Automated Glaucoma Detection from Fundus Eye Images Using Grey Level based Feature Extraction Methods and Supervised Learning Classification

Abstract

            Over 60 years, health organization’s survey exposed that glaucoma leads to permanent vision disorder in human eye. Although 79 million of people were affected by this second major cause in 2020, glaucoma detection technique exploits preventing scheme at the earlier stage and reducing the necessary disease treatment cost. This research is trained probe a best analysis for glaucoma prediction in segmented retinal fundus images to assist an ophthalmologist. Our model utilizes a many feature extraction approach to detect glaucoma from digital fundus image using a fusion feature set. Along-with extraction way, Gray Level Coherence Matrix, Gray Level Run Length Matrix, Gray Level Difference Method approaches are used to extract texture based features. The contributions of extracted spatial domain from these approaches are expected to formulate higher efficiency of fusion feature set using classification. Thus, proposed work is presented with the comparison of four machine learning algorithms like support vector machine (SVM), k-nearest neighbors (kNN), Decision Tree (DT) and Naive Bayes (NB). Diagnosis methodology is clearly demonstrated a way to detect glaucoma in early stage from fundus image. Performance of the classifier is analyzed by computing the accuracy value. Then effectiveness of the system is improved by the combination of texture with supervised learning techniques.