Video Instance Segmentation on HQ-YTVIS (test)
44.8APBVMT
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| VMTBackbone=Swin-L, Params=222.3M2022.07 | 44.8 | 43.4 | 43 | 64.8 | 70.1 | 59.3 | — | — | |
| SeqFormerBackbone=Swin-L, Params=220.0M2022.07 | 43.3 | 41.5 | 41.6 | 63.7 | 69.7 | 58.7 | — | — | |
| HQ-SAMZero-shot=true, Box prompt input=Swin-L-based Mask2Former, Object association prediction=reused from Mask2Former2023.06 | 34 | 24.3 | — | 63.6 | 70.5 | — | 79.5 | 91.1 | |
| HQ-SAMBackbone=ViT-L, Prompt Generator=Mask2Former [4], Evaluation Protocol=zero-shot2023.06 | 34 | — | — | 63.6 | — | — | — | — | |
| VMTBackbone=R101, Params=70.5M2022.07 | 32.5 | 28.9 | 32.6 | 51.2 | 55.1 | 49.3 | — | — | |
| VMTBackbone=R50, Params=51.5M2022.07 | 30.7 | 24.2 | 31.5 | 50.5 | 54.5 | 50.2 | — | — | |
| SeqFormerBackbone=R101, Params=68.4M2022.07 | 30.7 | 27.3 | 30.1 | 49 | 52.3 | 46.6 | — | — | |
| SAMZero-shot=true, Box prompt input=Swin-L-based Mask2Former, Object association prediction=reused from Mask2Former2023.06 | 30.2 | 19.1 | — | 60.7 | 68.1 | — | 72.9 | 90.5 | |
| SAMBackbone=ViT-L, Prompt Generator=Mask2Former [4], Evaluation Protocol=zero-shot2023.06 | 30.2 | — | — | 60.7 | — | — | — | — | |
| SeqFormerBackbone=R50, Params=49.3M2022.07 | 28.6 | 21.4 | 29.3 | 48.5 | 52.2 | 48.5 | — | — | |
| IFCBackbone=R101, Params=58.3M2022.07 | 27.2 | 23.6 | 28.3 | 48.2 | 51.8 | 47.6 | — | — | |
| IFCBackbone=R50, Params=39.3M2022.07 | 26.5 | 19.6 | 27.5 | 46.6 | 51.5 | 46.9 | — | — | |
| VisTrBackbone=R101, Params=76.3M2022.07 | 25.1 | 20.5 | 27.7 | 45.3 | 53.2 | 45.1 | — | — | |
| PCANBackbone=R101, Params=54.8M2022.07 | 24.8 | 20.1 | 27 | 44 | 52.1 | 44.3 | — | — | |
| CrossVISBackbone=R101, Params=56.6M2022.07 | 24.5 | 19.7 | 26.5 | 44.1 | 52.6 | 44.5 | — | — | |
| VisTrBackbone=R50, Params=57.2M2022.07 | 24 | 16.3 | 25.1 | 43.3 | 52.9 | 44.5 | — | — | |
| PCANBackbone=R50, Params=36.9M2022.07 | 23.9 | 16.1 | 25.2 | 42.2 | 51.8 | 43.9 | — | — | |
| CrossVISBackbone=R50, Params=37.5M2022.07 | 23.6 | 16.2 | 24.9 | 43 | 52.3 | 44 | — | — | |
| MaskTrackBackbone=R101, Params=77.2M2022.07 | 21.7 | 13.1 | 22.8 | 41.1 | 49.1 | 41.7 | — | — | |
| MaskTrackBackbone=R50, Params=58.1M2022.07 | 19.8 | 10.6 | 21.1 | 38.8 | 48.6 | 40.3 | — | — |