TRIFACTOR AI: AN OFFLINE MULTI-FACTOR BIOMETRIC AUTHENTICATION FRAMEWORK INTEGRATING ADAPTIVE LIVENESS AND BEHAVIORAL VERIFICATION
Keywords:
trifactor ai; multi-modal biometric authentication; adaptive liveness detection; dynamic time warping (dtw); crypt hashing.Abstract
The increasing prevalence of digital identity theft and credential-based attacks has exposed fundamental limitations of traditional password-based and two-factor authentication mechanisms. Although biometric authentication offers improved usability and security, most existing systems rely on a single modality, making them vulnerable to spoofing, replay, and behavioral forgery. This paper proposes TriFactor AI, a privacy-preserving, offline, multi-modal authentication framework designed for secure desktop access. The proposed system integrates three independent authentication factors: facial recognition with adaptive liveness detection, dynamic signature verification using Dynamic Time Warping (DTW), and knowledge-based secret-code validation secured through bcrypt hashing. Facial authentication employs ArcFace embeddings combined with randomized challenge–response liveness tasks, including blink and head-pose verification, to mitigate presentation attacks. Behavioral authentication captures temporal handwriting dynamics via a PyQt5-based interface and verifies user identity through DTW-based sequence alignment. All biometric templates and credentials are securely managed within a local SQLite database, ensuring complete data privacy and eliminating reliance on cloud-based processing. Experimental evaluation demonstrates that the facial module achieves an average accuracy of 98.2%, the dynamic signature module 87%, and secret-code verification 100% reliability. Ablation analysis confirms that combining physiological, behavioral, and knowledge-based factors reduces the False Acceptance Rate by more than 60× compared to single-factor authentication, yielding an overall system accuracy of 99.8% with an average authentication latency below 3 seconds. These results provide empirical evidence that security gains from tri-factor biometric fusion are multiplicative rather than additive. The proposed framework establishes a practical, low-latency, and fully local authentication paradigm suitable for privacy-sensitive desktop environments.References
[1]. Kumar, S., Rani, S., Jain, A., Verma, C., Raboaca, M. S., Illés, Z., and Neagu, B. C. (2022). Face spoofing, age, gender and
facial expression recognition using advance neural network architecture-based biometric system. Sensors, 22(14),
5160. https://doi.org/10.3390/s22145160
[2]. Mishra, A. R., Kumar, R., Gupta, V., Prabhu, S., Upadhyay, R., Chhipa, P. C., Rakesh, S., Mokayed, H., Das Chakladar, D.,
De, K., Liwicki, M., Simistira Liwicki, F., and Saini, R. (2024). SignEEG v1.0: Multimodal dataset with
electroencephalography and handwritten signature for biometric systems. Scientific Data, 11(1), 718.
https://doi.org/10.1038/s41597-024-03546-z
[3]. Ammour, N., Bazi, Y., and Alajlan, N. (2023). Multimodal approach for enhancing biometric authentication. Journal
of Imaging, 9(9), 168. https://doi.org/10.3390/jimaging9090168
[4]. Haider, S. A., Ashraf, S., Larik, R. M., Husain, N., Muqeet, H. A., Humayun, U., Yahya, A., Arfeen, Z. A., and Khan, M. F.
(2023). An improved multimodal biometric identification system employing score-level fuzzification of finger texture
and finger vein biometrics. Sensors, 23(24), 9706. https://doi.org/10.3390/s23249706
[5]. Yang, J., Lan, G., Xiao, S., Li, Y., Wen, J., and Zhu, Y. (2022). Enriching facial anti-spoofing datasets via an effective face
swapping framework. Sensors, 22(13), 4697. https://doi.org/10.3390/s22134697
[6]. Faundez-Zanuy, M., Diaz, M., and Ferrer, M. A. (2024). Online signature recognition: A biologically inspired feature
vector splitting approach. Cognitive Computation, 16(1), 265–277. https://doi.org/10.1007/s12559-023-10205-9
[7]. Che, Y., Du, L., Tang, G., and Ling, S. (2024). A biometric identification for multi-modal biomedical signals in geriatric
care. Sensors, 24(20), 6558. https://doi.org/10.3390/s24206558
[8]. Khairnar, S., Gite, S., Kotecha, K., and Thepade, S. D. (2023). Face liveness detection using artificial intelligence
techniques: A systematic literature review and future directions. Big Data and Cognitive Computing, 7(1), 37.
https://doi.org/10.3390/bdcc7010037
[9]. Zhang, S., and Nie, W. (2023). Multi-domain feature alignment for face anti-spoofing. Sensors, 23(8), 4077.
https://doi.org/10.3390/s23084077
[10]. Solomon, E., and Cios, K. J. (2023). FASS: Face anti-spoofing system using image quality features and deep learning.
Electronics, 12(10), 2199. https://doi.org/10.3390/electronics12102199
[11]. Xiao, J., Wang, W., Zhang, L., and Liu, H. (2024). A MobileFaceNet-based face anti-spoofing algorithm for low-quality
images. Electronics, 13(14), 2801. https://doi.org/10.3390/electronics13142801
[12]. Lai, Z., Guo, Y., Hu, Y., Su, W., and Feng, R. (2024). Evaluating and enhancing face anti-spoofing algorithms for light
makeup: A general detection approach. Sensors, 24(24), 8075. https://doi.org/10.3390/s24248075
[13]. Gong, L. Y., Li, X. J., and Chong, P. H. J. (2024). Facial anti-spoofing using “clue maps”. Sensors, 24(23), 7635.
https://doi.org/10.3390/s24237635
[14]. Huang, J., Xue, Y., and Liu, L. (2023). Dynamic signature verification technique for the online and offline
representation of electronic signatures in biometric systems. Processes, 11(1), 190.
https://doi.org/10.3390/pr11010190
[15]. Roszczewska, K., and Niewiadomska-Szynkiewicz, E. (2024). Online signature biometrics for mobile devices.
Sensors, 24(11), 3524. https://doi.org/10.3390/s24113524
[16]. Kothadiya, D., Bhatt, C., Soni, D., Gadhe, K., Patel, S., Bruno, A., and Mazzeo, P. L. (2023). Enhancing fingerprint
liveness detection accuracy using deep learning: A comprehensive study and novel approach. Journal of Imaging, 9(8),
158. https://doi.org/10.3390/jimaging9080158
[17]. Zhang, K., Huang, S., Liu, E., and Zhao, H. (2023). LFLDNet: Lightweight fingerprint liveness detection based on
ResNet and transformer. Sensors, 23(15), 6854. https://doi.org/10.3390/s23156854
[18]. Chen, L., Guo, T., Li, L., Jiang, H., Luo, W., and Li, Z. (2023). A finger vein liveness detection system based on multiscale spatial-temporal map and Light-ViT model. Sensors, 23(24), 9637. https://doi.org/10.3390/s23249637
[19]. Li, Y., Sun, W., Li, Z., and Guo, X. (2025). Face anti-spoofing based on adaptive channel enhancement and intra-class
constraint. Journal of Imaging, 11(4), 116. https://doi.org/10.3390/jimaging11040116
[20]. Xing, H., Tan, S. Y., Qamar, F., and Jiao, Y. (2025). Face anti-spoofing based on deep learning: A comprehensive
survey. Applied Sciences, 15(12), 6891. https://doi.org/10.3390/app15126891
[21]. Li, Y., Li, Y., and Wang, Z. (2025). FaceCloseup: Enhancing mobile facial authentication with perspective distortionbased liveness detection. Computers, 14(7), 254. https://doi.org/10.3390/computers14070254
[22]. Salturk, S., and Kahraman, N. (2024). Deep learning-powered multimodal biometric authentication: Integrating
dynamic signatures and facial data for enhanced online security. Neural Computing and Applications, 36(19), 11311–
11322. https://doi.org/10.1007/s00521-024-09690-2
[23]. Li, C., Li, Z., Sun, J., and Li, R. (2023). Middle-shallow feature aggregation in multimodality for face anti-spoofing.
Scientific Reports, 13(1), 9870. https://doi.org/10.1038/s41598-023-36636-w
[24]. Zanlorensi, L. A., Laroca, R., Lucio, D. R., Santos, L. R., Britto, A. S., and Menotti, D. (2022). A new periocular dataset
collected by mobile devices in unconstrained scenarios. Scientific Reports, 12(1), 17989.
https://doi.org/10.1038/s41598-022-22811-y
[25]. Qi, H., Han, R., Duan, K., Shi, Y., Qi, X., Gao, C., and Ren, L. (2025). A high-performance adaptive fusion network for
face anti-spoofing detection. Scientific Reports, 15(1), 37607. https://doi.org/10.1038/s41598-025-21461-0
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Intelligent Machine Learning and IoT Enabled Applications
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
