基于脑实质分割测量的脑萎缩辅助诊断研究

发布时间：2018-10-13 13:28

【摘要】：现如今,医学成像系统已成为临床与医学研究中不可或缺的辅助工具,随着医学成像技术的飞速发展,促进了计算机辅助诊断技术在医学研究和临床实验方面需求庞大且发展迅速,而计算机辅助诊断技术的基础是医学图像的处理,所以医学图像处理技术是一个计算机科学与临床医学多学科相互交叉的研究热点领域之一。随着现代社会人口老年化进程不断加快,而脑萎缩又是老年人十分常见的疾病,这导致了医务人员工作负担的不断增加。为了解决这个问题,可充分发挥和利用现代计算机速度快、效率高和成本低的优势。因此,本文对计算机辅助诊断脑萎缩技术展开研究。在提出脑萎缩辅助诊断系统的基础之上,先对原始脑部医学图像的预处理进行研究,然后讨论脑实质的提取方法,最后重点展开对脑实质分割和脑体积测量技术的研究讨论。论文主要研究成果包括:(1)利用现代图像去噪的技术,通过实验总结出一套适合脑部医学图像的去噪流程及方法。(2)通过分析脑部医学图像的特性和人脑的组织结构,创新使用多重阈值分割算法实现脑实质的提取。(3)通过分析高斯混合模型和K-means两种经典的聚类算法分割脑实质存在的不足来加以改进和融合,创新使用融合后的GKA算法分割脑实质。(4)提出两种不同的脑实质片面积和体积测量的方案,并针对两种不同的方案分别做出实验对比。在使用临床真实的脑部医学图像进行试验后得出结论:本文提出的关于医学图像去噪流程和脑实质提取算法均获得较理想的实验结果;在脑实质分割方面,GKA方法分割的结果更是在各项指标中比传统的高斯混合模型和K-means聚类算法要全面领先;在脑实质的片面积测量方面,由于论文提出的两种测量方法是基于对脑实质片面积的不同定义,从而导致两种方法得出的结果存在一定的差异,但是这两者的结果都具有一定的临床参考价值。
[Abstract]:Nowadays, medical imaging system has become an indispensable assistant tool in clinical and medical research. With the rapid development of medical imaging technology, It has promoted the development of computer-aided diagnosis technology in medical research and clinical experiment, and the basis of computer-aided diagnosis technology is medical image processing. Therefore, medical image processing technology is one of the hot research fields of computer science and clinical medicine. With the rapid aging of population in modern society, brain atrophy is a common disease in the elderly, which leads to the increasing burden of medical workers. In order to solve this problem, we can make full use of the advantages of high speed, high efficiency and low cost of modern computer. Therefore, the computer-aided diagnosis of brain atrophy is studied in this paper. On the basis of putting forward the assistant diagnosis system of brain atrophy, the preprocessing of the original brain medical image is studied, then the extraction method of the brain parenchyma is discussed. Finally, the segmentation of the brain parenchyma and the measurement of the brain volume are emphatically discussed. The main research results are as follows: (1) by using modern image denoising technology, a set of denoising processes and methods suitable for brain medical image are summarized through experiments. (2) by analyzing the characteristics of brain medical image and the structure of human brain, The multi-threshold segmentation algorithm is used to extract the brain parenchyma. (3) by analyzing the shortcomings of Gao Si's mixed model and K-means 's two classical clustering algorithms to improve and fuse the segmentation of brain parenchyma. The fused GKA algorithm is used to segment the brain parenchyma. (4) two different methods of measuring the area and volume of the brain parenchyma are proposed and compared with each other. After using the clinical real brain medical images, the conclusion is drawn: the proposed process of medical image denoising and the algorithm of extracting brain parenchyma have achieved satisfactory experimental results; In the aspect of brain parenchyma segmentation, the result of GKA segmentation is more advanced than the traditional Gao Si mixed model and K-means clustering algorithm in every index, and in the area measurement of brain parenchyma, Because the two measurement methods are based on the different definitions of the area of the brain parenchyma, there are some differences between the two methods, but the results of the two methods have some clinical reference value.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;R742

【参考文献】