Driver fatigue and distraction are among the leading causes of road accidents worldwide. Early detection of these conditions can significantly improve road safety and reduce accident risks. However, many existing driver monitoring systems rely on expensive hardware such as infrared cameras and physiological sensors, limiting their practical use. This project presents an AI-Based Driver Drowsiness Detection System that use computer vision and deep learning techniques to monitor driver alertness in real time using only a standard webcam. The proposed system combines multiple methods to achieve both accuracy and efficiency. Facial landmarks are detected using MediaPipe, and the Eye Aspect Ratio (EAR) is calculated to identify prolonged eye closure, a common indicator of drowsiness. To improve reliability, a lightweight Convolutional Neural Network (CNN) is used to classify eye states when EAR-based prediction is uncertain. This hybrid approach minimizes unnecessary computation while maintaining real-time performance. The system also incorporates head pose estimate to detect driver distraction and a dynamic alertness score to continuously evaluate the driver`s attention level. Developed using Python, OpenCV, MediaPipe, PyTorch, and Streamlit, the solution is lightweight, scalable, and easy to deploy. Experimental results demonstrate that the system effectively reduces false alarms while achieving real-time performance of over 20 frames per second on a standard CPU. Overall, the proposed approach provides a practical, non-intrusive, and cost-effective solution for enhancing road safety and supporting intelligent transportation systems.

Road accidents caused by driver fatigue and distraction remain a major concern worldwide, highlighting the need for intelligent and real-time monitoring systems. This research presented an AI-Based Driver Drowsiness Detection System that combines computer vision and deep learning to detect both drowsiness and distraction in a non-intrusive and cost-effective manner. The system uses a standard webcam and requires no specialized hardware. The first phase used Eye Aspect Ratio (EAR) and MediaPipe Face Mesh for real-time eye closure detection. This approach was fast and lightweight but sometimes produced false alerts due to natural blinking and environmental changes. In the second phase, a CNN-based model improved eye state detection accuracy under varying conditions. To improve efficiency, the third phase used a hybrid approach combining EAR-based filtering with CNN verification. The reduced unnecessary processing while maintaining high accuracy and smooth real-team performance. The final phase added head pose estimation and dynamic alertness scoring for distraction detection and continuous monitoring. Experimental results showed stable real-time performance around 20-25 FPS with fewer false positives.

1.Albadawi, Y., AlRedhaei, A., & Takruri, M. (2023). A real-time machine learning approach for driver drowsiness detection using visual features. Journal of Imaging, 9(5), 91. 2.An, Y.-J., Jeong, C., Park, Y.-H., Kim, G., & Kang, C. M. (2025). Combining MediaPipe and L2CSNet for driver drowsiness detection and gaze tracking. In Proceedings of the International Conference on Electronics, Information, and Communication (ICEIC) (pp. 1–4). 3.Badri, F., Ruhmana Sari, S. U., & Bin Hamzah, S. A. (2025). Analysis of driver drowsiness detection systems based on facial landmarks and MediaPipe. Inform: Jurnal Ilmiah Bidang Teknologi Informasi dan Komunikasi, 10(1), 21–28. 4.Dinges, D. F., Mallis, M. M., Maislin, G., & Powell, J. W. (1998). Evaluation of techniques for ocular measurement as an index of fatigue and the basis for alertness management. U.S. Department of Transportation. 5.Florez, R., Palomino-Quispe, F., Alvarez, A. B., Coaquira-Castillo, R. J., & Herrera-Levano, J. C. (2024). A real-time embedded system for driver drowsiness detection using eye and mouth analysis with CNN and MAR. Sensors, 24(19), 6261. 6.Florez, R., Palomino-Quispe, F., Coaquira-Castillo, R. J., Herrera-Levano, J. C., Paixão, T., & Alvarez, A. B. (2023). A CNN-based approach for driver drowsiness detection by real-time eye state identification. Applied Sciences, 13(13), 7849. 7.Kim, D., Park, H., Kim, T., Kim, W., & Paik, J. (2023). Real-time driver monitoring system with facial landmark-based eye closure detection and head pose recognition. Scientific Reports, 13, 18264. 8.Kose, N., Kopuklu, O., Unnervik, A., & Rigoll, G. (2019). Real-time driver state monitoring using a CNN-based spatio-temporal approach. In Proceedings of the IEEE Intelligent Transportation Systems Conference (ITSC) (pp. 3236–3242). 9.Lugaresi, C., Tang, J., Nash, H., McClanahan, C., Uboweja, E., Hays, M., Zhang, F., Chang, C.-L., Yong, M. G., Lee, J., Chang, W.-T., Hua, W., Georg, M., & Grundmann, M. (2019). MediaPipe: A framework for building perception pipelines. 10.Michelaraki, E., Katrakazas, C., Kaiser, S., Brijs, T., & Yannis, G. (2023). Real-time monitoring of driver distraction: State-of-the-art and future insights. Accident Analysis & Prevention, 192, 107241. 11.Safarov, F., Akhmedov, F., Abdusalomov, A. B., Nasimov, R., & Cho, Y. I. (2023). Real-time deep learning-based drowsiness detection: Leveraging computer vision and eye-blink analyses for enhanced road safety. Sensors, 23(14), 6459. 12.Sawant, A. G., Kamble, S. S., Kanade, R. S., Kanugo, R. N., Kapse, T. A., & Bhapse, K. A. (2025). A real-time driver drowsiness detection system using MediaPipe and eye aspect ratio. 13.Soukupová, T., & Čech, J. (2016). Real-time eye blink detection using facial landmarks. In Proceedings of the Computer Vision Winter Workshop (CVWW) (pp. 1–8). 14.World Health Organization. (2023). Global status report on road safety 2023. 15.Zaman, A. N. K., Chatterjee, P., & Sharma, R. (2024). Real-time drivers’ drowsiness detection and analysis through deep learning. In Proceedings of the 2nd International Conference on Artificial Intelligence, Blockchain, and Internet of Things (AIBThings).

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close