Temporal Action Localization On Fineaction
Metrics
Results
Performance results of various models on this benchmark
| Paper Title | |||||
|---|---|---|---|---|---|
| RDFA-S6 (InternVideo2-6B) | 29.6 | 46.4 | 29.5 | 7.6 | Enhancing Temporal Action Localization: Advanced S6 Modeling with Recurrent Mechanism |
| ActionMamba(InternVideo2-6B) | 29.04 | 45.44 | 28.82 | 6.79 | Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding |
| InternVideo2-6B | 27.7 | - | - | - | InternVideo2: Scaling Foundation Models for Multimodal Video Understanding |
| DyFADet (VideoMAE v2-g) | 23.8 | 37.1 | 23.7 | 5.9 | DyFADet: Dynamic Feature Aggregation for Temporal Action Detection |
| VideoMAE V2-g | 18.24 | 29.07 | 17.66 | 5.07 | VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking |
| InternVideo | 17.57 | - | - | - | InternVideo: General Video Foundation Models via Generative and Discriminative Learning |
| BMN (i3d feaure) | 9.25 | 14.44 | 8.92 | 3.12 | BMN: Boundary-Matching Network for Temporal Action Proposal Generation |
| G-TAD (i3d feature) | 9.06 | 13.74 | 8.83 | 3.06 | G-TAD: Sub-Graph Localization for Temporal Action Detection |
| DBG (i3d feature) | 6.75 | 10.65 | 6.43 | 2.50 | Fast Learning of Temporal Action Proposal via Dense Boundary Generator |
0 of 9 row(s) selected.