DocumentCode :
3410717
Title :
Blind Speech Separation Employing Laplacian Normal Mixture Distribution Model
Author :
Cai, Hua ; Sun, Junxi ; Ou, Shifeng
Author_Institution :
Changchun Univ. of Sci. & Technol., Jilin
fYear :
2007
fDate :
5-8 Aug. 2007
Firstpage :
3185
Lastpage :
3189
Abstract :
Careful choice of nonlinear function is necessary to obtain good performance from algorithms for blind source separation. In this paper, we propose a fast approach to perform blind speech separation based on natural gradient. The main ingredient is the use of a novel nonlinear function, which is accordant to the true PDF of speech signals. By appropriately choosing the shape parameter, we approximate a Laplacian normal mixture distribution to the source´s PDF in time domain, then a new form of nonlinear function more suitable for speech separation is derived using the given distribution model. Simulation results indicate the good convergence and steady-state performance of our proposed method.
Keywords :
blind source separation; gradient methods; nonlinear functions; normal distribution; speech processing; Laplacian normal mixture distribution model; blind source separation; blind speech separation; convergence; natural gradient; nonlinear function; probability density function; steady-state performance; Automation; Blind source separation; Convergence; Laplace equations; Mechatronics; Shape; Signal processing; Source separation; Speech; Steady-state; Laplacian normal mixture distribution; blind source separation; natural gradient algorithm; speech source;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Mechatronics and Automation, 2007. ICMA 2007. International Conference on
Conference_Location :
Harbin
Print_ISBN :
978-1-4244-0828-3
Electronic_ISBN :
978-1-4244-0828-3
Type :
conf
DOI :
10.1109/ICMA.2007.4304071
Filename :
4304071
Link To Document :
بازگشت