Automated Exam Proctoring using Support Vector Machine and Histogram of Oriented Gradients

Online learning has grown exponentially since the Covid-19 outbreak. Educational organizations are forced to stop their regular curriculum. Using remote communication software, students are able to pursue their educational curriculums. This paper proposes a multimedia exam proctoring system based on face recognition and object capturing systems. System hardware includes a webcam only. Dlib is used for facial detection and face landmark detection. The OpenCV Caffe model is then utilised to appropriately locate and sketch the face. YOLOv3 is used to detect objects. Onfocus event is used for tab detection. By combining all these, an exam proctoring system is developed which can classify if the student is found cheating and is then informed to the examiner.