Share Email Print
cover

Proceedings Paper

An active contour framework based on the Hermite transform for shape segmentation of cardiac MR images
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Early detection of cardiac affections is fundamental to address a correct treatment that allows preserving the patient’s life. Since heart disease is one of the main causes of death in most countries, analysis of cardiac images is of great value for cardiac assessment. Cardiac MR has become essential for heart evaluation. In this work we present a segmentation framework for shape analysis in cardiac magnetic resonance (MR) images. The method consists of an active contour model which is guided by the spectral coefficients obtained from the Hermite transform (HT) of the data. The HT is used as model to code image features of the analyzed images. Region and boundary based energies are coded using the zero and first order coefficients. An additional shape constraint based on an elliptical function is used for controlling the active contour deformations. The proposed framework is applied to the segmentation of the endocardial and epicardial boundaries of the left ventricle using MR images with short axis view. The segmentation is sequential for both regions: the endocardium is segmented followed by the epicardium. The algorithm is evaluated with several MR images at different phases of the cardiac cycle demonstrating the effectiveness of the proposed method. Several metrics are used for performance evaluation.

Paper Details

Date Published: 29 April 2016
PDF: 9 pages
Proc. SPIE 9896, Optics, Photonics and Digital Technologies for Imaging Applications IV, 98960U (29 April 2016); doi: 10.1117/12.2223450
Show Author Affiliations
Leiner Barba-J, Univ. Nacional Autónoma de México (Mexico)
Boris Escalante-Ramírez, Univ. Nacional Autónoma de México (Mexico)


Published in SPIE Proceedings Vol. 9896:
Optics, Photonics and Digital Technologies for Imaging Applications IV
Peter Schelkens; Touradj Ebrahimi; Gabriel Cristóbal; Frédéric Truchetet; Pasi Saarikko, Editor(s)

© SPIE. Terms of Use
Back to Top