Drowsiness Detection and Alertness Using Eye Motion Monitoring

Asghar, Rabia; Hassan, Ahmad; Junejo, Naveed Ur Rehman; Ikram, Farwa; Cheema, Abeera Mahfooz · 2023 · Crossref

DOI: 10.33317/ssurj.573

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the critical safety issue of driver drowsiness, a leading cause of fatal traffic accidents. Motivated by statistics indicating that a significant majority of truck accidents and driver fatigue incidents stem from diminished vigilance, the authors propose a non-intrusive, real-time alertness monitoring system. The primary objective is to detect drowsiness through eye state analysis and trigger an alarm to prevent accidents, offering a robust alternative to more complex or intrusive physiological monitoring methods. The proposed methodology utilizes computer vision and machine learning techniques implemented in Python using OpenCV and Dlib libraries. The system operates in four distinct stages: face detection, eye detection, eye state analysis, and drowsiness decision-making. A webcam captures video frames, which are processed to identify facial features using a 68-face landmarks algorithm. Specifically, the system isolates the eye regions (landmarks 37–42 for the right eye and 43–48 for the left eye). To determine whether the eyes are open or closed, the system calculates the Euclidean distance between specific vertical and horizontal points on the eyelids, deriving an Eye Aspect Ratio (EAR). If the calculated distance falls below a threshold value of 0.26 for a duration of one second, the system classifies the driver as drowsy. Upon detection, the system triggers an audible beep via the Win Sound library and displays a "Sleeping" message. The model employs a linear Support Vector Machine (SVM) with Histogram of Oriented Gradients (HOG) for classification. Experimental results demonstrate that the proposed system achieves a high accuracy of 95.5% with a low error rate of 4.5%. This performance surpasses several state-of-the-art methods, including a dual-camera system with 83.0% accuracy and eye blink monitoring systems with accuracies ranging from 90.7% to 94.0%. The system is capable of processing 30 frames per second on standard hardware and remains effective under low-light conditions and for drivers wearing spectacles. It successfully detects both fully closed and partially closed eyes. However, the authors note limitations, including potential inaccuracies caused by reflective objects behind the driver or extreme lighting conditions, and acknowledge that the threshold may require adjustment for diverse ethnic groups. The significance of this work lies in its provision of a cost-effective, single-camera solution for real-time drowsiness detection that outperforms existing benchmarks in accuracy. By reliably identifying fatigue through simple eye motion monitoring, the system offers a practical tool for enhancing road safety. The authors conclude that while the current model is effective, future work should focus on developing more autonomous systems capable of monitoring additional factors, such as pupil movement and distraction levels, to further improve reliability in varied driving environments.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• drowsiness detection algorithms
• gaze based attention detection
• dms validation
• microsleep
• drowsiness
• distraction detection algorithms

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: tool software, measurement protocol