"EXAM SHILED” USE FOR PROCTORED EXAM INVIGILATION…

This paper presented EXAM SHIELD, an AI-based automated exam invigilation system aimed at preserving academic integrity in online examination environments. By leveraging deep learning, computer vision, and temporal behaviour analysis, the proposed system addresses key challenges associated with remote assessments, including impersonation, unauthorized resource usage, and collaborative cheating. The modular system architecture enables seamless integration of identity verification, behavioural analysis, environment monitoring, and decision- making components, ensuring scalability and adaptability across diverse examination scenarios. The utilization of CNN-based architectures regarding facial recognition and object detection, combined with recurrent models such as LSTM for temporal behaviour analysis, allows EXAM SHIELD to achieve accurate and reliable detection while minimizing false positives. The multi-modal approach, which incorporates visual, audio, and system- level monitoring, further enhances robustness and ensures consistent performance under varying environmental and hardware conditions. Experimental observations discussed in existing literature confirm the effectiveness of such AI- driven invigilation systems in real-time deployment. Despite its strengths, automated invigilation presents ethical and technical challenges related to privacy, fairness, and user comfort. Addressing these concerns is essential for responsible adoption. Upcoming research shall concentrate on incorporating confidentiality-maintaining methodologies like collaborative machine learning to reduce centralized data storage and enhance user trust. Additionally, incorporating explainable AI mechanisms will improve transparency by providing interpretable justifications for system decisions. The development of standardized and diverse benchmarking datasets will further support fair evaluation and comparison of invigilation systems. Overall, EXAM SHIELD demonstrates the potential of artificial intelligence to create secure, scalable, and ethical online examination environments, contributing meaningfully to the future of digital education.

[1]                         S. Sharma and M. Jain, “AI-based online exam proctoring systems,” IEEE Access, vol. 10, pp. 45678–45690, 2022. [2]                         A. Singh & R. Yadav, “Real-time object detection for remote examination monitoring using YOLO,” in Proc. Int. Conf. Artificial Intelligence and Applications, 2023, pp. 112–118. [3]                         M. Abdelrahman, L. Chen, and K. Nguyen, “Ethical challenges in automated proctoring systems,” Information Processing in Education, vol. 5, no. 2, pp. 89– 101, 2021. [4]                         A. Norouzzadeh, M. A. Ghassemi, and T. M. Nguyen, “Deep learning for human behavior recognition: A survey,” IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 40, no. 10, pp. 2319–2340, 2018. [5]                         K. Zhang, Z. Zhang, Z. Li, and Y. Qiao, “Joint face detection & alignment using multitask cascaded convolutional networks,” IEEE Signal Processing Letters, vol. 23, no. 10, pp. 1499–1503, 2016. [6]                         J. REDMON & A. FARHADI, “YOLOV3: AN INCREMENTAL IMPROVEMENT,” ARXIV:1804.02767, 2018. [7]                         H. Wang and L. Wang, “Modeling temporal dynamics for human action recognition using LSTM networks,” IEEE DEALS on Circuits and Systems for VideoTAPE Technology, vol. 27, no. 6, pp. 1234–1246, 2017. [8]                         T. Baltrusaitis, C. Ahuja, and L. Morency, “Multimodal machine learning: A survey and taxonomy,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 41, no. 2, pp. 423–443, 2019. [9]                         R. R. Selvaraju et al., “Grad-CAM: PICTORIAL INTERPRETATION from deep networks via gradient-based localization,” IEEE International Conference on Computer Vision (ICCV), 2017, pp. 618–626. [10]                       Q. Yang, Y. Liu, T. Chen, and Y. Tong, “Federated machine literacy: Conception and operations,” ACM deals on Intelligent Systems and Technology, vol. 10, no. 2, pp. 1– 19,    2019.