Hybrid Approach for Enhancing Lesion Segmentation in Fundus Images

2025-09-29

AI Summary

Choroidal nevus segmentation in color fundus images faces challenges including low resolution, inter-observer annotation inconsistency, and heavy reliance on large-scale high-quality labeled data for deep learning models. To address these, this paper proposes a novel hybrid segmentation framework integrating mathematical clustering with a U-Net architecture, enhanced by an attention mechanism to optimize multi-scale feature fusion. The framework significantly reduces dependency on extensive annotated datasets while improving generalizability and robustness. Evaluated on 1024×1024 high-resolution fundus images, it achieves a Dice coefficient of 89.7% and an IoU of 80.01%, representing improvements of 38.4% and 45.8% over Attention U-Net, respectively. Moreover, it demonstrates superior cross-domain generalization on external datasets. This work provides an effective technical pathway toward high-accuracy, low-label-dependency AI-assisted ophthalmic diagnostic systems.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

Choroidal nevi are common benign pigmented lesions in the eye, with a small risk of transforming into melanoma. Early detection is critical to improving survival rates, but misdiagnosis or delayed diagnosis can lead to poor outcomes. Despite advancements in AI-based image analysis, diagnosing choroidal nevi in colour fundus images remains challenging, particularly for clinicians without specialized expertise. Existing datasets often suffer from low resolution and inconsistent labelling, limiting the effectiveness of segmentation models. This paper addresses the challenge of achieving precise segmentation of fundus lesions, a critical step toward developing robust diagnostic tools. While deep learning models like U-Net have demonstrated effectiveness, their accuracy heavily depends on the quality and quantity of annotated data. Previous mathematical/clustering segmentation methods, though accurate, required extensive human input, making them impractical for medical applications. This paper proposes a novel approach that combines mathematical/clustering segmentation models with insights from U-Net, leveraging the strengths of both methods. This hybrid model improves accuracy, reduces the need for large-scale training data, and achieves significant performance gains on high-resolution fundus images. The proposed model achieves a Dice coefficient of 89.7% and an IoU of 80.01% on 1024*1024 fundus images, outperforming the Attention U-Net model, which achieved 51.3% and 34.2%, respectively. It also demonstrated better generalizability on external datasets. This work forms a part of a broader effort to develop a decision support system for choroidal nevus diagnosis, with potential applications in automated lesion annotation to enhance the speed and accuracy of diagnosis and monitoring.

Problem

• Achieving precise segmentation of choroidal nevi in fundus images
• Overcoming limitations of low resolution and inconsistent labeling in datasets
• Reducing reliance on large annotated datasets for medical image segmentation

Innovation

• Hybrid model combines clustering with U-Net insights
• Reduces need for large-scale annotated training data
• Achieves high accuracy on high-resolution fundus images