EUVIMEDEUVIMED entdecken
Crossref

Real-Time Driver Drowsiness Detection System Using Computer Vision

Gungun Batham, Naina Engle, Purva Shukla

International Journal of Sciences and Innovation Engineering · 27.05.2026 · 3 (5) · 1482-1488

DOI: 10.70849/ijsci03052661544 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Sciences and Innovation Engineering, online erschienen am 27.05.2026. Verfasst von Gungun Batham und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Driver fatigue and drowsiness remain among the most critical contributing factors to road traffic accidents worldwide. Despite advances in automotive safety systems, a significant portion of fatal crashes continue to be linked to inattentiveness caused by sleepiness, particularly during long-haul or night-time driving. This paper presents the design and implementation of a non-intrusive, real-time driver drowsiness detection system leveraging computer vision techniques and deep learning principles. The proposed system uses a standard webcam or vehicle-mounted camera to capture continuous video frames, from which facial landmarks are extracted using the MediaPipe Face Mesh framework. Drowsiness is assessed through two primary physiological indicators: the Eye Aspect Ratio (EAR) and the Mouth Aspect Ratio (MAR), which quantify eyelid closure and yawning behavior, respectively. A threshold-based consecutive-frame analysis algorithm is applied to distinguish genuine fatigue from transient eye blinks or momentary mouth opening. When the system detects sustained drowsiness indicators over a predefined frame window, an audible alarm is triggered to alert the driver. The system runs entirely in Python using OpenCV and MediaPipe, achieving near-real-time performance without specialized hardware. Experimental evaluation under varied lighting conditions and subject demographics demonstrated reliable detection performance with minimal false positives. The proposed approach offers a computationally lightweight alternative to sensor-based and camera-based deep learning classifiers, making it well-suited for integration into commercially available vehicles and smart transportation systems.

Zitieren

APA

Batham, G., Engle, N., & Shukla, P. (2026). Real-Time Driver Drowsiness Detection System Using Computer Vision. International Journal of Sciences and Innovation Engineering, 3(5), 1482-1488. https://doi.org/10.70849/ijsci03052661544

Vancouver

Batham G, Engle N, Shukla P. Real-Time Driver Drowsiness Detection System Using Computer Vision. International Journal of Sciences and Innovation Engineering. 2026;3(5):1482-1488. doi:10.70849/ijsci03052661544

BibTeX
@article{Batham2026,
  author = {Batham, Gungun and Engle, Naina and Shukla, Purva},
  title = {Real-Time Driver Drowsiness Detection System Using Computer Vision},
  journal = {International Journal of Sciences and Innovation Engineering},
  year = {2026},
  volume = {3},
  number = {5},
  pages = {1482--1488},
  doi = {10.70849/ijsci03052661544}
}
Kostenlos starten →