Learning Label-Efficient Interpretable Medical Image Diagnosis via Semi-supervised Hypergraph Concept Bottleneck Model
2026-06-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address the problem of deep learning models being hard to interpret in medical image diagnosis, especially for tricky cases like Placenta Accreta Spectrum (PAS). They propose a new method that uses Concept Bottleneck Models enhanced with hypergraph learning, which helps the model understand complex relationships between medical concepts and makes better use of limited labeled data. Their approach improves transparency and diagnosis accuracy by allowing clinicians to check intermediate concept predictions. They tested this method on PAS, breast ultrasound, and skin lesion images, showing it works well across different datasets.
Deep learningMedical image analysisInterpretabilityConcept Bottleneck ModelsHypergraph learningSemi-supervised learningPlacenta Accreta SpectrumPseudo-labelingUltrasound imagingDermoscopic imaging
Authors
Yijun Yang, Ruiqiang Xiao, Lijie Hu, Angelica I Aviles-Rivero, Yunzhu Wu, Jing Qin, Lei Zhu
Abstract
Deep learning has revolutionized medical image analysis, delivering exceptional diagnostic accuracy across diverse applications. Yet, the lack of interpretability in its decision-making hinders clinical adoption, particularly in high-stakes medical contexts where transparency is paramount for trustworthiness. For example, in Placenta Accreta Spectrum (PAS), subtle cues in ultrasound imaging challenge reliable diagnosis, rendering black-box models untrustworthy for accurate scoring. To address this, Concept Bottleneck Models (CBMs) offer a promising avenue by embedding clinically meaningful intermediate concepts into the diagnosis pipeline, enabling clinicians to scrutinize and refine model outputs. However, conventional CBMs falter in capturing complex inter-concept dependencies and demand costly, expert-driven concept annotations, limiting their scalability. This study introduces a novel semi-supervised CBM framework designed for medical imaging, which leverages dual-level hypergraph learning to model high-order concept dependencies and generate domain-adaptive pseudo-labels. Our approach achieves superior interpretability and performance by integrating a concept-level hypergraph for enhanced reasoning and an image-level hypergraph for robust pseudo-label generation. Experiments on a newly annotated PAS ultrasound dataset and a breast ultrasound public dataset demonstrate the effectiveness of the proposed concept label-efficient interpretable framework. Its universality is further validated on the dermoscopic image dataset SkinCon. The code is available at https://github.com/scott-yjyang/HyperCBM.