Image Denoising Using Deblur Generative Adversarial Network Denoising U-Net

Image Denoising Using Deblur Generative Adversarial Network Denoising U-Net B. Usha Rani Department of Electronics and Communications Engineering, AMC Engineering College, Visvesvaraya Technological University, Belagavi 590018, India https://orcid.org/0009-0008-6098-0203 R. Aruna Department of Electronics and Communications Engineering, AMC Engineering College, Visvesvaraya Technological University, Belagavi 590018, India https://orcid.org/0000-0003-3367-9067 P. Velrajkumar Vels Institute of Science, Technology and Advanced Studies (VISTAS), Pallavaram, Chennai, India https://orcid.org/0000-0001-7900-3500 N. Amuthan Department of Electrical and Electronics Engineering, AMC Engineering College, Visvesvaraya Technological University, Belagavi 590018, India https://orcid.org/0000-0001-8830-3353 N. Sivakumar Department of Mechanical Engineering, AMC Engineering College, Visvesvaraya Technological University, Belagavi 590018, India https://orcid.org/0000-0002-6876-3253

Convolutional neural networks (CNNs) are becoming increasingly popular for image denoising. U-Nets, a type of CNN architecture, have been shown to be effective for this task. However, the impact of shallow layers on deeper layers decreases as the depth of the network increases. To address this issue, the authors propose a new image denoising method called DGANDU-Net. DGANDU-Net combines the DeblurGAN design with a U-Net architecture. This combination allows DGANDU-Net to effectively remove noise from images while preserving fine details. The authors also propose the use of two loss functions, mean square error (MSE) and perceptual loss, to improve the performance of DGANDU-Net. MSE is used to learn and improve the extracted features, while perceptual loss is used to produce the final denoised image. The authors evaluate the performance of DGANDU-Net on a variety of noise levels and find that it outperforms other state-of-the-art denoising algorithms in terms of both visual quality and two evaluation indices, including peak signal-to-noise ratio (PSNR) and Structural Similarity Index Measure (SSIM). Specifically, for extremely noisy environments with a noise standard deviation of 75, DGANDU-Net achieves an average PSNR of 37.39[Formula: see text]dB in the test dataset. The authors conclude that DGANDU-Net is a promising new method for image denoising that has the potential to significantly improve the quality of medical images used for diagnosis and treatment.
10 26 2024 02 2025 2550069 10.1142/S0218126625500690 10.1142/S0218126625500690 https://www.worldscientific.com/doi/10.1142/S0218126625500690 https://www.worldscientific.com/doi/pdf/10.1142/S0218126625500690 10.1137/040616024 2006 Int. Conf. Image Processing Wang J. 1429 10.1117/12.643267 10.1109/TIP.2007.906002 Adv. Neural Inf. Process. Syst. Jain V. 1 21 2008 10.1109/LNLA.2009.5278408 J. Theor. Appl. Inf. Technol. Saba T. 32 17 2010 2011 20th Eur. Conf. Circuit Theory and Design (ECCTD) Slavova A. 226 10.1002/tee.20684 10.1007/978-3-642-03718-4_12 2012 IEEE Conf. Computer Vision and Pattern Recognition Burger H. C. 2392 10.1109/TCYB.2013.2278548 10.1137/120874989 Int. J. Emerg. Technol. Adv. Eng. Saxena C. 878 4 2014 10.1109/IPTA.2014.7001976 Adv. Neural Inf. Process. Syst. Xu L. 1790 27 2014 10.1109/TCSVT.2015.2430711 10.1109/WCSP.2015.7341021 10.1109/ICDMW.2016.0041 10.1109/TIP.2017.2662206 10.1109/TIP.2018.2839891 10.1109/CVPR.2018.00338 10.1109/ICCAR.2018.8384706 10.1109/TGRS.2019.2952062 10.1049/iet-ipr.2019.0157 10.1016/j.image.2019.115727 10.1109/Confluence47617.2020.9057895 10.1007/s00530-021-00753-1 10.1109/TIP.2022.3181488 Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. Kupyn O. 8183 10.1109/TPAMI.2016.2596743 IEEE Int. Conf. Computer Vision and Pattern Recognition He K. IEEE Computer Society Conf. Computer Vision and Pattern Recognition (CVPR’05) Roth S. 860 2 10.1117/1.3600632 10.1109/CVPR.2017.300 10.1016/j.neunet.2019.12.024 10.1016/j.neunet.2019.08.022 10.1109/TIP.2007.901238 10.1166/asem.2019.2380