Secure Fingerprint Authentication Using Machine Learning and Deep Learning

This paper presented a comprehensive, security-centric comparative evaluation of CNN, Hybrid CNN, SVM with HOG, and Random Forest for fingerprint authentication. Under controlled SOCOFing conditions, the CNN achieves the best balanced profile (97.36% accuracy, FAR 8.11%, FRR 0.82%) while SVM delivers the strongest security with FAR 5.60%. The Hybrid CNN's elevated FAR (16.89%) despite regularization demonstrates that model complexity without proportionate data impairs security performance. The Random Forest's unacceptable FRR (26.46%) disqualifies it from deployment. External cross-dataset validation reveals universal distributional sensitivity, confirming that benchmark performance cannot proxy for real-world deployment robustness. Complementary model family profiles provide practical design guidance for hybrid deployment architectures. Future research directions include multi-sensor training, domain adaptation, adversarial training, threshold optimization, and integration with multimodal biometric frameworks for cloud and IoT environments. The experimental methodology and findings provide a reproducible, evidence-based foundation for practitioner model selection in security-sensitive fingerprint authentication deployments.

[1] A. Alduhailan, N. H. Kamarudin, S. N. H. S. Abdullah, and A. Dau, "Deep Learning in Biometric Authentication: Challenges, Recent Advancements, and Future Trends," Journal of Advances in Information Technology, vol. 16, no. 4, pp. 458-477, 2025. [2] M. M. Rahman, T. I. Mishu, and M. A. A. Bhuiyan, "Performance Analysis of a Parameterized Minutiae-Based Approach for Securing Fingerprint Templates," Journal of Information Security and Applications, vol. 67, p. 103209, 2022. [3] J. Saleem, U. Raza, M. Hammoudeh, and W. Holderbaum, "Machine Learning-Enhanced Attribute-Based Authentication for Secure IoT Access Control," Sensors, vol. 25, no. 9, p. 2779, 2025. [4] R. Garg, G. Singh, A. Singh, and M. P. Singh, "Fingerprint Recognition using Convolution Neural Network with Inversion and Augmented Techniques," Systems and Soft Computing, vol. 6, p. 200106, 2024. [5] A. Hussain, F. Murshed, M. N. Alandoli, and G. Aljafari, "A Hybrid Deep Learning Approach for Secure Biometric Authentication Using Fingerprint Data," Computers, vol. 14, no. 5, p. 178, 2025. [6] D. Maiti, M. Basak, and D. Das, "Fingerprint Biometric: Confronting Challenges, Embracing Evolution, and Extending Utility - A Review," Journal of Engineering Research and Sciences, vol. 3, no. 9, pp. 26-60, 2024. [7] N. Dalal and B. Triggs, "Histograms of Oriented Gradients for Human Detection," Proc. IEEE CVPR, pp. 886-893, 2005. [8] J. Rachel and E. Devarasan, "Robust Contactless Fingerprint Authentication Using Dolphin Optimization and SVM Hybridization," Frontiers in Big Data, vol. 8, 2025. [9] Sonal, A. Singh, and C. Kant, "Optimized Hybrid SVM-RF Multi-Biometric Framework for Enhanced Authentication," PeerJ Computer Science, vol. 11, p. e2699, 2025. [10] R. F. Nogueira, R. de A. Lotufo, and R. C. Machado, "Fingerprint Liveness Detection Using Convolutional Neural Networks," IEEE Trans. Information Forensics and Security, vol. 11, no. 6, pp. 1206-1213, 2016. [11] H. Li and R. Ramachandra, "Deep Learning Based Fingerprint Presentation Attack Detection: A Comprehensive Survey," arXiv:2305.17522, 2023.