Published August 3, 2026 | Version v1

AI Powered Offline Exam Invigilator

Description

The fast rise of artificial intelligence, along with computer vision tools, is changing how we monitor exams in places like schools. Old-school ways of watching students during tests usually come with mistakes, personal bias, or gaps in attention - especially when hundreds take the test at once. To fix this, our idea is a smart exam watcher using AI that spots odd actions, keeps tabs on what students do, flags forbidden items, while making sure rules are followed as things happen. It uses advanced neural networks to recognize faces, follow where someone’s eyes move, detect objects, check body positions - all helping make oversight more solid and quicker. This setup combines a smooth front-end made with up-to-date design tools plus a back-end running ML workflows and protected data storage. Live camera feeds get analyzed instantly to catch warning signs: sudden head turns, constant glancing off-screen, phone use, or multiple people showing up unexpectedly. Notifications pop up on their own for supervisors, so they spend less time checking things by hand while tests stay fairer. This smart exam helper works without help from people, runs smooth whether it’s used online or offline, scales easily when needed, keeps data safe, shows what's happening clearly, avoids favoritism, spots odd behavior fast, cuts down errors, saves effort across schools and boards alike.

Files

AI POWERED OFFLINE EXAM INVIGILATOR -HBRP Publication.pdf

Files (455.9 kB)

Additional details

References

  • 1. Salunkhe along with N. Shende tackled the project, while N. Shah joined in later; S. Ubale also played a key role. Their focus? Building an automatic exam monitoring tool powered by computer vision tech combined with a mixed machine learning model. The work got featured in an IEEE conference paper - no fluff, just solid results from real testing.
  • 2. Y. Atoum, along with L. Chen and A. Liu, teamed up with S. Hsu plus X. Liu to explore automated proctoring for online exams - published in IEEE Transactions on Multimedia.
  • 3. T. Potluri, alongside V. S. Phanikumar and V. K. Kishore, developed a web-based exam monitoring tool that uses Attentive-Net to detect suspicious actions by students during tests - published in Multimedia Tools and Applications, 2023.
  • 4. S. Kaddoura, "Towards effective and efficient online exam systems using real-time proctoring," (ScienceDirect), 2022. ScienceDirect
  • 5. Z. T. Hossain et al., "Automated Online Exam Proctoring System Using Eye Gaze, Head Pose and Voice Detection," BRAC University Thesis / Paper.
  • 6. S. Gopane, "Cheating Detection in Online Examinations Using Deep Learning and Face Tracking," (Architectural Education Journal), 2024.
  • 7. B. Erdem, M. Karabatak, "Cheating Detection in Online Exams Using Deep Learning and Machine Learning," Applied Sciences, 2025
  • 8. Y.-S. Shih, M. Liao, R. Liu, along with M. B. Baig put out a paper on using human-guided AI to spot cheating rings - posted on arXiv in 2024.
  • 9. G. Akçapınar, "Detecting AI-Assisted Cheating in Online Exams through Behavior Analytics," arXiv, 2025. 1
  • 10. Y. Liu, J. Ren, along with J. Xu and X. Bai, together with R. Kaur plus F. Xia explored ways to catch cheating during online tests using a method that checks multiple examples at once - shared via arXiv in 2024.