Attendance System with Face Recognition, Body Temperature, and Use of Mask using Multi-Task Cascaded Convolutional Neural Network (MTCNN) Method

Abstract

The application of health protocols in educational, office, or industrial environments can be made by changing old habits that can spread COVID-19. One of them is the habit of recording attendance, which still requires direct physical contact. In this research, an attendance system based on facial recognition, body temperature checks, and mask use using the multi-task cascaded convolutional neural network (MTCNN) has been developed. This research aims to integrate a facial recognition system, a mask detection system, and body temperature reading into an attendance recording system without the need for direct physical contact. The attendance system offered in this study can minimize the spread of COVID-19. So, it has enormous potential for use in educational, office, and industrial environments. The focus of this research is to create an attendance system by integrating the application of face recognition, body temperature, and the use of masks using a pre-trained model. Based on the research results, an attendance system was successfully developed where the results of face recognition, mask detection, and body temperature were displayed on the machine screen and attendance platform. Facial recognition testing on the original LFW dataset has an accuracy of 66.45%. The accuracy of the dataset reaches 92-100%. In addition, the intelligent attendance platform has been successfully developed with user management, machine service, and attendance service features. The results of the attendance record are successfully displayed on the platform or through the download feature.