中国医疗设备2016，Vol.31Issue(7)：41-44,4.DOI:10.3969/j.issn.1674-1633.2016.07.012
基于变差函数的肝癌超声图像纹理识别的研究
Study on Methods of Differentiating Liver Cancers from Normal Liver by Ultrasound Image Texture Features Based on Variogram
李秀丽 1孟晓涛1
作者信息
- 1. 包钢集团第三职工医院超声科，内蒙古包头014010
- 折叠
摘要
Abstract
Application of the variogram for characterizing ultrasound image texture features was proposed in this paper to differentiate abnormal liver from normal liver. And the coefficient of anisotropy was also deifned to further characterize ultrasound image texture features. Four abnormal livers and ifve liver cancers were involved in this study using this method to calculate the r(h) max value and analyze each image from eight directions. The results showed that r(h) max value of normal livers was smaller and r(h) max value of liver cancers was bigger; while, the coefficient of anisotropy for normal livers was more and that of liver cancers was smaller. Hence, both the parameters r(h) max and coefifcient of anisotropy can be used to differentiate liver cancers from normal livers.关键词
超声图像/纹理特征/变差函数/各向异性系数Key words
ultrasound images/texture features/variogram/anisotropy coefifcient分类
临床医学引用本文复制引用
李秀丽,孟晓涛..基于变差函数的肝癌超声图像纹理识别的研究[J].中国医疗设备,2016,31(7):41-44,4.
