METHOD OF SEMANTIC SEGMENTATION OF VIDEO IMAGES USING U-NET NEURAL NETWORK

Authors

  • Andrii Kostromytskyi Kharkiv National University of Radio Electronics, Kharkiv, Ukraine
  • Valery Bezruk Kharkiv National University of Radio Electronics, Kharkiv, Ukraine
  • Іgor Malinin V. N. Karazin Kharkiv National University, Kharkiv, Ukraine
  • Artem Panchuk Kharkiv National University of Air Force named after I. Kozhedub, Kharkiv, Ukraine
  • Bohdan Breslavets Lviv, Ukraine

DOI:

https://doi.org/10.18372/2310-5461.68.19045

Keywords:

video image, selective encryption, semantic segmentation, U-Net neural network, video information protection, machine learning, encoding, compression, cryptographic processing, video surveillance, resource optimization

Abstract

In today's information environment, video information has become one of the most massive and valuable types of data. Surveillance cameras, mobile devices, drones and online platforms generate huge amounts of visual data every day, a significant part of which contains confidential or personal information. Traditional approaches to protection, such as full encryption of the video stream, provide a high level of security, but have significant limitations: they require significant computing resources, increase transmission delays and do not allow for effective real-time operation. The article considers the scientific and applied problem of selective protection of video information in conditions of limited computing resources, which is relevant for video surveillance systems, medical records and other areas where it is necessary to ensure the confidentiality of only critically important areas of the image. The purpose of the research is to create a software tool for dynamic identification and encryption of significant fragments of a video image using semantic segmentation methods and machine learning technologies. The proposed system is based on the architecture of the U-Net neural network, which allows to accurately identify areas requiring cryptographic protection and apply selective encryption to them. The Stanford Background dataset was used as an experimental basis, the test results confirm the effectiveness of the chosen approach. The presented methods can be integrated into practical systems with limited resources and increased requirements for performance and security.

Author Biographies

Andrii Kostromytskyi, Kharkiv National University of Radio Electronics, Kharkiv, Ukraine

Candidate of Economic Sciences, Associate Professor

Valery Bezruk, Kharkiv National University of Radio Electronics, Kharkiv, Ukraine

Doctor of Technical Sciences, Professor

Іgor Malinin, V. N. Karazin Kharkiv National University, Kharkiv, Ukraine

Student

Artem Panchuk, Kharkiv National University of Air Force named after I. Kozhedub, Kharkiv, Ukraine

Senior lecturer

Bohdan Breslavets, Lviv, Ukraine

Senior Software Engineer

References

Ільяшов О. А., Бурячок В. Л. До питання захисту інформаційно-телекомунікаційної сфери від стороннього кібернетичного впливу // Наука и оборона. 2010. № 4. С. 35–41.

Image file type and format guide // MDN Web Docs. Media types and formats for image, audio, and video content. URL: https://developer.mozilla.org/en-US/docs/Web/Media/Guides/Formats/Image_types (дата звернення: 14.10.2025).

Codecs in common media types. MDN Web Docs. Media types and formats for image, audio, and video content. URL: https://developer.mozilla.org/en-US/docs/Web/Media/Guides/Formats/codecs_parameter (дата звернення: 14.10.2025).

Method of Coding Video Images Based on Meta-Determination of Segments / V. Barannik et al. // Digital Ecosystems: Interconnecting Advanced Networks with AI Applications : TCSET 2024. Cham : Springer, 2024. Vol. 1198. P. 566–589. DOI: https://doi.org/10.1007/978-3-031-61221-3_27.

Gimp raster to vector // Kitchenfity. URL: https://kitchenfity.weebly.com/blog/gimp-raster-to-vector (дата звернення: 14.10.2025).

Fotovvat A., Wahid K. A. Selective Encryption of VVC Encoded Video Streams for the Internet of Video Things // arXiv. 2021. URL: https://arxiv.org/pdf/2103.14844 (дата звернення: 14.10.2025).

PyTorch: An Imperative Style, High-Performance Deep Learning Library / A. Paszke et al. // arXiv. 2019. URL: https://arxiv.org/pdf/1912.01703 (дата звернення: 14.10.2025).

Barannik V., Karpenko S. Method of the 3-D image processing // Modern Problems of Radio Engineering, Telecommunications and Computer Science (TCSET): proceedings of Intern. Conf. (Lviv-Slavsko, Ukraine, February 19-23, 2008). Lviv-Slavsko, 2008. P. 378–380.

Chaudhari R. E., Dhok S. B. Review of Fractal Transform based Image and Video Compression // International Journal of Computer Applications. 2012. Vol. 57, No. 19. URL: https://www.ijcaonline.org/archives/volume57/number19/9223-3782/ (дата звернення: 14.10.2025).

Compression method in terms of ensuring the fidelity of video images in infocommunication networks / V. Barannik et al. // Radioelectronic and Computer Systems. 2022. No. 4 (100). P. 10–24. DOI: 10.32620/reks.2022.5/09.

Ficzere P. The possibilities of intelligent manufacturing methods. // Research Gate. 02.05.2020. URL: https://www.researchgate.net/publication/357738098_The_possibilities_of_intelligent_manufacturing_methods (дата звернення: 14.10.2025).

Method of Mini Segments Encoding in Difference Space Using Haar Wavelet / V. Barannik et al. // 2023 IEEE 5th International Conference on Advanced Information and Communication Technologies (AICT). Lviv, Ukraine, 2023. P. 1–4. DOI: 10.1109/AICT61584.2023.10452674.

Gould S., Fulton R., Koller D. Decomposing a Scene into Geometric and Semantically Consistent Regions. // ResearchGate. URL: https://www.researchgate.net/publication/224135950 (дата звернення: 14.10.2025).

A Method of Scrambling for the System of Cryptocompression of Codograms Service Components / V. Barannik et al. // Emerging Networking in the Digital Transformation Age : TCSET 2022. Cham : Springer, 2023. Vol. 965. P. 444–459. (Lecture Notes in Electrical Engineering).DOI: https://doi.org/10.1007/978-3-031-24963-1_26.

Метод аналізу взаємодії параметрів QOE та QOS на основі алгоритмів керування машинами / Р. Одарченко та ін. // Наукоємні технології. 2022. № 4 (56). С. 305–316. DOI: https://doi.org/10.18372/2310-5461.56.17130.

Krawczyk H., Bellare M., Canetti R. HMAC: Keyed-Hashing for Message Authentication. RFC 2104. 1997. URL: https://www.rfc-editor.org/rfc/rfc2104.html (дата звернення: 14.10.2025).

Метод кластеризації послідовності трансформант за структурними ознаками їх спектрально-параметричного опису / В. В. Бараннік та ін. // Наукоємні технології. 2024. № 2 (62). С. 185–192. DOI: https://doi.org/10.18372/2310-5461.62.18712.

Ranjith B., Raghu N. CryptoGAN: a new frontier in generative adversarial network-driven image encryption // IAES International Journal of Artificial Intelligence. 2024. Vol. 13, No. 4. P. 4813–4821. DOI: https://doi.org/10.11591/ijai.v13.i4.pp4813-4821.

Метод стиснення кластеризованих трансформант на основі блочного кодування з локально-монотонним визначенням довжини / В. В. Бараннік та ін. // Наукоємні технології. 2024. № 3 (63). С. 274–281. DOI: https://doi.org/10.18372/2310-5461.63.18971.

Strogatz S. H. Nonlinear Dynamics and Chaos. CRC Press, 2018, 513 p.

Recommendation for Password-Based Key Derivation Part 1: Storage Applications / M. Sönmez Turan et al. // NIST Special Publication 800-132. URL: https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-132.pdf (дата звернення: 14.10.2025).

Tang L. Methods for Encrypting and Decrypting MPEG Video Data Efficiently // Carnegie Mellon University. URL: https://dl.acm.org/doi/pdf/10.1145/244130.244209 (дата звернення: 14.10.2025).

A Comparative Analysis on Blockchain versus Centralized Authentication Architectures for IoT-Enabled Smart Devices in Smart Cities: A Comprehensive Review, Recent Advances, and Future Research Directions / A. M. Usman et al. // Sensors (MDPI). 2022. Vol. 22, Iss. 14. URL: https://www.mdpi.com/1424-8220/22/14/5168 (дата звернення: 14.10.2025).

Критерії вибору спектрально-ефективних сигналів у бездротових інформаційних мережах / В. Козловський та ін. // Наукоємні технології. 2022. № 4 (56). С. 273–286. DOI: https://doi.org/10.18372/2310-5461.56.17125 (дата звернення: 14.10.2025).

Xiaowu Li, Huiling Peng. Chaotic medical image encryption method using attention mechanism fusion ResNet model // Frontiers in Neuroscience. 2023. Vol. 17. Art. 1226154. DOI: https://doi.org/10.3389/fnins.2023.1226154.

Technology of Sliding Coding of Uneven Diagonal Sequences in Two-Dimensional Spectral Space of Transformants / V. V. Barannik et al. // Visnyk NTUU KPI Seriia - Radiotekhnika Radioaparatobuduvannia. 2023. No. 94. P. 13–23. DOI: 10.20535/RADAP.2023.94.13-23.

Saving Elements Methods for Service Components of Images Cryptocompression Codograms / V. V. Barannik et al. // Visnyk NTUU KPI Seriia - Radiotekhnika Radioaparatobuduvannia. 2023. No. 92. P. 28–40. DOI: 10.20535/RADAP.2023.92.28-40.

Significant Microsegment Transformants Encoding Method to Increase the Availability of Video Information Resource / V. Barannik et al. // IEEE Advanced Trends in Information Theory (ATIT) : proceedings of 2nd Intern. Conf. (Kyiv, Ukraine, November 25-27, 2020). Kyiv, 2020. P. 52–56. DOI: 10.1109/ATIT50783.2020.9349256.

Barannik V., Shiryaev A. Quadrature compression of images in polyadic space // Modern Problems of Radio Engineering, Telecommunications and Computer Science (TCSET'2012) : proceedings of Intern. Conf. (Lviv-Slavske, Ukraine, February 21–24, 2012). Lviv-Slavske, 2012. P. 422.

Development of Adaptive Arithmetic Coding Method to the Sequence of Bits / V. Barannik et al. Engineer of the XXI Century: EngineerXXI 2018. Cham: // Springer, 2020. Vol. 70. (Mechanisms and Machine Science). DOI: https://doi.org/10.1007/978-3-030-13321-4_18.

Video Segments Stamping Method Saving Their Reliability in the Spectral-Cluster Space / V. V. Barannik et al. // Visnyk NTUU KPI Seriia - Radiotekhnika Radioaparatobuduvannia. 2023. No. 92. P. 41–53. DOI: 10.20535/RADAP.2023.92.41-53.

A novel image encryption method based on improved two-dimensional logistic mapping and DNA computing / Y. Chen et al. // Frontiers in Physics. 2024. URL: https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2024.1469418/full (дата звернення: 14.10.2025).

Barannik V., Khimenko V., Barannik N. Method of indirect information hiding in the process of video compression. // Radioelectronic and Computer Systems. 2021. No. 4. P. 119–131. DOI: https://doi.org/10.32620/reks.2021.4.

Downloads

Published

2026-02-10

How to Cite

Kostromytskyi, A., Bezruk, V., Malinin І., Panchuk, A., & Breslavets, B. (2026). METHOD OF SEMANTIC SEGMENTATION OF VIDEO IMAGES USING U-NET NEURAL NETWORK. Science-Based Technologies, 68(4), 486–496. https://doi.org/10.18372/2310-5461.68.19045

Issue

Vol. 68 No. 4 (2025)

Section

Electronics, electronic communications, instrumentation and radio engineering

License

The scientific journal adheres to the principles of Open Access and provides free, immediate, and permanent access to all published materials without financial, technical, or legal barriers for readers.

All articles are published in Open Access under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

Copyright

Authors who publish their works in the journal:

  • retain the copyright to their publications;

  • grant the journal the right of first publication of the article;

  • agree to the distribution of their materials under the CC BY 4.0 license;

  • have the right to reuse, archive, and distribute their works (including in institutional and subject repositories), provided that proper reference is made to the original publication in the journal.