<![CDATA[Digital transformation in attendance data management still faces several challenges, particularly related to low accuracy, potential fraud, and inefficiency in conventional methods such as manual signatures or ID-based systems that are vulnerable to misuse. This study aims to develop an automated attendance system based on face recognition to improve accuracy, efficiency, and reliability in real-time attendance recording. The method used in this study is a deep learning approach employing the YOLOv4 algorithm, implemented using the Python programming language and supported by the OpenCV library for digital image processing. The system is designed to detect and recognize user faces directly through a camera device. The research stages include requirement analysis, system architecture design, model development, system implementation, and performance evaluation using metrics such as precision, recall, and accuracy. The analysis technique is based on a confusion matrix to evaluate the system's ability to classify facial data accurately and consistently. The experimental results show that the system can operate in real-time with a precision of 91.81%, recall of 100%, and accuracy of 92.03%, indicating a high level of performance in face detection and recognition. In addition, the system demonstrates good stability under various lighting conditions and face positions, making it suitable for implementation in educational institutions and workplaces as a modern, secure, and efficient attendance solution.]]>
Copyrights © 2026
