Blind Speech Separation Using Parafac Analysis and Integer Least Squares

Author

Mokios, Kleanthis N. ; Sidiropoulos, Nicholas D. ; Potamianos, Alexandros

Author_Institution

Dept. of Electron. & Comput. Eng., Crete Tech. Univ., Chania, Crete

Volume

5

fYear

2006

fDate

14-19 May 2006

Abstract

We propose a new two-step frequency domain algorithm for blind speech separation (BSS) for unknown channel order. This new approach employs parallel factor analysis (PARAFAC) to separate the speech signals and a novel integer-least-squares-based method for matching the arbitrary permutations in the frequency domain. The proposed algorithm offers guaranteed convergence and good separation performance, measured both quantitatively and in subjective tests. Performance gains in signal-to-interference ratio of up to 10 db are achieved for certain source-sensor geometries

Keywords

blind source separation; frequency-domain analysis; interference (signal); least squares approximations; matrix algebra; speech processing; PARAFAC analysis; arbitrary permutation matching; blind speech separation; integer-least-squares-based method; parallel factor analysis; signal-to-interference ratio; source-sensor geometries; two-step frequency domain algorithm; Convergence; Convolution; Frequency domain analysis; Geometry; Least squares methods; Performance gain; Signal analysis; Speech analysis; Speech processing; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 2006. ICASSP 2006 Proceedings. 2006 IEEE International Conference on

Conference_Location

Toulouse

ISSN

1520-6149

Print_ISBN

1-4244-0469-X

Type

conf

DOI

10.1109/ICASSP.2006.1661215

Filename

1661215