Object Discovery on YTVIS-HQ
46ARIRandSF.Q
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| RandSF.Q#slot=7, loss=tsim (time similarity), Input resolution=256×256 (224×224), Encoder=DINO2 ViT-S/142025.08 | 46 | 60.4 | 39.4 | 38.5 | |
| SmoothSAv#slot=7, Input resolution=224×224, Encoder=DINO2 ViT-S/142025.08 | 42.4 | 63 | 38.9 | 38.3 | |
| RandSF.Q#slot=7, loss=ssc (slot-slot contrast), Input resolution=256×256 (224×224), Encoder=DINO2 ViT-S/142025.08 | 40.1 | 58 | 37.6 | 37.2 | |
| RandSF.Q#slot=7, Input resolution=224×224, Encoder=DINO2 ViT-S/142025.08 | 40.1 | 58 | 37.6 | 37.2 | |
| SlotContrast#slot=7, Input resolution=256×256 (224×224), Encoder=DINO2 ViT-S/142025.08 | 37.2 | 49.4 | 33 | 32.8 | |
| SlotContrast#slot=7, Input resolution=224×224, Encoder=DINO2 ViT-S/142025.08 | 37.2 | 49.4 | 33 | 32.8 | |
| VideoSAUR#slot=7, Input resolution=256×256 (224×224), Encoder=DINO2 ViT-S/142025.08 | 33.8 | 49.2 | 29.9 | 29.7 | |
| VideoSAUR#slot=7, Input resolution=224×224, Encoder=DINO2 ViT-S/142025.08 | 33.8 | 49.2 | 29.9 | 29.7 |