Diabetic Retinopathy Detection Using Neural Networks and Ensemble Learning

Abstract

A very common complication of diabetes is Diabetic Retinopathy (DR) which affects the blood vessels in the retina and can result in vision loss if left untreated. For efficient treatment and the prevention of vision loss, timely identification of DR is essential. Manual retinal image screening takes time and is exorbitant, therefore, there is rising interest in automating DR screening through the use of machine learning techniques. In recent years, computerized methods for detecting DR from retinal scans have been developed using machine learning algorithms. A well-known Convolutional Neural Network(CNN) architecture is used for detecting diabetic retinopathy are discussed in this paper. The paper discusses feature extraction, classification, and image pre-processing techniques for detecting diabetic retinopathy.  Finally, a comprehensive evaluation of existing approaches is presented. The challenges and opportunities for future research in this area are highlighted as well.