摘要
提出了一种实时高效的心电图分类理论与方法。首先对心电图进行六尺度小波分解,将含有主要噪声的尺度进行系数置零,再将剩余层进行小波重构,从而达到除噪的目的。利用数学形态学定位心电图P、Q、R、S、T波位置,并提取计算各波间距离和斜率等12个特征值作为概率神经网络的输入向量,从而实现心电图的六分类。
An effective electrocardiogram（ECG） classification approach based on wavelet de-noising and probabilistic neural net- is proposed. Firstly the ECG signal is taken six scaling wavelet decomposition.The coefficient of the scales, which contain noise, are set to zero, inverse the wavelet transform. The mathematical morphology is used to find the positions of P,Q,R,S T wave, then the 12 feature values are taken as the input of probability neural network to realize the six classification of ECG.
出处
《电子技术应用》
北大核心
2013年第3期136-137,140,共3页
Application of Electronic Technique
关键词
心电图
分类
小波变换
概率神经网络
eleetroeardiogram（ECG）
classification
wavelet transform
probabilistic neural network
