Minimizing Face Spoofing Attacks with Liveness Detection in Cross-Platform Mobile Attendance Systems

Attendance systems have now implemented many of the most recent technologies, such as the use of facial recognition to verify and validate student attendance in lecture classes. However, face recognition technology in the attendance systems is still prone to fraud with attacks using artificial faces through various media, such as printed photos or pre-recorded videos. Therefore, a liveness detection system is needed that can minimize the fraud. The liveness detection system using the randomized challenge-response method, with several randomly given challenges within a time limit, can indicate whether the detected face is genuinely alive. This liveness detection system is implemented in a cross-platform mobile-based digital attendance application called “My Attendance” developed using the Flutter framework. Student attendance records are stored and can be viewed by lecturers as proof of attendance in lecture classes. The results of the study show that this liveness detection system can minimize the occurrence of fraudulent artificial face attacks, achieving 100% performance accuracy when tested on 30 respondents from students of the Software Engineering study program at the Indonesian Education University. This system could be an effective and efficient solution for applying face recognition systems to the attendance process.