Cataract Disease Classification using Convolutional Neural Network Architectures

The most common form of eye illness is a cataract. If a cataract is not detected and treated early, it can cause blindness. Early detection is the most beneficial way to reduce the risk and avoid painful surgery. This study primarily focuses on cataract disease classification from fundus images. To detect and classify the cataract or non-cataract disease from fundus images, an automatic cataract disease classification method is proposed. A different CNN architecture implemented the automatic cataract classification. The main purpose of the CNN model was to feature extraction and classifies cataract disease. The fundus images were gathered from different publicly accessible datasets and labeled as 0 or 1. Image pre-processing, extracting features and identification are the three basic steps in detecting and classifying cataract disease. The accuracy of the cataract disease classification obtained is 97.66%. According to the findings, the presented method is an effective technique for cataract classification.