Presentation 2004/3/11
Kernel Wiener Filter
Yoshikazu WASHIZAWA, Yukihiko YAMASHITA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Wiener filter is used widely for inverse problems. From a observed signal, it provides the best restored signal with respect to the square error averaged over the original signal and the noise among linear operators. In this paper, we provide the kernel Wiener filter which is a kernel based extension of the Wiener filter. When the kernel method is applied to the Wiener filter directly, the dimension of the space where the calculation has to be done is very large since samples of the noise have to be used. Then, by using the first order approximation of a kernel function, we provide a realistic solution. We also provide an approximated subspace information criteria (aSIC) that is an extension of SIC. Moreover, we provide an experimental result in order to show their advantages.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Wiener filter / kernel based method / inverse problem / image restoration
Paper # NC2003-177
Date of Issue

Conference Information
Committee NC
Conference Date 2004/3/11(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) Kernel Wiener Filter
Sub Title (in English)
Keyword(1) Wiener filter
Keyword(2) kernel based method
Keyword(3) inverse problem
Keyword(4) image restoration
1st Author's Name Yoshikazu WASHIZAWA
1st Author's Affiliation Graduate School of Science and Engineering, Tokyo Institute of Technology()
2nd Author's Name Yukihiko YAMASHITA
2nd Author's Affiliation Graduate School of Science and Engineering, Tokyo Institute of Technology
Date 2004/3/11
Paper # NC2003-177
Volume (vol) vol.103
Number (no) 733
Page pp.pp.-
#Pages 6
Date of Issue