Attendance System on Moving Objects through Face Recognition using MTCNN and CNN
Susetyo Bagas Bhaskoro, Siti Aminah, Khoutal Taqi · 2021
Face detection and recognition using the eigenfaces method accuracy is decreased when there are changes in object distance and lighting levels. The objective of this research is to propose an automatic presence system through face detection using the MTCNN method and facial image recognition using the CNN method. The CNN architecture used in this study is VGG16. Based on the test results, the MTCNN and CNN algorithm can handle the changes in object distance and lighting levels. The face detection system has an average error value of 17%, calculated using MAPE; the error rate is high because other objects cover faces, and some faces use face-covering attributes. The facial recognition system has an average accuracy value of 78% for the first architecture and 87.3% for the second architecture.