Smart Eye Health Monitoring System

Abstract

Most people today spend a considerable amount of time on their computers, mobiles and tablets. Computer Vision Syndrome (CVS) has become common among computer users due to the many symptoms they suffer from when working for long hours on a computer, such as eye strain, dryness, redness, blurry vision and headache. The purpose of this project is to develop an Eye Health Monitor system to help users keep track of their computer usage in order to ensure that they maintain good viewing habits. Application Overview: Eye Health Monitor The Eye Health Monitor application is a real-time application built upon an AI system that uses the webcam of the computer to monitor the user’s screen behavior. It uses computer vision and deep learning techniques to detect the face of the user, identify him using FaceNet and checks if the user is looking at the screen with open eyes. This application does not allow any inaccurate data to be captured. The application only captures the screen time of the verified user when and only if the user has their eyes open and is looking at the screen. The application uses MediaPipe for face landmark detection, FaceNet for face recognition. It uses a FastAPI based backend and MongoDB for the session data storage. To protect user privacy, visual analysis on the client side is the default behavior. Only a small portion of the encoded image is sent to our server for verification. We recommend that you follow the medically recommended 20-20-20 rule for screen time – which involves taking a 20 second break and looking away from the screen every 20 minutes for every 60 minutes of use. We provide a visual analysis and usage summary of your screen activity over time through our dashboard. The idea is for the “Eye Health Monitor” to be an application, simple to use and non-intrusive in relation to the personal data of each user. Its purpose is to protect eye health in the digital age. Based on artificial intelligence and real-time monitoring with personalised feedback, the application aims to help users to reduce eye strain and to lead a healthy lifestyle.

Introduction

The text presents an AI-based Eye Health Monitor system designed to reduce digital eye strain caused by prolonged screen usage on computers, mobiles, and tablets. It addresses problems such as eye fatigue, dryness, headaches, and reduced productivity by continuously tracking real user engagement with a screen.

The system uses a webcam along with computer vision and deep learning techniques to monitor whether a verified user is present, whether they are actively looking at the screen, and how long they are engaged. Key technologies include MediaPipe Face Detection and Face Mesh for detecting faces and eye landmarks, FaceNet with cosine similarity for user authentication, and the Eye Aspect Ratio (EAR) method to determine eye openness and detect blinking or inactivity.

Screen time is recorded only when three conditions are met: a recognized user is present, a face is detected, and the eyes are open. If any condition fails, tracking pauses automatically, ensuring accurate measurement of real engagement rather than device usage. The system also includes an alert mechanism based on the 20-20-20 rule, reminding users to take regular breaks to protect eye health.

All usage data is stored in a database and visualized through an analytics dashboard, helping users understand their screen habits. Performance results show high accuracy across modules (around 94–97%), confirming reliable face detection, identity verification, and eye monitoring.

Conclusion

The proposed Eye Health Monitor system provides an effective solution for monitoring screen usage and promoting better eye health habits. With the increasing use of digital devices for work, education, and entertainment, prolonged screen exposure has become a common cause of eye strain and discomfort. The developed system addresses this issue by using computer vision and deep learning techniques to monitor user activity in front of the screen. The system integrates several technologies such as MediaPipe Face Detection, Face Mesh, Eye Aspect Ratio (EAR), and FaceNet facial recognition to detect the user, verify identity, and analyze eye activity in real time. By combining these techniques, the system ensures that screen time is recorded only when the verified user is present and actively looking at the screen. This approach improves the accuracy of screen monitoring compared to traditional screen-time tracking applications.In addition, the system incorporates an alert mechanism based on the 20-20-20 rule, which reminds users to take short breaks after continuous screen usage. These reminders help reduce eye strain and encourage healthier digital habits. The collected session data is stored and analyzed to provide useful insights through an analytics dashboard, allowing users to better understand their screen usage patterns.Overall, the experimental results show that the proposed system performs efficiently in detecting faces, recognizing users, and monitoring eye activity with high accuracy. The Eye Health Monitor demonstrates the potential of artificial intelligence–based monitoring systems in supporting digital well-being. In the future, the system can be further enhanced by integrating mobile applications, fatigue detection techniques, and personalized health recommendations to provide a more comprehensive eye health monitoring solution.

References

[1] American Optometric Association, “Computer Vision Syndrome,” American Optometric Association (AOA), 2023. [2] A. L. Sheppard and J. S. Wolffsohn, “Digital eye strain: prevalence, measurement and amelioration,” BMJ Open Ophthalmology, vol. 3, no. 1, pp. 1–10, 2018. [3] M. Rosenfield, “Computer vision syndrome: A review of ocular causes and potential treatments,” Ophthalmic and Physiological Optics, vol. 36, no. 5, pp. 502–515, 2016. [4] Statista Research Department, “Average daily screen time worldwide,” Statista, 2023. [5] C. Blehm, S. Vishnu, A. Khattak, S. Mitra, and R. Yee, “Computer Vision Syndrome: A review,” Survey of Ophthalmology, vol. 50, no. 3, pp. 253–262, 2005. [6] Z. Zhang, P. Luo, C. C. Loy, and X. Tang, “Facial landmark detection by deep learning,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014. [7] P. Viola and M. Jones, “Rapid object detection using a boosted cascade of simple features,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2001. [8] F. Schroff, D. Kalenichenko, and J. Philbin, “FaceNet: A unified embedding for face recognition and clustering,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015. [9] T. Soukupová and J. ?ech, “Real-time eye blink detection using facial landmarks,” in 21st Computer Vision Winter Workshop, 2016. [10] Google Research, “MediaPipe: A framework for building perception pipelines,” Google AI Research, 2019. [11] Y. Bazarevsky, V. Kartynnik, A. Vakunov, and M. Grundmann, “BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs,” arXiv preprint arXiv:1907.05047, 2019. [12] O. M. Parkhi, A. Vedaldi, and A. Zisserman, “Deep Face Recognition,” in British Machine Vision Conference (BMVC), 2015. [13] M. Divjak and H. Bischof, “Eye Blink Based Fatigue Detection for Prevention of Computer Vision Syndrome,” International Journal of Human-Computer Studies, vol. 67, no. 3, pp. 256–267, 2009. [14] Y. Bazarevsky, V. Kartynnik, A. Vakunov, K. Raveendran, and M. Grundmann, “BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs,” arXiv preprint arXiv:1907.05047, 2019. [15] R. Szeliski, Computer Vision: Algorithms and Applications, Springer, 2011. [16] O. M. Parkhi, A. Vedaldi, and A. Zisserman, “Deep Face Recognition,” British Machine Vision Conference (BMVC), 2015. [17] M. Divjak and H. Bischof, “Eye Blink Based Fatigue Detection for Prevention of Computer Vision Syndrome,” International Journal of Human-Computer Studies, 2009. [18] J. Brooke, “Usability and User Engagement in Digital Systems,” Human–Computer Interaction Journal, vol. 30, no. 3, pp. 194–215, 2015. [19] A. Sheppard and J. Wolffsohn, “Digital Eye Strain: Prevalence, Measurement and Management,” BMJ Open Ophthalmology, 2018. [20] M. Chen, Y. Hao, K. Hwang, L. Wang, and L. Wang, “Artificial Intelligence in Health Monitoring Systems,” IEEE Access, vol. 8, pp. 181452–181465, 2020.

Copyright

Copyright © 2026 Sathya Priya J, Haaniyah Yeshfeen K , Devi Priya S , Deekshitha M . 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.