WDT-MD: Wavelet Diffusion Transformers for Microaneurysm Detection in Fundus Images

2025-11-12

AI Summary

Existing diffusion models for microaneurysm (MA) detection in fundus images suffer from three critical limitations: identity mapping, high false-positive rates, and inadequate reconstruction of normal anatomical structures. To address these issues, this paper proposes a novel diffusion-driven anomaly detection framework. First, a noise-encoded image-conditioning mechanism is introduced to suppress identity mapping. Second, image inpainting is employed to synthesize pseudo-normal samples, enabling pixel-level supervision for anomaly localization. Third, a hybrid architecture integrating diffusion-based Transformers with multi-scale wavelet analysis is designed to jointly enhance global contextual modeling and fine-grained local structure reconstruction. Evaluated on the IDRiD and e-ophtha MA benchmarks, the proposed method achieves state-of-the-art performance in both pixel-level MA segmentation and image-level classification, significantly improving detection accuracy and robustness. The framework demonstrates strong potential for automated, early-stage screening of diabetic retinopathy.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Microaneurysms (MAs), the earliest pathognomonic signs of Diabetic Retinopathy (DR), present as sub-60 $mu m$ lesions in fundus images with highly variable photometric and morphological characteristics, rendering manual screening not only labor-intensive but inherently error-prone. While diffusion-based anomaly detection has emerged as a promising approach for automated MA screening, its clinical application is hindered by three fundamental limitations. First, these models often fall prey to"identity mapping", where they inadvertently replicate the input image. Second, they struggle to distinguish MAs from other anomalies, leading to high false positives. Third, their suboptimal reconstruction of normal features hampers overall performance. To address these challenges, we propose a Wavelet Diffusion Transformer framework for MA Detection (WDT-MD), which features three key innovations: a noise-encoded image conditioning mechanism to avoid"identity mapping"by perturbing image conditions during training; pseudo-normal pattern synthesis via inpainting to introduce pixel-level supervision, enabling discrimination between MAs and other anomalies; and a wavelet diffusion Transformer architecture that combines the global modeling capability of diffusion Transformers with multi-scale wavelet analysis to enhance reconstruction of normal retinal features. Comprehensive experiments on the IDRiD and e-ophtha MA datasets demonstrate that WDT-MD outperforms state-of-the-art methods in both pixel-level and image-level MA detection. This advancement holds significant promise for improving early DR screening.

Problem

• Addressing identity mapping issues in diffusion models for medical imaging
• Reducing false positives by distinguishing microaneurysms from other anomalies
• Improving reconstruction of normal retinal features in fundus images

Innovation

• Noise-encoded conditioning prevents identity mapping
• Pseudo-normal synthesis discriminates anomalies via inpainting
• Wavelet diffusion Transformer enhances multi-scale reconstruction