Grounded Video Situation Recognition on VidSitu v1 (val)
46.79Verb Accuracy@1VidSitu
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| VidSituPrediction Verb=Predicted, Prediction Role=Predicted, Prediction SRL=Predicted2022.10 | 46.79 | 30.33 | 39.56 | 23.97 | 29.98 | 35.92 | — | — | |
| VideoWhispererPrediction Verb=Predicted, Prediction Role=Ground Truth, Prediction SRL=Predicted2022.10 | 45.06 | 68.23 | 74.15 | 61.79 | 45.58 | 48.22 | 0.27 | 0.15 | |
| VideoWhispererPrediction Verb=Predicted, Prediction Role=Predicted, Prediction SRL=Predicted2022.10 | 44.06 | 52.3 | 61.77 | 38.18 | 35.84 | 38 | 0.13 | 0.05 |