基于非线性动力学和GMM的病态嗓音识别与研究

广西师范大学学报（自然科学版） ›› 2011, Vol. 29 ›› Issue (3): 5-8.

基于非线性动力学和GMM的病态嗓音识别与研究

高俊芬, 胡维平

广西师范大学电子工程学院,广西桂林541004

收稿日期:2011-06-04 出版日期:2011-08-20 发布日期:2018-12-03
通讯作者: 胡维平(1963—),男,广西桂林人,广西师范大学教授,博士。E-mail:huwp@gxnu.edu.cn
基金资助:
广西自然科学基金资助项目(2010GXNSFA013128)

Recognition and Study of Pathological Voices Based on NonlinearDynamics Using GMM

GAO Jun-fen, HU Wei-ping

College of Electronic Engineering,Guangxi Normal University,Guilin Guangxi 541004,China

Received:2011-06-04 Online:2011-08-20 Published:2018-12-03

摘要/Abstract

摘要： 本文采用非线性动力学的分析方法,定量分析并提取了正常与病态嗓音的5维非线性特征:Hurst参数、香农熵、计盒维数、Kolmogorov熵和关联维数;使用来源于临床病例的151例数据,选用高斯混合模型GMM(gaussian mixture model)的模式识别方法,来评估基于非线性动力学分析方法所提取的特征参数的有效性。实验结果表明,非线性动力学的分析方法能够弥补传统分析方法的不足,较好分析正常与病态嗓音,取得96.05%的较好识别率。

关键词: 非线性动力学, GMM, 混沌理论, 病态嗓音

Abstract: The method of nonlinear dynamics analysis is used to quantitatively analyze and extractthe normal and pathological voice of the 5-dimensional nonlinear feature,Hurstparameter,Shannon entropy,box dimension,Kolmogorov entropy and correlation dimension.The data is from the clinical cases of 151 patients and the pattern recognition method of Gaussian mixture model is used to evaluate the validity of the parameters extracted by the method of nonlinear dynamics.Experimental results showthat this method can compensate for the deficiencyof traditional methods,and achievea better recognition rate of 96.05%.

Key words: nonlinear dynamics, GMM, chaos theory, pathological voices

中图分类号:

TP181

高俊芬, 胡维平. 基于非线性动力学和GMM的病态嗓音识别与研究[J]. 广西师范大学学报（自然科学版）, 2011, 29(3): 5-8.

GAO Jun-fen, HU Wei-ping. Recognition and Study of Pathological Voices Based on NonlinearDynamics Using GMM[J]. Journal of Guangxi Normal University(Natural Science Edition), 2011, 29(3): 5-8.

参考文献

[1] THOMPSON C,MULPUR A,MEHTA V.Trandition to chaos in acoustically driven flow (acoustic streaming)[J].Acoust Soc Am,1991,90:2097-2103.
[2] JACK J J,ZHANG Y,McGILLIGAN C.Chaos in voice,from modeling to measurement[J].Journal of Voice,2006,20(1):1-15.
[3] 于燕平,胡维平.病态嗓音特征的小波变换提取及识别研究[J].计算机工程与应用,2009,45(22):194-196.
[4] 姚赛芬.基于MATLAB自相似序列Hurst参数检测[J].信息与电脑,2010,4(1):177.
[5] VAZIRI G,ALMASGANJ F,JENABI M S.On the fractal self-similarity of laryngeal pathologies detection:the estimation of Hurst parameter[J].Electronics Letters,2005,41(16):383-386.
[6] 韦岗,陆以勤,欧阳景正.混沌、分形理论与语音信号处理[J].电子学报,1996,24(1):34-38.
[7] HENRIQUE P,ALONSO J B,FERRER M A,et al.Characterization of healthy and pathological voice through measures based on nonlinear dynamics[J].IEEE Transaction on Audio,Speech and Language Processing,2009,17(6):1186-1188.
[8] 赵贵兵,石炎福,段文锋,等.从混沌时间序列同时计算关联维和Kolmogorov熵[J].计算物理,1999,16(3):309-312.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed