Active Diffusion Matching: Score-based Iterative Alignment of Cross-Modal Retinal Images

2026-04-11

AI Summary

This work addresses the challenging cross-modal alignment between standard fundus images (SFIs) and ultra-widefield fundus images (UWFIs), which is hindered by significant differences in field of view and geometric distortions. To tackle this problem, the authors propose Active Diffusion Matching (ADM), the first method to introduce score-based diffusion models into SFI–UWFI alignment. ADM employs two coupled diffusion processes integrated with iterative Langevin dynamics to jointly optimize a global rigid transformation and local non-rigid deformations, enabling progressive stochastic optimal alignment. A tailored sampling strategy is further designed to enhance model adaptability. Evaluated on both a private SFI–UWFI dataset and a public SFI–SFI benchmark, ADM achieves state-of-the-art performance, improving mean AUC by 5.2 and 0.4 percentage points, respectively, thereby filling a critical methodological gap in this domain.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Objective: The study aims to address the challenge of aligning Standard Fundus Images (SFIs) and Ultra-Widefield Fundus Images (UWFIs), which is difficult due to their substantial differences in viewing range and the amorphous appearance of the retina. Currently, no specialized method exists for this task, and existing image alignment techniques lack accuracy. Methods: We propose Active Diffusion Matching (ADM), a novel cross-modal alignment method. ADM integrates two interdependent score-based diffusion models to jointly estimate global transformations and local deformations via an iterative Langevin Markov chain. This approach facilitates a stochastic, progressive search for optimal alignment. Additionally, custom sampling strategies are introduced to enhance the adaptability of ADM to given input image pairs. Results: Comparative experimental evaluations demonstrate that ADM achieves state-of-the-art alignment accuracy. This was validated on two datasets: a private dataset of SFI-UWFI pairs and a public dataset of SFI-SFI pairs, with mAUC improvements of 5.2 and 0.4 points on the private and public datasets, respectively, compared to existing state-of-the-art methods. Conclusion: ADM effectively bridges the gap in aligning SFIs and UWFIs, providing an innovative solution to a previously unaddressed challenge. The method's ability to jointly optimize global and local alignment makes it highly effective for cross-modal image alignment tasks. Significance: ADM has the potential to transform the integrated analysis of SFIs and UWFIs, enabling better clinical utility and supporting learning-based image enhancements. This advancement could significantly improve diagnostic accuracy and patient outcomes in ophthalmology.

Problem

• cross-modal alignment
• retinal image registration
• fundus image
• ultra-widefield imaging
• image matching

Innovation

• Active Diffusion Matching
• score-based diffusion models
• cross-modal alignment
• iterative Langevin dynamics
• retinal image registration