Title :
A consideration of blind source separation using wavelet transform
Author :
Hirai, Noriyuki ; Matsumoto, Hiroki ; Furukawa, Toshihiro ; Furuya, Kiyoshi
Author_Institution :
Dept. of information engineenng, Chuo Univ, Tokyo, Japan
Abstract :
A method of blind source separation in the environment of convolutive mixtures is proposed. To transform observed signals into time-frequency domain, the wavelet transform was used instead of the conventional short-time Fourier transform to obtain appropriate information related both with time and frequency. The improvement of the separation performance is based on a new methodology to estimate separation matrices. Results of computer simulation are shown for artificially controlled data.
Keywords :
blind source separation; convolution; matrix algebra; parameter estimation; time-frequency analysis; wavelet transforms; blind source separation; convolutive mixtures; separation matrix estimation; separation performance; time-frequency domain; wavelet transform; Blind source separation; Delay effects; Environmental management; Fourier transforms; Frequency domain analysis; Frequency estimation; Independent component analysis; Time frequency analysis; Wavelet domain; Wavelet transforms;
Conference_Titel :
Intelligent Signal Processing and Communication Systems, 2004. ISPACS 2004. Proceedings of 2004 International Symposium on
Print_ISBN :
0-7803-8639-6
DOI :
10.1109/ISPACS.2004.1439032
