Title :
Quantification of parametric images to assess segmental wall motion of the left ventricle in echocardiography
Author :
Frouin, F. ; Ruiz-Dominguez, C. ; Kalikian, T. ; Kachenoura, N. ; Delouche, A. ; Herment, A. ; Nardi, O. ; Diebold, B.
Author_Institution :
INSERM, UPMC, Paris
Abstract :
An automated method was developed in echocardiography for defining regional wall motion score (RWMS) using a 4 point scale. Eighty-six time image sequences were analyzed. The left ventricle was divided into 7 segments and the reference RWMS was given by experts. The image sequences were processed by factor analysis of the left ventricle in echocardiography (FALVE). Firstly, the FALVE parametric images were visually interpreted. Then a FALVE wall motion index was computed for each segment and a RWMS was deduced. When comparing the scores obtained by the reading of parametric images with the reference RWMS, an absolute agreement, Aa, of 65% a relative agreement, Ar, of 95% and a weighted Kappa coefficient kw of 0.67 were obtained. When comparing the scores obtained by the index with the reference RWMS, Aa of 64%, Ar of 92% and kw of 0.60 were reached. The index is fast and easy to compute, and results in defining RWMS are quite encouraging
Keywords :
biomechanics; echocardiography; image segmentation; image sequences; medical image processing; Kappa coefficient; automated method; echocardiography; factor analysis; image sequence; left ventricle; parametric image quantification; regional wall motion; segmental wall motion; Argon; Echocardiography; Image analysis; Image databases; Image segmentation; Image sequence analysis; Image sequences; Motion detection; Optical imaging; Valves;
Conference_Titel :
Computers in Cardiology, 2005
Conference_Location :
Lyon
Print_ISBN :
0-7803-9337-6
DOI :
10.1109/CIC.2005.1588053