COMPUTER AIDED DIAGNOSIS OF GLAUCOMA DETECTION USING DIGITAL FUNDUS IMAGE

A robust and cost-effective mass screening may help to detect glaucoma at the earliest which is a major cause of blindness. In this paper, a Computer Aided Diagnosis (CAD) approach for glaucoma detection using retinal fundus images based on clustering techniques is presented. The abnormalities in retinal fundus image are diagnosed using the physiological characteristics of Optic Cup (OC) and Optic Disc (OD). Based on the size of OC and OD, Cup to Disc Ratio (CDR) is computed for the diagnosis. Due to glaucoma, the size of OC increases which increases the CDR as well. In this study, the OD segmentation is achieved by K- Means clustering (KMC) and Hill Climbing Algorithm (HCA) for the selection of K value. Similarly, OC is extracted by exploiting fuzzy C-mean clustering. After segmentation of OC and OD, CDR is computed to diagnose glaucoma. The system is applied to a total of 45 images, and the results indicate the ability of the system for automated mass screening to diagnose glaucoma at the earliest.