Research map / Video / Video segmentation and tracking
Alignment Before Aggregation: Trajectory Memory Retrieval Network for Video Object Segmentation
Rui Sun, Yuan Wang, Huayu Mai, Tianzhu Zhang, Feng Wu
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 spatial alignment and temporal aggregation (trajectory modeling) (ablation) | STCN · DAVIS 2017 val | J&F: 85.4 -> 86.1 | Improved |
| add agent-level correlation to spatial alignment (ablation) | STCN + ST · DAVIS 2017 val | J&F: 86.1 -> 86.5 | Improved |
| add SVD initialization of agents (ablation) | STCN + ST + Agent · DAVIS 2017 val | J&F: 86.5 -> 87.0 | Improved |
| construct agents by SVD instead of all pixels, random sampling or Top-K (ablation) | TMRN · DAVIS 2017 val | J&F: 85.9 / 86.6 / 86.6 -> 87.0 | Improved |
| increase number of agents K from 64 to 128 (ablation) | TMRN · DAVIS 2017 val | J&F: 86.1 -> 87.0 | Improved |
| increase number of agents K from 128 to 192 (ablation) | TMRN · DAVIS 2017 val | J&F: 87.0 -> 86.4 | Hurt |
| plug TMRN into STM memory reading | STM · DAVIS 2017 test | J&F: 72.2 -> 75.3 | Improved |
| plug TMRN into XMem working memory reading | XMem · DAVIS 2017 val | J&F: 86.2 -> 87.2 | Improved |
| plug TMRN into STCN | STCN · YouTube-VOS 2018 val | G: 83.0 -> 84.2 | Improved |
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 4 of 4 shared results
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
- Video Object Segmentation With Dynamic Memory Networks and Adaptive Object Alignment (ICCV 2021 · similarity 0.90)
- Hierarchical Memory Matching Network for Video Object Segmentation (ICCV 2021 · similarity 0.89)
- Efficient Regional Memory Network for Video Object Segmentation (CVPR 2021 · similarity 0.89)
- Video Object Segmentation Using Space-Time Memory Networks (ICCV 2019 · similarity 0.89)
- Learning Segmentation from Point Trajectories (NeurIPS 2024 · similarity 0.87)
- LiVOS: Light Video Object Segmentation with Gated Linear Matching (CVPR 2025 · similarity 0.87)
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.
