International Research Journal of Education and Technology

Cataracts are one of the most prevalent eye conditions that lead to distorted vision, and early detection plays an essential role in controlling the risk and preventing blindness. However, the current techniques used for cataract detection are often labour-intensive and time-consuming. To address this issue, deep learning algorithms such as YOLO (You Only Look Once) have been proposed for the early prediction of cataracts. Deep learning is an advanced artificial intelligence approach that mimics the human brain's ability to organize data and create patterns for decision-making. The primary goal of the proposed plan is to detect cataracts in images by utilizing an effective deep learning model. For training the model, a dataset containing 100 cataract images, sourced from Kaggle and the UCI Machine Learning Repository, was used. The system enables users to upload a test image, which undergoes pre-processing steps such as image extraction and conversion from RGB to grayscale. Following this, image segmentation is applied, and for each segmented region, Convolutional Neural Networks (CNN) and YOLO neural networks are employed to predict and classify the image as either "Normal" or "Cataract." The training dataset consists of both cataract-affected images and healthy images, making the model capable of distinguishing between the two. This work can assist in faster and more accurate diagnoses, contributing to the on-going effort to combat cataract-related vision impairment.

Keywords—Cataract Detection, Early Diagnosis, YOLO Algorithm, Image Classification