Fully automatic diameter measurement method based on retinal vessel segmentation

Many systemic diseases can cause changes of the diameters of retinal vessels and Arteriolar-to-Venular diameter Ratios (AVR) is an important quantitative biomarker in the diagnosis of these diseases. In this paper, we proposed and validated a novel and automatically retinal vessel diameter measurement method. Given the binary segmentation from fundus of eye, the morphological method is firstly used to extract the centerline of the retinal vessels, which was further used to determine the position of special points in vessels. Then, Hessian matrix is used to determine the initial direction of the vascular skeleton point. Finally, the vessel diameter is measured based on the determined vessel centerline and direction information. The proposed method performed superiorly compared to several state-of-the art methods.