全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

基于皮尔逊系统的语音信号盲源分离

DOI: 10.3969/j.issn.1005-3026.2015.01.002, PP. 6-9

Keywords: 盲源分离,自然梯度,语音信号,皮尔逊系统,激活函数

Full-Text   Cite this paper   Add to My Lib

Abstract:

当利用传统自然梯度算法对所有语音信号都使用同一个激活函数进行分离时,对语音信号的盲源分离效果都不尽理想.针对这一问题,采用基于皮尔逊系统的分段激活函数对传统自然梯度算法进行改进.通过引入皮尔逊系统,将皮尔逊函数与传统激活函数相结合,再利用信号的矩估计方法,分段选择合适的激活函数代入分离矩阵,有效克服了传统语音分离算法的缺点和不足.仿真结果表明,在对实际的语音信号进行分离时,改进算法的性能明显优于传统自然梯度算法,并且在保持了良好收敛速度的同时大大减少了均方误差.

References

[1]  Choi C H,Chang W,Lee S Y.Blind source separation of speech and music signals using harmonic frequency dependent independent vector analysis[J].Electronics Letters,2012,48(2):124-125.
[2]  Liang Y,Chen G,Naqvi S M R,et al.Independent vector analysis with multivariate student''s t-distribution source prior for speech separation [J].Electronics Letters,2013,49(16):1035-1036.
[3]  Hesse C W,James C J.On semi-blind source separation using spatial constraints with applications in EEG analysis[J].IEEE Transactions on Biomedical Engineering,2006,53(12):2525-2534.
[4]  Karvanen J,Koivunen V.Blind separation methods based on Pearson system and its extensions[C]// Proceedings of the Second International Workshop on Independent Component Analysis and Blind Signal Separation.Espoo,2002:663-673.
[5]  Amari S.Natural gradient works efficiently in learning[J].Neural Computation,1998,10(2):251-276.
[6]  Karvanen J,Eriksson J,Koivunen V.Pearson system based method for blind separation[C]// Proceedings of the Second International Workshop on Independent Component Analysis and Blind Signal Separation.Espoo,2000:585-590.
[7]  季策,胡祥楠,朱丽春,等.改进的高阶收敛FastICA算法[J].东北大学学报:自然科学版,2011,32(10):1390-1393.(Ji Ce,Hu Xiang-nan,Zhu Li-chun,et al.Improved higher order convergent FastICA algorithm[J].Journal of Northeastern University:Natural Science,2011,32(10):1390-1393.)
[8]  程舒慧.动态盲源分离及其在生物医学信号处理中的应用研究[D].合肥:安徽大学,2011.(Cheng Shu-hui.Research of dynamic blind source separation and its application to biomedical signal processing[D].Hefei:Anhui University,2011.)

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133