Virtual Surveillance Oversight System

Online tests have gained popularity in recent years due to its adaptability, convenience, and user-friendliness. The prime issue with web based is the invigilator’s need to detect anomalous behavior while administering the test. The primary focus of traditional monitoring programs is on the testers' identities; anomalous behaviors of the testers are not effectively identified. In response to the issue of seeing aberrant behaviors during online exams. This study addresses the growing popularity of online tests, highlighting the challenge of detecting anomalous behavior during exams. Traditional monitoring focuses on identity verification, neglecting abnormal actions. To address this, our model employs features such as phone and person detection, mouth opening detection, eye tracking, face spoofing, voice-to-text detection, and number of person detection. Using the Dlib library's facial key point model, which detects 68 facial key points, our system undergoes data collection, processing, and model training. Evaluated in an online test environment, the proposed method outperforms existing systems, utilizing a webcam and microphone for automated proctoring. Testing implies that the given system works better than the existing systems. A webcam and microphone are used in an automated proctoring system to automatically monitor the user.