ASQuery

Reported on 22 benchmarks across 2 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision22 results

Action LocalizationonAssembly101
Edit
35.3
Action LocalizationonAssembly101
F1@10%
37.8
Action LocalizationonAssembly101
F1@25%
35.6
Action LocalizationonAssembly101
F1@50%
29.4
Action LocalizationonAssembly101
MoF
40.4
best: 41.2 (LTContext)
Action LocalizationonBreakfast
Acc
77.9
best: 78 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action LocalizationonBreakfast
Average F1
74.6
best: 76.2 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action LocalizationonBreakfast
Edit
78.4
best: 79.7 (FACT (efficient hybrid of convolution and transformer model))
Action LocalizationonBreakfast
F1@10%
80.7
best: 82.1 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action LocalizationonBreakfast
F1@25%
76.5
best: 79 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action LocalizationonBreakfast
F1@50%
66.5
best: 67.5 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action SegmentationonAssembly101
Edit
35.3
Action SegmentationonAssembly101
F1@10%
37.8
Action SegmentationonAssembly101
F1@25%
35.6
Action SegmentationonAssembly101
F1@50%
29.4
Action SegmentationonAssembly101
MoF
40.4
best: 41.2 (LTContext)
Action SegmentationonBreakfast
Acc
77.9
best: 78 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action SegmentationonBreakfast
Average F1
74.6
best: 76.2 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action SegmentationonBreakfast
Edit
78.4
best: 79.7 (FACT (efficient hybrid of convolution and transformer model))
Action SegmentationonBreakfast
F1@10%
80.7
best: 82.1 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action SegmentationonBreakfast
F1@25%
76.5
best: 79 (AdaFocus (newly extracted I3D-features, LT-Context model))
Action SegmentationonBreakfast
F1@50%
66.5
best: 67.5 (AdaFocus (newly extracted I3D-features, LT-Context model))