Intravascular ultrasound image segmentation: A fast-marching method

Abstract

Intravascular ultrasound (IVUS) is a medical imaging technique that not only provides three-dimensional information about the blood vessel lumen and wall, but also directly depicts atherosclerotic plaque structure and morphology. Automatic processing of large data sets of IVUS data represents an important challenge due to ultrasound speckle and technology artifacts. A new semi-automatic IVUS segmentation model, the fast-marching method, based on grayscale statistics of the images, is compared to active contour segmentation. With fast-marching segmentation, the lumen, intima plus plaque structure, and media contours are computed in parallel. Preliminary results of this new IVUS segmentation model agree very well with vessel wall contours. Moreover, fast-marching segmentation is less sensitive to initialization with average distance between segmentation performed with different initializations <0. 85% and Haussdorf distance <2.6%.