Morphology approach for features extraction in retinal images for diabetic retionopathy diagnosis

In this paper, a novel approach for vessels extraction edge-based image segmentation is proposed. Vessels segmentation and extraction play an important role in supporting computer assistance for diagnosis of Diabetic Retinopathy (DR). Diabetic Retinopathy is a severe and widely spread eye disease. The algorithms to detect and extract vessels from retinal images are mainly based on morphological filtering and segmentation methods. We proposed an image segmentation algorithm by integrating mathematical morphological edge detector with TopHat technique. In this paper, theoretical backgrounds and procedure illustrations of the proposed algorithm are presented. Furthermore, the proposed algorithm has been evaluated on several images of publicly available database. The results are compared with those obtained by other known methods as well as with the golden images. It is shown that the proposed method can yield a specificity value as high as 82%, which is comparable to the results obtained by other known methods.