Deep convolutional neural network for mixed random impulse and Gaussian noise reduction in digital images. Issue 15 (18th February 2021)

• Record Type: Journal Article
• Authors: Mafi, Mehdi; Izquierdo, Walter; Martin, Harold; Cabrerizo, Mercedes; Adjouadi, Malek
• Is Part Of: IET image processing. Volume 14:Issue 15(2020)
• Page Start: 3791
• Page End: 3801
• Publication Date: 2021-02-18
• Subjects: image denoising -- Gaussian noise -- learning (artificial intelligence) -- filtering theory -- convolution -- image segmentation -- neural nets -- image classification; digital images -- deep CNN -- minimal loss -- optimal estimation -- known noise mixtures -- unknown noise mixtures -- different structural metrics -- optimal denoising results -- 20‐layer network -- additional 12 images -- CNN‐based approach -- convolutional neural network -- mixed random impulse -- Gaussian noise reduction -- batch normalisation -- mixed impulse; Image processing -- Periodicals; 621.36705
• DOI: 10.1049/iet-ipr.2019.0931
• Languages: English
• ISSNs: 1751-9659

Abstract

Abstract : This study utilises a deep convolutional neural network (CNN) implementing regularisation and batch normalisation for the removal of mixed, random, impulse, and Gaussian noise of various levels from digital images. This deep CNN achieves minimal loss of detail and yet yields an optimal estimation of structural metrics when dealing with both known and unknown noise mixtures. Moreover, a comprehensive comparison of denoising filters through the use of different structural metrics is provided to highlight the merits of the proposed approach. Optimal denoising results were obtained by using a 20‐layer network with 40 × 40 patches trained on 400 180 × 180 images from the Berkeley segmentation data set (BSD) and tested on the BSD100 data set and an additional 12 images of general interest to the research community. The comparative results provide credence to the merits of the proposed filter and the comprehensive assessment of results highlights the novelty and performance of this CNN‐based approach.