基于局部频率和微分反演的MRE重建算法研究

发布时间：2018-03-31 21:46

本文选题：磁共振弹性成像(MRE)　切入点：剪切模量　出处：《东北大学》2014年硕士论文

【摘要】：磁共振弹性成像(Magnetic Resonance Elastography, MRE)是一种利用磁共振成像技术(MRI)新型的无创成像方法,能直观显示和量化组织弹性。它是传统触诊机器化、定量化的一种手段,客观且分辨率高,不受诊断部位的限制,MRE图像表达的是组织的生物力学特性参数,它能够检测体内组织在外部激励机械波作用下产生的质点位移。现在已经提出多种算法去确定物体的弹性性能,MRE面临的一个挑战性问题之一就是在重建剪切模量的图像中,如何减少噪声影响的同时并保持较高的对比度。所以在本文中,我们重点研究了磁共振弹性成像重建的相关算法。本文介绍了基于局部频率估计的弹性图重建算法,它有多种实现方式,传统的实现方式是基于双带宽高斯滤波器,而本文使用了一个新的滤波器进行局部频率估计,是对数正态正交小波,这也是本文的一个创新点,将我们提出的新方法与传统方法用Matlab平台实现并进行分析比较,结果证明使用对数正态正交小波得到的弹性图分辨率比传统方式高。接着本文又介绍了基于微分反演算法的磁共振弹性重建算法,因微分反演法的实现需要对相位位移数据进行平滑和二阶微分的计算,为了同时实现上述两个要求,本文选用Savitzky-Golay滤波器进行实现,它是一种实用性很强,除噪效果非常好的滤波器。将微分反演算法与局部频率估计算法得到的弹性图进行分析比较。最后本文又简单介绍了一下磁共振弹性重建的另一种方法剪切模量分解法,重点介绍了其原理和推导过程。
[Abstract]:Magnetic Resonance elasticity imaging (MREs) is a new noninvasive imaging method using magnetic resonance imaging (MRI), which can directly display and quantify tissue elasticity. It is a method of traditional palpation machine and quantification, which is objective and has high resolution. The MRE images, which are not restricted by the location of diagnosis, express the biomechanical characteristic parameters of the tissue. It can detect the displacement of particles caused by external mechanical waves. Now many algorithms have been proposed to determine the elastic properties of the body. One of the challenges for MRE is to reconstruct the image of shear modulus. How to reduce the noise effect while maintaining high contrast. Therefore, in this paper, we focus on the relevant algorithms of MRI reconstruction. In this paper, we introduce an elastic image reconstruction algorithm based on local frequency estimation. There are many ways to realize it. The traditional way is based on double bandwidth Gao Si filter. In this paper, a new filter is used to estimate the local frequency, which is the logarithmic normal orthogonal wavelet, which is also an innovation of this paper. The new method proposed by us is implemented with Matlab platform and compared with the traditional method. The results show that the resolution of the elastic image obtained by using logarithmic normal orthogonal wavelet is higher than that of the traditional method. Then, an elastic reconstruction algorithm of magnetic resonance based on differential inversion algorithm is introduced in this paper. The realization of differential inversion method requires smooth and second-order differential calculation of phase displacement data. In order to realize the above two requirements simultaneously, Savitzky-Golay filter is chosen to implement it, which is very practical. The differential inversion algorithm and the local frequency estimation algorithm are analyzed and compared with the elastic graph obtained by the local frequency estimation algorithm. Finally, another method of shear modulus decomposition for magnetic resonance elastic reconstruction is briefly introduced in this paper. The principle and derivation process are emphatically introduced.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R445.2

【参考文献】