AUTOMATED BLOOD VESSEL SEGMENTATION IN RETINAL IMAGE

Abstract

The First step in development of automated system for ophthalmic diagnosis is Automatic segmentation of blood vessels from retinal images.  Automated blood vessel segmentation in retinal image is used to provide the information about diseases like diabetes, stroke, hypertension and Arteriosclerosis. The abnormal image, cottonwoolspots and hard exudates makes the vessel extraction process as difficult task. A Support Vector Machine (SVM) based supervised method is proposed for extracting blood vessel in retinal fundus image. In this work, the green channel image is enhanced by Gabor filter, and then various features are extracted by Thresholding method for SVM classifier. The method will be implemented on publicly available digital retinal images for vessel extraction (DRIVE) database.