Graph-of-Differences: Anatomy-Structured Difference Alignment for Medical Image Re-Identification

2026-06-19

AI Summary

This work addresses the vulnerability of medical image re-identification to shortcut learning and its lack of auditability grounded in anatomical structure. It proposes the first approach that explicitly models named anatomical regions as graph nodes, anchoring identification signals to homologous anatomical structures through soft node matching and graph-level discrepancy alignment. By substituting unstable pixel-wise heatmaps with structural-level evidence, the method enables interpretable and verifiable decision-making and supports quantitative auditing via node insertion or deletion. Evaluated on internal benchmarks of fundus and chest X-ray images, the approach improves Rank-1 accuracy by 7.1% and 3.1%, respectively, while zero-shot external transfer further demonstrates substantially enhanced generalization capability.

Abstract

Medical image re-identification (MedReID) enables longitudinal patient linkage but remains vulnerable to shortcut learning and often produces decisions that clinicians cannot audit against named anatomy. We propose Graph-of-Differences (GoD), which grounds identity comparisons in explicit anatomical structure. Each image is represented as an anatomy graph whose nodes correspond to named anatomical regions; given an image pair, soft node correspondence is established, and differences are computed over matched anatomy. A graph-level difference alignment objective ties these anatomy-matched differences to the global backbone difference, ensuring the retrieval signal is anchored in homologous structures rather than arbitrary spatial tokens. Explanations are defined over named graph nodes and quantitatively audited via node insertion/deletion tests, replacing unstable pixel heatmaps with verifiable structure-level evidence. On internal benchmarks, GoD improves Rank-1 by +7.1 pp on fundus and +3.1 pp on CXR over a strong frozen-backbone baseline, with further gains on zero-shot external transfers confirming that anatomy grounding improves both accuracy and generalization. Code is available at https://github.com/GenMI-Lab/GoD.git.

Problem

• Medical image re-identification
• Shortcut learning
• Anatomical structure
• Clinical auditability
• Longitudinal patient linkage

Innovation

• anatomy-structured representation
• graph-based difference alignment
• medical image re-identification
• interpretable MedReID
• zero-shot generalization