基于小波分析和概率神经网络的心音诊断研究

Heart Sound Diagnosis Based on Wavelet Analysis and Probabilistic Neural Network

心音对大多数心血管疾病具有极高的临床诊断价值,对心音信号进行分析有助于临床上对心脏疾病的诊断。为了利用计算机智能分析心音信号,提出利用多尺度小波分解消除信号中的噪声,从各频带提取特征值,用概率神经网络(PNN)来进行心音信号的自动分析诊断。用Matlab仿真的方法测试了5种不同类型心音信号的分类情况,结果表明该方法可行。

To take advantage of computer analysis of heart sound signals intelligence,this paper proposes the use of multi-scale wavelet decomposition to eliminate signal noise,and extract features from the band value,and then use probabilistic neural network （PNN） to carry out automatic analysis and diagnosis of heart sound signals.With the Matlab simulation method to test 5 different types of classification of heart sound signals,the results show that the method is feasible.