Segmentation of optic disc on retinal fundus images using morphological reconstruction enhancement and active contour

Abstract

Glaucoma is one of the eye diseases characterised by increasing of fluid pressure in the eye. High fluid pressure can damage optic nerve head. It can lead to blindness if it is not treated properly. Segmentation of optic disc area is a technique to support the glaucoma diagnosis. This research contributes to the segmentation of optic disc. Red channel is extracted from colour images. Furthermore, noise is eliminated by average filter. Subsequently, the transformation of bottom hat is used to remove the vessel area in fundus image. The image quality is enhanced by morphological reconstruction operation. Before performing segmentation, Region of Interest (RoI) is determined based on the highest value of pixel intensity. Segmentation of optic disc is undertaken using active contour. This method has been tested on 50 colour retinal fundus images taken from DRISHTI-GS dataset. The propose scheme successfully achieves average accuracy of 99.55%, sensitivity of 89.92% and specificity of 99.85% in segmenting optic disc.