Automated Glaucoma Detection using CDR Method on Fundus Images using Fuzzy Logic

Abstract

Glaucoma is one of the severe visual diseases and the leading cause of irreversible blindness by affecting the optic nerve fibre and astrocytes in human eye. Early diagnosis and cure can play vital role in the lives of patients, and it is possible only if Glaucoma suspected patients are checked periodically by Ophthalmologist. Various clinical examination techniques are available that are time consuming, require multiple resources and ophthalmologist involvement. Therefore, an automated system is preferred that support Ophthalmologist by its minimum involvement and accurate diagnosis. Digital retinal based fundus imaging is one of the sources available in Ophthalmology that is used in detecting most of the eye related diseases including Glaucoma with the help of various image processing techniques. Fuzzy logic is a technique that can manage uncertainty and imperfection of an image in a better way by representing it as a fuzzy set, thus producing accurate results. In this research, a fully automated methodology for glaucoma detection from retinal fundus images using fuzzy logic is proposed. Cup-to-Disc Ratio (CDR), being the most accepted physiological parameter for glaucoma detection, is calculated by segmenting Optic Disc (OD) and Optic Cup (OC) using fuzzy logic, augmented with few morphological operations. Novel vector based approach is used for diameter calculation of OD and OC.