Abstract:

One of the important parts of daily classroom evaluation is attendance monitoring. The problem of attendance checking has been identified as one of the main problems for teachers, especially for large groups. Nowadays, biometric systems have become fairly ubiquitous and their use is increasingly spreading. Biometrics in education has become a regular occurrence and biometric methods are used for regulating access (building or classroom), paying in school dining rooms, tracking student and employee attendance, borrowing books in the library, or logging into e-learning systems. One of the increasingly popular research directions is using face recognition for attendance management.

This paper proposes the use of biometric face recognition for automatic registering students attending a lecture. Firstly, we give an overview of different face recognition methods with a focus on deep learning methods. Selected deep learning methods will be described in detail and tested. The main part of the paper will be the implementation of the selected face recognition method and its application for attendance management at a higher education institution. In the end, the advantages and challenges of the proposed method will be identified, as well as future research directions.

This research is part of a broader research topic dealing with the complex IoT-supported and AI-supported learning environment. Although educational environments are constantly becoming IoT playgrounds and different authors are researching the possibilities of using the IoT to enhance the teaching and learning processes, in our ongoing research we are not only combining software and hardware but also liveware objects with the aim of students’ activities tracking and success prediction. We call this environment an IoT ecosystem.

Keywords:

Attendance management, face recognition, deep learning, convolutional neural network, software development.