DEVELOPMENT OF SELF PROCTORING EXAM
Abstract
The "Development of a Self-Proctoring Exam System" project aims to create an AI-driven, web-based platform that enables secure, remote examinations without the need for human invigilators. By leveraging technologies like computer vision, machine learning (specifically OpenCV, TensorFlow, and YOLO models), and browser lockdown techniques, the system automates the monitoring of test-takers to ensure academic integrity. Abstract/Project Description:
The rapid shift to online education has made ensuring the integrity of remote assessments a critical challenge. This project develops a "self-proctoring" application that allows students to take exams from any location while maintaining a secure, honest environment. The system provides a multi-modal monitoring approach, continuously analyzing live video and audio feeds to detect anomalies. Key Features & Functionality:
AI-Powered Monitoring: Uses Computer Vision to track eye gaze, head movements, and facial expressions, identifying suspicious behaviors like looking away frequently, unauthorized materials, or leaving the seat.
Object & Face Detection: Detects the presence of multiple people or unauthorized objects (e.g., phones, books) in the testing area using YOLO algorithms.
Browser/Screen Lockdown: Restricts access to other websites, applications, or tab switching, preventing digital cheating.
Real-time Fraud Logging: Logs all detected anomalies with timestamps into a database for post-exam review, offering a "score" of the student's integrity during the test.
Automated Authentication: Verifies the candidate’s identity using facial recognition at the start of the exam to prevent impersonation.
Full-Text Access Notice
In accordance with the NERD Policy on promoting peer-reviewed publication, public access to the full text of a project, thesis or dissertation is restricted for three years, allowing the author and supervisors sufficient time to pursue peer-reviewed publication.
During this period, researchers with legitimate academic or research purposes may request authorisation directly from the author to enable NERD to release the indexed full texts of the work using the form below.