【作者】 胡莎莉；

【导师】 唐治德；

【作者基本信息】 重庆大学 ， 电气工程， 2004， 硕士

【摘要】 随着计算机技术的不断发展和成熟，图像存档与通讯系统(即PACS，Picture Archiving and Communications System)在医疗机构中的应用越来越广泛。PACS系统完成对医学图像的图像采集、存储归档、图像显示、图像处理和传输等功能。其中医学图像处理系统涉及多种图像处理、图形操作和三维可视化算法，是PACS系统的核心功能之一。 本文实现了一个基于微机的PACS系统中的医学图像处理系统，该系统实现了对符合医学数字图像与通讯标准DICOM的图像文件的解析功能；提供了对医学图像数据的基本处理功能和图像数据处理的用户界面；实现了由序列CT图像构造体数据，并提供体数据的预处理、表面绘制和体绘制功能，提供了可视化的控制界面。本文还对医学图像处理和分析原理及医学图像的可视化的各个过程进行了详细分析和说明。在实现基本功能的基础上，设计并实现了动态窗宽、窗位调节技术；设计并实现了基于目标图像的旋转算法，保证了图像旋转的目标图像面积不变性；提出并实现了一种改进的基于梯度的图像分割算法，通过利用图像的梯度信息和分割阈值相结合的算法，提高了图像分割的准确性；改进了常用的4邻域区域填充算法，使其符合具体的需要，使该算法在应用中对计算机内存的要求大为降低；分析和比较了常用的表面绘制和体绘制算法的优劣，基于对VTK(Visualization Toolkit)类库的研究和分析，实现了医学序列CT图像的三维可视化。更多还原

【Abstract】 With the development of computer, the usage of PACS is in great demand. PACS (Picture Archiving and Communications System) accomplishes the functions used in Medical image, including medical image acquisition, image storage and archiving, image displaying, image processing and communication. Among these, the medical image processing, which is one of the most important core functions in PACS, involves a lot of technology about the image processing, graphics manipulation, 3D data visualization.In this article, we realize a medical image processing system in PACS based on personal computer. In the system we accomplished the function of analyzing medical image which comply with digital image communication in medical, providing the basic function of medical image processing and user interface, constructing volume data from a serial medical CT image, providing the functions include volume data preprocessing, and volume data surface rendering and volume rendering as well, providing visualization control interface. In this article we also describe in detail the theory of medical image processing and analyzing. We described process of medical image visualization as well. In the procedure of implementation basic function, we designed a method to dynamic adjust Windows center/Windows width and image rotate algorithm, which don’t change target image dimension. We designed improved image segment algorithm according to image gradient, which improved the speed and accuracy of segment by combining threshold with gradient information. We improved 4-neighbor filling algorithm, decreased the spatial complexity, descent the demand for computer memory. We compared surface rendering and volume rendering and implement series CT image visualization based on VTK.更多还原