Artery and Vein Classification in Retinal Images by an Automated Graph-based Approach

Abstract

The characterization of retinal vessels into Artery/vein (A/V) is an essential stage for computerizing the recognition of vascular changes, and for the figuring of trademark signs connected with a few systemic infections, for example, diabetes, hypertension, and other cardiovascular conditions. This Project explains a programmed methodology for A/V order taking into account the investigation of a chart removed from the retinal vasculature. The proposed technique groups the whole vascular tree choosing the kind of every crossing point and relegating one of two names to every vessel portion. Last characterization of a vessel portion as A/V is performed through the mix of the chart based marking results with an arrangement of force elements.