The proposed work is an automated system for the detection of Glaucoma which is done by using the algorithms CDR and ISNT ratio of a retinal fundus image. In general for the Glaucoma patient the size of the optic cup increases but the size of the optic disc will be the same and so the cup to disc ratio value will increased for the patients with Glaucoma than the normal patients. The ROI of the green plane is done and that image is subjected to K Means clustering methodology repeatedly and the

more »

... tic disc is segmented using it. Area of the optic disc and optic cup is determined through elliptical fitting and hence CDR is calculated. For the identification of Glaucoma another parameter known as ISNT is used through the area of the blood vessels in the Inferior Superior to Nasal temporal side. The ratio will be lesser for the Glaucoma patient than the normal patient if the vessels shift to the Nasal side. To extract the blood vessels two techniques, Matched filtering and local entropy thresholding are applied. The programming language used to implement this method is C++ using OpenCV (Open Source Computer Vision Library) library functions. The library of micro and macro programming functions is known as OpenCV library functions which are developed by Intel. Core, highgui, image proc, ml are some of the main library functions used from OpenCV. These optimized functions in OpenCV improve the speed of the operation and hence they are very useful in real time mass screening purpose. A set of fifty retinal fundus images obtained (25 normal and 25 abnormal retinal images) from Aravind eye hospital has been used to determine the efficiency of the featured system.