Model-based estimation of coronary vessel diameter in angiographic images

An optimization method to detect edges of coronary vessels in angiographic images is described. The method is based on an elliptical model of a vessel's cross-section. The method consists of two stages: (1) rapid and crude estimation of centerline and radius, and (2) parameter estimation for elliptical vessel model based on densitometrical estimation of the vessel cross-sectional curvature. The optimization method is partly analytical and partly numerical resulting in short computational time. The algorithm was tested extensively using real and synthesized images. This paper includes an example of a real image analysis.