An Automated Cost-Effective Approach to Glaucoma Diagnosis in Human Using Artificial Neural Network

Abstract

Glaucoma is a term applied to eye condition that progressively brings about loss of vision
by harming the optic nerve that sends visual pictures to the mind. This paper provides an
automatic human eye diagnostic system utilized to detect glaucoma eye disease based on
Artificial Neural Network, it covers acquisition of fundus images, pre-processing the fundus
image, extracting useful features from the image, neural network training and classification
of eye status. In this work, the back propagation neural network was used to develop the
neuron model and the program developed using MATLAB. The developed system was
deployed to act as a standalone application using MATLAB guide. The acquired fundus
images were analyzed in MATLAB after they have been pre-processed and useful features
such as contrast, correlation, homogeneity and energy have been extracted from the preprocessed
fundus image. Here, 15% of acquired fundus images were utilized in the testing
of the model and an average of 97% accuracy was achieved.