EUS图像胰腺癌分类中粒子光斑的去除  

Light Spots'Removal in Classification of Pancreatic Cancer of EUS

作　　者：吴仪俊[1] 刘岩[2] 余建国[1] 金震东[2] 

机构地区：[1]复旦大学电子工程系,上海200433 [2]长海医院消化科,上海200433

出　　处：《生物医学工程学进展》2010年第3期135-137,共3页Progress in Biomedical Engineering

基　　金：上海市重点学科建设资助项目(No.B112)

摘　　要：已有基于图像纹理特征对胰腺超声内镜图像进行是否为癌症的分类方法。放射性粒子植入治疗是一种较常用的姑息治疗方法。治疗后的图像中由植入粒子所产生的光斑对分类的结果有一定的影响。通过使用改进的中值滤波方法去除光斑后分类结果的准确性有了一定的提高。There is a method to judge whether an EUS image stands for cancer or not depend on the texture features. The treatment of implanting radioactive particles is an ordinary way to treat cancer. The light spot in the after -treating images caused by the particles implanted have a considerable effect on the result of the classification. After removing the light spots by the way of modified median filtering the accuracy of the classification result is improved.

关 键 词：内镜超声 光斑 分类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]