Blind Source Separation of Post-Nonlinear Mixtures Using Evolutionary Computation and Gaussianization

  • Authors:

  • Tiago M. Dias;Romis Attux;João M. Romano;Ricardo Suyama

  • Affiliations:

  • Department of Microwave and Optics, and DSPCOM --- Laboratory of Signal Processing for Communications School of Electrical and Computer Engineering, University of Campinas (UNICAMP), Campinas, Bra ...;Department of Computer Engineering and Industrial Automation, and DSPCOM --- Laboratory of Signal Processing for Communications School of Electrical and Computer Engineering, University of Campina ...;Department of Microwave and Optics, and DSPCOM --- Laboratory of Signal Processing for Communications School of Electrical and Computer Engineering, University of Campinas (UNICAMP), Campinas, Bra ...;Department of Microwave and Optics, and DSPCOM --- Laboratory of Signal Processing for Communications School of Electrical and Computer Engineering, University of Campinas (UNICAMP), Campinas, Bra ...

  • Venue:
  • ICA '09 Proceedings of the 8th International Conference on Independent Component Analysis and Signal Separation
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this work, we propose a new method for source separation of post-nonlinear mixtures that combines evolutionary-based global search, gaussianization and a local search step based on FastICA algorithm. The rationale of the proposal is to attempt to obtain efficient and precise solutions using with parsimony the available computational resources, and, as shown by the simulation results, this aim was satisfactorily fulfilled.