ExamVision: A Dual-Task System for Academic Misconduct and Stress Detection in Online Exams

Online education and examination have become an integral part of learning. At the same time, online education offers flexibility and accessibility. Ensuring the integrity of online exams, as well as the psychological well-being of students, is a significant challenge. This research presents an ExamVision system designed to recognize student activities (such as academic misconduct) and detect facial expressions to analyze students’ stress levels during examinations. Although online proctoring tools are available, they primarily focus on detecting academic misconduct and lack the ability to detect stress. If the student is stressed, the proposed system will provide a motivational message in real-time and flag a frame in the activity log for the proctor’s review. This approach enhances online exam integrity and supports proactive improvement of student performance.