Motion tracking of left ventricular local myocardium in B-mode echocardiogram using DP-correlation method

Abstract

The echocardiogram is an essential tool for diagnosis and treatment of cardiovascular diseases. The use of echocardiograms for clinical diagnosis serves multifold advantages including noninvasiveness of diagnosis and real-time imaging. Recent technical innovations of ultrasonic diagnosis equipment require tracking a specific tissue of myocardium. An automated, computer-based objective tracking of myocardium is necessary for quantitative diagnosis. Several methods have been proposed for tracking of myocardial motion and quantitative assessment of the myocardial performance. However, speckle and noise contained in echocardiograms sometimes corrupt the tracking. In this study, to eliminate such errors and improve the tracking accuracy, we propose new tracking algorithm that employs the dynamic programming (DP) based correlation method. The proposed method assumes the cyclic movement of left ventricular myocardium. The objective function employed in correlation tracking is optimized using DP based tracking algorithm on this assumption. Tracking experiments using clinical echocardiography suggest that the proposed method is superior on myocardial motion tracking than the traditional correlation based tracking method. © 2010 International Federation for Medical and Biological Engineering.