Unsupervised Video Object Discovery on MOVi-E conditional (test)
43.9ARIRandSF.Q+xSSC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| RandSF.Q+xSSCEncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 43.9 | 75.2 | 26.7 | 25 | |
| SmoothSAEncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 36.7 | 73.6 | 28.6 | 27.4 | |
| SmoothSA+xSSCEncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 36.3 | 74.5 | 28.9 | 27.7 | |
| SlotContrast+xSSCEncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 34 | 68 | 21.9 | 20.8 | |
| RandSF.QEncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 30.5 | 82.1 | 23 | 21.6 | |
| SlotContrastEncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 29.9 | 70.6 | 20.7 | 19.3 | |
| VideoSAUREncoder=DINOv2 ViT-S/14, Input Resolution=224x224, #slot=24, Conditionality=conditional2026.05 | 17.4 | 34.6 | 8.3 | 7.5 |