A Novel Approach for Conducting Secure and Performance Proctored Examinations by Organization

Abstract

Conducting online quizzes and examinations have been done for several years in many ways. However, it became popular and unavoidable during this covid-19 pandemic. There are many types of solutions which are available predominantly in the web pages. Thorough study was carried out to enhance the prevailing methods with better security and performance. In existing online exam there are many ways to identify the malpractice, hence it is vital to develop an effective face detection algorithm to avoid malpractice. However, detecting faces continuously throughout the exam from the candidate is still challenging and memory consuming. In this paper, a computer vision algorithm named Haarcascade classification has been proposed for robust and continuous face detection for conducting proctor exams. One of the main malpractices is using the Virtual machines during the online exams. This work also focuses over an algorithm to detect the malpractice which has been done with the help of External or s Third-Party Virtual Machines. In future, a secured application can be developed for uploading the random questions for the respective organization to conduct Online Examinations.