Dictionary transfer for image denoising via domain adaptation

Gang Chen, Caiming Xiong, Jason J. Corso

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

12 Scopus citations

Abstract

The idea of using overcomplete dictionaries with prototype signal atoms for sparse representation has found many applications, among which image denoising is considered as an active research topic. However, the standard process to train a new dictionary for image denoising requires the whole image (or most parts) as input, which is costly; training the dictionary on just a few patches would result in overfitting. We instead propose a dictionary learning approach for image denoising via transfer learning. We transfer the source domain dictionary to a target domain for image denoising via a dictionary- regularization term in the energy function. Thus, we have a new dictionary that is trained from only a few patches of the target noisy image. We measure the performance on various corrupted images, and show that our method is fast and comparable to the state of the art. We also demonstrate cross-domain transfer (photo to medical image).

Original languageEnglish
Title of host publication2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings
Pages1189-1192
Number of pages4
DOIs
StatePublished - 2012
Event2012 19th IEEE International Conference on Image Processing, ICIP 2012 - Lake Buena Vista, FL, United States
Duration: 30 Sep 20123 Oct 2012

Publication series

NameProceedings - International Conference on Image Processing, ICIP
ISSN (Print)1522-4880

Conference

Conference2012 19th IEEE International Conference on Image Processing, ICIP 2012
Country/TerritoryUnited States
CityLake Buena Vista, FL
Period30/09/123/10/12

Keywords

  • Dictionary learning
  • domain adaptation
  • image denoising
  • sparse representations
  • transfer learning

Fingerprint

Dive into the research topics of 'Dictionary transfer for image denoising via domain adaptation'. Together they form a unique fingerprint.

Cite this