中文摘要：

目的：本研究运用现代声学技术,采集和分析五脏病变患者的声音信号,为中医声诊的脏腑辨证提供客观依据。方法：运用＂中医闻诊采集系统＂采集五脏病变患者声音样本803例,其中肺系139例,肝系48例,脾系86例,肾系66例,心系464例,另采集100名正常人声音样本作为对照,运用样本熵方法分析各组声音信号,提取与中医五脏分类相关的特征参数。结果：在嵌入维数为2时,各组样本熵特征比较发现：6组样本比较,6个时域频段的样本熵值差异均有统计学意义（P〈0.05）;正常组声音的样本熵特征均低于五脏患者声音的样本熵特征（P〈0.05）;肺系组声音的样本熵特征高于其余5组声音的样本熵特征（P〈0.05）。五脏病变患者声音各时域段的总样本熵值肺系组最高,其次脾系组、心系组、肝系组、肾系组和正常组。结论：根据五脏相音理论,运用现代语音信号采集分析方法,对803例五脏病变患者的声音进行样本熵分析,为中医依据声诊进行五脏分类辨证提供一定的客观参考依据。

英文摘要：

Objective： The purpose of this paper is to provide objective gist for organ differentiation of traditional Chinese medicine （TCM）. Methods： We collected 803 voice samples by ‘Voice Collecting System of TCM＇ , including 139 voice samples of patient with lung illness, 48 voice samples of patient with liver illness, 86 voice samples of patient with spleen illness, 66 voice samples of patient of kidney illness and 464 voice samples of heart illness. 100 voice samples of healthy people were included as normal control. We extracted the sample entropy characteristic of each group when embedded dimension was 2. Results： By analyzing the statistics result, the sample entropy characteristic of six groups were significantly different in multiple frequencies（P〈0.05）; the sample entropy characteristic of normal group was much lower than that of patients（P〈0.05）; the sample entropy characteristic of lung illness group was much higher than that of other illness groups and normal group（P〈0.05）. Conclusion： Using the method of sample entropy combined with the Wuyin theory, we have got some objective gist for organ differentiation of TCM auscultation.