| Paper: | SPTM-P9.12 (ICASSP 2003 Paper) | ||
| Session: | Nonlinear Systems and Signal Processing | ||
| Time: | Thursday, May 20, 13:00 - 15:00 | ||
| Presentation: | Poster (ICASSP 2003 Presentation) | ||
| Topic: | Signal Processing Theory and Methods: Sampling, Extrapolation, and Interpolation | ||
| Title: | A PROBABILISTIC APPROACH FOR BLIND SOURCE SEPARATION OF UNDERDETERMINED CONVOLUTIVE MIXTURES | ||
| Authors: | Michael Peterson; University of Southern California | ||
| Shubha Kadambe; HRL Laboratories, LLC | |||
| Abstract: | There are very few techniques that can separate signals from the convolutivemixture in the underdetermined case. We have developed a method that uses overcomplete expansion of the signal created with a time-frequency transformand that also uses the property of sparseness and a Laplacian source densitymodel to obtain the source signals from the instantaneously mixed signalsin the underderdetermined case. This technique has been extended here to separate signals (a) in the case of underdetermined convolutive mixtures, and (b) in the general case of more than 2 mixtures. Here, we also proposea geometric constrained based search approach to significantly reduce the computational time of our original ''dual update'' algorithm. Several examplesare provided. The results of signal separation from the convolutive mixturesindicate that an average signal to noise ratio improvement of 5.3 dB can beobtained. | ||
| Back | |||
Home -||-
Organizing Committee -||-
Technical Committee -||-
Technical Program -||-
Plenaries
Paper Submission -||-
Special Sessions -||-
ITT -||-
Paper Review -||-
Exhibits -||-
Tutorials
Information -||-
Registration -||-
Travel Insurance -||-
Housing -||-
Workshops