ML Research Wiki / Benchmarks / Semi-Supervised Video Object Segmentation / MOSE

MOSE

Semi-Supervised Video Object Segmentation Benchmark

Performance Over Time

📊 Showing 17 results | 📏 Metric: J&F

Top Performing Models

Rank Model Paper J&F Date Code
1 SAM2 📚 SAM 2: Segment Anything in Images and Videos 77.90 2024-08-01 📦 facebookresearch/segment-anything 📦 facebookresearch/sam2 📦 yangchris11/samurai
2 Cutie+ (base, MEGA) Putting the Object Back into Video Object Segmentation 71.70 2023-10-19 📦 hkchengrex/Cutie
3 Cutie+ (small, MEGA) Putting the Object Back into Video Object Segmentation 70.30 2023-10-19 📦 hkchengrex/Cutie
4 Cutie (base, MEGA) Putting the Object Back into Video Object Segmentation 69.90 2023-10-19 📦 hkchengrex/Cutie
5 Cutie (small, MEGA) Putting the Object Back into Video Object Segmentation 68.60 2023-10-19 📦 hkchengrex/Cutie
6 Cutie (base, with mose) Putting the Object Back into Video Object Segmentation 68.30 2023-10-19 📦 hkchengrex/Cutie
7 Cutie (small, with mose) Putting the Object Back into Video Object Segmentation 67.40 2023-10-19 📦 hkchengrex/Cutie
8 DEVA (with OVIS) Tracking Anything with Decoupled Video Segmentation 66.50 2023-09-07 📦 hkchengrex/Tracking-Anything-with-DEVA
9 Cutie (base) Putting the Object Back into Video Object Segmentation 64.00 2023-10-19 📦 hkchengrex/Cutie
10 Cutie (small) Putting the Object Back into Video Object Segmentation 62.20 2023-10-19 📦 hkchengrex/Cutie

All Papers (17)