Research map / Video / Video segmentation and tracking
SAM-MT: Real-Time Interactive Multi-Target Video Segmentation
Ruiqi Shen, Chang Liu, Henghui Ding
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 |
|---|---|---|---|
| use decoupled masked attention instead of full visibility among target queries (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 37.5 -> 43.0 | Improved |
| use decoupled masked attention instead of full masking of global queries (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 39.3 -> 43.0 | Improved |
| add identity-aware mask in the identity transformer (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 39.1 -> 43.0 | Improved |
| increase identity transformer depth from 1 to 3 (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 41.9 -> 43.0 | Improved |
| increase identity transformer depth from 3 to 5 (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 43.0 -> 43.3 | Improved |
| increase identity transformer depth from 3 to 5 (ablation) | SAM-MT · MOSEv2 val | FPS: 36.9 -> 36.4 | Hurt |
| increase sparse memory window from 8 to 32 (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 42.4 -> 43.1 | Improved |
| increase sparse memory window from 8 to 32 (ablation) | SAM-MT · MOSEv2 val | FPS: 37.1 -> 36.6 | Hurt |
| add static image pretraining (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 40.3 -> 43.0 | Improved |
| add strided sampling (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 41.7 -> 43.0 | Improved |
| add overlap prevention loss (ablation) | SAM-MT · MOSEv2 val | J&F-dot: 42.5 -> 43.0 | Improved |
| replace SAM2.1-B+ with SAM-MT as target count grows (>=5 targets) | SAM-MT vs SAM2.1-B+ · MOSEv2 FPS | FPS: 11.5 -> 35.8 | Improved |
Closest papers by meaning
- SAM2Long: Enhancing SAM 2 for Long Video Segmentation with a Training-Free Memory Tree (ICCV 2025 · similarity 0.89)
- RMP-SAM: Towards Real-Time Multi-Purpose Segment Anything (ICLR 2025 · similarity 0.89)
- SAM-I2V: Upgrading SAM to Support Promptable Video Segmentation with Less than 0.2% Training Cost (CVPR 2025 · similarity 0.87)
- SAM 2: Segment Anything in Images and Videos (ICLR 2025 · similarity 0.87)
- SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation (CVPR 2025 · similarity 0.86)
- FocSAM: Delving Deeply into Focused Objects in Segmenting Anything (CVPR 2024 · 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.
