Uncertainty-aware Cross-training for Semi-supervised Medical Image Segmentation

2025-08-12

AI Summary

In semi-supervised medical image segmentation, models suffer from significant cognitive bias and produce low-confidence pseudo-labels. To address this, we propose UC-Seg, an uncertainty-aware cross-training framework. Methodologically, UC-Seg employs a dual-branch network architecture to enhance feature robustness via cross-subnetwork consistency regularization. It introduces an uncertainty-map-guided pseudo-label generation mechanism that enables high-confidence pseudo-label selection and dynamic refinement. Additionally, it integrates multi-modal imaging information (e.g., MRI, CT, ultrasound, and endoscopy) to strengthen representation learning. Extensive experiments across diverse medical imaging modalities demonstrate that UC-Seg consistently outperforms state-of-the-art methods, achieving new SOTA performance in both Dice score and generalizability. Crucially, it significantly improves unlabeled data utilization efficiency and model robustness under limited supervision.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Semi-supervised learning has gained considerable popularity in medical image segmentation tasks due to its capability to reduce reliance on expert-examined annotations. Several mean-teacher (MT) based semi-supervised methods utilize consistency regularization to effectively leverage valuable information from unlabeled data. However, these methods often heavily rely on the student model and overlook the potential impact of cognitive biases within the model. Furthermore, some methods employ co-training using pseudo-labels derived from different inputs, yet generating high-confidence pseudo-labels from perturbed inputs during training remains a significant challenge. In this paper, we propose an Uncertainty-aware Cross-training framework for semi-supervised medical image Segmentation (UC-Seg). Our UC-Seg framework incorporates two distinct subnets to effectively explore and leverage the correlation between them, thereby mitigating cognitive biases within the model. Specifically, we present a Cross-subnet Consistency Preservation (CCP) strategy to enhance feature representation capability and ensure feature consistency across the two subnets. This strategy enables each subnet to correct its own biases and learn shared semantics from both labeled and unlabeled data. Additionally, we propose an Uncertainty-aware Pseudo-label Generation (UPG) component that leverages segmentation results and corresponding uncertainty maps from both subnets to generate high-confidence pseudo-labels. We extensively evaluate the proposed UC-Seg on various medical image segmentation tasks involving different modality images, such as MRI, CT, ultrasound, colonoscopy, and so on. The results demonstrate that our method achieves superior segmentation accuracy and generalization performance compared to other state-of-the-art semi-supervised methods. Our code will be released at https://github.com/taozh2017/UCSeg.

Problem

• Addresses cognitive biases in semi-supervised medical image segmentation
• Generates high-confidence pseudo-labels from uncertain perturbed inputs
• Improves feature consistency across dual subnets for better accuracy

Innovation

• Uncertainty-aware cross-training framework
• Cross-subnet consistency preservation strategy
• Uncertainty-aware pseudo-label generation component