.94 | |
| SPActNetwork=I3D2025.11 | 16.2 | |
| VITANetwork=I3D2025.11 | 16.1 | |
| Downsample-4xNetwork=I3D2025.11 | 15.72 | |
| BlackeningNetwork=I3D2025.11 | 15.72 |