Mesin Absensi Face Recognition Berbasis Raspberry Pi

Abstract

The fingerprint attendance machine failed to record attendance with injured, scratched, peeled finger skin and others so that attendance was not recorded. This research is a development of previous research with Radio Frequency Identification (RFID) attendance system because attendance system with RFID can be rigged. The incorporation of a facial recognition system and RFID on an attendance machine with a raspberry pi is expected to minimize failures during attendance. Because if there is a facial recognition failure, the user can make attendance with RFID. Attendance with RFID in this system can only be done when there is a face detection failure. To find out the percentage of success and accuracy of the machine, each user performs several trials. Furthermore, the number of facial recognition successes was recorded and the accuracy value was calculated using the accuracy calculation method. The results of the face identification experiment showed that the accuracy of the first user was 53%, the second user was 48%, the third user was 45% and the fourth user was 52%. The machine is able to predict 4 user images with 4 different face positions with an average identification process time of 7 seconds.