Temporal Action Localization on THUMOS '14
82.1mAP@0.3ActionFormer
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ActionFormerBackbone=I3D, End-to-End=false, Optical Flow=true2022.11 | 82.1 | — | — | 77.8 | 71 | 59.4 | 43.9 | — | — | |
| ActionFormer + Re²TALBackbone=Re²Slowfast-101, End-to-End=true, Optical Flow=false, Memory (GB)=6.82022.11 | 77.4 | — | — | 72.6 | 64.9 | 53.7 | 39 | — | — | |
| ActionFormer + Re²TALBackbone=Re²VSwin-Tiny, End-to-End=true, Optical Flow=false, Memory (GB)=2.42022.11 | 77 | — | — | 71.5 | 62.4 | 49.7 | 36.3 | — | — | |
| TaLLFormerBackbone=VSwin-Base, End-to-End=true, Optical Flow=false, Memory (GB)=292022.11 | 76 | — | — | — | 63.2 | — | 34.5 | — | — | |
| ActionFormerBackbone=Slowfast-101, End-to-End=false, Optical Flow=false2022.11 | 72.7 | — | — | 66.9 | 58.6 | 46.4 | 33.1 | — | — | |
| ActionFormerBackbone=VSwin-Tiny, End-to-End=false, Optical Flow=false2022.11 | 70.8 | — | — | 64.7 | 55.7 | 42.2 | 27 | — | — | |
| AFSDBackbone=I3D, End-to-End=true, Optical Flow=true, Memory (GB)=122022.11 | 67.3 | — | — | 62.4 | 55.5 | 43.7 | 31.1 | — | — | |
| VSGNBackbone=TSN, End-to-End=false, Optical Flow=true2022.11 | 66.7 | — | — | 60.4 | 52.4 | 41 | 30.4 | — | — | |
| DaoTADBackbone=I3D, End-to-End=true, Optical Flow=false, Memory (GB)=112022.11 | 62.8 | — | — | — | 53.8 | — | 30.1 | — | — | |
| TSIBackbone=TSN, End-to-End=false, Optical Flow=true2022.11 | 61 | — | — | 52.1 | 42.6 | 33.2 | 22.4 | — | — | |
| PBRNetBackbone=I3D, End-to-End=true, Optical Flow=false2022.11 | 58.5 | — | — | 54.6 | 51.3 | 41.8 | 29.5 | — | — | |
| GTANSupervision=Full2019.11 | 57.8 | 69.1 | 63.7 | 47.2 | 38.8 | — | — | — | — | |
| BC-GNNBackbone=TSN, End-to-End=false, Optical Flow=true2022.11 | 57.1 | — | — | 49.1 | 40.4 | 31.2 | 23.1 | — | — | |
| BMNBackbone=TSN, End-to-End=false, Optical Flow=true2022.11 | 56 | — | — | 47.4 | 38.8 | 29.7 | 20.5 | — | — | |
| G-TADBackbone=TSN, End-to-End=false, Optical Flow=true2022.11 | 54.5 | — | — | 47.6 | 40.2 | 30.8 | 23.4 | — | — | |
| BSNSupervision=Full2019.11 | 53.5 | — | — | 45 | 36.9 | 28.4 | 20 | — | — | |
| TAL-NetSupervision=Full2019.11 | 53.2 | 59.8 | 57.1 | 48.5 | 42.8 | 33.8 | 20.8 | — | — | |
| TAL-NetBackbone=I3D, End-to-End=false, Optical Flow=true2022.11 | 53.2 | — | — | 48.5 | 42.8 | 33.8 | 20.8 | — | — | |
| SSNSupervision=Full2019.11 | 51.9 | 66 | 59.4 | 41 | 29.8 | — | — | — | — | |
| CBRSupervision=Full2019.11 | 50.1 | 60.1 | 56.7 | 41.3 | 31 | 19.1 | 9.9 | — | — | |
| STARSupervision=Weak†2019.11 | 48.7 | 68.8 | 60 | 34.7 | 23 | — | — | — | — | |
| R-C3DSupervision=Full2019.11 | 44.8 | 54.5 | 51.5 | 35.6 | 28.9 | — | — | — | — | |
| R-C3DBackbone=C3D, End-to-End=true, Optical Flow=false2022.11 | 44.8 | — | — | 35.6 | 28.9 | — | — | — | — | |
| BaS-NetSupervision=Weak, Feature Extractor=I3D2019.11 | 44.6 | 58.2 | 52.3 | 36 | 27 | 18.6 | 10.4 | 3.9 | 0.5 | |
| TPCSupervision=Full2019.11 | 44.1 | — | — | 37.1 | 28.2 | 20.6 | 12.7 | — | — | |
| SSADSupervision=Full2019.11 | 43 | 50.1 | 47.8 | 35 | 24.6 | — | — | — | — | |
| BaS-NetSupervision=Weak, Feature Extractor=UNT2019.11 | 42.8 | 56.2 | 50.3 | 34.7 | 25.1 | 17.1 | 9.3 | 3.7 | 0.5 | |
| Liu et al.Supervision=Weak, Feature Extractor=I3D2019.11 | 41.2 | 57.4 | 50.8 | 32.1 | 23.1 | 15 | 7 | — | — | |
| MAANSupervision=Weak2019.11 | 41.1 | 59.8 | 50.8 | 30.6 | 20.3 | 12 | 6.9 | 2.6 | 0.2 | |
| CDCSupervision=Full2019.11 | 40.1 | — | — | 29.4 | 23.3 | 13.1 | 7.9 | — | — | |
| W-TALCSupervision=Weak, Feature Extractor=I3D2019.11 | 40.1 | 55.2 | 49.6 | 31.1 | 22.8 | — | 7.6 | — | — | |
| Liu et al.Supervision=Weak, Feature Extractor=UNT2019.11 | 37.5 | 53.5 | 46.8 | 29.1 | 19.9 | 12.3 | 6 | — | — | |
| Yuan et al.Supervision=Full2019.11 | 36.5 | 51 | 45.2 | 27.8 | 17.8 | — | — | — | — | |
| S-CNNSupervision=Full2019.11 | 36.3 | 47.7 | 43.5 | 28.7 | 19 | 10.3 | 5.3 | — | — | |
| Yeung et al.Supervision=Full2019.11 | 36 | 48.9 | 44 | 36 | 36 | 26.4 | 17.1 | — | — | |
| AutoLocSupervision=Weak2019.11 | 35.8 | — | — | 29 | 21.2 | 13.4 | 5.8 | — | — | |
| STPNSupervision=Weak, Feature Extractor=I3D2019.11 | 35.5 | 52 | 44.7 | 25.8 | 16.9 | 9.9 | 4.3 | 1.2 | 0.1 | |
| PSDF + T-SVMSupervision=Full2019.11 | 33.6 | 51.4 | 42.6 | 26.1 | 18.8 | — | — | — | — | |
| W-TALCSupervision=Weak, Feature Extractor=UNT2019.11 | 32 | 49 | 42.8 | 26 | 18.8 | — | 6.2 | — | — | |
| STPNSupervision=Weak, Feature Extractor=UNT2019.11 | 31.1 | 45.3 | 38.8 | 23.5 | 16.2 | 9.8 | 5.1 | 2 | 0.3 | |
| Action SearchSupervision=Full2019.11 | 30.8 | 51.8 | 42.4 | 20.2 | 11.1 | — | — | — | — | |
| Richard et al.Supervision=Full2019.11 | 30 | 39.7 | 35.7 | 23.2 | 15.2 | — | — | — | — | |
| UntrimmedNetSupervision=Weak2019.11 | 28.2 | 44.4 | 37.7 | 21.1 | 13.7 | — | — | — | — | |
| Hide-and-seekSupervision=Weak2019.11 | 19.5 | 36.4 | 27.8 | 12.7 | 6.8 | — | — | — | — |