Human Eye Disease Detection System Using Deep Learning

This study proposes an intelligent diagnostic system for human eye disease detection using a convolutional neural network (CNN)-based deep learning framework.With the rise in cases of diabetic retinopathy, glaucoma, cataracts, and other ocular conditions, early diagnosis through automated analysis has become increasingly critical.Traditional manual screening is time-consuming and subject to human error, especially in resource-limited environments.This work introduces a robust classification model capable of analysing high-resolution retinal fundus images to predict common eye disorders with a high degree of accuracy.The system is built using the TensorFlow and Keras deep learning frameworks, employing a custom-trained CNN derived from MobileNetV3 architecture, optimized for medical image classification.The dataset used includes thousands of labelled retinal images, ensuring a diverse representation of ocular conditions.Data preprocessing steps such as resizing, normalization, augmentation, and contrast enhancement were applied to improve learning outcomes and reduce overfitting.During training, multiple evaluation metrics-including accuracy, precision, recall, and F1-score-were tracked to monitor performance.The model achieved a validation accuracy exceeding 92%, demonstrating strong generalization on unseen data.Additionally, confusion matrix analysis highlighted areas where the model could distinguish between classes with high confidence.