STUDENT VIRTUAL CLASS ATTENDANCE BASED ON FACE RECOGNITION USING CNN MODEL

Abstract

Attendance records are an important tool that can be used to include and broadcast member participation in an activity, including the learning process. In online learning classrooms, the process of recording attendance becomes challenging to do manually, thus an automatic attendance recording system is needed. The authentication process is important in developing an existing recording system to guarantee the correctness of the recorded data. In this research, a face authentication system was built to create a system for recording online class attendance to help integrate participant activities and participation in online class learning. The face recognition approach uses a Convolutional Neural Network (CNN) model specifically designed to automate student attendance in virtual classes. Student image data is taken from virtual classroom sessions and used to train a CNN model. This model can recognize and verify student identity in various lighting conditions and head positions. This research consists of several stages, namely data collection, artificial neural networks, use of facial recognition, dataset application stage, and facial recognition in video frames. The experimental results showed that there were 11193 samples studied and of these 11193 samples the distribution was even, namely 6.7%. In addition, the model performance results show an accuracy of 76.28%.