中文摘要：

目的：本研究旨在应用现代声诊技术辅助病证的临床诊断，为中医声诊脏腑辨证理论提供一定的客观依据。方法：应用“中医闻诊采集系统”软件，采集肝郁脾虚、心脾两虚、心肾不交证型患者（231例）的语音信号，同时以正常人（100例）语音信号作为对照组，以小波包分解重构技术为基础，提取能量比例、扩展能量比例、能量梯度、扩展能量梯度4类特征参数。结果：肝郁脾虚、心脾两虚、心肾不交证型患者与正常4组样本的能量比例、扩展能量比例、能量梯度、扩展能量梯度特征比较，较多频段有明显差异（P〈0.05）；各个证型样本之间的4组特征较多频段也存在着明显差异（P〈0.05）；心脾两虚证型扩展能量比例特征倾向于高于其他证型患者及正常人；正常人扩展能量梯度特征倾向于低于3种证型，差异性主要集中在中部频段之后；特征筛选后进行3证与正常人的分类识别率较为理想。结论：运用现代声诊技术分析中医常见证型语音特征参数，研究结果为中医声诊的客观化研究提供参考，为实现中医声诊的客观检测和计算机智能化提供理论依据。

英文摘要：

This study was aimed to assist the clinic diagnosis and analysis with the application of modern ausculta-tion diagnosis technology. Meanwhile, this study provided certain objective evidences for traditional Chinese medicine （TCM） auscultation diagnosis in syndrome differentiation according to Zang-Fu. The ＂Voice Sampling System of TCM＂ was used in the collection of voice signals of patients with liver-depression and spleen-deficiency, deficiency of both heart and spleen, disharmony between heart and kidney （231 cases）. Meanwhile, voice signals of healthy peo-ple （100 cases） were used as the control group. Based on wavelet packet decomposition and reconstruction technique, four kinds of featured parameters, including energy proportion, generalized energy proportion, energy gradient and generalized energy gradient, were extracted. The results showed that there were significant differences in multiple fre-quencies on energy proportion, generalized energy proportion, energy gradient and generalized energy gradient be-tween the liver-depression and spleen-deficiency, deficiency of both heart and spleen, disharmony between heart and kidney, and healthy people （P〈 0.05）. There were significant differences in multiple frequencies of four features a-mong samples from each type （P 〈 0.05）. The generalized energy proportion of the deficiency of both heart and spleen type was higher than patients from other types and healthy people. The generalized energy gradient of healthy people was lower than the other three types. Differences were mainly posterior to the middle frequency segment. The recognition rate of three types and healthy people was ideal after feature screening. It was concluded that modern auscultation diagnosis technology can be used in the analysis of TCM phonetic feature parameters among common TCM syndrome types. The research results provided references for the objective study of TCM auscultation diagnosis. It provided theoretical evidences for the objective detection of TCM auscul