Presentation 2013-11-25
A Near-lossless Coding by Sparcification of DOT coefficients
Kazuki ITOU, Gene CHEUNG, Akira KUBOTA, Antonio ORTEGA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Near-lossless coding is of great interest in the fields of medical images, fine arts and all applications demanding high image quality. A general technique for Near-Lossless coding uses a prediction coding, and sets up the error in tolerance level in a pixel domain. Thus the optimization is performed in a pixel domain. In this paper, we propose a new Near-lossless Coding by sparcification of DCT coefficients in transform domain. We seek to minimize the weighted sum of l_0-norm of DCT coefficients. But minimizing conventional l_0-norm is in general difficult. We propose an alternative to solve the sparsification problem as an iterative reweighted l_2-norm instead.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) DCT coefficients / Sparsificaton
Paper # IE2013-62
Date of Issue

Conference Information
Committee IE
Conference Date 2013/11/18(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 Image Engineering (IE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Near-lossless Coding by Sparcification of DOT coefficients
Sub Title (in English)
Keyword(1) DCT coefficients
Keyword(2) Sparsificaton
1st Author's Name Kazuki ITOU
1st Author's Affiliation Faculty of Engineering, Chuo University()
2nd Author's Name Gene CHEUNG
2nd Author's Affiliation Digital Content and Media Sciences Research Division, National Institute of Informatics
3rd Author's Name Akira KUBOTA
3rd Author's Affiliation Faculty of Engineering, Chuo University
4th Author's Name Antonio ORTEGA
4th Author's Affiliation Department of Electrical Engineering, University of Southern California
Date 2013-11-25
Paper # IE2013-62
Volume (vol) vol.113
Number (no) 318
Page pp.pp.-
#Pages 4
Date of Issue