Automated Online Proctoring Using Artificial Intelligence

Abstract

Proctoring exams is the current need in social distancing and pandemic driven society. Education institutes and universities must explore alternate technologies to continue their essential educational exams. An Artificial Intelligence (AI) driven remote proctoring is the one method that will ensure end-to-end security for online Exams. The proposed method will help cheating and malpractices. Due to the COVID-19 pandemic, the world is under severe downfalls in terms of economy. Even daily activities get restricted due to enforcements like lockdown. With remote proctoring using AI-based supervision technologies, it will ensure that students do not get indulged in cheating or unfair means throughout the examination. Proctoring exams is the current need in social distancing and pandemic driven society. Education institutes and universities must explore alternate technologies to continue their essential educational exams. An Artificial Intelligence (AI) driven remote proctoring is the one method that will ensure end-to-end security for online Exams. The proposed method will help cheating and malpractices. Due to the COVID-19 pandemic, the world is under severe downfalls in terms of economy. Even daily activities get restricted due to enforcements like lockdown. With remote proctoring using AI-based supervision technologies, it will ensure that students do not get indulged in cheating or unfair means throughout the examination. We compared the working of The Multi-task Cascaded Convolutional Networks (MTCNN) and OpenCV's Caffe model of the Deep neural network (DNN). The MTCNN resulted in the best outcomes when there was contrasting the images. DNN gave a better performance with an impediment speed head movements and can detect side faces. In addition, it gave the speediest fps of all.