Design of Automated Exam Proctoring for User Authentication Through Face Augmentation and Verification

The uses of online exam continue to increase during covid-19 pandemic. The challenge of online examination is proctoring method to detect cheating possibility to maintain integrity in its implementation. Various design of proctoring methods had been proposed from remote proctoring using video surveillance until automatic proctoring using machine learning approaches. One of important things in proctoring is user authentication during exam session. Recently, face recognition algorithm using image are widely used in automatic proctoring method instead of video which use high load data for storage, transmission and processing. Despite the progress made to detect identity of examinee, these models very dependent on data and required iterative training process to get a robust model whereas there are many and various examinee facial image like pose, lighting, etc. This problem makes the training process require large resources and take a long time. In this paper, we proposed automatic proctoring design for examinee verification that more flexible and efficient through face verification with key-face augmentation instead of face recognition. This proctoring design verifies the examinee's image with variations of key-face augmentation so that it does not require repeated training processes. As a result, the design of the proposed method is presented in this paper.