Assesment Proctoring system

Abstract

This project introduces an AI Based Exam Proctoring System designed to maintain the integrity of online assessments by leveraging advanced technologies. The system employs a combination of artificial intelligence, machine learning, and computer vision techniques to monitor and secure online exams. Key functionalities include real-time video monitoring of test-takers, facial recognition, eye-tracking, and behavior analysis to detect any suspicious activities. The Online Proctoring System not only minimizes cheating but also reduces the administrative burden on educators by automating the proctoring process, allowing for scalable and efficient online assessments. In an era of remote learning, Assessment Proctoring System using AI/Ml is a vital tool for maintaining academic integrity and trust in online education. This project serves as a comprehensive guide for educators, institutions, and technologists interested in the development and deployment of effective online proctoring solutions. To implement this project, we leverage commonly available hardware components such as a webcam and a microphone which are connected to the student\'s PC or laptop. Real Time video processing capabilities are harnessed to continuously monitor the examination environment. Any suspicious activities or deviations from the expected behavior are promptly detected by the AI algorithm.

Introduction

The rise of online education has created challenges in maintaining academic integrity during remote exams. To address this, AI-based exam proctoring systems have been developed using advanced technologies like computer vision, machine learning, and speech detection. These systems monitor test-takers via webcams and microphones, using tools such as OpenCV and MediaPipe for face recognition and head pose estimation. Suspicious behaviors—like abnormal gaze, head movements, or speech—are detected using algorithms that assign a real-time cheating probability score.

The system integrates features such as:

• Live proctoring

• Real-time alerts

• Integration with LMS platforms

• Automated behavior analysis

• Cheating detection via conditional algorithms

During implementation, head pose angles are calculated using the PnP algorithm, and suspicious behaviors are flagged and analyzed dynamically. A graph tracks changes in cheating probability over time, helping educators respond promptly.

The motivation behind this system is to provide a scalable, reliable, and non-intrusive solution to exam integrity in digital learning environments. Traditional human proctoring methods are costly and less effective, prompting the shift toward AI-driven solutions.

In the future, these systems may include biometric authentication, eye-tracking, and behavioral analytics to improve accuracy and inclusivity. Ensuring data privacy, customization, and user trust will be essential as these tools become standard in online assessments, aiming to enhance both security and educational fairness.

Conclusion

This system is one of the popular revisited topics due to the pandemic and the need for people to conduct online tests. The system aimed to detect whether the user is showing suspicious paper using video and audio output. During the making of the system, we used various machine learning algorithms for head pose detection and successfully implemented head pose estimation using computer vision as well as speech detection using a microphone. We successfully developed a system that can detect suspicious behavior, and it is a lightweight, low-resource-consuming system. The integration of advanced algorithms enhances its efficiency, ensuring real-time monitoring without significant computational overhead. Additionally, the system is designed to be scalable, allowing institutions to implement it across various online examination platforms. Future improvements can focus on increasing detection accuracy by incorporating eye-tracking and facial expression analysis. Ensuring privacy and ethical considerations remains a crucial aspect, requiring continuous refinements in data handling and security measures. Moreover, adaptive learning mechanisms can be implemented to improve the system’s detection capabilities over time..

References

[1] Potluri, Tejaswi, and Venkata Krishna Kishore K. \"An automated online proctoring system using attentivenet to assess student mischievous behavior.\" Multimedia Tools and Applications (2023): 1-30. [2] Li, Brian, and Emma Li. \"Automated Online Proctoring System Using Gaze View Tracking and Custom Object Detection.\" 2022 IEEE MIT Undergraduate Research Technology Conference (URTC). IEEE, 2022. [3] Malhotra, Neil, et al. \"Smart artificial intelligence based online proctoring system.\" 2022 IEEE Delhi section conference (DELCON). IEEE, 2022. [4] Labayen, Mikel, et al. \"Online student authentication and proctoring system based on multimodal biometrics technology.\" IEEE Access 9 (2021): 72398-72411. [5] M. Abdelsalam, M. Shokry and A. M. Idrees, \"A Proposed Model for Improving the Reliability of Online Exam Results Using Blockchain,\" in IEEE Access, vol. 12,pp. 7719-7733, 2024, doi: 10.1109/ACCESS.2023.3304995. [6] I.D. Raga Priya, P. Sree Ramya, M. V. Vamshi, B. Chandana and M. K. Rao, \"Malpractice Detection in Online Proctoring using Deep Learning,\" 2023 Third International Conference on Artificial Intelligence and Smart Energy (ICAIS), Coimbatore, India, 2023, pp. 575-583, doi: 10.1109/ICAIS56108.2023.10073664 [7] I.D. Raga Priya, P. Sree Ramya, M. V. Vamshi, B. Chandana and M. K. Rao, \"Malpractice Detection in Online Proctoring using Deep Learning,\" 2023 Third International Conference on Artificial Intelligence and Smart Energy (ICAIS), Coimbatore, India, 2023, pp. 575-583, doi: 10.1109/ICAIS56108.2023.10073664

Copyright

Copyright © 2025 Rohit Havaldar, Prathmesh Gholap, Chaitanya More, Prof. Deepali Maste, Sarthak Chawande. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.