VESSEL DETECTION IN CAROTID ULTRASOUND IMAGES USING ARTIFICIAL NEURAL NETWORKS

Abstract

Carotid Doppler ultrasound and imaging are focused on the visualization, identification and measurement of vessels and blood flow providing critical diagnostic information on symptomatic or asymptomatic stenotic or embolic accidents. Ultrasound imaging is a complicated interplay between physical principles and signal processing methods. In this work the development of a new algorithm for vessel identification and image segmentation in ultrasound images is reported. A fully automatic technique based on pixel intensity distribution alleviates the laborious and time consuming manual measurement and classification of the carotid artery.