Designing a Modular Framework for Processing and Enhancing Scanned Documents Using Advanced Denoising Algorithms
DOI:
https://doi.org/10.29304/jqcsm.2026.18.12441Keywords:
Document Enhancement, Modular Framework, Scanned Documents, Image Processing, Denoising AlgorithmsAbstract
Scanned documents are often flawed such as background noise, skew and uneven illuminations which negatively affect reading and text recognition. The given study presents a two-step processing model that can serve to improve the quality of grayscale and color-scanned documents because of the use of combined denoising and deskewing methods. Three denoising methods were tested under various noise levels: DRUNet, DnCNN, and Total Variation (TV). To get the best results in restoring image quality, we used pre-trained models for the deep learning algorithms (DRUNet and DnCNN) through the DeepInv library. This allowed us to use powerful, ready-to-use features to clean the documents effectively. Their performance was then measured using standard quality scores and visual checks. The results showed that DRUNet produced the best and more reproducible performance, which was able to suppress noise and preserve fine structural and textual fidelity. In addition, preprocessing step of Otsu thresholding and minimum bounding rectangle estimation was also applied to automatically correct document skew to enhance text alignment and readability. Python and Gradio were used as the implementation language of the system to offer an interactive, transparent, and reproducible platform. In general, the suggested framework would significantly increase the clarity, alignment, and the overall quality of scanned documents and make them more reliable to use in OCR and digital archiving purposes.
Downloads
References
M. Smith and A. Brown, "Enhancing Document Image Processing: Correcting Skew in Printed Documents Using Deep Learning," Journal of Image and Graphics, vol. 13, no. 1, 2025.
R. Zhang, "End to End Unsupervised Document Image Blind Denoising," arXiv preprint arXiv:2101.00000, 2021.
R. Rotman et al., "A U-Net based pre-processing pipeline for robust OCR with synthetic noisy document datasets," in Proc. Int. Conf. Document Anal. Recognit. (ICDAR), 2022, pp. 154–168.
O. Boudraa, W. K. Hidouci, and D. Michelucci, "Using skeleton and Hough transform variant to correct skew in historical documents," Journal of Mathematics and Computers in Simulation, vol. 167, pp. 100–114, 2019.
R. Ahmad, S. Naz, and I. Razzak, "Efficient skew detection and correction in scanned document images through clustering of probabilistic hough transforms," Pattern Recognition Letters, vol. 152, pp. 93–99, 2021.
W. Kim et al., "A Systematic Review of Deep Learning-Based Image Denoising Methods," Frontiers in Medical Technology, vol. 6, Art. 134000, 2024.
H. M. Zangana and F. M. Mustafa, "Hybrid Image Denoising Using Wavelet Transform and Deep Learning," EAI Endorsed Transactions on AI and Robotics, vol. 3, pp. 1–10, 2024.
M. S. Tawfik et al., "Comparative Study of Traditional and Deep-Learning Denoising Approaches for Image-Based Petrophysical Characterization," Frontiers in Water, vol. 3, Art. 800369, 2022.
H. S. Abdulla, A. S. Shaheen, and N. M. Isaac, "Effectiveness of Image Curvelet Transform Coefficients for Image Denoising," Al-Rafidain Journal of Computer Science and Mathematics, vol. 18, no. 2, pp. 1–8, 2024.
H. H. Ali, "Development of Traditional Algorithms and Hybrid Approach for Denoising Color Images," Al-Rafidain Journal of Computer Science and Mathematics, vol. 18, no. 2, pp. 36–46, 2011.
A. Supriyono et al., "Advancements in NLP-driven OCR post-processing: A systematic review," Journal of Digital Document Processing, vol. 2, no. 1, pp. 45–60, 2024.
G. S. Hukkeri, R. H. Goudar, P. Janagond, and P. S. Patil, "Machine learning in OCR technology: Performance analysis of different OCR methods for slide-to-text conversion," International Journal of Advanced Computer Science and Applications, vol. 13, no. 8, 2022.
P. Mohta, "Total variation-based image denoising for edge-preserving smoothing," International Journal of Computer Vision and Image Processing, vol. 14, no. 2, pp. 45–58, 2024.
K. Zhang, W. Zuo, and L. Zhang, "Plug-and-play image restoration with deep denoiser prior," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 10, pp. 7005–7020, 2021.
K. Zhang, W. Zuo, Y. Chen, D. Meng, and L. Zhang, "Beyond a Gaussian denoiser: Residual learning of deep CNN for image denoising," IEEE Transactions on Image Processing, vol. 26, no. 7, pp. 3142–3155, 2017.
Y. Zhao et al., "Comprehensive Evaluation of IQA Metrics for Image Restoration Models," arXiv preprint arXiv:2403.10988, 2024.
MDPI, "Prospects of Structural Similarity Index for Medical Image Applications," Sensors, vol. 22, no. 18, Art. 6890, 2022.
R. Zhang et al., "Learning Perceptual Similarity for Image Restoration Using Deep Feature Representations," IEEE Access, vol. 11, pp. 45122–45136, 2023.
M. Safari et al., "MRI Super-Resolution Reconstruction Using Efficient Diffusion Probabilistic Model with Residual Shifting," Physics in Medicine & Biology, vol. 70, no. 12, p. 125008, 2025.
K. Loh et al., "A Generalized Quality Assessment Method for Gradient-Based Image Metrics," IET Image Processing, vol. 15, no. 12, pp. 2859–2871, 2021.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Zahraa Ali Mohamed Nather, Hasan Maher Ahmed
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
