Research map / Video / Video segmentation and tracking
Advancing Complex Video Object Segmentation via Progressive Concept Construction
Zhixiong Zhang, Shuangrui Ding, Xiaoyi Dong, Songxin He, Jianfan Lin, Junsong Tang, Yuhang Zang, Yuhang Cao, Dahua Lin, Jiaqi Wang
Read the paperCheck 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 concept guidance on top of pixel-level association (ablation) | SeC · SeCVOS | J&F: 62.2 -> 70.0 | Improved |
| add concept guidance on top of pixel-level association (ablation) | SeC · SA-V | J&F: 82.4 -> 82.7 | Improved |
| add pixel-level association (long-term memory) to SAM 2.1 (ablation) | SAM 2.1 · SA-V | J&F: 78.6 -> 82.4 | Improved |
| add pixel-level association (long-term memory) to SAM 2.1 (ablation) | SAM 2.1 · SeCVOS | J&F: 58.2 -> 62.2 | Improved |
| construct concept offline from the full video instead of online (ablation) | SeC · SeCVOS | J&F: 70.0 -> 71.8 | Improved |
| increase LVLM size from 1B to 4B (ablation) | SeC · SeCVOS | J&F: 68.4 -> 70.0 | Improved |
| increase LVLM size from 4B to 8B (ablation) | SeC · SeCVOS | J&F: 70.0 -> 70.3 | No clear effect |
| use 4 concept tokens instead of 1 (ablation) | SeC · SeCVOS | J&F: 70.0 -> 69.9 | No clear effect |
| use HSV scene detector instead of absolute difference (ablation) | SeC · SeCVOS | J&F: 69.0 -> 70.0 | Improved |
| use HSV scene detector instead of ORB (ablation) | SeC · SeCVOS | J&F: 68.7 -> 70.0 | Improved |
| use HSV scene detector instead of optical flow (ablation) | SeC · SeCVOS | J&F: 69.5 -> 70.0 | Improved |
| activate concept guidance on under 10% of frames rather than all frames (ablation) | SeC · SeCVOS | J&F: 7.4% guidance, 70.0 J&F | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- XMem — XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model (ECCV 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Breaking the "Object" in Video Object Segmentation (CVPR 2023 · similarity 0.89)
- M^3-VOS: Multi-Phase, Multi-Transition, and Multi-Scenery Video Object Segmentation (CVPR 2025 · similarity 0.87)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.87)
- The Devil is in Temporal Token: High Quality Video Reasoning Segmentation (CVPR 2025 · similarity 0.87)
- Point-VOS: Pointing Up Video Object Segmentation (CVPR 2024 · similarity 0.87)
- Distilling Spectral Graph for Object-Context Aware Open-Vocabulary Semantic Segmentation (CVPR 2025 · similarity 0.86)
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.
