<![CDATA[Employee attendance management is a crucial aspect in improving organizational efficiency and productivity. However, in practice, fraudulent activities such as buddy punching are still frequently encountered. Therefore, PT.XYZ requires an efficient and secure attendance system to address this issue. This study aims to implement a face recognition-based attendance system using the Convolutional Neural Network (CNN) method combined with liveness detection. The system is developed on both mobile and desktop platforms using Python, TensorFlow, and Firebase technologies. The research process includes collecting a dataset of 320 facial images from 32 employees, image preprocessing, CNN model training, and the integration of liveness detection based on facial movement analysis to verify user authenticity. System evaluation is conducted based on accuracy, response time, and robustness under varying conditions such as lighting and facial positions. The results show that the system is capable of recognizing faces in real-time with an accuracy rate of 97.81% and a response time ranging from 2 to 5 seconds. The system also demonstrates stability under various lighting conditions and shows good scalability. Therefore, the CNN and liveness detection-based attendance system is effective in improving accuracy, security, and supporting a more efficient, professional, and transparent employee attendance management system.]]>
Copyrights © 2026
