Classification of Optical Coherence Tomography Images with Customized Features Using Machine Learning Models

Macular degeneration due to old age is serious eye disease that causes heavy loss in vision unless it is diagnosed and treated on time. This technique is employed to diagnose inner retinal disorders such as macular degeneration. The paper focuses on the classification of OCT images. Machine learning techniques have been utilized in this attempt to make an automated distinction between Age-related Muscular Degeneration (AMD) affected and normal eyes. A total of 4200 OCT images were used, with 2100 images from normal eyes and 2100 from AMD-affected eyes. The classification was done using two simultaneous models that is K-Nearest Neighbors and Random Forest model. The process of classification using OCT includes a data set preparation, preprocessing, feature extraction, correlation matrix analysis, and application of two models. In this context, our primary objective was to check the suitability of the model in classifying images. This research implies that machine learning algorithms might detect AMD early, improving patient outcomes. The ability of both models to distinguish between AMD-affected and normal eyes are demonstrated and their strengths and limitations are precisely analyzed.