Back to articles
Article
Volume: 34 | Article ID: IMAGE-288
Image
Mix-loss trained bias-removed blind image denoising network
  DOI :  10.2352/EI.2022.34.8.IMAGE-288  Published OnlineJanuary 2022
Abstract
Abstract

We studied the modern deep convolutional neural networks used for image denoising, where RGB input images are transformed into RGB output images via feed-forward convolutional neural networks that use a loss defined in the RGB color space. Considering the difference between human visual perception and objective evaluation metrics such as PSNR or SSIM, we propose a data augmentation technique and demonstrate that it is equivalent to defining a perceptual loss function. We trained a network based on this and obtained visually pleasing denoised results. We also combine an unsupervised design and the bias-free network to deal with the overfitting due to the absence of clean images, and improve performance when the noise level exceeds the training range.

Subject Areas :
Views 52
Downloads 6
 articleview.views 52
 articleview.downloads 6
  Cite this article 

Yi Yang, Chih-Hsien Chou, Jan P. Allebach, "Mix-loss trained bias-removed blind image denoising networkin Electronic Imaging,  2022,  pp 288-1 - 288-7,  https://doi.org/10.2352/EI.2022.34.8.IMAGE-288

 Copy citation
  Copyright statement 
Copyright © 2022, Society for Imaging Science and Technology 2022
ei
Electronic Imaging
2470-1173
2470-1173
Society for Imaging Science and Technology
IS&T 7003 Kilworth Lane, Springfield, VA 22151 USA