Sheremet O. I., Sadovoi O. V., Sheremet K. S., Sokhina Yu. V. / Herald of Advanced Information Technology 2025; Vol. 8 No. 1: 43–53 ISSN 2663-0176 (Print) ISSN 2663-7731 (Online) Theoretical aspects of computer science, programming and data analysis 43 DOI: https://doi.org/10.15276/hait.08.2025.3 UDC 004.932.4 Using deep neural networks for image denoising in hardware-limited environments Oleksii I. Sheremet 1) ORCID: https://orcid.org/0000-0003-1298-3617; sheremet-oleksii@ukr.net. Scopus Author ID: 57170410800 Oleksandr V. Sadovoi 2) ORCID: https://orcid.org/0000-0001-9739-3661; sadovoyav@ukr.net. Scopus Author ID: 57205432765 Kateryna S. Sheremet 1) ORCID: https://orcid.org/0000-0003-3783-5274; artks@ukr.net. Scopus Author ID: 57207768511 Yuliia V. Sokhina 2) ORCID: https://orcid.org/0000-0002-4329-5182; jvsokhina@gmail.com. Scopus Author ID: 57205445522 1) Donbas State Engineering Academy, 39, Mashinobudivnykiv Blvd. Kramatorsk, 84313, Ukraine 2) Dniprovsky State Technical University, 2, Dniprobudivska Str. Kamyanske, 51918, Ukraine ABSTRACT Image denoising remains a vital topic in digital image processing, as it aims to recover visually clear content from observations compromised by random fluctuations. This article provides an overview of advanced deep neural network methods for image denoising and compares their performance with classical techniques. Emphasis is placed on the capacity of modern deep architectures to learn data-driven relationships that preserve structural details more effectively than traditional strategies. Implementation is conducted in a programming environment using open-source libraries, and the research is carried out in a cloud- based platform with Google Colab to facilitate reproducible and scalable experimentation. Both classical and deep learning-based solutions undergo quantitative and visual assessment, measured through standardized quality indices such as signal-to-noise ratio and a measure of structural similarity, alongside processing speed analysis. Results indicate that neural network-based approaches deliver superior restoration accuracy and detail preservation, although they typically require more computational resources. Classical methods, while simpler to implement and often feasible on hardware with minimal capabilities, frequently struggle when noise levels are high or exhibit complex characteristics. Methods based on block matching and three-dimensional filtering achieve competitive outcomes but impose higher computational overhead, limiting their practicality for time-sensitive applications. Potential future directions include hybrid techniques that merge the benefits of convolutional and transformer-inspired frameworks, along with refined training methodologies that extend applicability to scenarios lacking large volumes of clean reference data. By addressing these challenges, the evolving field of image denoising stands to offer more efficient and robust solutions for diverse real-world tasks. Keywords: Image denoising; deep neural networks; residual learning; transformer-inspired models; denoising quality; inference time For citation: Sheremet O. I., Sadovoi O. V., Sheremet K. S., Sokhina Yu. V. “Using deep neural networks for image denoising in hardware-limited environments”. Herald of Advanced Information Technology. 2025; Vol. 8 No. 1: 43–53. DOI: https://doi.org/10.15276/hait.08.2025.3 INTRODUCTION Image denoising is a fundamental challenge in digital image processing, aimed at restoring clear, noise-free images from degraded observations. Noise often appears as random fluctuations in pixel intensities and arises during image acquisition under adverse conditions. Consequently, images captured under these conditions frequently suffer from compromised clarity, diminished detail visibility, and reduced overall interpretability, which can adversely affect further processing tasks such as object recognition, scene analysis, and automated decision-making. The task of effectively suppressing noise and simultaneously preserving crucial image features, ______________________________________________ © Sheremet O., Sadovoi O., Sheremet K., Sokhina Yu., 2025 including fine textures, edges, and structural content, is very important across diverse application areas. These areas encompass digital photography, medical image diagnostics, satellite and aerial remote sensing, astronomical imaging, surveillance systems, industrial inspection, and various computer vision applications, where image quality significantly affects the accuracy and reliability of subsequent processing tasks. Historically, researchers have approached the denoising task using classical filtering and algorithmic strategies. Such traditional approaches include linear filtering methods, such as Gaussian smoothing and Wiener filtering, as well as nonlinear techniques, such as median filters and bilateral filters. Transform-domain methods, particularly those employing wavelets or Fourier transforms, have been widely utilized as well, benefiting from This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/deed.uk)