<![CDATA[The increasing demand for efficient and reliable student attendance systems highlights the limitations of conventional manual methods, which are often time-consuming, prone to fraud, and lack accuracy. This study develops a real-time attendance system based on facial recognition using the YOLOv8 deep learning model integrated with TensorFlow. The system is designed to automatically detect and recognize students’ faces through a webcam and record attendance data in real time. The research method involved three main stages: face registration, model training, and attendance recording. A dataset of 4,200 face images from 14 students was collected to train the model. The recognition process used cosine similarity with a threshold of 0.7 to balance accuracy and avoid false recognition. The results showed that the system could effectively recognize student faces under varying lighting and expression conditions with high accuracy, while also recording attendance automatically and reliably. The system proved to be efficient in reducing time consumption, minimizing fraud, and ensuring more accurate attendance records. This study demonstrates that YOLOv8-based face recognition can provide an effective solution for modernizing school attendance management.]]>
Copyrights © 2025
