The significance of the vessel registration for a reliable computation of arteriovenous ratio

Abstract

The arteriovenous ratio (AVR), this is, the relation between artery and vein widths, is a popular dimensionless measure to quantify changes in retinal microvasculature. However, its use in daily clinical practice has not been implanted due to the lack of reproducibility caused mainly by the laborious manual calculation and the dependence on the vessels selected for the estimation. This paper presents a vessel registration methodology in an AVR computation framework. The expert computes the AVR from a reference image in a semiautomatic manner and, after that, the AVR can be computed automatically from successive images of the same patient using the stored information from the reference image. The system has been evaluated in a large data set of 158 pairs of images and good correlation results between medical experts and system have been achieved. © 2012 Springer-Verlag.