Design of Intelligent Classroom Attendance System Based on Face Recognition

Abstract

It is time-consuming and laborious for classroom attendance methods in Chinese universities, and the attendance costs are too high. In this paper, we use the deep learning related ideas to improve the AlexNet convolutional neural network, and use the WebFace data set to improve the network training and test. The Top-5 error rate is only 6.73%. We applied this model to face recognition and combined with RFID card reading technology, which developed a smart classroom attendance system based on face recognition. Research shows that the system is efficient and stable, which effectively reduce classroom attendance costs.