Drowsy driver detection system using eye blink patterns

Danışman, Taner; Bilasco, Ioan Marius; Djeraba, Chabane; Ihaddadene, Nacim · 2010 · OpenAlex-citations

DOI: 10.1109/icmwi.2010.5648121

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 factor in fatal single-vehicle run-off-road crashes. While existing Drowsy Driver Detection Systems (DDDS) often rely on expensive hardware like infrared cameras or invasive physiological sensors (e.g., EEG, EOG), this study proposes a low-cost, fully automatic solution using a standard color webcam. The system aims to detect drowsiness by monitoring eye blink patterns, specifically the duration of eye-lid closures, to alert drivers before they fall asleep. The methodology employs a computer vision pipeline that processes video frames in real-time. First, face detection is performed using the Viola-Jones detector, followed by eye localization using a neural network-based detector derived from the STASM library. To handle head movement, the system estimates face orientation based on pupil positions and rotates the frame to correct for rolling effects up to ±25 degrees. The core innovation lies in the eye blink detection mechanism, which utilizes the horizontal symmetry property of the eye. The region of interest around the pupil is divided into upper and lower halves; an open eye exhibits horizontal symmetry, whereas a closed eye does not. By calculating the difference between the vertically flipped upper half and the lower half, the system determines the eye state from a single frame without requiring initialization or temporal analysis of previous frames. Drowsiness is classified into three states—Awake, Drowsy, and Sleeping—based on blink duration thresholds of 400ms and 800ms, respectively. Experimental results were evaluated using the JZU eye-blink database, which contains 80 videos with varying lighting conditions, head movements, and subjects wearing glasses. The system achieved a 94.8% accuracy in detecting eye blinks, with a precision of 90.7% and a recall of 71.4%. Notably, the false positive rate was only 1%. The system operates at 110 frames per second on a 320×240 resolution video stream using an Intel Xeon 2.9 GHz CPU, demonstrating suitability for real-time applications. The authors note that while the method is robust against head movements, performance is affected by the presence of glasses and high illumination changes, as these factors interfere with face and eye detection components. The significance of this work lies in providing a computationally efficient, non-invasive, and low-cost alternative to existing drowsiness detection technologies. By relying on spatial symmetry rather than temporal statistical data, the system reduces computational load and initialization requirements. Although the study focuses primarily on eye blink detection accuracy due to the lack of a common drowsiness database, the results suggest that monitoring blink duration via standard webcams is a viable strategy for preventing accidents caused by driver fatigue. The authors plan to further validate the approach using the Av@Car database to assess drowsiness detection performance in more complex driving scenarios.

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-prismaquant on 2026-06-26; verification: verified.

Topics

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

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

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