<![CDATA[This study proposes a facial biometric attendance system integrated with an anti-spoofing mechanism to enhance security and reliability. The system employs a Convolutional Neural Network (CNN) to classify real and spoofed faces. The dataset consists of the NUAA Photo Imposter dataset and additional samples collected using a webcam, including various spoofing attacks such as printed photos, masks, and screen displays During preprocessing, image augmentation techniques, including brightness adjustment and filtering, are applied to improve model robustness under different environmental conditions. Furthermore, several training parameters, such as the number of epochs, learning rate, optimizer type, and image size, are evaluated to determine their impact on model performance. Experimental results show that the proposed model achieves 100% accuracy during training with loss of 5,78% and validation accuracy of 97,84% with loss validation of 5,02%. In testing, the system attains 88% accuracy for real face detection and 92% for spoof detection, demonstrating its effectiveness for secure and reliable attendance applications. Keywords: Biometrics, Face Recognition, AntiSpoofing, Convolutional Neural Network (CNN), Attendance System. ]]>
Copyrights © 2026
