Estimation of cavity centroids and regions of interest in echocardiographic images

Author

De Sá, J. P Marques ; Monteiro, A.M.P. ; Abreu-Lima, C.

Author_Institution

Fac. of Eng., Oporto Univ., Portugal

Volume

i

fYear

1990

fDate

16-21 Jun 1990

Firstpage

659

Abstract

A method for automatically deriving feasible region of interest (ROI) and left-ventricle (LV) cavity mass center (centroid) estimates is described. The corresponding algorithm is used in a computer-based system for automatic echocardiographic image analysis. The centroid estimation takes place in an initially large ROI. The algorithm successively refines an estimate of the centroid and ROI, applying several criteria to a grid of line segments falling inside the cavity. Tests in 22 cases for several parameter values showed that in 11 closed-boundary cases the true centroid was estimated. In the remaining 11 open-boundary cases, a few errors were observed in three cases and viable estimates could be derived for the other ones

Keywords

acoustic imaging; biomedical ultrasonics; cardiology; computerised picture processing; medical diagnostic computing; automatic echocardiographic image analysis; cavity centroids; closed-boundary; computer-based system; computerised picture processing; left-ventricle; open-boundary; regions of interest; Filtering; Heart; Image analysis; Image processing; Image segmentation; Kernel; Low pass filters; Nonlinear filters; Testing; Yield estimation;

fLanguage

English

Publisher

ieee

Conference_Titel

Pattern Recognition, 1990. Proceedings., 10th International Conference on

Conference_Location

Atlantic City, NJ

Print_ISBN

0-8186-2062-5

Type

conf

DOI

10.1109/ICPR.1990.118184

Filename

118184