METHOD FOR DETERMINING THE EFFICIENCY OF ENSURING FUNCTIONAL STABILITY OF MULTIMODAL INFORMATION SYSTEMS FOR IMAGE RECOGNITION BASED ON EXPERT EVALUATION

DOI: 10.31673/2412-4338.2026.019015

Abstract

This paper proposes a method for determining the effectiveness of ensuring the functional stability of multimodal information systems for image recognition operating under conditions of uncertainty, interference, and partial degradation of sensor data. The relevance of the study is driven by the need for comprehensive evaluation of such systems functioning in dynamic environments with limited resources and unstable communication channels. The proposed approach is based on the formation of a profile of partial indicators that characterize key aspects of functional stability, including recognition accuracy, decision-making latency, robustness to noise, resistance to the loss of sensory modalities, energy efficiency, and adaptability of machine learning algorithms. The method involves interval expert evaluation, which allows taking into account the uncertainty and variability of expert judgments without strict normalization of input data. The layering method is applied to aggregate the results, ensuring the correct combination of heterogeneous indicators and minimizing information loss. The obtained results are approximated using trapezoidal membership functions, which enables the formalization of fuzzy assessments and the derivation of an integral efficiency indicator. The conducted computational experiment confirmed the possibility of obtaining a generalized quantitative criterion for evaluating the functional stability of the system. The practical significance of the results lies in the applicability of the proposed method for comparative analysis of alternative architectural and algorithmic solutions in multimodal image recognition systems and wireless sensor networks.

Keywords: functional stability, multimodal pattern recognition information system, wireless sensor network, expert evaluation, fuzzy sets, integral index

References

1. Boiko, Y. M., Druzhynin, V. A., & Toliupa, S. V. (2018). Theoretical aspects of improving noise immunity and efficiency of signal processing in radio engineering devices and telecommunication systems under interference conditions. Kyiv: Logos.
2. Boiko, Y. M., Polikarovskykh, O. I., Tkachuk, V. P., & Karpova, L. V. (2023). Software-defined systems for transmission, reception, and information processing. Khmelnytskyi: KhNU.
3. Toliupa, S. V. (2012). Method of multicriteria analysis of efficiency of functioning and ensuring information security of infocommunication systems. Information Protection, 14(3), 81–86. https://doi.org/10.18372/2410-7840.14.33715
4. Hnatiienko, O. (2025). Scheme for determining the effectiveness of ensuring functional stability of organizational systems. In V. Ye. Snytiuk (Ed.), Intelligent Solutions-S: Proceedings of the International Symposium (pp. 80–82). Kyiv: Karavela.
5. Hnatiienko, H. M., & Hnatiienko, O. H. (2024). Method for determining fuzzy values of the relative importance of alternative characteristics using the layering method. Scientific Bulletin of Uzhhorod University. Series: Mathematics and Informatics, 45(2), 172–187. https://doi.org/10.24144/2616-7700.2024.45(2).172-187
6. Pan, B., Hirota, K., Jia, Z., & Dai, Y. (2023). A review of multimodal emotion recognition from datasets, preprocessing, features, and fusion methods. Neurocomputing, 561, 126866. https://doi.org/10.1016/j.neucom.2023.126866
7. Zhang, S., Yang, Y., Chen, C., Zhang, X., Leng, Q., & Zhao, X. (2024). Deep learning-based multimodal emotion recognition from audio, visual, and text modalities: A systematic review. Expert Systems with Applications, 237, 121692. https://doi.org/10.1016/j.eswa.2023.121692
8. Chen, X., et al. (2022). Analysis of multimodal data fusion from an information theory perspective. Information Sciences, 608, 1030–1045.
9. Abdullah, M., et al. (2023). Enhanced multimodal biometric recognition systems based on deep learning techniques. Sensors, 23.

1. Liu, W., Qiu, J., Zheng, W., & Lu, B. (2021). Comparing recognition performance and robustness of multimodal deep learning models for emotion recognition. IEEE Transactions on Cognitive and Developmental Systems, 14, 715–729.