R$^{2}$Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and Statistical Rejection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shen, Shuaike, Liu, Ke, Xie, Jiaqing, Gao, Shangde, Shen, Chunhua, Liu, Ge, Crispin-Ortuzar, Mireia, Gao, Shangqi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909905845223424
author Shen, Shuaike
Liu, Ke
Xie, Jiaqing
Gao, Shangde
Shen, Chunhua
Liu, Ge
Crispin-Ortuzar, Mireia
Gao, Shangqi
author_facet Shen, Shuaike
Liu, Ke
Xie, Jiaqing
Gao, Shangde
Shen, Chunhua
Liu, Ge
Crispin-Ortuzar, Mireia
Gao, Shangqi
contents Foundation models for medical image segmentation struggle under out-of-distribution (OOD) shifts, often producing fragmented false positives on OOD tumors. We introduce R$^{2}$Seg, a training-free framework for robust OOD tumor segmentation that operates via a two-stage Reason-and-Reject process. First, the Reason step employs an LLM-guided anatomical reasoning planner to localize organ anchors and generate multi-scale ROIs. Second, the Reject step applies two-sample statistical testing to candidates generated by a frozen foundation model (BiomedParse) within these ROIs. This statistical rejection filter retains only candidates significantly different from normal tissue, effectively suppressing false positives. Our framework requires no parameter updates, making it compatible with zero-update test-time augmentation and avoiding catastrophic forgetting. On multi-center and multi-modal tumor segmentation benchmarks, R$^{2}$Seg substantially improves Dice, specificity, and sensitivity over strong baselines and the original foundation models. Code are available at https://github.com/Eurekashen/R2Seg.
format Preprint
id arxiv_https___arxiv_org_abs_2511_12691
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle R$^{2}$Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and Statistical Rejection
Shen, Shuaike
Liu, Ke
Xie, Jiaqing
Gao, Shangde
Shen, Chunhua
Liu, Ge
Crispin-Ortuzar, Mireia
Gao, Shangqi
Computer Vision and Pattern Recognition
Artificial Intelligence
Foundation models for medical image segmentation struggle under out-of-distribution (OOD) shifts, often producing fragmented false positives on OOD tumors. We introduce R$^{2}$Seg, a training-free framework for robust OOD tumor segmentation that operates via a two-stage Reason-and-Reject process. First, the Reason step employs an LLM-guided anatomical reasoning planner to localize organ anchors and generate multi-scale ROIs. Second, the Reject step applies two-sample statistical testing to candidates generated by a frozen foundation model (BiomedParse) within these ROIs. This statistical rejection filter retains only candidates significantly different from normal tissue, effectively suppressing false positives. Our framework requires no parameter updates, making it compatible with zero-update test-time augmentation and avoiding catastrophic forgetting. On multi-center and multi-modal tumor segmentation benchmarks, R$^{2}$Seg substantially improves Dice, specificity, and sensitivity over strong baselines and the original foundation models. Code are available at https://github.com/Eurekashen/R2Seg.
title R$^{2}$Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and Statistical Rejection
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2511.12691