A Comparative Investigation of Blood Vessels in Retinal Images

Angiography is a way towards picturing the retina of eye. Unfortunately, a specialist has to spend a lot of time to examine the picture. Segmentation is one of the processes to identify eye problems and decrease the specialist time. This paper proposes the performance comparison of vessel segmentation methodologies. Firstly, the fundus image is taken from the openly accessible datahise and pre-processing is done to adjust the contrast of input image so that the vessels are segmented Segmentation is the second step of algorithm that segments the image and post processing is done to enhance the segmented binary image. The methods are involved with ISODATA, adaptive thresholding and vessel location map. The results declare that the method dominates over the prior proposed strategies like segmentation of vessels and recovering edges. Additionally, the evaluated algorithms achieving 94% accuracy of extracting normal fundus vessels.