Abstract
Diabetic retinopathy is that the single largest explanation for sight loss and visual impairment in eighteen to sixty five year olds. Screening programs for the calculable 1 to 6 % of the diabetic population are incontestable to be value and sight saving, but unfortunately there are inadequate screening resources. An automatic screening system might facilitate to solve this resource short fall.The retinal vasculature consists of the arteries and veins with their tributaries that are visible at intervals in the retinal images.This paper proposes a graphbased artery vein classification system inretinal images for diabetic retinopathybased on the structural informationextracted from the retinalvasculature. The method at first extracts agraph from the vascular tree and then makes a decision on the typeof each intersection point (graph node).Based on this node types one of the twolabels are assigned to each vessel segment.Finally, the A/V classes are assigned tothe sub graph labels by extracting a set ofintensity features and using artificialneural network
Users
Please
log in to take part in the discussion (add own reviews or comments).