ENHANCING ACCURACY AND EFFICIENCY IN ONLINE MULTI-OBJECT TRACKING VIA DEEP LEARNING: APPLICATIONS IN CYBERSECURITY
- Authors
-
-
Dilmurod Mirzaaxmedov
Department of Digital Economy, Uzbekistan Tashkent State University of Economics
Author
-
- Keywords:
- Multi-Object Tracking; Deep Learning; Cybersecurity; Intrusion Detection; Anomaly Detection; Transformer Networks; Network Traffic Analysis.
- Abstract
-
Modern cybersecurity systems face a growing challenge: simultaneously monitoring multiple threats, anomalous behavioral patterns, and malicious network entities across large-scale, dynamic environments. This paper introduces a novel framework Cyber MOT that adapts the methodological foundations of Online Multi-Object Tracking (MOT) to the domain of cybersecurity threat detection and network intrusion monitoring. By establishing a rigorous structural analogy between visual object tracking and the persistent tracking of cyber threat actors across sequential network telemetry, we demonstrate that state-of-the-art deep learning architectures including transformer-based association models and dual-stream appearance-motion encoders can be effectively repurposed for tracking lateral movement, advanced persistent threats (APTs), and coordinated attack campaigns within enterprise-scale networks. Experimental evaluations conducted on both standard MOT benchmarks and simulated network intrusion datasets confirm that CyberMOT achieves superior tracking accuracy, substantially reduced identity-switch errors, and improved real-time processing efficiency compared to conventional intrusion detection baselines. The proposed framework introduces an identity-consistency loss function designed to explicitly penalize tracking failures attributable to adversarial identity obfuscation tactics,such as IP rotation and MAC address spoofing. Results indicate that deep learning-driven MOT paradigms represent a promising and underexplored frontier for advancing next-generation cybersecurity architectures, warranting further investigation.
- References
-
1. Bewley, A., Ge, Z., Ott, L., Ramos, F., & Upcroft, B. (2016). Simple online and realtime tracking. Proceedings of the IEEE International Conference on Image Processing (ICIP), 3464–3468.
2. Moustafa, N., & Slay, J. (2015). UNSW-NB15: A comprehensive data set for network intrusion detection systems. Proceedings of the Military Communications and Information Systems Conference (MilCIS), 1–6.
3. Wojke, N., Bewley, A., & Paulus, D. (2017). Simple online and realtime tracking with a deep association metric. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 3645–3649.
4. Meinhardt, T., Kirillov, A., Leal-Taixé, L., & Feichtenhofer, C. (2022). TrackFormer: Multi-object tracking with transformers. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 8844–8854.
5. Zeng, F., Dong, B., Zhang, Y., Wang, T., Zhang, X., & Wei, Y. (2022). MOTR: End-to-end multiple-object tracking with transformer. Proceedings of the European Conference on Computer Vision (ECCV), 659–675.
6. Mirsky, Y., Doitshman, T., Elovici, Y., & Shabtai, A. (2018). Kitsune: An ensemble of autoencoders for online network intrusion detection. Proceedings of the Network and Distributed System Security Symposium (NDSS).
7. Lo, W. W., Yang, X., & Wang, Y. (2022). An xgboost-based intrusion detection system using graph neural network. Proceedings of the IEEE International Conference on Trust, Security and Privacy in Computing and Communications, 1–8.
8. Sharafaldin, I., Habibi Lashkari, A., & Ghorbani, A. A. (2018). Toward generating a new intrusion detection dataset and intrusion traffic characterization. Proceedings of the International Conference on Information Systems Security and Privacy (ICISSP), 108–116.
- Downloads
- Published
- 2026-06-09
- Issue
- Vol. 2 No. 6 (2026)
- Section
- Articles
- License
-
This work is licensed under a Creative Commons Attribution 4.0 International License.
