Saved in:
Bibliographic Details
Main Authors: Brouwers, Jesse, Xing, Xiaoyan, Timans, Alexander
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.23427
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918265356288000
author Brouwers, Jesse
Xing, Xiaoyan
Timans, Alexander
author_facet Brouwers, Jesse
Xing, Xiaoyan
Timans, Alexander
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23427
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Integrating Uncertainty for Domain-Agnostic Segmentation
Brouwers, Jesse
Xing, Xiaoyan
Timans, Alexander
Computer Vision and Pattern Recognition
Machine Learning
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.
title Towards Integrating Uncertainty for Domain-Agnostic Segmentation
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2512.23427