The Segment Anything Model (SAM) has established itself as a powerful zero-shot image segmentation model, employing interactive prompts such as points to generate masks. This paper presents SAM-PT, a method extending SAM’s capability to tracking and segmenting anything in dynamic videos. SAM-PT leverages robust and sparse point selection and propagation techniques for mask generation, demonstrating that a SAM-based segmentation tracker can yield strong zero-shot performance across popular video object segmentation benchmarks, including DAVIS, YouTube-VOS, and MOSE. Compared to traditional object-centric mask propagation strategies, we uniquely use point propagation to exploit local structure information that is agnostic to object semantics. We highlight the merits of point-based tracking through direct evaluation on the zero-shot open-world Unidentified Video Objects (UVO) benchmark. To further enhance our approach, we utilize K-Medoids clustering for point initialization and track both positive and negative points to clearly distinguish the target object. We also employ multiple mask decoding passes for mask refinement and devise a point re-initialization strategy to improve tracking accuracy.
@article{sam-pt,
title = {Segment Anything Meets Point Tracking},
author = {Rajič, Frano and Ke, Lei and Tai, Yu-Wing and Tang, Chi-Keung and Danelljan, Martin and Yu, Fisher},
journal = {arXiv:2307.01197},
year = {2023}
}
