Estimator of independent sources from degenerate mixtures

Details Estimator of independent sources from degenerate mixtures Estimator of independent sources from degenerate mixtures

1998-12-01

2002-01-29

Dorvil, Richemond (Department: 2641)

Data processing: speech signal processing, linguistics, language

Speech signal processing

For storage or transmission

C704S226000, C704S233000, C381S094100, C381S094200, C706S022000

Reexamination Certificate

active

06343268

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to blind source separation problems and more particularly to estimating independent Auto Regressive (AR) processes from their sum.
2. Description of the Prior Art
A generic Blind Source Separation (BSS) problem is defined by the following: given m measurements x
1
, . . . ,x
m
obtained from n independent signals (sources) s
1
, . . . ,s
n
, estimate the original signals through n estimators ŝ
1
, . . . , ŝ
n
, based on the time-series x
1
(·), . . . ,x
m
(·).
Current BSS literature addresses only the case when the number of sources is equal to the number of microphones, in example m=n. This is discussed by S. Amari in “Minimum Mutual Information Blind Separation”,
Neural Computation,
1996, by A. J. Bell and T. J. Sejnowski in “An Information-maximization Approach To Blind Separation And Blind Deconvolution”,
Neural Computation,
7:1129-1159, 1995, by J. F. Cardoso in “Infomax And Maximum Likelihood For Blind Source Separation”,
IEEE Signal Processing Letters,
4(4):112-114, April 1997, by P. Comon in “Independent Component Analysis, A New Concept?”,
Signal Processing,
36(3):287-314, 1994, by C. Jutten and J. Herault in “Blind Separation Of Sources, Part I: An Adaptive Algorithm Based On Neuromimetic Architecture”,
Signal Processing,
24(l):1-10, 1991, by B. A. Pearlmutter and L. C. Parra in “A Context-sensitive Generalization Of ICA”, In International Conference on
Neural Information Processing,
Hong Kong, 1996, by K. Torkkola in “Blind Separation Of Convolved Sources Based On Information Maximization”, In
IEEE Workshop on Neural Networks for Signal Processing,
Kyoto, Japan 1996. The case when the number of measurements m is strictly smaller than the number of sources n is called the degenerate case.
It is an object of the present invention to reconstruct independent signals from degenerate mixtures. More specifically, it is an object of the present invention to estimate independent Auto Regressive (AR) processes from their sum.
SUMMARY OF THE INVENTION
The present invention is a system that reconstructs independent signals from degenerate mixtures. More specifically, the present invention estimates independent Auto Regressive (AR) processes from their sum. The invention addresses the identification subsystem for such a degenerate case, particularly the case of two AR processes of known finite dimension (m=1 and n=2).
The present invention includes an identification system and an estimator. A mixture signal and noise are inputted into the system and through noise separation, a near pure signal is outputted. The identification system includes an ARMA identifier, a computation of autocovariance coefficients, an initializer and a gradient descent system. The estimator includes filtering.


REFERENCES:
patent: 5539832 (1996-07-01), Weinstein et al.
patent: 5706402 (1998-01-01), Bell
patent: 5721694 (1998-02-01), Graupe
patent: 5808913 (1998-09-01), Choi et al.
patent: 5943429 (1999-08-01), Handel
patent: 6067513 (2000-05-01), Ishimitsu
Tugnair, J.K. “Linear identification of non-Gaussian noncausal auto regressive signal-in-noise processes” Aug. 26, 1989, p. 447-450.*
Taleb, A. et al. “On underdetermined source separation” IEEE Mar. 10, 1999, p. 1445-1448.*
Verbout, S.M. et al. “Parameter estimation for autoregressive Gaussian-mixture processes: the EMAX algprithm” IEEE Apr. 24, 1997 p. 3549-3552.*
Tugnait, J.K. “Approaches of FIR system identification with noisy data using higher-order statistics” IEEE Jul. 1990. p. 1307-1317.*
Tugnait, J.K. “On blind separation of convolutive mixtures of independent linear signals in unknown addative noise” IEEE 11/98 p. 3117-3123.

Affiliated with

Also associated with

Rating

Estimator of independent sources from degenerate mixtures does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Estimator of independent sources from degenerate mixtures, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Estimator of independent sources from degenerate mixtures will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2837448