Premium
How to discretize the Total Variation of an image?
Author(s) -
Moisan Lionel
Publication year - 2007
Publication title -
pamm
Language(s) - English
Resource type - Journals
ISSN - 1617-7061
DOI - 10.1002/pamm.200700424
Subject(s) - deblurring , discretization , variation (astronomy) , image (mathematics) , mathematics , sampling (signal processing) , algorithm , statistics , computer science , image processing , image restoration , mathematical analysis , artificial intelligence , physics , computer vision , astrophysics , filter (signal processing)
Because they are based on finite differences, usual discretizations of the Total Variation lead to aliased images. We propose a new discretization called spectral total variation that agrees with Shannon sampling principles and produces images that can be exactly interpolated. The quality improvement is illustrated experimentally in the case of image deblurring. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)