Rapid development of information technology has greatly changed the pattern of teaching and learning in education, and e-learning is the most widely used mode of learning nowadays. In e-learning systems, some cheating phenomena can occur because the teacher cannot see the students' learning process. User authentication by user names and passwords only when logging on in most e-learning systems might cause cheating. In other words, someone else may substitute for learners, and the learner may leave before the lecture is over.
If a student enrolled in e-learning system is able to complete a course without actually learning, the reliability of the course and e-learning system is diminished, and since lectures or learning materials are intellectual assets, such deception should be prevented. Therefore, in e-learning systems, it is necessary to identify the learner during the learning activities as well. To detect spoofing in e-learning is of great significance in enhancing the reliability of e-learning systems and protecting the copyright of learning content.
Therefore, biometric authentication techniques such as face recognition, fingerprint recognition and speech recognition have been introduced into user identification in e-learning systems, and prevention of spoofing of learning by face recognition in e-learning process is in active use. However, face recognition is usually conducted at the beginning of user engagement, lectures and tests, and there has been no intense research to prevent spoofing of learning in the learning process.
Ryu Chang Sik, a section head at the Faculty of Distance Education, has proposed a new method to prevent spoofing, based on implicit face matching at randomly selected time intervals during online learning.
In the proposed method, face-matching identification is performed at the time of system login and the beginning of learning, and in each learning process (e.g., learning a section). In addition, implicit face-matching is performed several times without interaction with students at random time intervals to calculate the similarity value and detect the impostor. If not detected in the random check, explicit user identification would be required.
The proposed method has been applied to an e-learning system that provides a graduate qualification. The result shows that the system can effectively distribute the server load and prevent the spoofing of learning, while significantly reducing the number of stoppage of learning due to face contrasts.
You can find more information in his paper “Identification of Spoofing in e-Learning system with the implicit face recognition” in “INFORMATICA”.
...