Blood vessel detection based on fractional Hessian matrix with non-singular Mittag–Leffler Gaussian kernel

Abstract

In this work, we proposed a novel method based on fractional Gaussian kernel involving the Atangana–Baleanu fractional derivative to detect blood vessels in retinal images. This method is based on the fractional Gaussian matrix eigenvalues analysis to obtain the principal curvature. The non-integer orders of the fractional Gaussian mask for retinal image denoising and the fractional Gaussian matrix were optimized by Cuckoo Search algorithm to get more accurate results. To evaluate the effectiveness and the robustness of the method it was compared with manual segmentation and other works reported in the literature.