【作者】 岳泉；

【导师】 王金海；

【作者基本信息】 天津工业大学 ， 信号与信息处理， 2007， 硕士

【摘要】 随着计算机技术、信息技术、医学影像技术的不断发展，众多图形图像工作站被用于医疗事业。内窥镜技术是一种重要的临床检查和治疗的手段，而数字化内窥镜影像系统也是现代医疗体系的重要组成部分。该系统的应用便于内窥镜图像存储、管理、传输、组织和分析，为医生进行临床诊断、病情跟踪、教学和研究提供了的重要依据。本文研究和分析了内窥镜影像系统的功能要求和特点，将数据库技术、数字图像处理和分析方法应用到系统之中。因此本文首先考虑到该系统的各项功能和工作流程，依据多媒体数据库理论，建立满足现实需求的数据库关系模型，进而构成完整内窥镜信息管理系统，对内窥镜图像和文本信息进行有效的组织、管理和维护，并且实现了报告的规范化书写和病例查询；结合数字图像处理的相关理论，针对内窥镜图像的特点，讨论了多种图像增强的方法，在HSI空间用小波分析的方法进行彩色内窥镜图像增强，获得较好的视觉效果；并且对几种图像分割的方法进行了讨论，并将其应用于图像病灶标记；将基于内容的图像检索技术用于内窥镜图像检索，提取图像的颜色和纹理特征，并对单一特征检索和组合特征检索的结果进行了分析，利用组合特征检索的结果更令人满意。更多还原

【Abstract】 With the developing of computer technology, information engineering and medical imaging, more and more image workstations are used in the medical area. Endoscope technology is one of the important technologies in clinical diagnosis and treatment, and digital endoscope image system is an important part of modern medical system. With the application of the system, it facilitates the storing, managing, transmitting, organizing and analyzing of the endoscope image. So it provides an important and objective basis to doctor carrying on clinical diagnosis, condition following, teaching and study.Based on the study and analysis of characteristic and requirement in digital endoscope image system, database technology and digital image processing and analysis are applied in that system. Considering the functions and tasks of the system, we establish the relation data model of the database to meet the need of requirement on the basis of database theory. And then an integrate endoscope information management system is established. In the system we can realize the efficient organizing, managing and maintaining of endoscope image and information that is related. At the same time, we can realize writing reports in a standardized way and case query. Combining the related theories of digital image processing and characteristics of endoscope image, many image enhancement methods are discussed. In order to obtain better visual quality, a method based on wavelet analysis in HSI color space was applied in color medical endoscope image enhancement. Several image segmentation techniques are used to mark the focus on the image. Content-based color endoscope image retrieval is studied. Color and texture feature are extracted from the color endoscope image, and then we analyze the result of single feature image retrieving techniques and multi-feature fusion. The last method is the best one.更多还原