Video Object Discovery and Tracking on YouTube-VIS HQ
76.6ARIfgTSA
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TSAK=7, Frames=Up to 36, Backbone=DINOv2 ViT-S/14, Resolution=256x2562026.06 | 76.6 | 53.3 | 43 | 44.6 | |
| RandSF.QtsimK=7, Frames=Up to 36, Backbone=DINOv2 ViT-S/14, Resolution=256x2562026.06 | 60.4 | 39.4 | 23.8 | 19.3 | |
| RandSF.QsscK=7, Frames=Up to 36, Backbone=DINOv2 ViT-S/14, Resolution=256x2562026.06 | 58 | 37.6 | 21.6 | 15.1 | |
| SlotContrastK=7, Frames=Up to 36, Backbone=DINOv2 ViT-S/14, Resolution=256x2562026.06 | 49.4 | 33 | 18.8 | 8.7 | |
| VideoSAURK=7, Frames=Up to 36, Backbone=DINOv2 ViT-S/14, Resolution=256x2562026.06 | 49.2 | 29.9 | 16.9 | 6.3 |