摘要
本文利用高阶谱分析方法，提出以非高斯AR模型对心音信号进行参数化双谱估计。采用双谱互相关系数估计心音信号的非高斯AR模型阶次。另外，根据模型参数构成特征向量，对正常心音和病理性心音进行二类模式分类．通过对实际心音信号的分析，结果表明利用高阶谱技术分析与处理心音信号是一种新的有效和实用的方法。
In the paper, non-Gaussian parametric model is proposed for parametric bispectral estimation of heart sound signals based on higher-order spectra. The approach for determining non-Gaussian AR model order of the normal heart sound signal by employing bispectral cross correlation is investigated. Moreover, the characteristic vector by using the parameters of the AR model are obtained for the pattern classification of two class of heart sound signals. The results show that the analysis and processing of heart sound signals by applying higher-order speetral technique are both of applicable and effective
出处
《信号处理》
CSCD
1997年第3期250-255,共6页
Journal of Signal Processing
