Scholarlink

Fundus imaging techniques, such as Fundus Fluorescein Angiography (FFA) and Optical Coherence Tomography (OCT), serve as pivotal diagnostic tools for retinal disease detection. These modalities reveal intricate variations in retinal tissue structure, thereby aiding clinicians in accurate diagnosis and treatment planning. Recent advancements in Multiple Instance Learning (MIL) have transformed the analysis of fundus images datasets by effectively extracting discriminative features from image-level key instances. However, a critical limitation of conventional MIL methods lies in their neglect of region-level key instances (e.g., localized lesions) within the 2-D spatial domain, particularly in fundus imaging where pathological regions are often sparse. To address this gap, we propose a Contrastive Hierarchical Graph based Multiple Instance Learning (CHG-MIL) framework, which integrates three novel components: (1) a Spatial Instance Graph (SIG) module that preserves 2-D spatial topology and mines contextual relationships among neighboring region-level instances; (2) a Hierarchical Instance Interaction (HII) module that utilizes deeper semantic representations to refine shallow-layer features through cross-hierarchy guidance; and (3) a Hierarchical Contrastive Loss (HCL) designed to suppress redundant, non-disease-related features in nodes and instances. Extensive experiments on the APTOS2023 and GAMMA datasets demonstrate the superiority of our proposed method over existing state-of-the-art MIL methods.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Contrastive Hierarchical Graph Based Multiple Instance Learning for Fundus Screening

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yubo Tan,
  • Shiye Wang,
  • Wenda Shen,
  • Yong-Jie Li

摘要

Fundus imaging techniques, such as Fundus Fluorescein Angiography (FFA) and Optical Coherence Tomography (OCT), serve as pivotal diagnostic tools for retinal disease detection. These modalities reveal intricate variations in retinal tissue structure, thereby aiding clinicians in accurate diagnosis and treatment planning. Recent advancements in Multiple Instance Learning (MIL) have transformed the analysis of fundus images datasets by effectively extracting discriminative features from image-level key instances. However, a critical limitation of conventional MIL methods lies in their neglect of region-level key instances (e.g., localized lesions) within the 2-D spatial domain, particularly in fundus imaging where pathological regions are often sparse. To address this gap, we propose a Contrastive Hierarchical Graph based Multiple Instance Learning (CHG-MIL) framework, which integrates three novel components: (1) a Spatial Instance Graph (SIG) module that preserves 2-D spatial topology and mines contextual relationships among neighboring region-level instances; (2) a Hierarchical Instance Interaction (HII) module that utilizes deeper semantic representations to refine shallow-layer features through cross-hierarchy guidance; and (3) a Hierarchical Contrastive Loss (HCL) designed to suppress redundant, non-disease-related features in nodes and instances. Extensive experiments on the APTOS2023 and GAMMA datasets demonstrate the superiority of our proposed method over existing state-of-the-art MIL methods.