A new morphology based approach for blood vessel segmentation in retinal images

Diabetic retinopathy (DR) disease occurs due to leakage of blood and protein from small diseased vessels into retina which serves as main cause of blindness among diabetic patients. The early detection of diabetes in the retinal vessels is useful for the prevention of disease. Therefore, the accurate extraction of blood vessels from retinal images helps in diagnosis of such eye diseases. In this paper a new approach is proposed for the blood vessel detection from digital retinal images. The morphological based approach is used for background elimination and blood vessel enhancement with phase preserving noise removal algorithm. Vessel silhouette is then extracted with fixed threshold scheme. Post-processing is done to remove unwanted regions, eliminate spur pixels and fill gaps within detected vessel. The proposed method is evaluated on two publically available datasets, STARE and DRIVE because both datasets provides the ground truth of retinal images precisely marked by experts. Experimental results show that proposed method is comparable to other state of art techniques due to high detection rate.