Analysis of Segmentation Algorithms in Colour Fundus and OCT Images for Glaucoma Detection

Glaucoma is the largesteye disease which affects the optic nerve head and results in visual impairment. In this paper, we analyze the various segmentation algorithms for glaucoma detection using color fundus images and spectral domain Optical Coherence Tomography (OCT) images of same subjects. In fundus images, the disc and the cup regions are segmented separately with four different segmentation algorithms namely Otsu method, Region growing, Hill climbing and Fuzzy C-means clustering algorithms. In OCT images, the cup and the disc diameter were measured by segmenting the retinal nerve fibre and retinal pigment epithelium layers. From both the analysis, the Cup to Disc Ratio (CDR) is calculated and compared with the clinical values. The experimental results show that the performance error in the OCT image analysis is less when compared to the fundus image analysis. Thus, it can be concluded that glaucoma detection can be done more effectively using OCT image analysis.