Presentation 2003/3/10
Multichannel Blind Deconvolution Using Overcomplete representaions
Tomoya OKUMURA, Satoshi MAEKAWA, Manabu KOTANI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Multichannel blind deconvolution method based on the linear filter tries to make the output both spatially and temporally independent. However, the number of source signals must be less than or eaual to the number of observed signals. On the other hand, signal separation algorithms based on overcomplete representations, may deal with exceed number of source signals, but these existing algorithms are not applicable to spatial and temporal mixture signals. Therefore, we propose a method to realize multichannel blind deconvolution using overcomplete representations. In the method, phase symmetry of bases are explicitly introduced to deal with temporal signals.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Multichannel blind deconvolution / Overcomplete / Independent component analysis / Phase symmetry
Paper # NC2002-151
Date of Issue

Conference Information
Committee NC
Conference Date 2003/3/10(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 Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Multichannel Blind Deconvolution Using Overcomplete representaions
Sub Title (in English)
Keyword(1) Multichannel blind deconvolution
Keyword(2) Overcomplete
Keyword(3) Independent component analysis
Keyword(4) Phase symmetry
1st Author's Name Tomoya OKUMURA
1st Author's Affiliation Faculty of Engineering, Kobe University()
2nd Author's Name Satoshi MAEKAWA
2nd Author's Affiliation Communications Research Laboratory
3rd Author's Name Manabu KOTANI
3rd Author's Affiliation Faculty of Engineering, Kobe University
Date 2003/3/10
Paper # NC2002-151
Volume (vol) vol.102
Number (no) 729
Page pp.pp.-
#Pages 6
Date of Issue