EYE-DEX: Eye Disease Detection and EXplanation System

2025-09-28

AI Summary

Early diagnosis of retinal diseases is critical for preventing vision loss and mitigating socioeconomic burden; however, conventional manual fundus image interpretation suffers from time consumption and subjectivity. This study proposes an interpretable automated classification framework for ten retinal disease categories. Leveraging pre-trained CNN architectures—including VGG16, VGG19, and ResNet50—we employ transfer learning with fine-tuning and, for the first time in this domain, systematically integrate Grad-CAM to generate visual explanations highlighting pathological regions. Evaluated on a large-scale dataset of 21,577 fundus images, the fine-tuned VGG16 model achieves a test accuracy of 92.36%, establishing a new state-of-the-art (SOTA) for multi-class retinal disease classification. Crucially, attention heatmaps significantly enhance model transparency and clinical interpretability, advancing AI-assisted diagnosis toward the synergistic realization of high performance and high explainability.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Transparent, Interpretable, Explainable ML
• Philosophy and Ethics of AI: Accountability, Interpretability & Explainability

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Retinal disease diagnosis is critical in preventing vision loss and reducing socioeconomic burdens. Globally, over 2.2 billion people are affected by some form of vision impairment, resulting in annual productivity losses estimated at $411 billion. Traditional manual grading of retinal fundus images by ophthalmologists is time-consuming and subjective. In contrast, deep learning has revolutionized medical diagnostics by automating retinal image analysis and achieving expert-level performance. In this study, we present EYE-DEX, an automated framework for classifying 10 retinal conditions using the large-scale Retinal Disease Dataset comprising 21,577 eye fundus images. We benchmark three pre-trained Convolutional Neural Network (CNN) models--VGG16, VGG19, and ResNet50--with our finetuned VGG16 achieving a state-of-the-art global benchmark test accuracy of 92.36%. To enhance transparency and explainability, we integrate the Gradient-weighted Class Activation Mapping (Grad-CAM) technique to generate visual explanations highlighting disease-specific regions, thereby fostering clinician trust and reliability in AI-assisted diagnostics.

Problem

• Automating retinal disease diagnosis from fundus images using deep learning
• Addressing time-consuming and subjective manual grading by ophthalmologists
• Providing visual explanations for AI decisions to enhance clinical trust

Innovation

• Fine-tuned VGG16 model for retinal disease classification
• Integrated Grad-CAM technique for visual explanations
• Automated framework using pre-trained CNN models