cs.CVNov 3, 2026

ProGeo-RC: Prototype-guided Semantic Reliability and Geometric Error Calibration for Semi-supervised Cardiac Image Segmentation

Authors: Yunfei Zhu, Wenfeng Zhang, Pan Zeng, Xin Huang, Tao Jia, Weiting Shao, Wei Hu, Jianming Hu

Organizations: College of Computer and Information Science, Research Center of Biomimetic Intelligence and Computational Biomedicine, Chongqing Normal University, Chongqing, China · School of Educational Sciences, Chongqing Normal University, Chongqing, China · School of Physics and Electrical Engineering, Chongqing Normal University, Chongqing, China

Abstract

Semi-supervised medical image segmentation aims to train reliable segmentation models using limited labeled data and abundant unlabeled data. However, existing methods typically rely on prediction confidence or model consistency to estimate pseudo-label reliability, but these criteria are insufficient to characterize both semantic reliability and geometric plausibility. To address this issue, we propose ProGeo-RC, a prototype-guided semantic reliability and geometric error calibration framework for semi-supervised medical image segmentation. ProGeo-RC consists of two key modules: Prototype-guided Semantic Reliability Estimation (PSRE) and Structured Semantic-Geometric Error Calibration (S-SGEC). PSRE uses probabilistic class prototypes to estimate the semantic reliability of unlabeled predictions and calibrate pixel-wise teacher pseudo labels, while S-SGEC introduces signed-distance-field-based geometric constraints to refine ambiguous boundaries and suppress spurious foreground responses outside target structures. Experiments on the ACDC and MVAA-CT datasets show that ProGeo-RC achieves better performance compared with representative semi-supervised segmentation methods. Under the 10% labeled setting on ACDC, ProGeo-RC improves DSC and Jaccard by 0.37% and 0.49% over the corresponding second-best results, while reducing 95HD and ASD by 0.61 and 0.21, respectively. These results indicate that joint semantic reliability estimation and geometric error calibration can provide more reliable pseudo supervision for semi-supervised medical image segmentation.

Figures & tables

Explore similar work

CardsList
  1. Reliability-Role Collaborative Learning for Semi-Supervised Cardiac Image Segmentation

    Nov 3, 2026Zhongrui Xie, Wenfeng Zhang, Wei Hu +5Semi-Supervised Medical Image SegmentationMedical Image Segmentation

  2. Are We Overconfident in Models and Results for Semi-Supervised 3D Medical Image Segmentation?

    May 25, 2026Jun Li, Ziwei QinBenchmark ContaminationSemi-Supervised Medical Image Segmentation

  3. Peripheral prototypical contrastive loss for supervised medical image segmentation

    Nov 1, 2026Takuro Shimaya, Ryosaku Shino, Masahiro SaikoPrototypical Contrastive LearningMedical Image Segmentation