首页 | 本学科首页   官方微博 | 高级检索  
     检索      

混合型径向基函数神经网络检测PVC
引用本文:王涛,周荷琴,江朝晖,冯焕清.混合型径向基函数神经网络检测PVC[J].数据采集与处理,2000,15(1):1-5.
作者姓名:王涛  周荷琴  江朝晖  冯焕清
作者单位:中国科学技术大学生物医学工程研究所,合肥,230026
基金项目:国家自然科学基金!(编号 :3 95 70 2 14 )
摘    要:采用径向基函数(RBF)神经网络作为室性早搏(PVC)的检测器,将QRS波的模板以及其他不同物理意义的特征参数综合在一个检测网络中,构成一种由Gauss函数和Sigmoid函数作为隐层点基函数的混合型RBF网络,并且给出确定网络结构参数的方法。用MIT心电数据库对算法进行验证,单层心电信号的室性早搏检出率达到98.46%。

关 键 词:神经网络  检测  室性早搏  径向基函数  心电信号

Detecting PVC Using Hybrid RBF Networks
Wang Tao,Zhou Heqin,Jiang Zhaohui,Feng Huanqing.Detecting PVC Using Hybrid RBF Networks[J].Journal of Data Acquisition & Processing,2000,15(1):1-5.
Authors:Wang Tao  Zhou Heqin  Jiang Zhaohui  Feng Huanqing
Abstract:Detecting premature ventricular contraction (PVC) is one of the main tasks in electrocardiogram (ECG) arrhythmia analysis. This paper presents a PVC detection method based on radial basis function (RBF) network. This paper combines the prototypes of QRS complexes and some other features into a kind of hybrid RBF network with Gauss and Sigmoid basis function in hidden nodes, and presents a method to determine its structural parameters. The differences among patients are considered in constructing the hidden nodes using self organizing clustering algorithm and fuzzy logical inference, so that the supervised training processes for each record are unnecessary. The detection rate reaches 98.46% using single lead signals from MIT arrhythmia database.
Keywords:neural networks  clustering  detection  premature ventricular contraction
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号