基于神经网络集成的舌苔分类方法  

A Method for Tongue Coat Classification Based on Neural Network Integration

作　　者：刘关松[1] 徐建国[2] 高敦岳[1] 

机构地区：[1]华东理工大学信息科学与工程学院,上海200237 [2]上海中医药大学,上海200032

出　　处：《计算机工程》2003年第14期100-102,共3页Computer Engineering

基　　金：上海市高等学校科学技术发展基金项目(2000C10)

摘　　要：提出一种基于神经网络集成的舌苔自动分类方法。该方法把经单独训练的具有一定差异度的单个BP神经网络加以集成，构成舌苔分类器。试验结果表明，该分类器比现有的舌苔分类方法识别率更高、分类更细、更符合中医舌诊要求。This paper presents a method for tongue coat automatic classification based on the neural network integration. In the method, those single BP neural networks which have been trained solely are integrated averagely in order to build up the tongue coat classifier. Experimental results show that the classifier had better performance than the others proposed previously did, for instance, it can classify tongue coats more detailedly and more accurately. So, it meets better the needs of the traditional Chinese medicine tongue diagnosis.

关 键 词：BP神经网络 神经网络集成 舌苔 分类 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]