<![CDATA[Technological advancements have significantly influenced many aspects of life, including education. Student attendance systems are one area that can be improved through modern technology, as many schools still rely on manual methods that are inefficient and prone to errors and fraud. This study proposes an automated attendance system using classroom surveillance cameras combined with facial recognition technology through the pre-trained FaceNet model. The aim is to develop an accurate and efficient solution that overcomes the limitations of traditional attendance practices. Surveillance cameras enable continuous and non-intrusive collection of students’ facial data during learning activities. FaceNet, based on deep convolutional neural networks, is expected to recognize faces accurately in real time. Beyond attendance efficiency, the system may also enhance security and support classroom monitoring. Results show that the FaceNet-based system with transfer learning performs well under adequate lighting. Individual testing achieved a 100% success rate, while dim lighting reduced performance to 88.9% in individual tests and 11.1% in group tests. Group testing under sufficient lighting reached 77.8%. Camera positioning, facial overlap, and image noise influenced recognition outcomes.]]>
Copyrights © 2026
