Maximum Likelihood For Blind Separation And Deconvolution Of Noisy Signals Using Mixture Models (1997)
| Citations: | 53 - 3 self |
BibTeX
@INPROCEEDINGS{Moulines97maximumlikelihood,
author = {Éric Moulines and Jean-François Cardoso and Elisabeth Gassiat},
title = {Maximum Likelihood For Blind Separation And Deconvolution Of Noisy Signals Using Mixture Models},
booktitle = {},
year = {1997},
pages = {3617--3620}
}
Years of Citing Articles
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
In this paper, an approximate maximum likelihood method for blind source separation and deconvolution of noisy signal is proposed. This technique relies upon a data augmentation scheme, where the (unobserved) input are viewed as the missing data. In the technique described in this contribution, the input signal distribution is modeled by a mixture of Gaussian distributions, enabling the use of explicit formula for computing the posterior density and conditional expectation and thus avoiding Monte-Carlo integrations. Because this technique is able to capture some salient features of the input signal distribution, it performs generally much better than third-order or fourth-order cumulant based techniques. 1. INTRODUCTION This contribution is devoted to blind source separation and blind deconvolution. In these models, the observed data fx(t)g (a m \Theta 1 vector) is assumed to be given by x(t) = L X k=0 A(k)s(t \Gamma k) + v(t) (1) where fs(t)g (n \Theta 1) is the (unobserved) inp...
