Towards a blob-based presence verification system in summative e-assessments
Traditionally, authentication systems are required to verify a claimed identity one time only at the initial login. However, in high-stake environments such as a summative e-assessment environment, a one-time authentication session is insufficient to guarantee security. Hence, the security of online summative assessments goes beyond ensuring that the 'right' student is authenticated at the initial login. More is required to verify the presence of an authenticated student for the duration of the test. In this paper, we explore potential approaches to achieving presence verification. However, these approaches have limitations that make them unsuitable for verifying presence in e-assessments. Hence, we propose an object tracking approach using a blob analysis solution. The blob analysis solution is a video processing technique that attempts to detect, verify and classify a student's presence throughout the test session thus indicating the likelihood of acceptable or unacceptable activities. By employing the blob analysis operation, we propose a novel blob-based presence verification system which uses the geometric statistics of binary images to make inferences about an object's presence in the video sequence. The proposed system is designed to verify the student's presence in a non-interruptive and non-distracting fashion. Furthermore, by simulating possible student activities in test conditions, we carried out experiments to investigate the feasibility of using blob analysis for presence verification. In addition, the decisions made about a student's presence in the test environment were driven by a set of well-defined fuzzy logic rules. The results show that the verification of a student's presence presents valuable improvements to preserving e-assessment user security.