An Analytical Descriptor of Three-Dimensional Geometry: Application to the Analysis of the Left Ventricle Shape and Contraction

Haim Azhari, Samuel Sideman, Rafael Beyar, Ehud Grenadier, Uri Dinnar

Research output: Contribution to journalArticle

Abstract

A novel method to describe the instantaneous three-dimensional (3-D) geometry of the left ventricular (LV) endocardial surface by an analytical time-varying scalar function of Fourier spectrum coefficients is suggested. The method utilizes experimental echocardiography data and uses a helical coordinate system to transform the 3-D data into a unidimensional numerical function. The instantaneous numerical function is then represented by its Fourier sine series which serves as an analytical shape descriptor from which the 3-D shape is reconstructed. The procedure can also be applied to data compression (spatial low-pass filtering), spectral analysis, and the evaluation of geometric similarity of 3-D shapes. When applied to the endocardial surface of the LV at end-diastole (ED) and end-systole (ES) this technique gives a quantitative analysis of the global LV contraction of the real heart. The descriptor was tested by utilizing echocardiographic LV data from 10 normal subjects (age 15± 4.5 years), showing that the average LV geometrical information is much more distributed in the frequency domain than the volumic information, and that the ES geometry is of higher irregularity than the ED one. It is noted that LV contraction can be described by a combination of homogeneous lateral and longitudinal contractions of the first five harmonics.

Original languageEnglish (US)
Pages (from-to)345-355
Number of pages11
JournalIEEE Transactions on Biomedical Engineering
VolumeBME-34
Issue number5
DOIs
StatePublished - 1987
Externally publishedYes

ASJC Scopus subject areas

  • Biomedical Engineering

Fingerprint Dive into the research topics of 'An Analytical Descriptor of Three-Dimensional Geometry: Application to the Analysis of the Left Ventricle Shape and Contraction'. Together they form a unique fingerprint.

  • Cite this