Research map / Video / Video segmentation and tracking
Spatial-Temporal Multi-level Association for Video Object Segmentation
Deshui Miao, Xin Li, Zhenyu He, Huchuan Lu, Ming-Hsuan Yang
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 object features to the STML module (ablation) | STMA · DAVIS 2017 test | J&F: 83.8 -> 85.6 | Improved |
| add reference frames from spatial memory to STML input (ablation) | STMA · DAVIS 2017 val | J&F: 87.8 -> 88.9 | Improved |
| add temporal memory with ID association (ablation) | STMA · DAVIS 2017 test | J&F: 81.0 -> 85.6 | Improved |
| add temporal memory with ID association (ablation) | STMA · DAVIS 2017 val | FPS: 3.5 -> 10.7 | Improved |
| use decoupled three-stream STML attention instead of joint self-attention (ablation) | STMA · DAVIS 2017 test | J&F: 82.7 -> 85.6 | Improved |
| use MAE pre-trained weights instead of training from scratch (ablation) | STMA ConvViTb · DAVIS 2017 val | J&F: 83.6 -> 88.9 | Improved |
| use MAE instead of MoCoV3 pre-trained weights (ablation) | STMA ViTb · DAVIS 2017 val | J&F: 84.1 -> 88.2 | Improved |
| add static-image pre-training | STMA · DAVIS 2017 val | J&F: 88.9 -> 90.4 | Improved |
| add BL30K pre-training | STMA · DAVIS 2017 test-dev | J&F: 87.1 -> 87.7 | Improved |
| evaluate on DAVIS 2016 against SimVOS-B | STMA · DAVIS 2016 val | J&F: 92.9 -> 91.6 | Hurt |
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
- Target-Aware Object Discovery and Association for Unsupervised Video Multi-Object Segmentation (CVPR 2021 · similarity 0.91)
- Video Object Segmentation Using Space-Time Memory Networks (ICCV 2019 · similarity 0.89)
- Space-Time Memory Networks for Video Object Segmentation With User Guidance (TPAMI 2020 · similarity 0.89)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.89)
- Learning Video Object Segmentation With Visual Memory (ICCV 2017 · similarity 0.88)
- Betrayed by Attention: A Simple yet Effective Approach for Self-supervised Video Object Segmentation (ECCV 2024 · 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.
