一种基于序列知识的左心室分割算法  

A NOVEL LEFT VENTRICULAR SEGMENTATION ALGORITHM BASED ON SEQUENCE KNOWLEDGE

作　　者：王豫峰[1,3] 徐春雨[2] WANG Yu-feng;XU Chun-yu(School of Software y Nanyang Institute of Technology,Nanyang 473000,China;School of Computer and Information Engineering,Nanyang Institute of Technology,Nanyang 473000,China;School of Computer Science,Wuhan University,Wuhan 430012,China)

机构地区：[1]南阳理工学院软件学院,河南南阳473000 [2]南阳理工学院计算机与信息工程学院,河南南阳473000 [3]武汉大学计算机学院,湖北武汉430072

出　　处：《南阳理工学院学报》2019年第6期38-41,共4页Journal of Nanyang Institute of Technology

基　　金：国家自然科学基金(61672024);河南省科技厅重点科研项目(192102310477);南阳市基础前沿项目(JCQY2018012)。

摘　　要：由于心脏核磁共振图像容易出现图像偏移场,并且对于单张图像的分割缺少相邻像素间的空间信息,从而导致传统基于高斯混合模型的左心室分割算法在成分聚类过程中容易出现错判情况。本文提出一种基于序列知识的左心室分割算法,该算法通过学习短轴心脏核磁共振序列图像的序列知识,将图像关键点元素的序列空间信息引入到自适应混合高斯模型中,从而提高该算法在左心室分割过程中的准确性。通过公共测试图像数据库验证了该分割算法的有效性。Due to the image offset field and the lack of spatial information between adjacent pixels in cardiac magnetic resonance images,the traditional left ventricular segmentation algorithm based on Gauss mixture model is prone to misjudgement in the process of component clustering.LVS-SK learns the sequence knowledge of short-axis cardiac magnetic resonance sequence images,and introduces the sequence spatial information of key point elements into the adaptive Mixture Gauss model,so as to improve the accuracy of the algorithm in the process of left ventricular segmentation.The validity of the segmentation algorithm is verified by the common test image database.

关 键 词：左心室分割 混合高斯模型 序列知识 最大估计算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]