Hypertensive Retinopathy Diagnosis from Fundus Images by Estimation of Avr.

AbstractHypertensive retinopathy is a disease that damages the retina of the eye and results in loss of vision and is closely associated with high blood pressure. Severe case of hypertensive retinopathy causes systematic aliments that may cause cardiovascular diseases, heart and renal failure, loss of vision and finally death. Thus the timely diagnosis and treatment of the disease is vital. Arteriovenous ratio is used to diagnose hypertensive retinopathy. In this paper we proposed an algorithm in which, the blood vessels are segmented out initially, from the pre processed retinal images. Gray level and moment based features are extracted to classify the detected pixels as belonging to the blood vessel class or not. Intensity variation and colour information is used to classify the vessel as arteries or veins. Vessel width estimation method is used to measure the arteriovenous ratio from which various stages of hypertensive retinopathy can be identified. Retinal images were obtained from the VICAVR database, along with images collected form Deepam eye hospital Chennai. From the images that were collected 25 were normal images and 76 images of hypertensive retinopath