Test-time Adaptation of Pelvic Bone Segmentation Models via Dynamic Reliability-Guided

2026-08-01

AI Summary

This work addresses the significant performance degradation of pelvic CT segmentation models when deployed in new hospitals due to cross-center domain shift. To tackle this issue, the authors propose ReGA, a dynamic reliability-guided test-time adaptation framework that operates without access to source-domain data. ReGA introduces an innovative SICE reliability criterion to generate high-quality pseudo-labels and integrates confidence-weighted boundary refinement, anatomically consistent region contrastive learning, and a teacher-student mechanism designed to alleviate voxel-level class imbalance for online feature adaptation. Extensive experiments on three heterogeneous 3D pelvic CT datasets demonstrate that ReGA substantially outperforms existing test-time adaptation methods, effectively enhancing segmentation performance in unseen clinical domains.

Abstract

Reliable pelvic bone segmentation (PBS) from CT is essential for robot-assisted pelvic trauma surgery, yet deploying a source-trained model to a new hospital suffers from severe performance degradation due to cross-center domain shifts. While test-time adaptation (TTA) enables online model adaptation without accessing source data, existing methods show limited effectiveness for PBS, facing challenges including boundary degradation, anatomical inconsistency under domain shifts, and voxel-level class imbalance. To address these challenges, we propose a novel closed-loop dynamic Reliability-Guided TTA framework (ReGA) for PBS. Specifically, we introduce a pseudo-label reliability criterion termed Segmentation Inference Consistency Evaluation (SICE), which jointly measures region overlap and boundary deviation via dropout-based ensemble predictions. Based on SICE, a trust-weighted refinement module adaptively updates features to mitigate boundary errors in pseudo-labels. Furthermore, a confidence-weighted region-level contrastive learning strategy is proposed to enforce anatomical consistency. Finally, ReGA follows the teacher-student (TS) scheme to alleviate voxel-level class imbalance. Experiments on three heterogeneous 3D pelvic CT datasets demonstrate that ReGA consistently outperforms state-of-the-art TTA methods, enabling effective adaptation of the source-trained PBS model to unseen clinical domains. The code is available at https://github.com/Ren-ling/ReGA.

Problem

• pelvic bone segmentation
• test-time adaptation
• domain shift
• anatomical consistency
• class imbalance

Innovation

• Test-time Adaptation
• Reliability-guided Learning
• Segmentation Inference Consistency Evaluation
• Contrastive Learning
• Domain Shift