Research map / Video / Video segmentation and tracking
YouMVOS: An Actor-Centric Multi-Shot Video Object Segmentation Dataset
Donglai Wei, Siddhant Kharbanda, Sarthak Arora, Roshan Roy, Nishant Jain, Akash Palrecha, Tanav Shah, Shray Mathur, Ritik Mathur, Abhijay Kemkar, Anirudh Chakravarthy, Zudi Lin and 6 more
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add the single-shot module design (SMD) to ObjProp (ablation) | ObjProp · YouMVOS val and test | J: 21.8 -> 25.8 (val); 21.2 -> 25.0 (test) | Improved |
| Add multi-shot module design (MMD) to SMD (ablation) | ObjProp + SMD · YouMVOS val and test | J: 25.8 -> 31.8 (val); 25.0 -> 30.9 (test) | Improved |
| Add HTC (ablation) | ObjProp · YouMVOS val | J: 21.8 -> 24.3 | Improved |
| Add a memory bank (ablation) | ObjProp + HTC · YouMVOS val | J: 24.3 -> 25.8 | Improved |
| Add shot-aware object linking (ablation) | ObjProp + HTC + memory · YouMVOS val | J: 25.8 -> 28.7 | Improved |
| Add CoSAM pose features (ablation) | ObjProp + HTC + memory + SOL · YouMVOS val | J: 28.7 -> 29.6 | Improved |
| Add face Re-ID features (ablation) | ObjProp + HTC + memory + SOL · YouMVOS val | J: 28.7 -> 29.9 | Improved |
| Use both pose and face features (ablation) | ObjProp + HTC + memory + SOL · YouMVOS val | J: 28.7 -> 31.8 | Improved |
| Set SOL clustering threshold to 0.88 instead of 0.90 (ablation) | full method · YouMVOS val | J: 31.8 -> 30.6 | Hurt |
| Set SOL clustering threshold to 0.92 instead of 0.90 (ablation) | full method · YouMVOS val | J: 31.8 -> 31.5 | Hurt |
| Replace CoSAM with LightMBN for pose features (ablation) | full method · YouMVOS val | J: 31.8 -> 31.2 | Hurt |
| Replace CoSAM with ABDNet for pose features (ablation) | full method · YouMVOS val | J: 31.8 -> 31.1 | Hurt |
Closest papers by meaning
- YouTube-VOS: Sequence-to-Sequence Video Object Segmentation (ECCV 2018 · similarity 0.90)
- LVOS: A Benchmark for Large-Scale Long-Term Video Object Segmentation (TPAMI 2025 · similarity 0.89)
- LVOS: A Benchmark for Long-term Video Object Segmentation (ICCV 2023 · similarity 0.89)
- Segment Anything Across Shots: A Method and Benchmark (AAAI 2026 · similarity 0.88)
- OneVOS: Unifying Video Object Segmentation with All-in-One Transformer Framework (ECCV 2024 · similarity 0.88)
- XMem++: Production-level Video Segmentation From Few Annotated Frames (ICCV 2023 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
