Beyond Human Vision: AI-Powered Eye Tracking for Safety and Performance System

B. Venkata Charan Kumar, B. Megha Shyam Kumar, Y. Subbarayudu, S. Sandeep, U. Prashanth, M. Mahindra Reddy

2025

Abstract

This study presents a novel real-time approach to detecting eye blinks by utilizing computer vision techniques and geometric analysis of facial landmarks. The system employs OpenCV and Dib libraries to process video input, recognize facial features, and accurately identify instances of eye closure. By analyzing the ratio between the vertical and horizontal distances of specific eye landmarks, we develop a robust metric for blink detection that remains effective across diverse lighting conditions and facial angles. Experimental testing demonstrates a detection accuracy of 94.3% at 27 frames per second on standard hardware, making this solution viable for practical applications. The implemented system shows particular promise in driver fatigue monitoring systems, assistive technology interfaces, and clinical assessment of blinking patterns. This work contributes to the growing field of non-intrusive behavioral monitoring by providing an efficient, accessible method for eye blink detection that balances computational demands with real-time performance requirements.

Paper Citation

in Harvard Style

Kumar B., Subbarayudu Y., Sandeep S., Prashanth U. and Reddy M. (2025). Beyond Human Vision: AI-Powered Eye Tracking for Safety and Performance System. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 579-587. DOI: 10.5220/0013886700004919

in Bibtex Style

@conference{icrdicct`2525,
author={B. Kumar and Y. Subbarayudu and S. Sandeep and U. Prashanth and M. Reddy},
title={Beyond Human Vision: AI-Powered Eye Tracking for Safety and Performance System},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={579-587},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013886700004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Beyond Human Vision: AI-Powered Eye Tracking for Safety and Performance System
SN - 978-989-758-777-1
AU - Kumar B.
AU - Subbarayudu Y.
AU - Sandeep S.
AU - Prashanth U.
AU - Reddy M.
PY - 2025
SP - 579
EP - 587
DO - 10.5220/0013886700004919
PB - SciTePress