Journal of Soft Computing Paradigm
view PDF
How to Cite
P., Pushpa, Varsha A., and Mahalakshmi S. 2026. “Reliability-Weighted Softmax Fusion for Continuous Multimodal Authentication”. Journal of Soft Computing Paradigm 8 (3): 310-28. https://doi.org/10.36548/jscp.2026.3.009.
Keywords
Continuous authentication
Multimodal Authentication
Reliability-Aware Fusion
Softmax Fusion
Wi-Fi RSSI (Received Signal Strength Indicator)
Ambient Audio
Mouse Dynamics
Abstract
Continuous authentication refers to the continued process of identifying individuals after the login stage. On the other hand, the individual behaviors and context-based modalities can be inaccurate based on the operating condition. This research work proposes a lightweight multimodal continuous authentication system using laptop Wi-Fi RSSI, smartphone Wi-Fi RSSI, ambient audio after filtering, and mouse dynamics. Every modality gives a match score as well as instant reliability depending on the signal strength, data age, or interaction frequency. Reliability-Weighted Softmax Fusion (Softmax-R) is adopted for a dynamic weighting scheme, where the weight of each modality is calculated based on its reliability. The proposed continuous authentication system was tested against eight fusion approaches by utilizing the real datasets of five users including both genuine and impostor sessions. The Softmax-R approach has resulted in a Equal Error Rate (EER) of 3.2% and Area Under Curve (AUC) of 0.986, which is noticeably better than all the other baseline approaches. The ablation study revealed that all four modalities complement each other in the process.References
- AlQahtani, Ali Abdullah S., Thamraa Alshayeb, Mahmoud Nabil, and Ahmad Patooghy. ”Leveraging Machine Learning for Wi-Fi-based Environmental Continuous Two-Factor Authentication.” IEEE Access 2024, vol. 12, 13277-13289.
- Wang, Tao, Dandan Yang, Shunqing Zhang, Yating Wu, and Shugong Xu. ”Wi-Alarm: Low-Cost Passive Intrusion Detection Using WiFi.” Sensors 2019, vol. 19, no. 10, 2335.
- Yang, Xiaopeng, Qiang Wang, Hao Li, Ziyi Jia, Lizhe Liu, and Xiaobo Guo. ”A Lightweight Multi-Factor Continuous Authentication Method Based on IBE.” 24th International Conference on Communication Technology (ICCT), IEEE, 2024, 360-365.
- Feng, Huan, Kassem Fawaz, and Kang G. Shin. ”Continuous Authentication for Voice Assistants.” In Proceedings of the 23rd Annual International Conference on Mobile Computing and Networking 2017, 343-355.
- Pusara, Maja, and Carla E. Brodley. ”User Re-Authentication via Mouse Movements.” In Proceedings of the ACM workshop on Visualization and Data Mining for Computer Security 2004, 1-8.
- Dwivedi, Chaitanya, Divyanshu Kalra, Divesh Naidu, and Swati Aggarwal. ”Keystroke Dynamics Based Biometric Authentication: A Hybrid Classifier Approach.” IEEE Symposium Series on Computational Intelligence (SSCI) 2018, 266-273.
- Ahmed, Ahmed Awad E., and Issa Traore. ”A New Biometric Technology Based on Mouse Dynamics.” IEEE Transactions on Dependable and Secure Computing 2007, vol. 4, no. 3, 165-179.
- Niinuma, Koichiro, Unsang Park, and Anil K. Jain. ”Soft Biometric Traits for Continuous User Authentication.” IEEE Transactions on Information Forensics and Security 2010, vol. 5, no. 4, 771-780.
- Ammour, Nassim, Yakoub Bazi, and Naif Alajlan. ”Multimodal Approach for Enhancing Biometric Authentication.” Journal of Imaging 2023, vol. 9, no. 9, 168.
- Kittler, Josef, Mohamad Hatef, Robert PW Duin, and Jiri Matas. ”On Combining Classifiers.” IEEE Transactions on Pattern Analysis and Machine Intelligence 1998, vol. 20, no. 3, 226-239.
- Shafer, Glenn. ”A Mathematical Theory of Evidence.”, Princeton University Press 2020, 1-314.
- L. Kuncheva, Combining Pattern Classifiers: Methods and Algorithms. Wiley, 2004. https://doi.org/10.1002/0471660264
- Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. ”Attention is All You Need.” Advances in Neural Information Processing Systems 2017, 30.