A Fast Approach to Retinal Vessel Segmentation

Automated segmentation of blood vessels in retinal images will help eye care specialists screen larger populations for vessel abnormalities. However, automated retinal segmentation is complicated by the fact that a number of vessels are very thin and the local contrast is low. We propose the radial projection method to locate the vessel centerlines which contains the thin vessels. Then the aggregate gradient is used for extraction the major structure of vessels. The final segmentation is obtained by the union of the two types of vessels. Our approach is tested on the STARE database, the results demonstrate that our algorithm is capable of detecting low-contrast and narrow vessels while computational cost is significantly reduced.