Research map / Video / Video segmentation and tracking
Delving Deep Into Many-to-Many Attention for Few-Shot Video Object Segmentation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace full-rank attention with domain-agent attention (ablation) | DAN on YouTube-VIS FSVOS · YouTube-VIS FSVOS, mean over folds | F: 54.2 -> 55.6 | Helped |
| Replace full-rank attention with domain-agent attention (ablation) | DAN on YouTube-VIS FSVOS · YouTube-VIS FSVOS, mean over folds | J: 56.1 -> 57.2 | Helped |
| Add online learning by fine-tuning encoders on support images (ablation) | DAN · YouTube-VIS FSVOS, unseen classes | F: 45.0 -> 49.1 | Helped |
| Add online learning by fine-tuning encoders on support images (ablation) | DAN · YouTube-VIS FSVOS, unseen classes | J: 38.8 -> 42.7 | Helped |
| Use domain-agent attention instead of full-rank attention (ablation) | DAN · YouTube-VIS FSVOS | memory cost: 8.79 GiB; 10.76 GiB | Helped |
| Use meta-learning plus online learning rather than meta-learning alone (ablation) | DAN on unseen classes · YouTube-VIS FSVOS | F: 45.0 -> 49.1 | Helped |
| Use domain-agent attention rather than per-frame prototype-based segmentation | DAN · YouTube-VIS FSVOS, mean over folds | F: 47.9 -> 56.3 | Helped |
| Use domain-agent attention rather than per-frame prototype-based segmentation | DAN · YouTube-VIS FSVOS, mean over folds | J: 57.1 -> 58.0 | Helped |
Closest papers by meaning
- Beyond Pixel and Object: Part Feature as Reference for Few-Shot Video Object Segmentation (AAAI 2025 · similarity 0.90)
- Dual-Agent Optimization framework for Cross-Domain Few-Shot Segmentation (CVPR 2025 · similarity 0.87)
- Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining (CVPR 2024 · similarity 0.86)
- Adapt Before Comparison: A New Perspective on Cross-Domain Few-Shot Segmentation (CVPR 2024 · similarity 0.86)
- Foreground-Covering Prototype Generation and Matching for SAM-Aided Few-Shot Segmentation (AAAI 2025 · similarity 0.85)
- Better to Teach than to Give: Domain Generalized Semantic Segmentation via Agent Queries with Diffusion Model Guidance (ICML 2025 · similarity 0.85)
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.