Machine learning classification of normal vs. age-related macular degeneration OCT images

Abstract

The macula is a small area of the retina which is especially very important for good eyesight. The age- related macular degeneration (AMD) is a type of visual impairment that can cause blindness or even loss of eyesight. AMD was a dangerous and progressing chronic disease that affects people over the age of 60 years. One of the most common symptoms of this condition is the appearance of a type of extracellular material called the drusen. Detecting this condition using an imaging technique known as optical coherence tomography (OCT) can help to prevent further damage to the eyes. The motivation behind this work is to test machine learning (ML) with OCT images for the identification of retinal disease. OCT retinal images consist of normal retina and AMD. ML algorithms were used to classify 261 OCT images to determine if the person was normal or macular. The proposed method for diagnosing between diseased and healthy conditions has a classification accuracy that considerably exceeds beyond the current state of the art. This work will be used in developing further ML concepts in the diagnosis of ocular disorders and diseases.