<![CDATA[Employee attendance is an important aspect of attendance management in office environments. Face recognition technology offers a modern approach to employee attendance systems; however, its implementation remains vulnerable to spoofing attacks. On the other hand, many companies still rely on traditional methods such as manual signatures, which are inefficient, prone to fraud, and time-consuming. To address these issues, this study develops a real-time face recognition-based employee attendance system aimed at improving the efficiency, accuracy, and security of attendance recording by integrating blink detection as an anti-spoofing mechanism. The proposed system utilizes MTCNN for face detection, ArcFace for facial feature extraction in the form of vector embeddings, and Support Vector Machine (SVM) as the classification algorithm. Experimental results show that the system is able to recognize employee faces with an accuracy of 95%, achieving precision, recall, and F1-score values of 95.86%, 93.58%, and 93.89%, respectively. The blink detection mechanism is proven to prevent attendance spoofing using static photos. In addition, the system supports dynamic addition of new employee data without requiring long waiting times through a controlled data update process. With the implementation of this system, the employee attendance process becomes faster, more accurate, and security-verified, making it a practical solution for modern and efficient attendance data management in companies.]]>
Copyrights © 2026
