SAM 3: Segment Anything with Concepts
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910080846266368 |
|---|---|
| author | Carion, Nicolas Gustafson, Laura Hu, Yuan-Ting Debnath, Shoubhik Hu, Ronghang Suris, Didac Ryali, Chaitanya Alwala, Kalyan Vasudev Khedr, Haitham Huang, Andrew Lei, Jie Ma, Tengyu Guo, Baishan Kalla, Arpit Marks, Markus Greer, Joseph Wang, Meng Sun, Peize Rädle, Roman Afouras, Triantafyllos Mavroudi, Effrosyni Xu, Katherine Wu, Tsung-Han Zhou, Yu Momeni, Liliane Hazra, Rishi Ding, Shuangrui Vaze, Sagar Porcher, Francois Li, Feng Li, Siyuan Kamath, Aishwarya Cheng, Ho Kei Dollár, Piotr Ravi, Nikhila Saenko, Kate Zhang, Pengchuan Feichtenhofer, Christoph |
| author_facet | Carion, Nicolas Gustafson, Laura Hu, Yuan-Ting Debnath, Shoubhik Hu, Ronghang Suris, Didac Ryali, Chaitanya Alwala, Kalyan Vasudev Khedr, Haitham Huang, Andrew Lei, Jie Ma, Tengyu Guo, Baishan Kalla, Arpit Marks, Markus Greer, Joseph Wang, Meng Sun, Peize Rädle, Roman Afouras, Triantafyllos Mavroudi, Effrosyni Xu, Katherine Wu, Tsung-Han Zhou, Yu Momeni, Liliane Hazra, Rishi Ding, Shuangrui Vaze, Sagar Porcher, Francois Li, Feng Li, Siyuan Kamath, Aishwarya Cheng, Ho Kei Dollár, Piotr Ravi, Nikhila Saenko, Kate Zhang, Pengchuan Feichtenhofer, Christoph |
| contents | We present Segment Anything Model (SAM) 3, a unified model that detects, segments, and tracks objects in images and videos based on concept prompts, which we define as either short noun phrases (e.g., "yellow school bus"), image exemplars, or a combination of both. Promptable Concept Segmentation (PCS) takes such prompts and returns segmentation masks and unique identities for all matching object instances. To advance PCS, we build a scalable data engine that produces a high-quality dataset with 4M unique concept labels, including hard negatives, across images and videos. Our model consists of an image-level detector and a memory-based video tracker that share a single backbone. Recognition and localization are decoupled with a presence head, which boosts detection accuracy. SAM 3 doubles the accuracy of existing systems in both image and video PCS, and improves previous SAM capabilities on visual segmentation tasks. We open source SAM 3 along with our new Segment Anything with Concepts (SA-Co) benchmark for promptable concept segmentation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_16719 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SAM 3: Segment Anything with Concepts Carion, Nicolas Gustafson, Laura Hu, Yuan-Ting Debnath, Shoubhik Hu, Ronghang Suris, Didac Ryali, Chaitanya Alwala, Kalyan Vasudev Khedr, Haitham Huang, Andrew Lei, Jie Ma, Tengyu Guo, Baishan Kalla, Arpit Marks, Markus Greer, Joseph Wang, Meng Sun, Peize Rädle, Roman Afouras, Triantafyllos Mavroudi, Effrosyni Xu, Katherine Wu, Tsung-Han Zhou, Yu Momeni, Liliane Hazra, Rishi Ding, Shuangrui Vaze, Sagar Porcher, Francois Li, Feng Li, Siyuan Kamath, Aishwarya Cheng, Ho Kei Dollár, Piotr Ravi, Nikhila Saenko, Kate Zhang, Pengchuan Feichtenhofer, Christoph Computer Vision and Pattern Recognition Artificial Intelligence We present Segment Anything Model (SAM) 3, a unified model that detects, segments, and tracks objects in images and videos based on concept prompts, which we define as either short noun phrases (e.g., "yellow school bus"), image exemplars, or a combination of both. Promptable Concept Segmentation (PCS) takes such prompts and returns segmentation masks and unique identities for all matching object instances. To advance PCS, we build a scalable data engine that produces a high-quality dataset with 4M unique concept labels, including hard negatives, across images and videos. Our model consists of an image-level detector and a memory-based video tracker that share a single backbone. Recognition and localization are decoupled with a presence head, which boosts detection accuracy. SAM 3 doubles the accuracy of existing systems in both image and video PCS, and improves previous SAM capabilities on visual segmentation tasks. We open source SAM 3 along with our new Segment Anything with Concepts (SA-Co) benchmark for promptable concept segmentation. |
| title | SAM 3: Segment Anything with Concepts |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2511.16719 |