摘要
心电信号(ECG)识别是重病特别护理中一个非常重要的课题,自动检测和分类心电节律信号是诊断心脏异常的一项重要任务。基于小波变换理论,小波神经网络已经被广泛的应用于信号的表达和分类。文中介绍了一种具有一层感知机的小波神经网络,对小波神经网络收敛性影响较大的网络初始化值提出了一种改进的初始化算法,并研究得出了隐含层的选取与网络收敛性的关系。应用该网络分类心电信号的正常心跳和室前收缩,取得了很好的效果。文中使用到的心电数据是从MIT-BIH心电失常数据库中下载的。
Recognition of electrocardiogram （ECG） is an important area in biomedical signal processing. Based on the wavelet transform theory, the wavelet neural network has been wildly used for signal representation and classification. A new adaptive wavelet neural network with one perceptron was introduced for ECG signal recognition. The initialization and training approaches were proposed and the relation between the number of the hidden layers and the astringency of the network was found. The network used for distinguishing the normal beat and the premature ventricular contraction and high performance was obtained. In present work, the ECG data was taken from MIT- BIH Arrhythmia database.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2007年第14期3281-3282,3297,共3页
Journal of System Simulation
基金
国家自然科学基金资助项目(50477015)
关键词
心电信号
小波变换
小波神经网络
室前收缩
ECG
wavelet transform
wavelet neural network
premature ventricular contraction
