Strabismus Classification using Convolutional Neural Networks

Abstract

Early diagnosis and treatment of amblyopia, including strabismus, is important to prevent permanent blindness in infants and toddlers. In this paper, we design an convolutional neural networks that classifies exotropia, esotropia, and normal eyes. Designed model uses the front view of 9-photo as input data, and evaluates its performance using prediction accuracy according to the number of training epochs.