Action Recognition on Kinetics 600 (AURC, AUROC, FPR95, ACC)
81.45AccuracyACR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ACRInput Modalities=video + optical flow2026.03 | 81.45 | 41.85 | 88.99 | 55.89 | |
| MaxLogitInput Modalities=video + optical flow2026.03 | 81.24 | 63.9 | 81.25 | 69.85 | |
| EnergyInput Modalities=video + optical flow2026.03 | 81.24 | 66.48 | 80.12 | 76.58 | |
| EntropyInput Modalities=video + optical flow2026.03 | 81.24 | 47.3 | 86.85 | 65.22 | |
| MSPInput Modalities=video + optical flow2026.03 | 81.24 | 46.29 | 87.33 | 61.29 | |
| DOCTORInput Modalities=video + optical flow2026.03 | 81.24 | 46.37 | 87.28 | 62.27 | |
| RegMixupInput Modalities=video + optical flow2026.03 | 81.16 | 51.44 | 86.06 | 62.71 | |
| OpenMixInput Modalities=video + optical flow2026.03 | 80.79 | 46.73 | 87.69 | 60.27 | |
| MixupInput Modalities=video + optical flow2026.03 | 80.58 | 50.56 | 86.87 | 60.57 | |
| CRLInput Modalities=video + optical flow2026.03 | 80.47 | 49.16 | 87.29 | 61.73 | |
| A2DInput Modalities=video + optical flow2026.03 | 79.97 | 49.26 | 87.97 | 59.79 | |
| STDDPublication=AAAI’25, Training Paradigm=Vision-Language Training2026.04 | 75.1 | — | — | — | |
| Motion-Guided Semantic Alignment with Negative PromptsPublication=Ours, Training Paradigm=Vision-Language Training2026.04 | 72.4 | — | — | — | |
| ZARPublication=EI’25, Training Paradigm=Vision-Language Training2026.04 | 70.5 | — | — | — | |
| X-CLIPPublication=ECCV’22, Training Paradigm=Vision-Language Training2026.04 | 67.1 | — | — | — | |
| ER-ZSARPublication=ICCV’21, Training Paradigm=Vision Training2026.04 | 42.1 | — | — | — |