Video Instance Segmentation on YouTube-VIS long videos 2022 (val)
45.9APDVIS
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DVISBackbone=Swin-L2023.06 | 45.9 | 69 | 48.8 | 37.2 | 51.8 | |
| GenVISBackbone=Swin-L, Evaluation Protocol=Online2022.11 | 45.1 | 69.1 | 47.3 | 39.8 | 48.5 | |
| GenVISBackbone=Swin-L, Evaluation Protocol=Semi-online2022.11 | 44.3 | 69.9 | 44.9 | 39.9 | 48.4 | |
| VITABackbone=Swin-L, Evaluation Protocol=Semi-online / Offline, Notes=Evaluated using official repository2022.11 | 41.1 | 63 | 44 | 39.3 | 44.3 | |
| VITABackbone=Swin-L2023.06 | 41.1 | 63 | 44 | 39.3 | 44.3 | |
| IDOLmotion module=InstMove2023.03 | 40.6 | 67.2 | 45.1 | 35 | 48.2 | |
| GenVISBackbone=ResNet-50, Evaluation Protocol=Online2022.11 | 37.5 | 61.6 | 41.5 | 32.6 | 42.2 | |
| GenVISBackbone=ResNet-50, Evaluation Protocol=Semi-online2022.11 | 37.2 | 58.5 | 42.9 | 33.2 | 40.4 | |
| IDOL2023.03 | 35.7 | 62.4 | 35.7 | 32 | 44.7 | |
| MinVISBackbone=Swin-L, Evaluation Protocol=Online, Notes=Evaluated using official repository2022.11 | 33.1 | 54.8 | 33.7 | 29.5 | 36.6 | |
| MinVISBackbone=Swin-L2023.06 | 33.1 | 54.8 | 33.7 | 29.5 | 36.6 | |
| VITABackbone=ResNet-50, Evaluation Protocol=Semi-online / Offline2022.11 | 32.6 | 53.9 | 39.3 | 30.3 | 42.6 | |
| MinVISmotion module=InstMove2023.03 | 25.9 | 46.7 | 24.8 | 24.5 | 30 | |
| MinVISBackbone=ResNet-50, Evaluation Protocol=Online2022.11 | 23.3 | 47.9 | 19.3 | 20.2 | 28 | |
| MinVIS2023.03 | 22.3 | 44.6 | 20.2 | 19.2 | 26.4 |