Published June 24, 2026 | Version v1

Eye Disease Detection Using Multiclass Classification of Retinal Images

  • 1. Sinhgad Institute Of Technology And Science, Narhe

Description

The increasing global prevalence of myopia and its progression to Pathologic Myopia (PM) pose a significant risk of irreversible vision loss, making early and accurate detection essential. This paper presents a deep learning-based framework for multi-class classification of retinal fundus images into Normal, Myopia, and Pathologic Myopia categories. The proposed approach employs EfficientNet-B3 with a custom classification head and a two-phase transfer learning strategy that combines AdamW optimization, weighted loss, label smoothing, MixUp augmentation, and cosine annealing learning rate scheduling. To improve interpretability, Gradient-weighted Class Activation Mapping (Grad-CAM) is integrated to visualize the regions influencing model predictions. The model was trained and evaluated on a large-scale retinal image dataset drawn from multiple sources, including the PALM dataset, using a custom 380 × 380 input resolution. Experimental results show that EfficientNet-B3 achieved an overall test accuracy of 96.08%, outperforming EfficientNet-B0 (89.3%) and ResNet50 (82%). The model also demonstrated strong class-wise performance, including perfect detection of Pathologic Myopia with 100% precision, recall, and F1-score, while significantly improving discrimination between Normal and Myopia cases. The framework is further deployed as a Django-based web application supporting real-time prediction and automated report generation. These results indicate that the proposed system is a clinically relevant decision-support tool for automated retinal screening.

Files

eye-disease-detection-using-multiclass-classification-of-retinal-images-IJERTV15IS060901.pdf

Additional details