[1]陈立伟,王文姝,袁頔.自适应高斯混合模型语音增强方法[J].应用科技,2009,36(07):11-15.[doi:doi:10.3969/j.issn.1009-671X.2009.07.004]
 CHEN Li-wei,WANG Wen-shu,YUAN Di.A speech enhancement method based on adaptive Gaussian mixture model[J].Applied science and technology,2009,36(07):11-15.[doi:doi:10.3969/j.issn.1009-671X.2009.07.004]
点击复制

自适应高斯混合模型语音增强方法
分享到:

《应用科技》[ISSN:1009-671X/CN:23-1191/U]

卷:
第36卷
期数:
2009年07期
页码:
11-15
栏目:
现代电子技术
出版日期:
2009-07-05

文章信息/Info

Title:
A speech enhancement method based on adaptive Gaussian mixture model
文章编号:
1009-671X(2009)07-0011-05
作者:
陈立伟王文姝袁頔
(哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001)
Author(s):
CHEN Li-wei WANG Wen-shu YUAN Di
(College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China)
关键词:
语音增强小波变换自适应高斯混合模型贝叶斯萎缩函数
Keywords:
speech enhancement wavelet transform adaptive Gaussian mixture model Bayesian shrinkage function
分类号:
TN911.7
DOI:
doi:10.3969/j.issn.1009-671X.2009.07.004
文献标志码:
A
摘要:
语音增强是解决噪声污染的有效方法,它的首要目标是在接收端尽可能从带噪语音中恢复纯净的语音信号.针对噪声环境下的语音增强问题,提出了一种语音增强新方法.该方法利用小波子带的方向性特点以及小波系数尺度内的相关性,将小波系数的概率分布建模为一种自适应高斯混合模型,在贝叶斯框架中采用这种概率模型可以得到一种具有空间自适应性的贝叶斯萎缩函数.利用这种萎缩函数可以实现对小波系数的修正.仿真实验表明,该算法对于噪声有较好的抑制作用, 该算法在主观和客观测试中都具有良好的语音增强效果,可以在语音识别、语音编码中获得应用.
Abstract:
Speech enhancement is an efficient method in removing noises. The aim of speech enhancement is resuming the pure speech signal from the speech signals containing noise to a great extent. In order to solve the problem of speech enhancement in noise environment, a new speech enhancement method was proposed. This method models the distribution of wavelet coefficients as an adaptive Gaussian mixture mode1. This model takes into account intra scale dependencies between wavelet coefficients. Based on this model in a Bayesian framework, a spatially adaptive Bayesian shrinkage function was obtained and each modified coefficient was decided separately. Simulation results show this algorithm is effective in reducing the noise; this algorithm possesses good performance both in objective and subjective tests, so it can be used in speech recognition and speech coding.

参考文献/References:

[1]SUN Fuxong, XIANG Jidong. Algorithm of lifting wavelet package in realtime fault diagnosis system [C]//International Conference on Machine Learning and Cybernetics. Xi’an, China,2003:340-344.
[2]JANSEN M.Wavelet thresholding and noise reduction [D].Leuven: Katholieke Universiteit Leuven, 2000.
[3]CHIPMAN H A,KOLACZYK E D,McCulloch R E.Adaptive Bayesian wavelet shrinkage[J].J Amer Statist AsSOC,1997,92(440):1413-1421.
[4]ACHIM A,KOLACZYK E E.Image denoising using bivariate —stable distributions in the complex wayelet domain[J].IEEE Signal Processing Letters,2005,12(1):17-20.
[5]CROUSE M S, NOWAK R D, BARANIUK R G. Wavelet based statistical on Signal processing using hidden markov models[J]. IEEE Transaction on Signal Processing, 1998, 46(4): 886-902.
[6]吴叶军,袁希亮, 金克武.基于平稳小波变换的X射线衍射信号消噪研究[J]. 现代制造工程, 2005(5):134-136.
[7]LOMBARDI M,GODSILL S J. Online Bayesian estimation of AR signals in symmetric alphastable noise[R]. Boston: Working Paper, 2004.

相似文献/References:

[1]王鹏飞,张兴周,胡文飞.基于小波变换的火车车轮扁疤信号能量分析[J].应用科技,2009,36(06):25.
 WANG Peng-fei,ZHANG Xing zhou,HU Wen-fei.he energy analysis of wheelflat signal based on wavelet transform[J].Applied science and technology,2009,36(07):25.
[2]徐伟,张帅,王克家.拉曼光谱预处理中几种小波去噪方法的分析[J].应用科技,2009,36(11):27.[doi:10.3969/j.issn.1009-671X.2009.11.007]
 XU Wei,ZHANG Shuai,WANG Ke-jia.Denoising of raman spectra based on wavelet transform[J].Applied science and technology,2009,36(07):27.[doi:10.3969/j.issn.1009-671X.2009.11.007]
[3]席志红,金甲,肖易寒.一种改进的基于小波偏微分方程的图像去噪方法[J].应用科技,2010,37(01):12.[doi:3969/j.issn.1009-671X.2010.01.007]
 XI Zhi-hong,JIN Jia,XIAO Yi-han.An improved algorithm for image noise removal based on wavelet transform and partial differential equations[J].Applied science and technology,2010,37(07):12.[doi:3969/j.issn.1009-671X.2010.01.007]
[4]席志红,郭亮,肖易寒.一种基于Contourlet变换的图像边缘检测方法[J].应用科技,2010,37(04):35.[doi:10.3969/j.issn.1009-671X.2010.04.009]
 XI Zhi-hong,GUO Liang,XIAO Yi-han.An image edge detection scheme based on Contourlet transform[J].Applied science and technology,2010,37(07):35.[doi:10.3969/j.issn.1009-671X.2010.04.009]
[5]佘科,谢红.基于WNN的双目摄像机标定方法[J].应用科技,2010,37(11):35.[doi:10.3969/j.issn.1009-671X.2010.11.009]
 SHE Ke,XIE Hong.Camera calibration of binocular vision system based on wavelet neural network[J].Applied science and technology,2010,37(07):35.[doi:10.3969/j.issn.1009-671X.2010.11.009]
[6]孙明珠,盖强.基于双阈值小波变换在雷达测试数据处理中的应用[J].应用科技,2011,38(04):57.[doi:doi:10.3969/j.issn.1009-671X.2011.04.13]
 SUN Mingzhu,GAI Qiang.The application of dual threshold wavelet transform in data processing of radar test data[J].Applied science and technology,2011,38(07):57.[doi:doi:10.3969/j.issn.1009-671X.2011.04.13]
[7]钟山,孙凤,丁福光,等.基于中值滤波的全垫升气垫船运动仿真参数滤波方法研究[J].应用科技,2005,32(01):31.
 ZHONG Shan,SUN Feng,DING Fuguang,et al.Air cushion vehicle signal de—noising methods based on Daubechies wavelet and median filter[J].Applied science and technology,2005,32(07):31.
[8]李 佶,付永庆,王咏胜.一种基于模糊聚类的小波图像压缩方法[J].应用科技,2005,32(03):3.
 LI Ji,FU Yong-qing,WANG Yong-sheng.Wavelet image compression based on potential fuzzy clustering[J].Applied science and technology,2005,32(07):3.
[9]赵春晖,王克成,陈万海.一种基于小波变换的超光谱图像融合方法[J].应用科技,2005,32(09):1.
[10]陈立伟,赵春晖,杨洪.一种基于模糊系统的语音增强方法[J].应用科技,2005,32(10):13.
 CHEN Li_wei,ZHAO Chun_hui,YANG Hong_li.Research on speech enhancement method based on fuzzy system[J].Applied science and technology,2005,32(07):13.

备注/Memo

备注/Memo:
基金项目:国家自然科学基金资助项目(607702053).
作者简介:陈立伟(1974-),女,副教授,博士,主要研究方向:信号处理、模式识别,E-mail:chenliwei@hrbeu.edu.cn.