Informatics and Applications
2024, Volume 18, Issue 2, pp 60-71
IMAGE DECOMPOSITION WITH DISCRETE WAVELET TRANSFORM TO DESIGN A DENOISING NEURAL NETWORK
Abstract
Reducing noise in digital images is one of the most common tasks in image processing. At the moment, noise reduction approaches based on the applying of convolutional neural networks are widely used. In this case, as a rule, model training is based on minimizing the error function between the result of the network operation and the expected reference image and, additionally, various representations of the two-dimensional image signal and their properties are not used to optimize the training of noise reduction network architectures. The paper proposes an approach to training neural networks to suppress noise. The described approach is based on the usage of the N-fold fast Haar wavelet transform. This representation of a discrete image signal allows one to discard the classical architecture of the autoencoder and to use only its part that encodes the signal which leads to a significant reduction in model parameters and speeds up the network.
[+] References (32)
- Ronneberger, O., P Fischer, and T. Brox. 2015. U-Net: Convolutional networks for biomedical image segmentation. Medical Image Computing and Computer-Assisted Intervention Proceedings. Cham: Springer International Publishing. 234-241.
- Komatsu, R., and T. Gonsalves. 2020. Comparing U-Net based models for denoising color images. AI 1(4):465- 486. doi: 10.3390/ai1040029.
- Hasinoff, S.W 2014. Saturation (imaging). Computer vision: A reference guide. Boston, MA: Springer. 699-701. doi: 10.1007/978-0-387-31439-6_483.
- Wang, Z., A. Bovic, H. Sheikh, and E. Simoncelli. 2004. Image quality assessment: From error visibility to structural similarity. IEEE T. Image Process. 13(4):600-612. doi: 10.1109/TIP.2003.819861.
- Zhao, H., O. Gallo, I. Frosio, and J. Kautz 2017. Loss functions for image restoration with neural networks. IEEE Trans. Computational Imaging 3(1):47-57. doi: 10.1109/TCI.2016.2644865.
- Woo, S., P. Jongchan, L. Joon-Young, and K. In-So. 2018. CBAM: Convolutional block attention module. 17 p. Available at: https://arxiv.org/abs/1807.06521 (accessed April 28, 2024).
- Jiang, J., H. Xiangming, Y. Zhao, X. Xu, and Y. Cui. 2023. SDAUNet: A simple dual attention mechanism UNet for mixed noise removal. IET Image Process. 17(13):3884- 3896. doi: 10.1049/IPR2.12905.
- Liu, P, H. Zhang, K. Zhang, L. Lin, and W. Zuo. 2018. Multi-level wavelet-CNN for image restoration. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops Proceedings. Los Alamitos, CA: IEEEComputer Society. 886-895. doi: 10.1109/CVPRW 2018.00121.
- Batziou, E., K. Ioannidis, I. Patras, S. Vrochidis, and I. Kompatsiaris. 2023. Low-light image enhancement based on U-Net and Haar wavelet pooling. MultiMedia modeling. Cham: Springer Nature Switzerland. 510-522.
- Tian, C., M. Zheng, W Zuo, B. Zhang, Y. Zhang, and D. Zhang. 2023. Multi-stage image denoising with the wavelet transform. Pattern Recogn. 134:109050. doi: 10.1016/j.patcog.2022.109050.
- Huang, J.-J., and P L. Dragotti. 2021. WINNet: Wavelet- inspired invertible network for image denoising. IEEE T. Image Process. 31:4377-4392.
- Guo, T., H. Mousavi, T. Vu, and V. Monga. 2017. Deep wavelet prediction for image super-resolution. Conference on Computer Vision and Pattern Recognition Workshops Proceedings. Los Alamitos, CA: IEEE Computer Society. 1100-1109. doi: 10.1109/CVPRW.2017.148.
- Qin, X., H. Dai, X. Hu, D.-P Fan, L. Shao, and G. Van. 2022. Highly accurate dichotomous image segmentation. Computer vision. Cham: Springer Nature Switzerland. 38-56.
- He, K., H. Zhang, S. Ren, and J. Sun. 2016. Deep residual learning for image recognition. Conference on Computer Vision and Pattern Recognition Proceedings. Los Alamitos, CA. 770-778.
- Scherer, D., A. Muller, and S. Behnke. 2010. Evaluation of pooling operations in convolutional architectures for object recognition. Artificial neural networks. Berlin, Heidelberg: Springer. 92-101. doi: 10.1007/978-3-642- 15825-4.10.
- Makarenko, A. V. 2020. Glubokie neyronnye seti: zarozhdenie, stanovlenie, sovremennoe sostoyanie [Deep neural networks: Origins, development, and current status]. Problemy upravleniya [Control Sciences] 2:3-19. doi: 10.25728/pu.2020.2.1.
- Buy, T. T. C., and V.G. Spitsyn. 2011. Razlozhenie tsifrovykh izobrazheniy s pomoshch'yu dvumernogo diskretnogo veyvlet-preobrazovaniya i bystrogo preobra- zovaniya Khaara [Digital image decomposition using twodimensional discrete wavelet transform and fast Haar transform]. Izvestiya Tomskogo politekhnicheskogo uni- versiteta [Bulletin of the Tomsk Polytechnic University] 318(5):73-76.
- Pavlov, A. N. 2008. Detektirovanie informatsionnykh signalov na osnove rekonstruktsii dinamicheskikh sistem i diskretnogo veyvlet-preobrazovaniya [Detection of information signals based on reconstruction of dynamical systems and discrete wavelet-transform]. Izvestiya vysshikh uchebnykh zavedeniy. Prikladnaya nelineynaya dinami- ka [Bulletin of Higher Educational Institutions. Applied Nonlinear Dynamics] 16(6):3-17.
- Pronkin, A. V. 2022. Otsenivanie urovnya shuma v sostave izobrazheniya s ispol'zovaniem veyvletov Khaara [Noise level estimation in images using Haar wavelets]. 22nd Con - ference (International) on Computer Graphics and Computer Vision Proceedings. Moscow: IPM im. M.V. Keldysha. 32:442-448. doi: 10.20948/graphicon-2022-442-448.
- Bradski, G. 2000. The OpenCV Library. Dr. Dobbs J. 25(11):120-125.
- Abdelhamed, A., S. Lin, and M. S. Brown. 2018. A high-quality denoising dataset for smartphone cameras. IEEE/CVF Conference on Computer Vision and Pattern Recognition Proceedings. Los Alamitos, CA: IEEE Computer Society. 1692-1700. doi: 10.1109/CVPR. 2018.00182.
- Huang, J.-B., A. Singh, and N. Ahuja. 2015. Single image super-resolution from transformed self-exemplars. Conference on Computer Vision and Pattern Recognition Proceedings. Los Alamitos, CA: IEEE Computer Society. 5197-5206.
- Agustsson, E., and R. Timofte. 2017. NTIRE 2017 challenge on single image super-resolution: Dataset and study. Conference on Computer Vision and Pattern Recognition Workshops Proceedings. Los Alamitos, CA: IEEEComputer Society. 1110-1121. doi: 10.1109/ CVPRW2017.149.
- Martin, D., C. Fowlkes, D. Tal, and J. Malik. 2001. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. 8th Conference (International) on Computer Vision Proceedings. Piscataway, NJ: IEEE Computer Society. 2:416-423. doi: 10.1109/ICCV. 2001.937655.
- Plotz, T., and S. Roth. 2017. Benchmarking denoising algorithms with real photographs. 2017. Conference on Computer Vision and Pattern Recognition Proceedings. Los Alamitos, CA: IEEE Computer Society. 2750-2759. doi: 10.1109/CVPR.2017.294.
- Paszke, A., S. Gross, F Massa, and A. Lerer. 2019. PyTorch: An imperative style, high-performance deep learning library. Adv. Neur. Inf. 32:8024-8035.
- Wightman, R. 2019. PyTorch image models. Available at: https://github.com/rwightman/pytorch-image- models (accessed April 28, 2024).
- Wang, Z., X. Cun, J. Bao, W Zhou, J. Liu, and H. Li. 2022. Uformer: Ageneral U-shaped transformer for image restoration. IEEE/CVFConference on Computer Vision and Pattern Recognition Proceedings. Los Alamitos, CA: IEEE Computer Society. 17683-17693.
- Lu, J. 2023. AdaSmooth: An adaptive learning rate method based on effective ratio. Sentiment analysis and deep learning. Singapore: Springer Nature Singapore. 273-293.
- Girshick, R. B. 2015. Fast R-CNN. 9 p. Available at: https://arxiv.org/abs/1504.08083 (accessed April 28, 2024).
- Dabov, K., A. Foi, V. Katkovnik, and K. Egiazarian. 2007. Image denoising by sparse 3-D transform-domain collaborative filtering. IEEE T. Image Process. 16:2080-2095.
- Zhang, K., W. Zuo, Y. Chen, D. Meng, and L. Zhang. 2017. Beyond a Gaussian denoiser: Residual learning of deep CNN for image denoising. IEEE T. Image Process. 26(7):3142-3155. doi: 10.1109/TIP2017.2662206.
[+] About this article
Title
IMAGE DECOMPOSITION WITH DISCRETE WAVELET TRANSFORM TO DESIGN A DENOISING NEURAL NETWORK
Journal
Informatics and Applications
2024, Volume 18, Issue 2, pp 60-71
Cover Date
2024-06-20
DOI
10.14357/19922264240209
Print ISSN
1992-2264
Publisher
Institute of Informatics Problems, Russian Academy of Sciences
Additional Links
Key words
neural networks; deep learning; image denoising; image processing
Authors
A. S. Kovalenko
Author Affiliations
Institute of Mathematics, Mechanics, and Computer Science named after 1.1. Vorovich, Southern Federal University, 105/42 Bolshaya Sadovaya Str., Rostov-on-Don 344006, Russian Federation
|