生物医学信号模式的状态评估法分析方法研究

The study on analysis method of state representation methodology of biomedical signal pattern

目的探索一种生物医学信号模式(BSP)系统状态评估(SRM)的新方法。方法基于心音信号、心电信号、呼吸信号,针对BSP描述的重要研究问题,在比较支持向量机(SVM)、响应面法(RSM)等几种传统方法之后,采用频率切片小波变换(FSWT)的方法提取BSP信号的动态阻尼特征,从而提出了一种SRM分析的新思路。以心音信号分析为例,给出了SRM评估方法的一般步骤。结果以40例正常心音病例建立SRM模型,以80例异常心音病例进行SRM状态比较,发现两组人群存在明显的状态分布差异。结论将SRM与FSWT相结合可以为BSP分析提供一种新方法,为BSP分析提供强有力的开发工具。

Objective To explore a kind of biomedical signal pattern （BSP） with a new method called as state representation methodology （SRM ） .Methods Based on the heart sound signals ,ECG signals ,breathing ,as the important research problem for BSP description ,with some comparisons on several traditional methods ,in which support vector machines （SVM ） and response sur-face methodology （RSM ） etc .,using frequency slice wavelet transform （FSWT ） method to extract the BSP signal dynamic damping characteristics ,thus ,this paper proposes a new idea of SRM analysis .In the case of heart sound signal analysis ,the general steps of SRM evaluation method is given .Results In 40 cases of normal heart sounds SRM model is set up ,with 80 cases of abnormal heart sounds are compared ,the obvious differences of the SRM state distributions of the two groups are found .Conclusion The combi-nation of SRM with FSWT can provide a novel approach for BSP analysis ,and provide powerful development tool for the analysis of BSP .