List your company

Temporal Action Detection on ActivityNet v1.3 (val)

61.72mAP@0.5AdaTAD
47.763251.386655.0158.6334Jul 23, 2019May 3, 2020Feb 13, 2021Nov 26, 2021Sep 7, 2022Jun 20, 2023Apr 1, 2024

Evaluation results

MethodLinks
AdaTADBackbone=VideoMAEv2-g, E2E=true, Flow=false, Mem=50.6, Resolution=1536x224^2, Note=Last half of backbone layers adapted2023.1161.7243.3510.8541.93
AdaTAD†Backbone=VideoMAEv2-g, E2E=true, Flow=false, Mem=43.6, Resolution=1536x224^22023.1160.8242.699.8441.15
AdaTADBackbone=VideoMAEv2-g, E2E=true, Flow=false, Mem=29.92023.1158.4541.1610.4539.79
AdaTADBackbone=VideoMAE-H, E2E=true, Flow=false, Mem=19.22023.1158.0440.559.7539.37
AdaTADBackbone=VideoMAE-L, E2E=true, Flow=false, Mem=11.02023.1157.6940.5610.1339.22
ADI-DiffFeature=R(2+1)D2024.0456.938.99.138.3
AdaTADBackbone=VideoMAE-B, E2E=true, Flow=false, Mem=4.92023.1156.7739.359.7138.39
TAGSFeature=I3D2024.0456.336.89.636.5
AdaTADBackbone=VideoMAE-S, E2E=true, Flow=false, Mem=2.52023.1156.1538.999.0737.85
DiffTADFeature=I3D2024.0456.136.9936.1
ContextLocFeature=I3D2024.045635.23.634.2
AdaTADBackbone=SlowFast-R50, E2E=true, Flow=false, Mem=4.32023.1155.2838.118.8737.11
Re2TALBackbone=Re2 VideoSwin-T, E2E=true, Flow=false, Mem=242023.1154.7537.819.0336.8
ActionFormerFeature=R(2+1)D2024.0454.737.88.436.6
TriDetFeature=R(2+1)D2024.0454.7388.436.8
TriDetBackbone=I3D, E2E=false, Flow=true2023.1154.7388.436.8
AOE-NetFeature=C3D2022.1054.4235.4310.3535.48
TCANetProposal Source=BMN, Feature=SlowFast2021.0354.3339.138.4137.56
ActionFormerBackbone=SlowFast-R50, E2E=false, Flow=true2023.1154.2637.048.1336.02
ASLBackbone=I3D, E2E=false, Flow=true2023.1154.137.4836.2
TALLFormerBackbone=VideoSwin-B, E2E=true, Flow=false, Mem=292023.1154.136.27.935.6
PBRNetFeature=I3D2024.045435935
ActionFormerBackbone=I3D, E2E=false, Flow=true2023.1153.536.28.235.6
VSGNFeature=R(2+1)D2024.0453.336.88.135.9
GTANVenue & Year=CVPR19, Feature=P3D2022.1052.6134.148.9134.31
GTANVideo-level classification=[43], Post-processing=NMS2021.1052.634.18.934.3
GTANFeature=P3D2024.0452.634.18.934.3
AEI-GArchitecture=GCN-based BMM, Video-level classification=[43], Post-processing=NMS2021.1052.434.59.634.7
AFSDType=Anchor-based, FLOPs(G)=15.32022.0752.435.26.534.3
AFSDBackbone=I3D, RGB-Only=false2022.0552.435.36.534.4
BMN-CSAFeature=TSN2024.0452.436.25.235.4
AFSDFeature=I3D2024.0452.435.36.534.4
AFSDBackbone=I3D, E2E=true, Flow=true, Mem=122023.1152.435.36.534.4
AEI-BArchitecture=CNN-based BMM, Video-level classification=[43], Post-processing=NMS2021.1052.334.59.734.7
VSGNType=Actioness2022.0752.335.28.334.7
TCANetType=Anchor-based2022.0752.336.76.935.5
AEI-BVenue & Year=BMVC21, Feature=C3D2022.1052.334.59.734.7
TCANetFeature=TSN2024.0452.336.76.935.5
TCANetProposal Source=BMN, Feature=2stream2021.0352.2736.736.8635.52
BMNFeature=SlowFast2021.0352.2435.898.3335.28
Self-DETRFeature=I3D2024.0452.233.68.433.7
TCANetProposal Source=BSN, Feature=2stream2021.0351.9134.927.4634.43
ABNVideo-level classification=[43], Post-processing=NMS2021.1051.834.210.334.2
DCANVideo-level classification results=Zhao et al. 2017b2021.1251.7835.989.4535.39
ABNVenue & Year=IEEE-Access21, Feature=C3D2022.1051.7834.1810.2934.22
STPTFeature=STPT2024.0451.433.76.833.4
BSN++Video-level classifier=Xiong et al. 20162020.0951.2735.78.3334.88
BSN++Feature=2stream2021.0351.2735.78.3334.88
BSN++Video-level classification results=Zhao et al. 2017b2021.1251.2735.78.3334.88
TSI++Video-level classification=[43], Post-processing=NMS2021.1051.2356.634.2
PlusTAD (Anchor-free)Backbone=R50-SlowOnly, RGB-Only=true, Resolution=224, Input Frames=962022.0551.233.417.5733.12
BasicTADBackbone=SlowOnly-R50, E2E=true, Flow=false, Mem=122023.1151.233.417.5733.12
TSI++Venue & Year=ACCV20, Feature=2Stream2022.1051.18356.5934.15
SSTAPLabel Percentage=100%, Classifier=[55]2021.0450.7235.287.8734.48
SSTAPLabel Percentage=90%, Classifier=[55]2021.0450.6435.127.834.35
BC-GNNFeature=TSN2024.0450.634.89.434.3
BC-GNNReference=ECCV'20, Classifier=[55]2021.0450.5634.759.3734.26
BC-GNNVideo-level classification results=Zhao et al. 2017b2021.1250.5634.759.3734.26
BC-GNNVenue & Year=ECCV20, Feature=2Stream2022.1050.5634.759.3734.26
BC-GNNclassifier=[33]2020.0850.5634.759.3734.26
Ji et al.Label Percentage=100%, Reference=ICCV'19, Classifier=[55]2021.0450.5535.017.5834.23
E2E-TADBackbone=SlowFast-R50, E2E=true, Flow=false, Mem=122023.1150.4735.9910.3335.1
GTADVideo-level classification=[43], Post-processing=NMS2021.1050.434.6934.1
G-TADType=Actioness, FLOPs(G)=45.72022.0750.434.6934.1
G-TADFeature=TSN2024.0450.434.6934.1
G-TAD2019.1150.3634.69.0234.09
G-TADVideo-level classifier=Xiong et al. 20162020.0950.3634.69.0234.09
G-TADFeature=2stream2021.0350.3634.69.0234.09
G-TADReference=CVPR'20, Classifier=[55]2021.0450.3634.69.0234.09
G-TADVideo-level classification results=Zhao et al. 2017b2021.1250.3634.69.0234.09
GTADVenue & Year=CVPR20, Feature=2Stream2022.1050.3634.69.0234.09
G-TADBackbone=TSN, RGB-Only=false2022.0550.3634.69.0234.09
Ji et al.Label Percentage=90%, Reference=ICCV'19, Classifier=[55]2021.0450.2434.977.3534.13
SSTAPLabel Percentage=60%, Classifier=[55]2021.0450.1434.927.4334.01
BMNVideo-level classification=[43], Post-processing=NMS2021.1050.134.88.333.9
BMNType=Actioness, FLOPs(G)=45.62022.0750.134.88.333.9
BMNFeature=TSN2024.0450.134.88.333.9
BMNclassifier=[36]2019.0750.0734.788.2933.85
BMN2019.1150.0734.788.2933.85
BMNVideo-level classifier=Xiong et al. 20162020.0950.0734.788.2933.85
BMNFeature=2stream2021.0350.0734.788.2933.85
BMNLabel Percentage=100%, Reference=ICCV'19, Classifier=[55]2021.0450.0734.788.2933.85
BMNVideo-level classification results=Zhao et al. 2017b2021.1250.0734.788.2933.85
BMNVenue & Year=ICCV19, Feature=2Stream2022.1050.0734.788.2933.85
BMNBackbone=TSN, RGB-Only=false2022.0550.0734.788.2933.85
BMNBackbone=TSN, E2E=false, Flow=true2023.1150.0734.788.2933.85
BMNclassifier=[33]2020.0850.0734.788.2933.85
SP-TADBackbone=I3D, RGB-Only=false2022.0550.0632.928.4432.99
PlusTAD (Anchor-based)Backbone=R50-SlowOnly, RGB-Only=true, Resolution=224, Input Frames=962022.0550.0433.792.7532.13
MUSESFeature=I3D2024.0450356.634
BMNLabel Percentage=90%, Reference=ICCV'19, Classifier=[55]2021.0449.9433.738.2333.74
Ji et al.Label Percentage=60%, Reference=ICCV'19, Classifier=[55]2021.0449.8234.537.0133.52
ReActType=DETR-based, FLOPs(G)=0.382022.0749.6338.632.6
ReActFeature=3DCNN2024.0449.6338.632.6
BMNLabel Percentage=60%, Reference=ICCV'19, Classifier=[55]2021.0449.533.688.1533.17
TadTrType=DETR-based, FLOPs(G)=0.382022.0749.132.68.532.3
TadTRFeature=I3D2024.0449.132.68.532.3
TadTRBackbone=I3D, E2E=false, Flow=true2023.1149.132.68.532.3
TSA-NetFeature=P3D2024.0448.732931.9
PGCNType=Anchor-based, FLOPs(G)=5.02022.0748.333.23.331.1