{"ID":2824325,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.23427","arxiv_id":"2512.23427","title":"Towards Integrating Uncertainty for Domain-Agnostic Segmentation","abstract":"Foundation models for segmentation such as the Segment Anything Model (SAM) family exhibit strong zero-shot performance, but remain vulnerable in shifted or limited-knowledge domains. This work investigates whether uncertainty quantification can mitigate such challenges and enhance model generalisability in a domain-agnostic manner. To this end, we (1) curate UncertSAM, a benchmark comprising eight datasets designed to stress-test SAM under challenging segmentation conditions including shadows, transparency, and camouflage; (2) evaluate a suite of lightweight, post-hoc uncertainty estimation methods; and (3) assess a preliminary uncertainty-guided prediction refinement step. Among evaluated approaches, a last-layer Laplace approximation yields uncertainty estimates that correlate well with segmentation errors, indicating a meaningful signal. While refinement benefits are preliminary, our findings underscore the potential of incorporating uncertainty into segmentation models to support robust, domain-agnostic performance. Our benchmark and code are made publicly available.","short_abstract":"Foundation models for segmentation such as the Segment Anything Model (SAM) family exhibit strong zero-shot performance, but remain vulnerable in shifted or limited-knowledge domains. This work investigates whether uncertainty quantification can mitigate such challenges and enhance model generalisability in a domain-ag...","url_abs":"https://arxiv.org/abs/2512.23427","url_pdf":"https://arxiv.org/pdf/2512.23427v1","authors":"[\"Jesse Brouwers\",\"Xiaoyan Xing\",\"Alexander Timans\"]","published":"2025-12-29T12:46:21Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.LG\"]","methods":"[]","has_code":false}