通过半监督超图概念瓶颈模型实现标签高效的医学图像可解释诊断
Learning Label-Efficient Interpretable Medical Image Diagnosis via Semi-supervised Hypergraph Concept Bottleneck Model
- HKUST(GZ)（香港科技大学（广州））
- Joy Future Academy（京东探索研究院）
- MBZUAI（穆罕默德·本·拉希德智能研究院）
- Tsinghua University（清华大学）
- Sichuan University（四川大学）
- PolyU
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
提出一种半监督超图概念瓶颈模型，利用双层超图学习建模高阶概念依赖并生成领域自适应伪标签，在胎盘植入谱系等医学图像诊断中实现高可解释性和性能。
AI中文摘要：
深度学习在医学图像分析中取得了革命性进展，在多种应用中提供了卓越的诊断准确性。然而，其决策缺乏可解释性阻碍了临床采纳，特别是在高风险医疗场景中，透明度对可信度至关重要。例如，在胎盘植入谱系（PAS）中，超声图像中的细微线索挑战了可靠诊断，使得黑盒模型难以获得准确的评分信任。为了解决这一问题，概念瓶颈模型（CBM）通过将临床上有意义的中间概念嵌入诊断流程，提供了一种有前景的途径，使临床医生能够审查和优化模型输出。然而，传统的CBM在捕捉复杂的概念间依赖关系方面表现不佳，并且需要昂贵、专家驱动的概念注释，限制了其可扩展性。本研究引入了一种新颖的半监督CBM框架，专为医学成像设计，利用双层超图学习来建模高阶概念依赖并生成领域自适应伪标签。我们的方法通过集成概念级超图以增强推理和图像级超图以生成鲁棒的伪标签，实现了卓越的可解释性和性能。在新标注的PAS超声数据集和乳腺超声公共数据集上的实验证明了所提出的概念标签高效可解释框架的有效性。其通用性在皮肤镜图像数据集SkinCon上得到了进一步验证。代码可在https://github.com/scott-yjyang/HyperCBM获取。
英文摘要：
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.
