Received: April 12, 2018;   Published: April 24, 2018

DOI: 10.26717/BJSTR.2018.04.000995

Full Text PDF

To view the Full Article   Peer-reviewed Article PDF

Abstract

Objective:To develop an automated system to classify extra ocular diseases

Study Design: Retrospective cohort

Method: The entire dataset consists of about 7,244 labelled images of patients from Drashti Netralaya Eye Hospital in Gujarat, India. Five diseases were selected for classification: Corneal scars, Dermoid Cyst, Strabismus, Ptosis, and Ocular Surface Disease. Histogram of Oriented Gradient feature descriptors were utilized with Support Vector Machines and Logistic Regression. Modern Neural Network architectures were also applied. Bottleneck CNN and Logistic Regression (Balanced) both performed well according to different error measurements. This work outlines the development of a classifier for extra ocular conditions that uses natural, noisy images of faces taken with point-and-shoot cameras.

Result: The Bottleneck CNN achieved the highest test accuracy of 77% but Logistic Regression had the highest true positive rate averaged across all classes. We have found accuracy of 92% for strabismus.

Conclusion: Thus, it is possible to develop a classifier from images for a variety of eye diseases.

Abstract| Introductiont| Methods| Results| Discussion| Conclusion| References|