A Mathematical Model for the Analysis of Eye Fundus Images in Healthy and Diabetic Patients.

In this paper, we provide a study on eye fundus images of healthy and diabetic patients. Taking benefits from its reconstruction and enhancing properties, the sampling Kantorovich algorithm is used to process the considered images, after registration and averaging processes. Moreover, a hybrid segmentation procedure applied on superficial capillary plexus images (SCP) and one using the local Phansalkar method on choriocapillary images (CC) are exploited in order to asses a cluster counting process which is based on finding connected regions according to the 8–adjacency criterion. The results achieved on the healthy and diabetic patients show that the novel strategy allows to obtain accurate data from both a mathematical and a clinical point of view.