10-06-2017, 10:20 AM
Segmentation of the retinal blood vessel image is a process that can help predict and diagnose cardiovascular disease, such as hypertension and diabetes, which are known to affect the appearance of blood vessels in the retina. This work proposes an unsupervised method for the segmentation of retinal vessel images using a combined combined filter, the Frangi filter and the Gabor Wavelet filter to improve the images. The combination of these three filters in order to improve segmentation is the main motivation of this work. We investigated two approaches to perform the combination of filters: weighted mean and median of the classification. Segmentation methods are tested after vessel improvement. Medium-graded enhanced images are segmented using a single-threshold criterion. Two segmentation procedures are applied when considering retinal images enhanced using the weighted mean method. The first method is based on deformable models and the second uses diffuse C media for image segmentation. The procedure is evaluated using two public image databases, Drive and Stare. Experimental results demonstrate that the proposed methods work well for vessel segmentation compared to prior art methods.