Towards All-in-One Medical Image Re-Identification

2025-03-11

AI Summary

This paper formally introduces and addresses the medical image re-identification (MedReID) problem—enabling cross-device, cross-modality identity matching of medical images to support personalized diagnosis and privacy-preserving healthcare. To this end, we establish the first unified multi-modal MedReID benchmark and propose the Continual Modality Parameter Adapter (ComPA) alongside a differential feature alignment paradigm, enabling a single model to dynamically adapt to and robustly re-identify CT, MRI, and X-ray images. Our approach integrates continual modality representation learning, dynamic parameter adaptation, cross-modal metric learning, and medical foundation model fusion. Extensive evaluation across 11 real-world medical datasets demonstrates significant improvements over 25 medical foundation models and 8 large multimodal language models. The framework has been successfully deployed in two practical applications: historical enhancement for diagnostic decision support and privacy-preserving medical data management.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Security and Privacy: Data transparency and provenance

Abstract

Medical image re-identification (MedReID) is under-explored so far, despite its critical applications in personalized healthcare and privacy protection. In this paper, we introduce a thorough benchmark and a unified model for this problem. First, to handle various medical modalities, we propose a novel Continuous Modality-based Parameter Adapter (ComPA). ComPA condenses medical content into a continuous modality representation and dynamically adjusts the modality-agnostic model with modality-specific parameters at runtime. This allows a single model to adaptively learn and process diverse modality data. Furthermore, we integrate medical priors into our model by aligning it with a bag of pre-trained medical foundation models, in terms of the differential features. Compared to single-image feature, modeling the inter-image difference better fits the re-identification problem, which involves discriminating multiple images. We evaluate the proposed model against 25 foundation models and 8 large multi-modal language models across 11 image datasets, demonstrating consistently superior performance. Additionally, we deploy the proposed MedReID technique to two real-world applications, i.e., history-augmented personalized diagnosis and medical privacy protection. Codes and model is available at href{https://github.com/tianyuan168326/All-in-One-MedReID-Pytorch}{https://github.com/tianyuan168326/All-in-One-MedReID-Pytorch}.

Problem

• Develops unified model for medical image re-identification.
• Proposes ComPA for adaptive multi-modality medical data processing.
• Integrates medical priors to enhance image difference modeling.

Innovation

• Continuous Modality-based Parameter Adapter (ComPA)
• Integration of medical priors with foundation models
• Superior performance across multiple datasets and models