Glaucoma Detection and its classification using Image Processing and Fuzzy Classification

Manual diagnosis needs a great deal of time for ophthalmologists to analyse and review retinal images of the eye obtained by fundus camera. Digital image processing techniques enable ophthalmologists to detect and treat several eye diseases like diabetic retinopathy and glaucoma. Glaucoma, the most common cause of blindness is the disease of the optic nerve of the eye and can lead to ultimate blindness if not treated at an early stage. Raised intraocular pressure, increase in cup to disk ratio and visual field test are some of the measures for such a disease. The main objective of this paper is to find an automated tool to detect glaucoma at an early stage and to classify this disease based on its severity and damage of the optic fibre. The objective of this study is pre-processing of retinal fundus image for enhancing the quality which is required for further processing and to design a novel algorithm to measure the cup to disc ratio of retinal fundus image from the online database and classify the disease according to its severity using fuzzy classification toolbox in MATLAB. We use the method of principle component analysis wherein we extract the dominant Eigen values and Eigen vectors to classify the images into various stages of glaucoma with high accuracy