A level set method for retina image vessel segmentation based on the local cluster value via bias correction

Abstract

Segmentation of the retina vessels plays an important role in detecting earlier diseases, i.e. diabetic and hypertension. These diseases may cause the shape change of the vasculature. Due to the inhomogeneous of the retina image, it's very difficult to use region static information to get the result. We use a new level set method based on local region cluster information which proves to be well performed the inhomogeneous of the retina image. Experiment shows that our method is more robust to the initial contour C. In the experiment, our method can get a desirable segmentation result in the segmentation of retina images.