FACIAL RECOGNITION SYSTEM WITH ANTI SPOOFING MACHANISM | IJCT Volume 13 – Issue 2 | IJCT-V13I2P35
International Journal of Computer Techniques
ISSN 2394-2231
Author
S. Hariprakash Reddy, Shaik. Mahaboob Basha, V. Vijay Simha Reddy, Mr.C.Ramachandran
Abstract
Facial recognition has become one of the most common techniques of biometric authentication which is highly accurate and convenient. Nonetheless, it is still highly vulnerable to the presentation attacks like photo and video spoofing or mask-based spoofing that may undermine the security of the system. The paper proposes a Facial Recognition System with an Anti- spoofing Mechanism using the MobileNetV3 which is a lightweight convolutional neural network that allows it to be used on a mobile and a device where it has high performance. The suggested system combines the element of a facial extraction and the detection of spoof in a single deep learning architecture. The system works well to differentiate authentic attempts and fake attempts in real time by training the model on real and spoof face datasets. Trial performances prove the idea that MobileNetV3 offers an optimal trade-off conclusions in respect to accuracy, computations, and velocity, therefore, it is most appropriate to be utilized on a practical setting in terms of mobile authentication, access control, and surveillance. This method maximizes the effectiveness of facial recognition as well as providing a high level of protection against a spoofing attack.
Keywords
Anti-Spoofing, Liveness Detection, Deep Learning, Presentation Attack Detection, Facial Recognition, and Convolutional Neural Network. Anti- Spoofing, Liveness Detection, Deep Learning, Presentation Attack Detection, Facial Recognition, and Convolutional Neural Network.
Conclusion
This paper has created and assessed a Facial Recognition System using Anti-Spoofing Mechanism based on MobileNetV3 on CASIA Face Anti-Spoofing Dataset. The model was developed to effectively differentiate between live and spoof faces using deep learning methods thus reaching an astounding accuracy of 85.76% and an AUC of
0.95. The findings also suggest that MobileNetV3 due to its lightness architecture and high representational efficiency is effective to be used in real-time applications with low computational capability. This model was very effective to distinguish between real and spoofed images since it learnt more complex patterns of facial texture and fine variations in illumination and depth- related clues. Improved generalization and robustness were helped by the data augmentation, class balancing and adaptive schedule of learning rate. The suggested technique can be simply embedded into the existing biometric authentication system to augment the security level in face- based access control systems, Internet verifications, and surveillance systems. In spite of good performance witnessed with the proposed system, some restrictions were realized when poor light occurs, side poses, and motion blur. Such instances sometimes resulted in inaccurate classification of a case because of a lack of time information in one- frame classification. Because MobileNetV3 is more geared towards spatial features, the model cannot easily capture motion features, which can be crucial to the application of detecting dynamic instances of spoofing like video replay or 3D mask spoofing. Future Work
References
1.A. H. Pawar, N. Kadam, Y. Dadas, S. Kakade, and K. Kamble, “Face Anti-Spoofing Using Deep Learning Approach,” International Journal of Scientific Research And Engineering Development (IJSRED), vol. 6, no. 11, pp. 1– 3, 2023.
2.M. S. Sagar and C. Saravanan, “Face Spoof Detection Using Image Processing,” International Journal of Creative Research Thoughts (IJCRT), vol. 9, no. 6, pp. 478–481, 2021. 3.T.-C. Tseng, T.-F. Shih, and C.-S. Fuh, “Anti- Spoofing of Live Face Authentication on Smartphones,” Journal of Information Science and Engineering, vol. 37, no. 3, pp. 605– 616, 2021. 4.S. Sayed, S. Fatima, Q. A. Kader, M. Ebrahim, and S. M. Ibrahim, “Face Recognition Attendance System with AntiSpoofing Detection,” International Journal of Advanced Research in Science, Communication and Technology (IJARSCT), vol. 5, no. 11, pp. 230–238, 2025. 5.A. Dwivedi and S. Verma, “SCNN-Based Classification Technique for Face Spoof Detection Using Deep Learning Concept,” The Scientific
Temper, vol. 13, no. 2, pp. 165– 172, 2022. 6.P. Mhasakar, S. Mandal, and S. K. Mitra, “Multi- stream CNN for Face Anti-Spoofing Using Color Space Analysis,” Computer Vision and Image Processing (CVIP), vol. 2, no. 1, pp. 45–53, 2020.
7.T. N. Prasad and B. Anuradha, “Face Anti- Spoofing and Liveness Detection Using MobileNet and Haar Cascade Algorithm,” International Journal of Scientific Research in Science and Technology (IJSRST), vol. 10, no. 2, pp. 34– 39, 2023. 8.R. Kumar, A. Sharma, and M. Singh, “Hybrid EfficientNetB0 and Vision Transformer Model for Face Spoof Detection,” Springer Nature Computer Science, vol. 5, no. 3, pp. 215– 227, 2024. 9.J. Patel and N. Mehta, “Lightweight CNNs for Real-Time Face Liveness Detection,” IEEE Transactions on Information Forensics and Security, vol. 15, no. 4, pp. 1021–1030, 2020. 10.Y. Zhang, Y. Liu, J. Wan, and G. Guo, “CASIA- SURF: A Benchmark for Face Anti-Spoofing,” in Proc. IEEE CVPR Workshops, pp. 1000–1008, 2020.
11.X. Li, H. Wang, and T. Zhang, “Generalized Face Presentation Attack Detection with Cross- Domain Learning,” Pattern Recognition Letters, vol. 152, no. 2, pp. 76–85, 2021. 12.J. Kim and S. Park, “MobileNetV3-Based Lightweight Architecture for Real-Time Face AntiSpoofing,” Sensors, vol. 22, no. 5, pp. 2101–
2112, 2022. 13.T. Nguyen, H. Le, and P. Tran, “Temporal Shift Module for Face Anti-Spoofing in Videos,” IEEE Access, vol. 10, pp. 78112–78123, 2022. 14.H. Gao, X. Ren, and L. Liu, “Depth and RGB Fusion for Face Spoofing Detection,” Neural Computing and Applications, vol. 35, no. 4, pp. 3251–3265, 2023. 15.R. Zhao and K. Liu, “Transformer-Based Framework for Face Presentation Attack Detection,” in Proc. IEEE CVPR Workshops, pp. 540–548, 2023. 16.C. Luo, Y. Chen, and F. Zhang, “Enhancing Spoof Detection Under Low Illumination Using Adaptive Feature Enhancement,” Pattern Analysis and Applications, vol. 26, no. 6, pp. 489–500, 2023.
How to Cite This Paper
S. Hariprakash Reddy, Shaik. Mahaboob Basha, V. Vijay Simha Reddy, Mr.C.Ramachandran (2026). FACIAL RECOGNITION SYSTEM WITH ANTI SPOOFING MACHANISM. International Journal of Computer Techniques, 13(2). ISSN: 2394-2231.
Related Posts:
