权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

DEVELOPMENT OF EXPERT SYSTEM FOR HIGH-SPEED 3D COMPUTED TOMOGRAPHY

高速 3D 计算机断层扫描专家系统的开发

基本信息

批准号：
03557048
负责人：
OHYAMA Nagaaki
金额：
$ 4.35万
依托单位：
TOKYO INSTITUTE OF TECHNOLOGY
依托单位国家：
日本
项目类别：
Grant-in-Aid for Developmental Scientific Research (B)
财政年份：
1991
资助国家：
日本
起止时间：
1991 至 1992
项目状态：
已结题

项目摘要

A practical investigation on 3-D X-ray computed tomography (CT) using con-beam projections is done and the system for the visualization of 3-D data is developed in this project. Whole 3-D data are acquired only with a rotation of a X-ray source and detector pair. 3-D images are experimentally reconstructed from the projections of phantoms and live subject. 3-D images of high-resolution bone structures are obtained using magnified projection geometry.The 3-D imaging technique of coronary artery is also investigated. The 3-D image of blood vessels is reconstructed using simulated annealing algorithm through the experiment. It is possible to image coronary artery from venography by this method, which is valuable for the diagnosis of heart disease.We construct the expert system to observe 3-D volume data that are acquirecd from 3-D CT systems. A new method for the visualization of volume data using optical model is developed and implemented in the system. High speed interactive operation is possible with electro-optical hybrid processing. The operator can change the mapping table between CT numbers to optical constants, and can generate images of fuzzy surfaces. In the experiment, the images of skull, muscle, soft tissues, and brain are reconstructed from CT data. 3-D structures of such subjects, which is difficult to observe three-dimensionally through conventional method, are visualized with rotation through this system.

本项目对共束投影三维X射线计算机断层扫描（CT）进行了实际研究，并开发了三维数据可视化系统。整个3-D数据采集仅与X射线源和探测器对的旋转。实验中，我们从人体模型和活体的投影中重建了三维图像。利用放大投影几何学获得了高分辨率的骨骼结构三维图像，并对冠状动脉三维成像技术进行了研究。通过实验，采用模拟退火算法重建血管三维图像。该方法可以通过静脉造影对冠状动脉进行成像，这对心脏病的诊断有价值。我们构建了专家系统来观察从三维CT系统获取的三维体数据。提出并实现了一种基于光学模型的体数据可视化方法。高速交互式操作是可能的电光混合处理。操作员可以改变CT数到光学常数之间的映射表，并且可以生成模糊表面的图像。在实验中，颅骨，肌肉，软组织和大脑的图像重建从CT数据。通过该系统，通过旋转可视化难以通过常规方法三维观察的此类对象的3-D结构。