基于改进区域生长法的肝脏血管分割算法

发布时间：2018-11-21 17:06

【摘要】：随着现代医学成像技术的不断进步,根据CT图像进行分析和诊断病情,已经成为当今最主要的方法之一。肝脏血管疾病是影响人们健康的重要疾病之一。所以,在肝脏CT图像中分析血管具有非常关键的意义。但是,肝脏血管的系统非常复杂,要想得到比较完整的血管信息,需要采用合理的图像分割技术对肝脏血管实施分割,才能对肝脏血管具有更加完整的认识与理解。在各种肝脏血管分割的方法里,区域生长算法是应用比较广泛的。所以本文在传统区域生长算法的基础上,以原始腹部CT图像作为输入,采用改进的区域生长算法对肝脏血管进行了分割。进行的研究工作主要有:(1)理论基础的研究。对医学图像的相关知识,特别是CT图像的特点进行了学习和研究。对区域生长算法的相关知识和应用进行了学习。并且研究了基于区域生长算法的肝脏血管分割的研究背景和近几年的研究成果,通过阅读文献,了解了现阶段国内外关于采用区域生长算法分割肝脏血管的研究现状,并结合肝脏血管的特点,构建了本文的算法框架。(2)图像预处理方法的选择。简单介绍了CT图像的特点与CT图像中常用的肝脏血管分割方法,根据肝脏内血管的特点,选择合适的图像预处理方法,以及肝脏血管的分割方法。(3)区域生长算法的改进思路。首先简单的介绍了传统区域生长算法的基本理论和方法。通过分析传统区域生长算法的优缺点,提出粗分割和细分割相结合的改进区域生长算法,然后以改进区域生长算法的基本原理为基础,设计并实现了肝脏血管的三维分割,其中详细描述了分割的具体流程、分割的方法以及所用到的相关技术。最后,对分割后的血管进行了研究和分析,并且选择相应的优化和可视化方法,对血管进行了进一步的处理。(4)实验验证。利用VS2010和Matlab软件平台,以分割后的肝脏CT图像序列为输入,编写算法的仿真实验程序,采用改进的区域生长算法进行了分割实验。并且借助于可视化工具包VTK,对血管进行了三维重建,呈现出血管的三维结构,使人们更加便于观察血管。本文的算法主要是针对过去分割血管的一次性问题,采用粗分割和细分割相结合的分割方法,并且在最后的实验结果和分析中,不仅对本文算法的实验结果进行了分析,而且把改进算法和常规算法的实验分割结果进行了比较与分析,证明了改进的区域生长方法在肝脏血管的分割中获得的血管信息更全面,分割结果更准确,同时,对本文算法也进行了性能的分析和比较,验证了本文算法的分割精度和可行性。
[Abstract]:With the development of modern medical imaging technology, it has become one of the most important methods to analyze and diagnose the disease according to CT images. Hepatic vascular disease is one of the most important diseases affecting people's health. Therefore, it is very important to analyze blood vessels in CT images of liver. However, the system of hepatic vessels is very complex. In order to obtain more complete vascular information, it is necessary to use reasonable image segmentation technology to segment hepatic blood vessels in order to have a more complete understanding and understanding of liver blood vessels. Among the methods of hepatic vascular segmentation, the region growth algorithm is widely used. Therefore, based on the traditional region growth algorithm, the original abdominal CT image is used as input, and the improved region growth algorithm is used to segment the liver vessels. The main research work is as follows: (1) theoretical basis research. The related knowledge of medical images, especially the characteristics of CT images, is studied and studied. The knowledge and application of region growth algorithm are studied. And the research background of liver vascular segmentation based on regional growth algorithm and the research results in recent years are studied. Through reading the literature, the current research status of using regional growth algorithm to segment liver blood vessels at home and abroad is understood. Combined with the characteristics of hepatic vessels, the algorithm framework is constructed. (2) the selection of image preprocessing methods. This paper briefly introduces the characteristics of CT image and the commonly used methods of hepatic vascular segmentation in CT image. According to the characteristics of intrahepatic vessels, the appropriate image preprocessing method is selected. And the segmentation method of hepatic vessels. (3) the improvement of region growth algorithm. Firstly, the basic theory and method of traditional region growth algorithm are introduced briefly. By analyzing the advantages and disadvantages of the traditional region growth algorithm, an improved region growth algorithm combining coarse segmentation and fine segmentation is proposed. Based on the basic principle of the improved region growth algorithm, the 3D segmentation of hepatic vessels is designed and realized. Detailed description of the specific process of segmentation, segmentation methods and the use of relevant technologies. Finally, the segmented blood vessels are studied and analyzed, and the corresponding optimization and visualization methods are selected to further deal with the blood vessels. (4) Experimental verification. Based on the VS2010 and Matlab software platform and taking the segmented liver CT image sequence as input, the simulation program of the algorithm is compiled, and the improved region growth algorithm is used to carry out the segmentation experiment. The 3D reconstruction of blood vessels is carried out with the aid of VTK, which presents the three-dimensional structure of blood vessels and makes it easier for people to observe blood vessels. The algorithm of this paper mainly aims at the one-time problem of segmenting blood vessels in the past, and adopts the method of combining coarse segmentation and fine segmentation. In the final experiment and analysis, not only the experimental results of this algorithm are analyzed. Moreover, the experimental results of the improved algorithm and the conventional algorithm are compared and analyzed. It is proved that the improved region growth method has more comprehensive information and more accurate segmentation results in the segmentation of liver blood vessels. The performance of this algorithm is also analyzed and compared to verify the segmentation accuracy and feasibility of this algorithm.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：R816.5;TP391.41

【参考文献】