Classification for Driver’s Distraction and Drowsiness Through Eye Closeness Using Receiver Operating Curve (ROC)

Anis Hazirah Rodzi; Zalhan Bin Mohd Zin; Norazlin Ibrahim

doi:10.32734/jocai.v4.i1-3516

Authors

Anis Hazirah Rodzi anis.rodzi@s.unikl.edu.my
Zalhan Bin Mohd Zin Universiti Kuala Lumpur Malaysia France Institute, Selangor, Malaysia
Norazlin Ibrahim norazlin@unikl.edu.my

DOI:

https://doi.org/10.32734/jocai.v4.i1-3516

Keywords:

Classification, Distraction, Drowsiness, Driver, Advanced Driver Assistance Systems, Receiver Operating Curve, PERCLOS, SVM

Abstract

In Malaysia, driver inattention and drowsiness becomes one of the causes of road accidents which sometime could lead to fatal ones. From the data provided by Malaysian Police Force, Polis Di Raja Malaysia or PDRM in 2016, deaths from road accidents increased from 6,706 in 2015 to 7,512 in 2016. Some accidents were caused by human factor such as driver's inattention and drowsiness. This problem motivates many parties to look for better solution in dealing with this human factor. Some of the car manufacturers have introduced to their certain models of car with an assistant system to oversee driver’s condition. The assistant system is in fact part of the main safety system known as Advanced Driver Assistance Systems (ADAS). The kind of system has been developed to strengthen vehicle systems for safety and conducive driving. The system has been contemplated to congregate accurate input, rapid processing data, precisely predict context, and respond in real time. In addition to that, suitable method is also needed to detect and classify driver drowsiness and inattention using computer vision as the latter become more and more important in any intelligent system development. In this paper, the proposed system introduces a method to classify drowsiness into three different classes of eye state; open, semi close and close. The classification has been done by using feature extraction method, percentage of eye closure (PERCLOS) technique and Support Vector Machine (SVM) classifier. The performances of the methods have been then measured and represented by using confusion matrix and ROC performance graph. The results have show that the proposed system has been able to achieve high performance of distraction and drowsiness detection according to driver's eye closeness level.

Downloads

Download data is not yet available.

References

G. K. M. MBA, “Drowsiness:Causes,Symptoms and Diagnosis.” [Online]. Available: https://www.healthline.com/symptom/hemorrhage. [Accessed: 20-Nov-2017].

Cholinesterase - Blood, “MedlinePlus Medical Encyclopedia: Drowsiness.” [Online]. Available: http://www.nlm.nih.gov/medlineplus/ency/article/003358.htm. [Accessed: 20-Nov-2017].

MIROS, “2016 MIROS Annual Report,” 2017.

T. Pradhan, A. N. Bagaria, and A. Routray, “Measurement of PERCLOS using eigen-eyes,” 4th Int. Conf. Intell. Hum. Comput. Interact. Adv. Technol. Humanit. IHCI 2012, pp. 0–3, 2012.

C.-F. Tsai, “Bag-of-Words Representation in Image Annotation: A Review,” ISRN Artif. Intell., vol. 2012, pp. 1–19, 2012.

K. Kesorn and S. Poslad, “An enhanced bag-of-visual word vector space model to represent visual content in athletics images,” IEEE Trans. Multimed., vol. 14, no. 1, pp. 211–222, 2012.

Z. Lu and H. H. S. Ip, “Image categorization with spatial mismatch kernels,” 2009 IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. Work. CVPR Work. 2009, vol. 2009 IEEE, pp. 397–404, 2009.

X. Chen, X. Hu, and X. Shen, “Spatial weighting for bag-of-visual-words and its application in content-based image retrieval,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 5476 LNAI, pp. 867–874, 2009.

Bosch, X. Muñoz, and R. Martí, “Which is the best way to organize/classify images by content?,” Image Vis. Comput., vol. 25, no. 6, pp. 778–791, 2007.

Y. Junsong, W. Ying, and Y. Ming, “Discovery of collocation patterns: From visual words to visual phrases,” Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., 2007.

F. Monay, P. Quelhas, J. M. Odobez, and D. Gatica-Perez, “Integrating co-occurrence and spatial contexts on patch-based scene segmentation,” Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., vol. 2006, 2006.

D. Gökalp and S. Aksoy, “Scene classification using bag-of-regions representations,” Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., 2007.

P. Eliasson and N. Rosén, “Efficient K-means clustering and the importance of seeding.”