FACE AUTHENTICATION SYSTEM USING TRANSFER LEARNING OF NEURAL NETWORKS
DOI:
https://doi.org/10.35546/kntu2078-4481.2025.3.2.39Keywords:
biometric authentication, face recognition, computer vision, convolutional neural networks, transfer learning, information securityAbstract
The use of convolutional neural networks in the process of image analysis has also had an impact on biometric face authentication systems, which play an important role in ensuring information security. Modern models for solving this problem achieve high accuracy and have a complex architecture that requires a lot of time to design. The use of transfer learning helps to simplify the process of training neural network models, while ensuring a sufficiently high level of accuracy. The paper proposes an approach to building face authentication systems, which uses a pre-trained convolutional neural network model, which was adapted to the face authentication task through the use of transfer learning. This approach simplifies the process of designing the neural network architecture and significantly reduces the time for its training. During the study, two convolutional neural network architectures were developed, which differed in the settings of the base model layers for feature extraction, the number of fully connected layers and their parameters, and regularization methods. Two sets of face images were collected, and subsequently they were used to train and test neural network models, graphs of the dependence of the loss function on the number of epochs during training were obtained, and verification was performed on the test subset. To verify the adequacy of the solution, a prototype of the system was developed and tested, the obtained authentication accuracy was high on both sets of images, so the proposed approach can be used in real biometric face authentication systems, and can also potentially be supplemented with security mechanisms, such as a module for checking face spoofing. Further research in this direction is proposed to be directed at improving the methods of pre-processing images from the camera and supplementing the training dataset.
References
Kortli Y., Jridi M., Al Falou A., Atri M. Face recognition systems: A survey. Sensors. 2020. Vol. 20, no. 2. P. 342. URL: https://doi.org/10.3390/s20020342
Turk M. A., Pentland, A. Face recognition using eigenfaces. Computer Vision and Pattern Recognition : Proceedings of the IEEE Computer Society Conference, Maui, 03–06 June 1991. Maui, 1991. P. 586–591. URL: https://doi.org/10.1109/CVPR.1991.139758
Belhumeur P. N., Hespanha J. P., Kriegman D. J. Eigenfaces vs. Fisherfaces: Recognition using class specific linear projection. IEEE Transactions on pattern analysis and machine intelligence. 1997. Vol. 19, no. 7. P. 711–720. URL: https://doi.org/10.1109/34.598228
Dalal N., Triggs B. Histograms of oriented gradients for human detection. Computer Vision and Pattern Recognition : Proceedings of the IEEE Computer Society Conference, San Diego, 20–25 June 2005. San Diego, 2005. P. 886–893. URL: https://doi.org/10.1109/CVPR.2005.177
Bartlett M. S., Movellan J. R., Sejnowski T. J. Face recognition by independent component analysis. IEEE Transactions on neural networks. 2002. Vol. 13, no. 6. P. 1450–1464. URL: https://doi.org/10.1109/TNN.2002.804287
Ghostfacenets: Lightweight face recognition model from cheap operations / Alansari M. et al // IEEE Access. 2023. Vol. 11. P. 35429–35446. URL: https://doi.org/10.1109/access.2023.3266068
Ghostnet: More features from cheap operations / Han K. et al // Computer vision and pattern recognition: Proceedings of the IEEE/CVF conference, Seattle, 13–19 June 2020. Seattle, 2020. P. 1580–1589. URL: https://doi.org/10.1109/CVPR42600.2020.00165
GhostNetv2: Enhance cheap operation with long-range attention / Tang Y. et al // Neural Information Processing Systems : Proceedings, New Orleans, 28 November – 09 December 2022. New Orleans, 2022. P. 9969–9982. URL: https://doi.org/10.48550/arXiv.2211.12905
Wen Y., Zhang K., Li Z., Qiao Y. A discriminative feature learning approach for deep face recognition. European conference on computer vision : Proceedings. Cham, 2016. P. 499–515. URL: https://doi.org/10.1007/978-3-319-46478-7_31
Deng J., Guo J., Xue N., Zafeiriou S. Arcface: Additive angular margin loss for deep face recognition. Computer Vision and Pattern Recognition : Proceedings of the IEEE Computer Society Conference, Long Beach, 15–20 June 2019. Long Beach, 2019. P. 4690–4699. URL: https://doi.org/10.1109/CVPR.2019.00482
Cosface: Large margin cosine loss for deep face recognition / Wang H. et al // Computer vision and pattern recognition : Proceedings of the IEEE/CVF conference, Salt Lake City, 18-23 June 2018. Salt Lake City, 2018. P. 5265–5274. URL: https://doi.org/10.1109/CVPR.2018.00552
Kim M., Jain A. K., Liu X. Adaface: Quality adaptive margin for face recognition. Computer vision and pattern recognition : Proceedings of the IEEE/CVF conference, New Orleans, 18–24 June 2022. New Orleans, 2022. P. 18750–18759. URL: https://doi.org/10.1109/CVPR52688.2022.01819
Boutros F., Damer N., Kirchbuchner F., Kuijper A. Elasticface: Elastic margin loss for deep face recognition. Computer vision and pattern recognition : Proceedings of the IEEE/CVF conference, New Orleans, 18–24 June 2022. New Orleans, 2022. P. 1578–1587. URL: https://doi.org/10.1109/CVPRW56347.2022.00164
Unitsface: Unified threshold integrated sample-to-sample loss for face recognition / Jia X. et al // Neural Information Processing Systems : Proceedings, New Orleans, 10–16 December 2023. New Orleans, 2023. P. 32732–32747. URL: https://doi.org/10.48550/arXiv.2311.02523
Wang X., Xue H., Liu X., Pei Q. A privacy-preserving edge computation-based face verification system for user authentication. IEEE Access. 2019. Vol. 7. P. 14186–14197. URL: https://doi.org/10.1109/ACCESS.2019.2894535
Lin W. H., Wang P., Tsai C. F. Face recognition using support vector model classifier for user authentication. Electronic Commerce Research and Applications. 2016. Vol. 18. P. 71–82. URL: https://doi.org/10.1016/j.elerap.2016.01.005
Your face your heart: Secure mobile face authentication with photoplethysmograms / Chen Y. et al // Computer Communications : Proceedings of the IEEE Conference, Atlanta, 01–04 May 2017. Atlanta, 2017. P. 1–9. URL: https://doi.org/10.1109/INFOCOM.2017.8057220
Smith-Creasey M., Albalooshi F. A., Rajarajan M. Continuous face authentication scheme for mobile devices with tracking and liveness detection. Microprocessors and Microsystems. 2018. Vol. 63. P. 147-157. URL: https://doi.org/10.1016/j.micpro.2018.07.008
Zhou B., Xie Z., Ye F. Multi-modal face authentication using deep visual and acoustic features. Communications : Proceedings of the IEEE International Conference, Shanghai, 20–24 May 2019. Shanghai, 2019. P. 1–6. URL: https://doi.org/10.1109/ICC.2019.8761776
Aface: Range-flexible anti-spoofing face authentication via smartphone acoustic sensing / Xu Z. et al // Interactive, Mobile, Wearable and Ubiquitous Technologies. 2024. Vol. 8, no. 1. P. 1–33. URL: https://doi.org/10.1145/3643510
Imagenet: A large-scale hierarchical image database / Deng J. et al. // Computer Vision and Pattern Recognition : Proceedings of the IEEE Conference, Miami, 20–25 June 2009. Miami, 2009. P. 248–255. URL: https://doi.org/10.1109/CVPR.2009.5206848
Simonyan K., Zisserman A. Very deep convolutional networks for large-scale image recognition. Learning Representations : Proceedings of the International Conference, San Diego, 07-09 May 2015. San Diego, 2015.
Huang G. B, Mattar M., Berg T., Learned-Miller E. Labeled faces in the wild: A database forstudying face recognition in unconstrained environments. Faces in’Real-Life’Images: detection, alignment, and recognition : Proceedings. 2008
