Detection of vascular diameters by ultrasonic imaging
Roland Michaely
Micro-engineering
Student Project
Semester Project, February 2002
Introduction
We treat an image sequence of an artery, so to detect the variation of the diameter in time.
Initialization
We initialize the boundaries of the artery on the first image using an automatic scheme based on local histogram. The detected points are interpolated by cubic B-spline..
Snake
We detect the boundaries in successive images using the estimated boundaries of the previous image as initialization. A snake algorithm, based on the gradient, is used to refine the initialization.
Results
We evaluate the average diameter of the artery for every image and plot the evolution of the diameter in time.
