An ELU Network with Total Variation for Image Denoising

Tianyang Wang, Zhengrui Qin, Michelle Zhu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

12 Scopus citations

Abstract

In this paper, we propose a novel convolutional neural network (CNN) for image denoising, which uses exponential linear unit (ELU) as the activation function. We investigate the suitability by analyzing ELU’s connection with trainable nonlinear reaction diffusion model (TNRD) and residual denoising. On the other hand, batch normalization (BN) is indispensable for residual denoising and convergence purpose. However, direct stacking of BN and ELU degrades the performance of CNN. To mitigate this issue, we design an innovative combination of activation layer and normalization layer to exploit and leverage the ELU network, and discuss the corresponding rationale. Moreover, inspired by the fact that minimizing total variation (TV) can be applied to image denoising, we propose a TV regularized L2 loss to evaluate the training effect during the iterations. Finally, we conduct extensive experiments, showing that our model outperforms some recent and popular approaches on Gaussian denoising with specific or randomized noise levels for both gray and color images.

Original languageEnglish
Title of host publicationNeural Information Processing - 24th International Conference, ICONIP 2017, Proceedings
EditorsDerong Liu, Shengli Xie, El-Sayed M. El-Alfy, Dongbin Zhao, Yuanqing Li
PublisherSpringer Verlag
Pages227-237
Number of pages11
ISBN (Print)9783319700892
DOIs
StatePublished - 1 Jan 2017
Event24th International Conference on Neural Information Processing, ICONIP 2017 - Guangzhou, China
Duration: 14 Nov 201718 Nov 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10636 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other24th International Conference on Neural Information Processing, ICONIP 2017
CountryChina
CityGuangzhou
Period14/11/1718/11/17

Keywords

  • Convolutional neural network
  • Deep learning
  • ELU
  • Image denoising
  • Image processing
  • Total variation

Fingerprint Dive into the research topics of 'An ELU Network with Total Variation for Image Denoising'. Together they form a unique fingerprint.

  • Cite this

    Wang, T., Qin, Z., & Zhu, M. (2017). An ELU Network with Total Variation for Image Denoising. In D. Liu, S. Xie, E-S. M. El-Alfy, D. Zhao, & Y. Li (Eds.), Neural Information Processing - 24th International Conference, ICONIP 2017, Proceedings (pp. 227-237). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10636 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-70090-8_24