Cataracts Detection Using Convolution Neural Network

Cataracts are a major cause of vision loss, affecting millions worldwide. This study presents a novel approach for early cataract detection using a Deep Convolutional Neural Network (DCNN) implemented in MATLAB. The system utilizes fundus images, preprocessed with techniques like grayscale conversion and resizing, to extract critical texture features such as contrast, correlation, and entropy. The DCNN model, with a single hidden layer of sixteen neurons and two output neurons, was trained on the Kaggle dataset, achieving a detection accuracy of 92.78% and an execution time of 5.33 seconds. The study compares the DCNN's performance to other classification techniques, highlighting its superior accuracy. This expert system, while not intended to replace ophthalmologists, aims to assist in early diagnosis, potentially reducing the cost of healthcare and improving patient outcomes by leveraging the power of deep learning in medical imaging analysis.