Abstract: Automated feature extraction like cup and disk of retina has great significance in automatic retina analysis and diagnosis. This leads computer aided diagnosis of the eye diseases, especially Glaucoma. The major challenge in Retina image analysis is to obtain the boundary and the area mask of cup and disk. Automated detection of cup and disk area demands sophisticated morphological and image processing applications to make the boundary more prominent and hence detectable with higher accuracy. This paper proposes a unique technique of Disk and Cup detection technique with the help of morphological and gray level processing of the image and improves the performance of disk detection using Hough circle method. This paper analyzes the performance of the techniques under the presence of noise and proposes a suitable adaptive median filtering method for suppressing the noise and detecting the accuracy of the image.

Keywords: Cup to Disk Ratio (CDR), Fundus, Glaucoma, Hough transform and Region of Interest (ROI).