Real-Time Driver Drowsiness and Distraction Detection System

This paper described a real-time driver drowsiness and distrac- tion detection system built on OpenCV and dlib facial landmark analysis. The pipeline detects eye closure through EAR, yawn- ing through MAR and head orientation through PnP-based pose estimation, issuing an audible alert whenever a sustained drowsi- ness or distraction event is detected. Evaluation on the NTHU Drowsy Driver Detection dataset yielded an average accuracy of 89.0% across four conditions, and generalisation testing on UTA-RLDD confirmed an F1-score of 82.4% for drowsy de- tection in naturalistic recordings. The system operates at over 25 fps on a mid-range CPU, making it a viable candidate for deployment in low-cost embedded platforms. Future work will focus on personalised calibration, more robust face detection under sunglasses, and multi-modal fusion with vehicle telemetry data.

[1]        T. Soukupová and J. Cˇ ech, “Real-Time Eye Blink Detec- tion Using Facial Landmarks,” in Proc. 21st Computer Vision Winter Workshop (CVWW), Rimske Toplice, Slove- nia, Feb. 2016. [2]        C.-H. Weng, Y.-H. Lai, and S.-H. Lai, “Driver Drowsi- ness Detection via a Hierarchical Temporal Deep Belief Network,” in Proc. Asian Conference on Computer Vision (ACCV) Workshops, 2016, pp. 117–133. [3]         R. Ghoddoosian, M. Galib, and V. Athitsos, “A Realistic Dataset and Baseline Temporal Model for Early Drowsi- ness Detection,” in Proc. IEEE CVPR Workshop on Au- tonomous Driving, 2019, pp. 11341–11349. [4]         C. Zhao, H. Zhang, M. Zhang, B. Liu, and H. Sun, “A Review on Driver Drowsiness and Distraction Detection Using Computer Vision Techniques,” IET Intelligent Trans- port Systems, vol. 16, no. 5, pp. 604–623, 2022. [5]         E. Murphy-Chutorian and M. M. Trivedi, “Head Pose Esti- mation in Computer Vision: A Survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, no. 4, pp. 607–626, Apr. 2009. [6]        J.-M. Guo, H. Markoni, and J.-D. Lee, “Driver Drowsiness Detection Using Hybrid Convolutional Neural Network and Long Short-Term Memory,” Multimedia Tools and Applications, vol. 78, no. 20, pp. 29059–29087, 2019. [7]        B. Reddy, Y.-H. Kim, S. Yun, C. Seo, and J. Jang, “Real- Time Driver Drowsiness Detection for Embedded System Using Model Compression of Deep Neural Networks,” in Proc. IEEE CVPR Workshops, 2017, pp. 438–445. [8]         A. Maggioni, L. Liu, R. Siegwart, and M. Chli, “Towards Real-Time Drowsiness Detection on Embedded Hardware,” IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 3893–3900, 2021.