Estimation of two dimensional regional wall motion and two dimensional vector flow

传统的评价局部室壁运动的方法需要对室壁运动的方向进行判断，如向腔中心或连续帧中最近点的运动。然而，实际壁运动包括心脏旋转或更复杂的运动，并且该实际运动可以反映心脏状况。因此，我们尝试使用高帧速率超声心动图（60至89帧/秒）在没有假设的情况下跟踪二维室壁运动，因为高帧速率回波中两个连续帧之间的心室壁变形预计很小，这使得跟踪图像变得容易。我们应用复合二维相关方法，它使用适当加权的三个指标，运动的连续性，周期性变化和大的感兴趣区域。每个指标的权重是通过最小化人类跟踪的累积平方误差来确定的。将该方法应用于另一幅图像，并与简单相关方法、传统的假设运动方向的方法和传统的帧率图像进行了比较。这种复合方法是最准确的。将该方法应用于临床，检测到的室壁运动客观地反映了疾病特异性室壁运动。第二个研究目标是从常规彩色多普勒图像中，基于连续性方程估计心室腔内的二维矢量流。用B超二维血流显像、微气泡注入及计算机模拟三维血流的误差进行了验证。目前正在准备提交这一结果。

Conventional methods to evaluate regional wall motion need assumptios on the direction of wall motion such as motion toward the center of cavity, or the nearest points in the successive frame. However, actual wall motion includes cardiac rotation or more complicated motions and this actual motion may reflect cardiac conditions. Therefore we tried to track two dimensional wall motion without assumptions using high frame rate echocardiography (60 to 89 frame/sec) since defprmation of ventricular wall between two successive frames in high frame rate echo is expected to be small and this makes it easy to track images. We applied composite 2D correlation method, which uses appropriately weighted three indices, continuation of motion, cyclic change and large region of interest. The weight of each indices was determined by minimizing the accumulated squared errors from human tracing. We applied this method to another image and compared the result with simple correlation method, conventional methods with assumption on moving direction and conventional frame rate images. This composite method was the most accurate. This methods was applied to clinical diseases and detected wall motion revealed objectively disease specific wall motions. In the second research aim, we estimated two dimensional vector flow from conventional color Doppler image in ventricular cavity based on continuity equation. This result was examined with visualized 2D flow image by B-mode echo, injecting micro bubble and error derived from 3D flow was estimated computer simulation. This result is now in preparation for submission.