{"ID":2834562,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.01789","arxiv_id":"2512.01789","title":"SAM3-UNet: Simplified Adaptation of Segment Anything Model 3","abstract":"In this paper, we introduce SAM3-UNet, a simplified variant of Segment Anything Model 3 (SAM3), designed to adapt SAM3 for downstream tasks at a low cost. Our SAM3-UNet consists of three components: a SAM3 image encoder, a simple adapter for parameter-efficient fine-tuning, and a lightweight U-Net-style decoder. Preliminary experiments on multiple tasks, such as mirror detection and salient object detection, demonstrate that the proposed SAM3-UNet outperforms the prior SAM2-UNet and other state-of-the-art methods, while requiring less than 6 GB of GPU memory during training with a batch size of 12. The code is publicly available at https://github.com/WZH0120/SAM3-UNet.","short_abstract":"In this paper, we introduce SAM3-UNet, a simplified variant of Segment Anything Model 3 (SAM3), designed to adapt SAM3 for downstream tasks at a low cost. Our SAM3-UNet consists of three components: a SAM3 image encoder, a simple adapter for parameter-efficient fine-tuning, and a lightweight U-Net-style decoder. Prelim...","url_abs":"https://arxiv.org/abs/2512.01789","url_pdf":"https://arxiv.org/pdf/2512.01789v1","authors":"[\"Xinyu Xiong\",\"Zihuang Wu\",\"Lei Lu\",\"Yufa Xia\"]","published":"2025-12-01T15:27:35Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false,"code_links":[{"ID":606422,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2834562,"paper_url":"https://arxiv.org/abs/2512.01789","paper_title":"SAM3-UNet: Simplified Adaptation of Segment Anything Model 3","repo_url":"https://github.com/WZH0120/SAM3-UNet","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}