Action Recognition on AVA 2.1 (test)
24.93mAPAction Transformer (Tx+I3D+96f)
Evaluation results
| Method | Links | |
|---|---|---|
| Action Transformer (Tx+I3D+96f)Modalities=RGB, Architecture=I3D, Transformer, Input Frames=96f2018.12 | 24.93 | |
| Action Transformer (Tx+I3D head)Modalities=RGB, Architecture=I3D, Transformer2018.12 | 24.6 | |
| Action Transformer (Tx-only head)Modalities=RGB, Architecture=I3D, Transformer2018.12 | 24.3 | |
| Tsinghua/MegviiModalities=RGB, Flow, Architecture=I3D, Faster R-CNN, Non-local, TSN, C2D, P3D, C3D, FPN2018.12 | 21.08 | |
| YH TechnologiesModalities=RGB, Flow, Architecture=I3D, Faster R-CNN, Non-local, TSN, C2D, P3D, C3D, FPN2018.12 | 19.6 | |
| Fudan UniversityModalities=RGB, Flow, Architecture=P3D, Faster R-CNN2018.12 | 17.16 |