The Analysis of CNN Structure for Image Denoising

Jae Hyeon Park, Jeong Hyeon Kim, Sung In Cho

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

13 Scopus citations

Abstract

This paper proposes an optimal structure of a convolutional neural network (CNN) for image denoising by analyzing the conventional CNN denoisers. There are three main factors that can determine the denoising performance of the CNN denoiser: the number of feature dimensions of each convolution layer, the number of convolution layers, and the usage of dilated convolution. We analyze the denoising performance variations of the conventional CNN denoiser depending on the above three factors and propose the optimal structure of the CNN denoiser. Experimental results showed that the above three factors have a high correlation with the denoising performance. Based on the experimental results, we could provide the optimal structure of the CNN denoiser.

Original languageEnglish
Title of host publicationProceedings - International SoC Design Conference 2018, ISOCC 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages220-221
Number of pages2
ISBN (Electronic)9781538679609
DOIs
StatePublished - 2 Jul 2018
Event15th International SoC Design Conference, ISOCC 2018 - Daegu, Korea, Republic of
Duration: 12 Nov 201815 Nov 2018

Publication series

NameProceedings - International SoC Design Conference 2018, ISOCC 2018

Conference

Conference15th International SoC Design Conference, ISOCC 2018
Country/TerritoryKorea, Republic of
CityDaegu
Period12/11/1815/11/18

Keywords

  • Convolution neural network
  • Dilated convolution
  • Feature dimension
  • Image denoising

Fingerprint

Dive into the research topics of 'The Analysis of CNN Structure for Image Denoising'. Together they form a unique fingerprint.

Cite this