医学超声图像的特征提取方法初探  

A Study on the Feature Extraction Algorithm of Medical Ultrasound Image

作　　者：陈刚[1] 冯登超[1] 吴新颖[1] 杜文雅[2] 肖楷乐[1] 

机构地区：[1]北华航天工业学院电子工程系,河北廊坊065000 [2]太原科技大学电子信息学院,山西太原030024

出　　处：《北华航天工业学院学报》2012年第4期1-3,共3页Journal of North China Institute of Aerospace Engineering

基　　金：河北省自然科学基金项目(F2011409008);河北省教育厅青年基金项目(2010208);北华航天工业学院博士基金项目(KY-2008-02-B)

摘　　要：本文在分析了医学超声图像特殊性的基础上,构建了基于灰度共生矩阵的四个纹理特征描述算子,对正常肝和非正常肝的超声图像进行了特征提取。MATLAB仿真实验结果验证了该算法的有效性。In the paper, the particularity of medical ultrasound image was analyzed and four texture feature operators were constructed based on gray leve co-occurrence matrix. Accordingly, the ultrasound image feature extraction of normal and non-normal liver were executed. Finally, the simulation experiment under the envirnoment of Matlab was carried on and verified the validity of the above algorithm.

关 键 词：灰度共生矩阵 超声图像 纹理 特征提取 

分 类 号：TB559[理学—物理]