| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Video Object Segmentation | 17 video datasets (EndoVis 2018, ESD, LVOSv2, LV-VIS, UVO, VOST, PUMaVOS, Virtual KITTI 2, VIPSeg, Wildfires, VISOR, FBMS, Ego-Exo4D, Cityscapes, Lindenthal Camera, HT1080WT Cells, and Drosophila Heart) zero-shot | Zero-shot J&F Accuracy79.3 | 25 | |
| Video Forgery Detection | Video Datasets ID (In-Domain) GenBuster++, LOKI | GenBuster++ Score93.1 | 16 | |
| Semi-supervised video object segmentation | 17 video datasets (test) | J&F Accuracy79.3 | 15 | |
| Video Classification | Video Datasets K400, SSv2 | K400 Accuracy75.53 | 8 |