Tracking Anything with Decoupled Video Segmentation

Benchmark Model Rank Results
open-world-video-segmentation-on-burst-valDEVA (Mask2Former)#1OWTA (all): 69.9OWTA (com): 75.2OWTA (unc): 41.5
open-world-video-segmentation-on-burst-valDEVA (EntitySeg)#2OWTA (all): 69.5OWTA (com): 73.3OWTA (unc): 50.5
unsupervised-video-object-segmentation-on-davis-2016-valDEVA (DIS)#2G: 88.9J: 87.6F: 90.2
referring-expression-segmentation-on-davis-2017-valDEVA (ReferFormer)#3J&F 1st frame: 66.3
unsupervised-video-object-segmentation-on-davis-2017-test-devDEVA (EntitySeg)#3J&F: 62.1
unsupervised-video-object-segmentation-on-davis-2017-valDEVA (EntitySeg)#3J&F: 73.4Jaccard (Mean): 70.4F-measure (Mean): 76.4
semi-supervised-video-object-segmentation-on-davis-2017-test-devDEVA#7J&F: 83.2Jaccard (Mean): 79.6F-measure (Mean): 86.8FPS: 25.3
video-panoptic-segmentation-on-vipsegDEVA (Mask2Former - SwinB)#10VPQ: 55.0STQ: 52.2
semi-supervised-video-object-segmentation-on-moseDEVA (with OVIS)#13J&F: 66.5J: 62.3F: 70.8FPS: 25.3
semi-supervised-video-object-segmentation-on-moseDEVA (no OVIS)#16J&F: 60.0J: 55.8F: 64.3FPS: 25.3
visual-object-tracking-on-davis-2017DEVA#13J&F: 87.6Jaccard (Mean): 84.2F-measure (Mean): 91.0…
referring-expression-segmentation-on-refer-youtube-vos-2021-public-validationDEVA (ReferFormer)#27J&F: 66.0
referring-video-object-segmentation-on-ref-davis17DEVA + ReferFormer Swin-L (offline) (extra data)#27J&F: 66.3