心音在时频两域中解析方法的研究  

Study on a New Method for Heart Sound Analysis in Time and Frequency Domains

作　　者：胡玉良[1] 王海滨[1] 陈健[1] 魏秀波[2] 

机构地区：[1]西华大学电气信息学院四川省信号与信息处理重点实验室,四川成都610039 [2]解放军第452医院心血管疾病中心,四川成都610039

出　　处：《西华大学学报（自然科学版）》2009年第5期5-8,26,共5页Journal of Xihua University:Natural Science Edition

基　　金：四川省教育厅自然基金重点项目(09209025);西华大学人才引进项目(R0620907)

摘　　要：为实现心音的快速准确解析,对心音信号分别在时域和频域中的解析方法进行了研究。时域里,建立单自由度模型(SDOF)以仿真人耳听音鼓膜的特性,并从该模型中提取心音特征波形(CSCW),并从CSCW中提取4个有效的特征诊断参数T1、T2、T11、T12,它们是心音特征波形与阈值THV之间形成的交叉点所确定的时间间隔。频域里,基于自回归-功率谱密度(NAR-PSD)曲线定义了新的形态特征,提取一些心脏杂音可识别参数指标,提出了两个特征诊断参数fm ax和fwidth,分别描述了NAR-PSD最大频率峰值和频带宽度。通过采集大量的临床心音数据对正常/异常的心音案例进行解析。实验结果表明,在时域和频域中,这些特征诊断参数有助于判别心音的正常与异常。This paper presents a new method for heart sound analysis in time and frequency domains in order to investigate heart sound effectively and quickly. In time domain, an analytical model of single degree-of-freedom（SDOF） is established to simulate the tympanic membrane＇s characteristics, and cardiac sound characteristic waveform method（CSCW） is extracted out. Four characteristic parameters T1, T2, T11, T12 are defined by the time intervals between the crossed points of the CSCW and a threshold value （THV）. In frequency domain , two diagnostic parameters fmax and fwdth are proposed based on new morphological characteristics , which are defined on NAR-PSD curve to discriminate normal and abnormal heart sounds . Two diagnostic parameters are appointed to describe the maximum peak of NAR-PSD curve and frequency width respectively between the crossed points of NAR-PSD curve based on a selected THV. Furthermore, large clinical cardiac sounds are used to validate the efficiency of this proposed methods , the results show that these parameters are useful for detect normal/abnormal heart sound in time and frequency domains.

关 键 词：心音 单自由度模型(SDOF) 心音特征波形(CSCW) 自回归-功率谱密度(NAR-PSD) 

分 类 号：TN911.6[电子电信—通信与信息系统]