Heterotropia Diagnosis with Smartphone using Machine Learning and Computer Vision

In this paper, a deviation-based algorithm is proposed which exploits the power of TensorFlow, OpenCV, and machine learning to diagnose four forms of heterotropia - esotropia, exotropia, hypertropia and hypotropia. The mechanism developed for lateral, vertical, and angular deviation has been tested for K-Nearest-Neighbors (KNN), Support Vector Machine (SVM), and a trained neural network, based on 6 features. Finally, the accuracy, specificity, and sensitivity of the 2 MLC and convolutional neural network (CNN) was tested in a study of 337 users with and without heterotropia. Eyeris, an Android based application, has been developed as a proof of concept for this paper. The results yielded highest accuracy for SVM-based classification.