Real-Time Three-Dimensional Ultrasound Methods for Shape Analysis and Visualization - Robotics Institute Carnegie Mellon University

Real-Time Three-Dimensional Ultrasound Methods for Shape Analysis and Visualization

Journal Article, Methods, Vol. 25, No. 2, pp. 221 - 230, October, 2001

Abstract

Real-time three-dimensional (RT3D) ultrasound is a relatively new imaging modality that uses a special ultrasound transducer consisting of a matrix array of elements. The array electronically steers an ultrasound beam to interrogate a 3D volume of tissue. The real-time nature of RT3D ultrasound differentiates it from reconstructed 3D ultrasound, in which a conventional ultrasound transducer is moved mechanically through the third dimension. RT3D ultrasound is considerably faster than reconstructed 3D ultrasound, making it suitable for capturing continuous rapid motion such as that of the beating heart. Although RT3D ultrasound has not yet found widespread clinical use, these scanners are presently employed in more than 20 locations worldwide, primarily for cardiac research. The author helped develop the RT3D ultrasound technology as well as specialized analysis and visualization methods for the resulting data. In developing such methods, it has been necessary to consider the physical and mathematical processes by which the ultrasound data are collected. Difficulties arise because of high noise, variation in contrast and intensity between scans, ultrasound's nonrectilinear coordinate system, and the anisotropic nature of the echoes themselves. This article reviews these specific difficulties and provides solutions that are applicable to generalized analysis and visualization of RT3D ultrasound data. Some of the methods presented can also be applied to other imaging modalities with nonrectilinear coordinates.

BibTeX

@article{Stetten-2001-122538,
author = {George Stetten and Robert Tamburo},
title = {Real-Time Three-Dimensional Ultrasound Methods for Shape Analysis and Visualization},
journal = {Methods},
year = {2001},
month = {October},
volume = {25},
number = {2},
pages = {221 - 230},
}