论文信息 - Delayed AMUSE - A Tool for Blind Source Separation and Denoising

Delayed AMUSE - A Tool for Blind Source Separation and Denoising

In this work we propose a generalized eigendecomposition (GEVD) of a matrix pencil computed after embedding the data into a high-dim feature space of delayed coordinates. The matrix pencil is computed like in AMUSE but in the feature space of delayed coordinates. Its GEVD yields filtered versions of the source signals as output signals. The algorithm is implemented in two EVD steps. Numerical simulations study the influence of the number of delays and the noise level on the performance.

[1] Nuno Ferreira,et al. On-line source separation of temporally correlated signals , 2002, 2002 11th European Signal Processing Conference.

[2] Andrzej Cichocki,et al. Second-order statistics based blind source separation using a bank of subband filters , 2003, Digit. Signal Process..

[3] Michael Ghil,et al. ADVANCED SPECTRAL METHODS FOR CLIMATIC TIME SERIES , 2002 .

[4] James V. Stone. Blind Source Separation Using Temporal Predictability , 2001, Neural Computation.

[5] T. Schreiber,et al. Nonlinear projective filtering I: Application to real time series , 1998 .

[6] Fabian J. Theis,et al. AUTOMATIC DENOISING USING LOCAL INDEPENDENT COMPONENT ANALYSIS , 2004 .

[7] V. Moskvina,et al. Approximate Projectors in Singular Spectrum Analysis , 2002, SIAM J. Matrix Anal. Appl..

[8] Ana Maria Tomé. Separation of a mixture of signals using linear filtering and second order statistics , 2002, ESANN.

[9] Andrzej Cichocki,et al. Adaptive blind signal and image processing , 2002 .