Date of Award

Summer 1999

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)

Department

Biomedical Engineering

First Advisor

Hock, Jeffrey

Second Advisor

Riedel, Susan

Third Advisor

Kelnhofer, Richard

Abstract

Segmentation of the left ventricular epicardium is a fundamental step in analyzing the performance of the heart. Accurate determination of the epicardium is key to important measurements such as wall thickening, wall motion, and myocardial mass. Currently, algorithms are available for segmenting the epicardium, but they lack robustness and are difficult to use. The epicardial segmentation technique presented in this thesis solves several of the problems facing existing techniques. The only user interaction required is the placement of one seed point for the dataset. Even with this limited information from the user, the error in the resulting epicardial contours is similar to the variability between doctors. In addition, the new segmentation technique easily enables more advanced visualization methods. Three dimensional surface models of the epicardium can be created quickly from the segmentation data. Combining these models with clinically relevant cardiac measurements will provide doctors with more advanced diagnostic techniques.

Share

COinS

Restricted Access Item

Having trouble?