Myocardial Motion Analysis Using Modified Radial Direction Distribution Based on Magnitude Criteria

Abstract

Accurate information on myocardial motion is important in diagnosing cardiac abnormalities. In this research, the motion vectors are computed using an optical flow technique and further analyzed based on the magnitude and angle. We utilize a relative motion direction with respect to the center of the cavity, which is more useful for diagnosis because a physician can observe whether a segment is moving to the center or not more clearly. Instead of analizing individual vectors, it is more helpful to visualize the overall trend by performing their angular distribution. However, a common angular distribution considers motions with small and large magnitudes to have only an equal contribution and often fails to represent an accurate motion direction of a segment if there is a wide distribution of vector angles in the respective segment. In this paper, we propose a new method of providing an accurate radial direction profile by modifying the contribution of the motion vectors to the radial direction distribution based on a certain magnitude function. This method has been tested on clinical echocardiography sequences and shown to be successful in providing a more accurate radial direction profile compared with the common angular distribution.