Optic Disc Identification Methods for Retinal Images

Abstract

Presented are the methods proposed by authors to identify and model the optic disc in colour retinal images. The first three our approaches localized the optic disc in two steps: a) in the green component of RGB image the optic disc area is detected based on texture indicators and pixel intensity variance analysis; b) on the segmented area the optic disc edges are extracted and the resulted boundary is approximated by a Hough transform. The last implemented method identifies the optic disc area by analysis of blood vessels network extracted in the green channel of the original image. In the segmented area the optic disc edges are obtained by an iterative Canny algorithm and are approximated by a circle Hough transform.

Keywords

• optic disc
• retinal images
• vessel segmentation
• Hough transform