弹性成像算法比较研究

	弹性成像算法比较研究
其他题名	Stduy of Ultrasound Elastography Algorithms in different methods
	崔亚琨
学位类型	硕士
导师	崔崤峣
	2015-10
学位授予单位	中国科学院大学
学位专业	机械工程
关键词	弹性成像运动估计二维互相关先验估计加权相位分离
摘要	弹性成像能够检测组织的弹性模量信息，进而描述组织的生理、病理状态，对疾病的检测和诊断具有重要的应用价值。例如，肿瘤在组织中生长或扩散，组织的弹性会发生变化，在肿瘤出现形态学改变前，它的硬度已经发生了变化，导致正常组织与病变组织之间存在较大的弹性差异。在弹性成像出现之前，医生主要通过触诊来检测组织之间硬度的差别。这种方法有很大的局限性，如果肿块位于组织深处或者病变部位较小时，这种方法不能取得较好的结果。目前，弹性成像已经成为医用超声领域的一个研究热点。本文通过仿真和仿体实验进行了比较研究，提出了一种基于加权相位分离和二维互相关的混合位移估计算法。具体工作如下：首先，介绍了弹性力学的基本知识，为生物组织仿真奠定了理论基础。研究了弹性成像的原理。使用有限元法对生物组织建模仿真，研究了不同压缩率下的组织的位移、应变、应力分布。使用Field 仿真声场，得到了不同压缩率下的超声RF信号。其次，研究了一维弹性成像方法。一维算法可以分为基于梯度运算的应变估计算法和直接应变估计算法。本文研究了经典的一维互相关算法，针对相关函数的最大值只能精确到采样点上，采用亚采样时延估计技术来提高位移估计精度。并且对算法进行了仿真、仿体实验。最后，针对一维弹性成像方法只考虑组织轴向运动的弱点，提出了一种混合位移估计算法，它基于加权相位分离和二维互相关技术，首先利用二维时域互相关技术进行粗估计，然后再利用相位加权分离技术进行精估计。同时，通过仿真和仿体实验对算法的准确性和效率进行了验证。结果表明，算法具有较好的鲁棒性，能够有效的提高运算效率和图像信噪比，上述研究对高性能弹性成像系统的研究与设计具有重要的指导意义。
其他摘要	Elastography is an imaging method which is able to display the distribution of elastic properties of soft tissue, and provide valuable diagnostic information. For example, with tumor growing, the elasticity of tissue will change. Elasgotraphy can detect the disease early. The doctor detected the stiffness information of biological tissue by palpation method which is subjective. This method may be less effective when tumor is small. In recent year, elastography has become a hot spot in ultrasound imaging areas. In this paper, we compared 1D motion estimation with 2D motion estimation for elastography and present a framework combined with weighted phase separation (WPS) and 2D cross-correlation for displacement estimation. Our work is shown as follow: Firstly, we studied the basic knowledge of Theory of elasticity. Elastography was studied in this paper. And we use the finite element method to study the simulation of compression organization. We also use the Field II to simulate the soft tissue. Secondly, We studied 1D algorithm. In this paper, We studied the classical 1D cross-correlation algorithm and overcame its weaknesses that the maximum value of correlation was located in sampling point by using sub-sample time delay estimation. Thirdly, taking into account the influence of tissue side-slipping, we presented a 2D cross-correlation method firstly to apply to the RF signals and then used WPS method to calculate more precise axial displacements. We also compared images qualities and efficiency of 1D algorithm and 2D algorithm by simulations and phantom experiments. It is shown that this method is fast and robust to relatively complex tissue motions and could improve SNRe. This method will have some potential for the latter experiments and clinical applications.
语种	中文
文献类型	学位论文
条目标识符	http://ir.ciomp.ac.cn/handle/181722/49311
专题	中科院长春光机所知识产出
推荐引用方式 GB/T 7714	崔亚琨. 弹性成像算法比较研究[D]. 中国科学院大学,2015.

条目包含的文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
崔亚琨.pdf（2050KB）	学位论文		开放获取	CC BY-NC-SA	请求全文