A Comparative Study of Retinal Image Processing Technique for Blood Vessel Segmentation

Blood vessel assessment from retinal images is an important factor for many eye related disease. Retinal blood vessel segmentation is done for identifying ophthalmic disease like glaucoma, diabetic retinopathy, and macular degeneration. Glaucoma is a condition of elevated pressure within the eyeball, which causes gradual loss of sight. Diabetic retinopathy occurs due to damage of the tiny blood vessels that nourish the retina. It results in blood and other fluids leakage that swells retinal tissue that causes clouding of vision. Macular Degeneration happens when central portion of the retina deteriorates. Thus, blood vessel segmentation efficiently reduces the risk of blindness. The manual comparison of fundus images taken at different moments is a very time-consuming job and prone to human error because of the distortions between images that make superposition very difficult, and also there are large numbers of lesions that need to be compared. Hence, computer assisted approach is needed. We have compared different methods of retinal blood vessel segmentation based on parameters like accuracy, sensitivity, specificity and algorithms that have been used for blood vessel segmentation method.