Segment Any Motion in Videos
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912324151934976 |
|---|---|
| author | Huang, Nan Zheng, Wenzhao Xu, Chenfeng Keutzer, Kurt Zhang, Shanghang Kanazawa, Angjoo Wang, Qianqian |
| author_facet | Huang, Nan Zheng, Wenzhao Xu, Chenfeng Keutzer, Kurt Zhang, Shanghang Kanazawa, Angjoo Wang, Qianqian |
| contents | Moving object segmentation is a crucial task for achieving a high-level understanding of visual scenes and has numerous downstream applications. Humans can effortlessly segment moving objects in videos. Previous work has largely relied on optical flow to provide motion cues; however, this approach often results in imperfect predictions due to challenges such as partial motion, complex deformations, motion blur and background distractions. We propose a novel approach for moving object segmentation that combines long-range trajectory motion cues with DINO-based semantic features and leverages SAM2 for pixel-level mask densification through an iterative prompting strategy. Our model employs Spatio-Temporal Trajectory Attention and Motion-Semantic Decoupled Embedding to prioritize motion while integrating semantic support. Extensive testing on diverse datasets demonstrates state-of-the-art performance, excelling in challenging scenarios and fine-grained segmentation of multiple objects. Our code is available at https://motion-seg.github.io/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_22268 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Segment Any Motion in Videos Huang, Nan Zheng, Wenzhao Xu, Chenfeng Keutzer, Kurt Zhang, Shanghang Kanazawa, Angjoo Wang, Qianqian Computer Vision and Pattern Recognition Moving object segmentation is a crucial task for achieving a high-level understanding of visual scenes and has numerous downstream applications. Humans can effortlessly segment moving objects in videos. Previous work has largely relied on optical flow to provide motion cues; however, this approach often results in imperfect predictions due to challenges such as partial motion, complex deformations, motion blur and background distractions. We propose a novel approach for moving object segmentation that combines long-range trajectory motion cues with DINO-based semantic features and leverages SAM2 for pixel-level mask densification through an iterative prompting strategy. Our model employs Spatio-Temporal Trajectory Attention and Motion-Semantic Decoupled Embedding to prioritize motion while integrating semantic support. Extensive testing on diverse datasets demonstrates state-of-the-art performance, excelling in challenging scenarios and fine-grained segmentation of multiple objects. Our code is available at https://motion-seg.github.io/. |
| title | Segment Any Motion in Videos |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.22268 |