Segmentation of Optic Disc by Localized Active Contour Model in Retinal Fundus Image

Abstract

Analysis based on retinal fundus image has been progressed on computer-aided diagnostic system, as it provides the primary screening for early identification of retinal based diseases such as glaucoma, etc. This paper suggests a novel method for localization and segmentation of Optic Disc (OD) based on image processing operations, Circular Hough Transform (CHT) and localized active contour model (ACM). The morphological closing and opening operations are applied consecutively for removal of noise, blood vessels and small gaps presented in grey-scaled image. The CHT was employed for the detection of OD region and it is segmented out by using localized region-based ACM. All these processes were conducted on RIM-ONE database and the average accuracy and overlap score were estimated, respectively. Thus, this introduced method has been remarkable for effective detection and segmentation of OD which can be outrated by other methods.