Presentation 2003/11/15
Blind Source Separation by Minimum Nonlinear Distortion
Tetsuya MINATOHARA, Tetsuo FURUKAWA, Shozo YASUI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In the new BSS method presented here, a linear combination of the sensor signals is fed to a nonlinear unit. The weight vector associated with the linear combination is optimized in the sense of minimizing the average nonlinear distortion, i.e., average difference between the input and output of the nonlinear unit. Such optimization made by the steepest descent algorithm is shown to be able to extract the blind signal sources one-by-one if the sources are sub-Gaussian.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Blind Source Separation / Independent Component Analysis / Nonlinear Transformation
Paper # NLP2003-122
Date of Issue

Conference Information
Committee NLP
Conference Date 2003/11/15(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Nonlinear Problems (NLP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Blind Source Separation by Minimum Nonlinear Distortion
Sub Title (in English)
Keyword(1) Blind Source Separation
Keyword(2) Independent Component Analysis
Keyword(3) Nonlinear Transformation
1st Author's Name Tetsuya MINATOHARA
1st Author's Affiliation Department of Brain science and Sys.Eng.,Grad.School of Life Science and Sys.Eng.,Kyushu Institute of Technology()
2nd Author's Name Tetsuo FURUKAWA
2nd Author's Affiliation Department of Control Eng.and Science,Faculty of Computer Science and Sys. Eng.,Kyushu Institute of Technology
3rd Author's Name Shozo YASUI
3rd Author's Affiliation Department of Brain science and Sys.Eng.,Grad.School of Life Science and Sys.Eng.,Kyushu Institute of Technology
Date 2003/11/15
Paper # NLP2003-122
Volume (vol) vol.103
Number (no) 464
Page pp.pp.-
#Pages 6
Date of Issue