Aid to the medical diagnosis by retinal analysis

Abstract

To diagnose serious eyes diseases, ophthalmologists use the color retinal images of a patient acquired from the digital fundus camera, such as the diabetic retinopathy that affects the morphology of the blood vessels tree, so an automatic detection and extraction of blood vessels in retinal images is important. In this paper, we present a method to extract blood vessels and delineate vascular intersection and crossovers. The proposed algorithm consists of three steps: first, the appearance of the blood vessels is enhanced and background noise is suppressed using Gaussian filters, in second step, entropic thresholding is used to detecting pixels in the enhanced image. Finally, the vascular intersections and bifurcations are obtained via the crossing number (CN) technique.