Browsing by Author "Udovyk, Iryna M."

Now showing 1 - 3 of 3

Applying Machine Learning Techniques to Analyze Forest Fire Impacts on Sentinel-2 Imagery across Ukraine
(Український державний університет науки і технологій, ННІ ≪Інститут промислових та бізнес технологій≫, ІВК ≪Системні технології≫, Дніпро, 2026) Hnatushenko Viktoriia V.; Udovyk, Iryna M.; Heipke, Christian; Hnatushenko, Maksym V.
ENG: Forest fires pose severe ecological and socio-economic threats, necessitating efficient tools for rapid damage assessment. This study presents a machine learning approach for detecting burnt forest areas in Ukraine using multispectral Sentinel-2 imagery. A new manually annotated dataset was developed for training semantic segmentation models, addressing the scarcity of open data for the region. The proposed convolutional neural network, based on an encoder–decoder architecture with Xception blocks, effectively captures spectral patterns associated with fire damage. Experiments conducted on Sentinel-2 Level-2A imagery of the Kinburn Peninsula (October 2022) demonstrate high detection performance, achieving an Intersection over Union (IoU) of 95%. The results confirm the model’s capability for accurate burnt-area mapping and highlight its potential for broader applications in regional fire monitoring and environmental management.
A Neural Network Approach to Semantic Segmentation of Vehicles in Very High Resolution Images
(National University "Zaporizhzhia Polytechnic", Zaporizhzhia, 2025) Kashtan, Vita Yu.; Hnatushenko, Volodymyr V.; Udovyk, Iryna M.; Kazymyrenko, O. V.; Radionov, Y. D.
ENG: Context. The semantic segmentation of vehicles in very high resolution aerial images is essential in developing intelligent transportation systems. It allows for the automation of real-time traffic management and the detection of congestion and emergencies. Objective. This work aims to develop and evaluate the effectiveness of a neural network approach to semantic segmentation in very high resolution aerial images, which provides high detail and correct reproduction of object boundaries. Method. The DeepLab architecture with ResNet-101 as a backbone is used for gradient preservation and multiscale feature analysis. We trained on DOTA data and retrained on specialized sets with classes: vehicles, green areas, buildings, and roads. A loss function based on the Dice coefficient was applied to reduce the imbalance of classes. It effectively solves the class imbalance problem and improves the accuracy of segmenting objects of different sizes. Using ResNet-101 instead of Xception in the backbone network allows us to maintain the gradient as the network depth increases. Results. Experimental studies have confirmed the effectiveness of the proposed approach, which achieves a segmentation accuracy of more than 90%, outperforming existing analogs. The use of multiscale feature analysis allows for preserving the texture features of objects, reducing false classifications. A comparative study with U-Net, SegNet, FCN8s, and other methods confirms the higher performance of the proposed approach in terms of mIoU (82.3%) and Pixel Accuracy (95.1%). Conclusions. The experiments confirm the effectiveness of the proposed method of semantic segmentation of vehicles in ultrahigh spatial resolution images. Using DeepLab v3+ResNet-101 significantly improves the quality of vehicle segmentation in an urbanized environment. Excellent metric performance makes it promising for infrastructure monitoring and traffic planning tasks. Further research will focus on adapting the model to new datasets.
Simulation-Driven Assessment of Cryptographic Algorithms for Resource-Constrained Infocommunication Networks
(Dnipro University of Technology, Dnipro, 2025) Laktionov, I. S.; Hnatushenko, Volodymyr V.; Udovyk, Iryna M.; Olevskyi, V. I.
ENG: Purpose. To conduct a multi-criteria evaluation and analysis of the performance of encryption algorithms that may be potentially resistant to contemporary cyberattacks, including quantum attacks. The evaluation takes into account the ability of the algorithms to be deployed on devices with limited computational resources within the infocommunication networks during the transmission of information messages. Methodology. Software implementation, testing and validation of selected cryptographic algorithms based on Python, considering the impact of limited resources and destabilising factors, such as signal noise components, based on computer experiments were applied. The performance of the studied cryptographic algorithms was analysed using statistical data processing methods and a multi-criteria evaluation approach. Findings. The symmetric algorithms AES-256-GCM and ChaCha20-Poly1305 demonstrated the highest accuracy in signal recovery following encryption and decryption (MSE ranges from 1.95 · 10-6 to 5.12 · 10-5). The time taken to encrypt and decrypt I/Q signals using symmetric algorithms was found to be around 2.5 times faster than that required by the Kyber family. Computer experiments confirmed the existence of a trade-off between processing speed and security level. Symmetric algorithms are optimal for scenarios with critical processing speed requirements. However, Kyber provides greater protection reliability, albeit at the cost of additional resources. The correctness of the proposed computer model, which enables the computational and information-functional characteristics of cryptographic algorithms to be evaluated, has been proven. Originality. Patterns of the destabilising influence of signal-to-noise ratio indicators and signal length on the accuracy of digital signal recovery after encryption have been established for different cryptographic algorithms (AES, ChaCha20 and the Kyber) in the context of their use in resource-constrained infocommunication systems. Practical value. Implementing the computer model proved its suitability for studying cryptographic algorithms in resource-constrained environments, as well as its potential for improving information security protocols and selecting optimal algorithms based on processing speed requirements and desired security levels.