English

SAM3-DMS: Decoupled Memory Selection for Multi-target Video Segmentation of SAM3

Computer Vision and Pattern Recognition 2026-01-15 v1

Abstract

Segment Anything 3 (SAM3) has established a powerful foundation that robustly detects, segments, and tracks specified targets in videos. However, in its original implementation, its group-level collective memory selection is suboptimal for complex multi-object scenarios, as it employs a synchronized decision across all concurrent targets conditioned on their average performance, often overlooking individual reliability. To this end, we propose SAM3-DMS, a training-free decoupled strategy that utilizes fine-grained memory selection on individual objects. Experiments demonstrate that our approach achieves robust identity preservation and tracking stability. Notably, our advantage becomes more pronounced with increased target density, establishing a solid foundation for simultaneous multi-target video segmentation in the wild.

Keywords

Cite

@article{arxiv.2601.09699,
  title  = {SAM3-DMS: Decoupled Memory Selection for Multi-target Video Segmentation of SAM3},
  author = {Ruiqi Shen and Chang Liu and Henghui Ding},
  journal= {arXiv preprint arXiv:2601.09699},
  year   = {2026}
}

Comments

Code: https://github.com/FudanCVL/SAM3-DMS

R2 v1 2026-07-01T09:04:41.669Z