<![CDATA[Face recognition-based attendance systems are vulnerable to spoofing attacks without effective liveness detection. This study proposes an employee attendance system that integrates CNN-based liveness detection with MagFace-based face recognition to enhance security. The liveness module serves as a preliminary filter to distinguish live faces from spoof attempts before identity verification. Experimental results show that the liveness detection module achieved accuracies of 98%, 96.28%, and 87.27% on training, validation, and testing datasets, respectively, with a False Positive Rate (FPR) of 6.0% on the testing dataset. The MagFace-based recognition module achieved an accuracy of 95.24%, with a False Acceptance Rate (FAR) of 4.64% and an Equal Error Rate (EER) of approximately 4.76%. These results indicate that the proposed system is suitable for employee attendance applications. However, the liveness detection module is intended as a baseline prototype and is not yet designed for high-security biometric authentication scenarios.]]>
Copyrights © 2026
