Author : Rutuja Shinde and Girish Potdar

Optic cup and disc segmentation is critical for automated cup-to-disk ratio measurement and hence computeraided diagnosis of glaucoma. Earlier, segmentation using the concept of r-bends resulted Various methods used for Optic Cup and Disc segmentation are CNN, mean curvature and Gabor texture energy features, intensity based cup segmentation, deformable models, template matching and morphological procedures, statistical measures and Superpixel Classification. As a result, a new methodology using Colour based fundus image segmentation and segmentation approach using Adaptive threshold based segmentation is proposed. The segmentation of Optic Disc has been done using red channel whereas that of Optic Cup has been done using green channel of an RGB image. The threshold is ascertained from the smoothed histogram of the pre-processed image. In colour based segmentation approach, a mask based on intensity values of input image is created and image is segmented accordingly. These approaches would help overcome the challenges faced using previous methodologies. The dataset comprises of a set of 50 retinal fundus images alongwith its ground truth provided by the doctors.

Keywords: Glaucoma;     Cup-to-Disc          Ratio     (CDR);    Colour Spaces Segmentation; Adaptive Threshold based Segmentation