【作者】 王飞；

【导师】 秦志光；

【作者基本信息】 电子科技大学 ， 工程硕士（专业学位）， 2017， 硕士

【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。更多还原

【Abstract】 Nowadays,medical imaging systems have become an indispensable accessory tool for clinical and medical research.With the rapid development of medical imaging technology,computer-aided diagnostic technology in medical research and clinical experiments has been hugely prompted towards high demand and rapid development because medical image processing forms the basis of computer-aided diagnostic techniques.Therefore,medical image processing technology remains one of the hotspots in multidisciplinary research in the field of computer science and medicine.The high increase in the number of the aging population battling with the Alzheimer’s disease makes it an increasing burden for the medical community.In solving this problem,it can take full advantage of modern computer capability such as high processing speed,low cost and high efficiency.As such,this thesis focus on the basis of computer-aided diagnosis system of Alzheimer’s disease technology research.This involves research on original brain medical image pretreatment and then the method of brain tissue extraction fully illustrated.Finally,the thesis focus on analysis of brain tissue segmentation and brain volume measurement technology.As such,the main research results of the thesis include the following:(1)Using modern image denoising technology to create a set of denoising process for brain medical image.(2)By analyzing the characteristics of brain medical images and the structure of the human brain,and then use of multiple threshold segmentation algorithm to extraction of brain tissue.(3)By analyzing the shortcomings of two classical clustering algorithms;Gaussian mixture model and the k-means algorithms in segmenting the brain tissue.Also fusing both algorithms to achieve a more improved GKA algorithm for brain image segmentation its tissues.(4)Two different methods of measuring the brain tissue area and volume are proposed in this thesis,then a comprehensive experimental result comparison of the two schemes are adequately discussed.After engaging tests using Magnetic resonance images from clinical procedures,it can be concluded that the medical image denoising process and brain tissue extraction algorithm obtained ideal and accurate experimental results.In the brain tissue segmentation,the results of the GKA method provided better results compared to the traditional Gaussian mixture model and k-means clustering algorithm in each index.In terms of the measurement of brain tissue area,because the two methods proposed in this thesis are based on different definitions of brain tissue area,the results of the two methods are different but both of the results have certain distinguishable clinical reference value.更多还原