Application of machine learning methods to restore damaged objects in a photograph
- 1,2,3,4,5 Financial University under the Government of the Russian Federation (Moscow, Russia)
Abstract
Restoring damaged objects in photographs is an urgent task in the field of image processing and computer vision. Traditional methods often do not provide sufficient quality for serious injuries, so more effective approaches are required.
Improve the quality of restoration of damaged objects in photographs by applying machine learning methods, including deep neural networks, convolutional neural networks, and generative adversarial networks.
An analysis of previous studies has been carried out, confirming the prospects of using machine learning to solve this problem. A data source containing pairs of corrupted and original images has been selected; their preprocessing and augmentation have been performed. A model based on a neural network was built, its training and comparison with other methods were carried out. It has been experimentally shown that the proposed approach provides high accuracy of reconstruction and surpasses traditional image processing methods.
The results obtained can be used in medicine to repair damaged images, in the restoration of old photographs, in industry to assess defects, as well as in other areas where automatic restoration of damaged objects in images is required.
- Received
- 04.09.2026
- Approved after review
- 14.09.2026
- Accepted
- 01.10.2026
How to cite
Киселев А.Р., Окунева Э.А., Кочкаров Р.А., Сальников Е.А., Рябов П.Е. Применение методов машинного обучения для восстановления поврежденных объектов на фотографии // Динамика сложных систем - XXI век. 2026. Т. 20. № 5. С. 31−39. DOI: https://doi.org/10.18127/j19997493-202605-05
References
- Gonzalez R.C., Woods R.E. Digital Image Processing. 4th ed. Pearson. 2018.
- Koklev P.S. Otsenka stoimosti kompanii s ispolzovaniem metodov mashinnogo obucheniya. Finansy: teoriya i praktika. 2022. T. 26. № 5. S. 132–148. (in Russian).
- Dozhdikov A.V. Prognozirovanie finansovoy effektivnosti rossiyskogo kinematografa s pomoshchyu mnogofaktornoy ansamblevoy modeli mashinnogo obucheniya, trenirovannoy na dannykh proshlykh periodov. Finansy: teoriya i praktika. 2025. T. 29. № 6. S. 243–268. (in Russian).
- Goodfellow I., Bengio Y., Courville A. Deep Learning. MIT Press, 2016.
- Zhang K., Ren W., Luo W. et al. Deep Image Deblurring: A Survey. International Journal of Computer Vision. 2022. Vol. 130. P. 2103–2130.
- LeCun Y., Bengio Y., Hinton G. Deep learning. Nature. 2015. V. 521. P. 436–444.
- Huang G., Liu Z., van der Maaten L., Weinberger K.Q. Densely Connected Convolutional Networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2017. P. 2261–2269.
- Zhu J.-Y., Park T., Isola P., Efros A.A. Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks. Proceedings of the IEEE International Conference on Computer Vision (ICCV). 2017. P. 2223–2232.
- Mao X.-J., Shen C., Yang Y.-B. Image Restoration Using Convolutional Auto-encoders with Symmetric Skip Connections. Proceedings of the 30th International Conference on Neural Information Processing Systems (NIPS). 2016. P. 2810–2818.
- Zamir S.W., Arora A., Khan S. et al. Restormer: Efficient Transformer for High-Resolution Image Restoration. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2022. P. 5718–5729.
- Liu D., Wen B., Fan Y. et al. Non-Local Recurrent Network for Image Restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2020. V. 42. № 5. P. 1200–1214.
- Loshchilov I., Hutter F. Decoupled Weight Decay Regularization. International Conference on Learning Representations (ICLR), 2019.
- Zamir S.W., Arora A., Khan S., Hayat M., Khan F.S., Yang M.-H., Shao L. Learning Enriched Features for Fast Image Restoration and Enhancement. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2022. V. 44. № 2. P. 1934–1948.
- Lin T.-Y., Maire M., Belongie S. et al. Microsoft COCO: Common Objects in Context. European Conference on Computer Vision (ECCV). 2014. P. 740–755.
- Kochkarov R.A., Borodina A.D. Sistematizatsiya orkhono-eniseyskikh runicheskikh pismennostey i perspektivy ikh avtomatizirovannogo analiza metodami mashinnogo obucheniya. Dinamika slozhnykh sistem – XXI vek. 2025. T. 19. № 5. S. 97–103. DOI: 10.18127/j19997493-202505-11 (in Russian).
- Zvyagin L.S. Voprosy razvitiya i modelirovaniya informatsionno-izmeritelnykh i upravlyayushchikh sistem kak instrumenta razvitiya potentsiala promyshlennykh predpriyatiy. Dinamika slozhnykh sistem – XXI vek. 2025. T. 19. № 5. S. 87–96. DOI: 10.18127/j19997493-202505-10 (in Russian).
- Prokopchina S.V., Zvyagin L.S. Aktualnye napravleniya i perspektivy razvitiya prikladnykh informatsionno-izmeritelnykh sistem dlya resheniya zadach avtomatizatsii monitoringa i upravleniya v slozhnykh tekhnicheskikh kompleksakh. Dinamika slozhnykh sistem – XXI vek. 2025. T. 19. № 5. S. 110–125. DOI: 10.18127/j19997493-202505-13 (in Russian).
- Pelevin P.A., Petrosyan A.A., Chernyshenko V.S., Koneva V.V. Razrabotka sistemy monitoringa ekologii lesoparkov i sokhraneniya deponirovannogo ugleroda s mobilnymi tekhnologiyami, geymifikatsiey i geodannymi. Dinamika slozhnykh sistem – XXI vek. 2025. T. 19. № 5. S. 47–56. DOI: 10.18127/j19997493-202505-06 (in Russian).