{"ID":6620454,"CreatedAt":"2026-07-15T01:01:48.440468303Z","UpdatedAt":"2026-07-15T03:28:55.185153975Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.12292","arxiv_id":"2607.12292","title":"Semantic-Edge Response Decoding of SAM3 for Zero-Shot Crack Segmentation","abstract":"Crack segmentation is essential for infrastructure inspection and structural health assessment, but existing high-performance methods typically require task-specific pixel-level annotations and training. Text-promptable vision foundation models enable zero-shot deployment, yet their final mask proposals are poorly suited to thin, fragmented, and low-contrast cracks, whose evidence may be suppressed, truncated, or over-expanded during mask generation. We find that language-conditioned semantic responses within the SAM3 decoder preserve more continuous and complete crack evidence than its final masks. Based on this observation, we propose Semantic-Edge Response Decoding (SERD), which interprets internal responses as a dense crack-likelihood field, calibrates them with a lightweight edge prior, and generates crack masks using a unified global threshold, without annotation or fine-tuning. Experiments on six public datasets show that SERD consistently improves over native SAM3 and outperforms the compared zero-shot and open-vocabulary segmentation methods, achieving an average Crack IoU of 61.14\\%, 4.63 points higher than SAM3. Further analyses show that most gains arise from directly decoding internal semantic responses, while edge calibration improves structural recovery and false-positive control without increasing end-to-end inference overhead. These results suggest that, for thin and non-compact targets, internal continuous responses can provide a more transferable interface than the final masks of foundation models. Code is available at: https://github.com/xauat-liushipeng/SERD","short_abstract":"Crack segmentation is essential for infrastructure inspection and structural health assessment, but existing high-performance methods typically require task-specific pixel-level annotations and training. Text-promptable vision foundation models enable zero-shot deployment, yet their final mask proposals are poorly suit...","url_abs":"https://arxiv.org/abs/2607.12292","url_pdf":"https://arxiv.org/pdf/2607.12292v1","authors":"[\"Shipeng Liu\",\"Zhanping Song\",\"Liang Zhao\",\"Dengfeng Chen\"]","published":"2026-07-14T03:01:12Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false,"code_links":[{"ID":614235,"CreatedAt":"2026-07-15T01:01:48.440468303Z","UpdatedAt":"2026-07-15T01:01:48.440468303Z","DeletedAt":null,"paper_id":6620454,"paper_url":"https://arxiv.org/abs/2607.12292","paper_title":"Semantic-Edge Response Decoding of SAM3 for Zero-Shot Crack Segmentation","repo_url":"https://github.com/xauat-liushipeng/SERD","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}