A new image denoising and enhancement method combining the nonsubsampled contourlet transform and improved total variation

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Transform-based denoising methods are very popular in recent years. However, they often suffer from unwanted artifacts like pesudo-Gibbs phenomena. In this paper, we propose a new hybrid image denoising by combining the nonsubsumpled contourlet transform (NSCT) with improved total variation. First, an improved stark function which integrates noise reduction with feature enhancement is developed to nonlinearly shrink and stretch the NSCT coefficients. Then an improved Total variation is introduced to reduce the pseudo-Gibbs artifacts of the enhanced image which are caused by the elimination of small NSCT coefficients. Numerical experiments show that this approach improves the image quality by enhancing the shape of edges and important detailed features while suppressing noise in comparison to many well known methods.

Original languageEnglish
Title of host publicationIntelligence Science and Big Data Engineering - 4th International Conference, IScIDE 2013, Revised Selected Papers
PublisherSpringer Verlag
Pages860-867
Number of pages8
ISBN (Print)9783642420566
DOIs
StatePublished - 2013
Event4th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2013 - Beijing, China
Duration: 31 Jul 20132 Aug 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8261 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference4th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2013
Country/TerritoryChina
CityBeijing
Period31/07/132/08/13

Keywords

  • image denoising
  • image enhancement
  • nonsubsumpled contourlet transform
  • total variation

Fingerprint

Dive into the research topics of 'A new image denoising and enhancement method combining the nonsubsampled contourlet transform and improved total variation'. Together they form a unique fingerprint.

Cite this