We describe a method for recovering the left intracardiac cavities from 3D Transesophageal Echocardiography (3D TEE). 3D TEE is an important modality for cardiac applications because of its ability to do fast and non-ionizing 3D imaging of the left heart complex. Segmentation based on 3D TEE can be used to characterize pathophysiologies of the valve and myocardium, and as input to patient-specific biomechanical models and preoperative planning tools. The segmentation employed here is based on a dynamic surface evolution. This is performed under a growth inhibition function that incorporates information from several sources including k-means clustering, 3D gradient magnitude, and a morphological structure tensor intended to locate the mitral valve leaflets. We report experiments using intraoperative 3D TEE data, showing good agreement between the segmented structures and ground truth. © 2011 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Burlina, P., Mukherjee, R., Juang, R., & Sprouse, C. (2011). Recovering endocardial walls from 3D TEE. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6666 LNCS, pp. 284–293). https://doi.org/10.1007/978-3-642-21028-0_35
Mendeley helps you to discover research relevant for your work.