Development of Ultrasonic 4-D Imaging Method

本研究的目的是开发一种利用超声图像显示靶器官三维运动图像的新方法。为了更好地显示三维图像，准确提取器官边界是至关重要的。但是从超声图像中准确、自动地检测器官边界是一个困难的问题。因此，我们开发了使用模糊推理和松弛技术的边界提取方法。将该实验应用于胎儿临床资料，获得了高质量的胎儿三维图像。此外，我们还将该算法应用于心脏临床数据。在心脏临床资料中，b型图像的质量较差，因此很难检测到器官的边界。b型图像为超声回波图像。顺便说一下，m型图像在心脏诊断中应用广泛。因为它很容易识别边界。然后将改进后的方法应用于临床资料。通过以下两种方式获取4D数据。首先，我们获得了心电图间隔2.5度的四维数据。另外，我们每隔一次心跳通过经食管扫描获得4-D数据。我们将改进的方法应用于临床数据，提取器官边界，并使用深度提示技术显示4D图像。结果表明，应用该算法可获得显著的四维图像。

The purpose of this study is to develop a new method for displaying 3D motion image of target organ from ultrasound images. To display 3D image finely, it is essential to extract the organ boundaries exactly. But it is difficult to detect the organ boundaries from ultrasound images exactly and automatically. Therefore, we developed the boundary extraction method using fuzzy reasoning and relaxation techniques. As a result of applying the experiment to fetal clinical data, we obtained the high quality 3D image of fetus. And besides, we applyed this algorithm to cardiac clinical data. In the case of cardiac clinical data, the quality of B-mode image is worse, and so it is much difficult to detect the organ boundaries. B-mode image is ultrasound echo image. By the way, M-mode image is used widely in cardiac diagnosis. Because it is much ease to recognize the boundary. Then we applyed this improved method to clinical data. 4D data were acquired by following two ways. First we acquired 4-D data at an intervals of 2.5 degrees with ECG.Other we acquired 4-D data with transesophageal scan every a heartbeat. We applyed the improved method to the clinical data, and extracted organ boundaries, and displayd 4D image using depth cueing technique. It was found that significant 4D image obtained by applying the algorithm.