| dc.description.abstract | The traditional attendance system using manual signatures is still widely used in various educational institutions, including the Computer Science Department at Universitas Sumatera Utara. However, this method has drawbacks such as the potential for fraud, data inaccuracies, and disruptions to lectures. Several innovations, such as RFID-based and facial recognition attendance systems, have been implemented, but they often require expensive additional hardware. This study develops an innovative attendance system that integrates location tracking and facial recognition technology without the need for additional hardware. The system utilizes the Haversine formula for student location verification and Euclidean Distance for facial recognition. Testing results show that the system successfully recognizes student faces with 100% accuracy under normal conditions. However, accuracy decreases to 72% when the face is tilted and 36% in low-light conditions. Meanwhile, the location verification feature produced varying results depending on the room conditions. In rooms with minimal partitions, the system achieved 92% accuracy, whereas in rooms with many partitions, accuracy dropped significantly to 16%. This is due to GPS signal limitations, which struggle to penetrate building structures, making location verification accuracy highly dependent on the device’s signal reception. Although the system has limitations in indoor location tracking within rooms with many partitions, the integration of facial recognition still provides a more efficient solution compared to traditional attendance methods. With future advancements in location tracking technology, this system has the potential to become a more reliable solution for ensuring student attendance automatically and accurately. | en_US |
