A NOVEL AUTOMATIC METHOD FOR VESSEL TORTUOSITY EVALUATION

Abstract

Tortuosity evaluation of retinal or conjunctival vessels is one of the<br><br>significant steps in early treatment of diabetic retinopathy. Despite<br><br>the importance of this field, a few techniques have been proposed.<br><br>In this paper, we proposed a new automatic algorithm for<br><br>measuring vessel tortuosity based on Non Subsampled Contourlet<br><br>Transform (NSCT). Major vessels and their directional information<br><br>are extracted using NSCT in the first step. Then local vessel<br><br>curvature is computed using obtained NSCT information and entire<br><br>vessel network tortuosity is computed by combination of these<br><br>local curvature values. Accuracy of our algorithm is evaluated by<br><br>spearman correlation of our result and a set of images which are<br><br>ordered by an ophthalmologist in ascending manner of tortuosity.<br><br>We have shown that our algorithm achieves high accuracy in<br><br>evaluation of vessels network tortuosity beside less computational<br><br>time by reducing major steps of traditional tortuosity evaluation<br><br>algorithm.