Adaptation and comparison of two facial recognition methods for driver drowsiness detection
DOI:
https://doi.org/10.26439/ciis2018.5495Keywords:
drowsiness detection, facial recognition, traffic accident, effectiveness rate, recognition rate, Viola-Jones, Regression Based Facial Landmark Detection, EARAbstract
This article seeks to make a comparison between two facial recognition methodologies: Viola-Jones and Regression Based Facial Landmark Detection and find the one that performs better in the face of different variable problems such as changing illumination, occlusion, face rotation, among others. The methodology that has a better rate of effectiveness for the detection of facial expressions is sought. Two facial detection methodologies will be adapted for the detection of drowsiness in drivers. After an analysis of advantages and disadvantages, a comparison of the results obtained will be shown. Different effects occur due to lack of sleep, such as decreased reaction time, eye fatigue, blurred vision, reduced concentration, etc.; These factors directly influence the increase in traffic accidents.
Downloads
References
Manu, B. N. (noviembre de 2016). Facial features monitoring for real time drowsiness detection. En: Innovations in Information Technology (IIT), 2016 12th International Conference, pp. 1-4. IEEE.
Rosenbrock, A. Soukupová, T., y Cech, J. (Febrero de 2016). Real-Time Eye Blink Detection Using Facial Landmarks. En: 21st Computer Vision Winter Workshop.
Viola, P., y Jones, M. J. (2004). Robust Real-Time Face Detection. International Journal of Computer Vision 57(2), pp. 137–154.
