Research map / Video / Video segmentation and tracking
SAM2Long: Enhancing SAM 2 for Long Video Segmentation with a Training-Free Memory Tree
Shuangrui Ding, Rui Qian, Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Yuwei Guo, 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 constrained memory tree search with P=3 pathways (ablation) | SAM2-L · SA-V val | J&F: 76.3 -> 80.8 | Improved |
| add constrained memory tree search with P=3 pathways (ablation) | SAM2-L · LVOS v2 val | J&F: 83.0 -> 85.4 | Improved |
| increase pathways from 3 to 4 (ablation) | SAM2Long · SA-V val | J&F: 80.8 -> 80.7 | No clear effect |
| use 2 pathways instead of 1 (ablation) | SAM2-L · SA-V val | J&F: 76.3 -> 80.1 | Improved |
| use 3 pathways instead of 1 (cost) (ablation) | SAM2-L · SA-V val | FPS: 22 -> 19 | Hurt |
| add temporal decay to IoU-based memory frame selection (ablation) | SAM2Long · SA-V val | J&F: 80.8 -> 80.4 | Hurt |
| add spatial similarity to IoU-based memory frame selection (ablation) | SAM2Long · LVOS v2 val | J&F: 85.4 -> 84.8 | Hurt |
| round predicted IoU to select diverse masks (ablation) | SAM2Long · LVOS v2 val | J&F: 84.4 -> 85.4 | Improved |
| round predicted IoU to select diverse masks (ablation) | SAM2Long · SA-V val | actual pairwise IoU: 84.5 -> 51.4 | Improved |
| add IoU threshold of 0.3 for memory frames (vs none) (ablation) | SAM2Long · SA-V val | J&F: 80.0 -> 80.8 | Improved |
| set uncertainty threshold to 2 instead of 0 (ablation) | SAM2Long · LVOS v2 val | J&F: 84.4 -> 85.4 | Improved |
| use modulation weights [0.95,1.05] instead of none [1,1] (ablation) | SAM2Long · LVOS v2 val | J&F: 84.9 -> 85.4 | Improved |
Later papers that compared with this method
2 later accepted papers report SAM2Long as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- PanoSAM2: Lightweight Distortion- and Memory-aware Adaptions of SAM2 for 360 Video Object Segmentation (ECCV 2026 · better on 4 of 4 shared results)
- DiffuPrompt: Adapting Video Foundation Models to 3D Medical Volumes via Latent Trajectory Priors (ECCV 2026 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SAM 2 — SAM 2: Segment Anything in Images and Videos (ICLR 2025) · better on 7 of 7 shared results
- XMem — XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model (ECCV 2022) · better on 2 of 2 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
- SAM 2: Segment Anything in Images and Videos (ICLR 2025 · similarity 0.93)
- SAM-I2V: Upgrading SAM to Support Promptable Video Segmentation with Less than 0.2% Training Cost (CVPR 2025 · similarity 0.93)
- SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation (CVPR 2025 · similarity 0.90)
- Unlocking the Power of SAM 2 for Few-Shot Segmentation (ICML 2025 · similarity 0.88)
- Robust Ego-Exo Correspondence with Long-Term Memory (NeurIPS 2025 · similarity 0.88)
- SAM2Object: Consolidating View Consistency via SAM2 for Zero-Shot 3D Instance Segmentation (CVPR 2025 · 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.
