心音的降噪预处理和基于HHT的特征分析研究  

Research on Noise Reduction of Heart Sound and Features Analysis Based on HHT

作　　者：尹明 王燕[2] 刘欣 

机构地区：[1]解放军95538部队 [2]索贝数码科技股份有限公司

出　　处：《电声技术》2013年第6期58-63,共6页Audio Engineering

摘　　要：为了准确地提取心音信号的病理特征,在研究小波阈值收缩降噪法的基础上提出改进的阈值函数以弥补传统软、硬阈值函数的不足,将其应用于临床采集的心音信号降噪中获得了很好的降噪效果,有助于特征参数的准确提取和分析。在深入研究HHT的基础上,对降噪后的心音信号利用EMD分解、HT变换及归一化香农能量定理等从多个角度提取心音的特征值,通过对采集的多组正常和异常心音数据进行分析,验证了该方法提取特征的有效性,具有较强的临床实用价值。In order to extract the pathological features of heart sound （HS） accurately, based on the wavelet threshold shrinkage denoising method, an improved threshold function is proposed to compensate for the deficiencies of the traditional soft and hard threshold functions, which is applied to reduce the noise of HS, and help to extract and analyze the fea- tures. Based on the study of HHT, the Empirical Mode Decomposition （EMD）, Hilbert Transform （HT） and Normalized Shannon energy theorem are applied to extract the features. Many cases of normal and abnormal HS are collected and analyzed, and the experimental results show that the feature extraction method based on HHT is effective with a strong clinical practical value.

关 键 词：心音 降噪 阈值函数 HHT 香农能量 特征提取 

分 类 号：O428[理学—声学]