基于Cascade组合分类器的医学图像分类方法研究  

Research on medical image classification based on Cascade combined classifiers

作　　者：张春芬[1] 朱玉全[1] 陈耿[2] 王敏[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013 [2]南京审计学院,南京270029

出　　处：《计算机工程与应用》2007年第36期211-213,共3页Computer Engineering and Applications

基　　金：国家自然科学基金(the National Natural Science Foundation of China under Grant No.60572112)

摘　　要：提出了利用Cascade组合方法生成基于贝叶斯、神经网络与决策树的组合分类器,并将之应用到肝脏图像的分类中。实验结果表明,与现有医学图像分类方法相比,该组合方法可以有效地提高医学图像分类的准确性和稳定性。Based on Cascade combination algorithm,two combined classifiers constructed by naive bayes,BP neural network and decision tree are developed and applied to lung images classification.Compared with existing medical image classifications,experiment results indicate that this method can obviously improve the accuracy and stability of medical image classification.

关 键 词：多分类器组合 朴素贝叶斯 神经网络 决策树 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]