Convolutive blind source separation by fourth-order statistics

Author

Wang, Fuxiang ; Zhang, Jun

Author_Institution

Sch. of Electron. & Inf. Eng., Beihang Univ., Beijing, China

fYear

2009

fDate

8-10 Dec. 2009

Firstpage

1

Lastpage

4

Abstract

In this paper, we present a new method using fourth-order statistics for convolutive blind source separation. After a whitening process of the observed data, the separation of convolutive mixtures are transformed to find a semi-unitary demixing matrix. Then the semi-unitary demixing matrix is determined by joint diagonalization for a set of fourth-order joint cumulant matrices. Simulation results of speech separation demonstrate the effectiveness of the new approach.

Keywords

blind source separation; convolution; matrix algebra; statistical analysis; convolutive blind source separation; convolutive mixtures; fourth-order joint cumulant matrices; fourth-order statistics; semi-unitary demixing matrix; Blind source separation; Delay; Finite impulse response filter; Frequency estimation; IIR filters; Mutual information; Source separation; Speech; Statistics; Time domain analysis; Blind source separation; fourth-order statistics; joint diagonalization;

fLanguage

English

Publisher

ieee

Conference_Titel

Information, Communications and Signal Processing, 2009. ICICS 2009. 7th International Conference on

Conference_Location

Macau

Print_ISBN

978-1-4244-4656-8

Electronic_ISBN

978-1-4244-4657-5

Type

conf

DOI

10.1109/ICICS.2009.5397743

Filename

5397743