OFD-Net: Teacher-Free Reliable Semi-supervised Medical Image Segmentation with Orthogonal Feature Disentanglement Net of Foreground-Background

2026-07-18

AI Summary

This work addresses the challenges of error accumulation and confirmation bias in semi-supervised medical image segmentation, which often arise from pseudo-labels lacking structural guidance. To this end, the authors propose the first single-network framework that eliminates the need for a teacher network by introducing an Orthogonal Feature Disentanglement Module (OFDM) to explicitly separate foreground and background features. A Disentanglement-Guided Module (DGM) further injects structural priors to enable reliability-aware pseudo-label learning. Notably, this approach pioneers foreground-background orthogonal disentanglement within a single-network architecture and designs a structure-consistency-based strategy for pseudo-label selection. Extensive experiments demonstrate that the proposed method significantly improves both segmentation accuracy and training stability across four benchmarks: ISIC-2016, Kvasir-SEG, Synapse, and ACDC.

Abstract

Semi-supervised learning (SSL) is an effective solution for medical image segmentation with limited annotations. Existing SSL methods mainly rely on pseudo-labels generated by teacher-student supervision or cross-network consistency. However, these methods lack an explicit structural reference for judging pseudo-label quality. Low-quality pseudo-labels may lead to unreliable training, error accumulation and confirmation bias when processing unlabeled data with substantial appearance variations. To address this issue, we proposed OFD-Net, a teacher-free single-network framework for reliable semi-supervised medical image segmentation. OFD-Net employs an Orthogonal Feature Disentanglement Module (OFDM) to capture OFD features for reliable SSL by disentangling unlabeled data into background and foreground representations with a reliable structural distribution, thereby effectively reducing error accumulation and alleviating confirmation bias among unlabeled data. Specifically, OFD-Net explicitly employs a Disentanglement Guidance Module (DGM) to inject the resulting structural priors of foreground-background into the decoder by deformable convolution processing, and outputs predictions with clearer foreground representations. Based on DGM and the OFDM, we further develop a reliability-aware pseudo-label learning mechanism that evaluates unlabeled supervision according to the structural consistency between the main prediction and the disentangled foreground-background responses, and then down-weights unreliable regions during training. Extensive experiments on four public medical image segmentation benchmarks, namely ISIC-2016, Kvasir-SEG, Synapse, and ACDC, validate the effectiveness of OFD-Net. These results confirm that orthogonal foreground-background disentanglement enables OFD-Net to establish an efficient and reliable training paradigm within a teacher-free single-network framework.

Problem

• semi-supervised learning
• medical image segmentation
• pseudo-label quality
• error accumulation
• confirmation bias

Innovation

• Orthogonal Feature Disentanglement
• Teacher-Free Semi-supervised Learning
• Foreground-Background Disentanglement
• Reliability-Aware Pseudo-Labeling
• Medical Image Segmentation