See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/259964148

A Survey on Driver's Drowsiness Detection Techniques

Article · November 2013

CITATIONS READS

13 1,343

2 authors, including:

Dulari Bhatt
Nirma University
5 PUBLICATIONS   18 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Damage indication Algorithm for Structural health monitoring View project

All content following this page was uploaded by Dulari Bhatt on 31 January 2014.

The user has requested enhancement of the downloaded file.

International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 1 Issue: 11 816 – 819

______________________________________________________________________________
A Survey on Driver’s Drowsiness Detection Techniques

Jay D. Fuletra Dulari Bosamiya

M.E. Scholar, Department of Information Technology, Assistant Prof., Department of Information Technology,
Shantilal Shah Government Engineering College, Shantilal Shah Government Engineering College,
Bhavnagar, Gujarat (India) Bhavnagar, Gujarat (India)
[email protected] [email protected]

Abstract— Nowadays, there are many systems are available in market like navigation systems, warning alarm systems etc. to make
driver’s work easy. Traffic accidents due to human errors cause many deaths and injuries around the world. Drowsiness and
sleeping while driving are now identified as one of the reasons behind fatal crashes and highway accidents caused by drivers.
Various drowsiness detection techniques researched are discussed in this paper. These techniques are classified and then
compared using their features. Computer vision based image processing techniques is one of them. This uses various images of
driver to detect drowsiness states using his/her eyes states and facial expressions. This technique is on the focus of this survey
paper.

Keywords- Drowsiness detection techniques, Driver fatigue, Image processing, Face detection, Eye Detection, Accidents, Alert

________________________________________________________*****_____________________________________________________

artificial neural network based techniques. And image

I. INTRODUCTION processing based techniques can be divided in three categories.
The increasing number of traffic accidents due to a driver’s These categories are template matching technique, eye blinking
diminished vigilance level has become a serious problem for technique, yawning based technique. These techniques are
society. Some of these accidents are the result of the driver’s based on computer vision using image processing. In the
medical condition. However, a majority of these accidents are computer vision technique, facial expressions of the driver like
related to driver fatigue, drowsiness of drivers. Car accidents eyes blinking and head movements are generally used by the
associated with driver fatigue are more likely to be serious, researchers to detect driver drowsiness. Various drowsiness
leading to serious injuries and deaths. Fletcher et al. in [9] has detection techniques researched are discussed in this paper.
mentioned that 30% of all traffic accidents have been caused by
drowsiness. It was demonstrated that driving performance
deteriorates with increased drowsiness with resulting crashes II. VARIOUS DROWSINESS DETECTION TECHNIQUES
constituting more than 20% of all vehicle accidents As shown in fig. 1 there are three various techniques being
[10]Traditionally transportation system is no longer sufficient. used by researchers to detect drowsiness wiz. I) Images
One can use a number of different techniques for analyzing Processing based techniques II) Artificial neural network based
driver’s drowsiness. These techniques are Image Processing techniques III) EEG (electroencephalograph) based techniques.
based techniques, Electroencephalograph based techniques, and

Figure 1: Various Drowsiness Detection Techniques

816
IJRITCC | November 2013, Available @ http://www.ijritcc.org
____________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 1 Issue: 11 816 – 819

______________________________________________________________________________
A. Images Processing Based Techniques
In image processing based techniques, drivers face images
are used for processing so that one can find its states. From the
face image one can see that driver is awake or sleeping. Using
same images, they can define drowsiness of driver because in (a) (c)
face image if driver is sleeping or dozing then his/her eyes are
closed in image. And other symptoms of drowsiness can also
detected from the face image. We can classify these techniques
in three sub-categories.
1) Template Matching Technique: In this technique, one
can use the states of eye i.e. if driver closes eye/s for some (b) (d)
particular time then system will generate the alarm.Becasue in Figure3: (a) Normal mouth image (b) Staring yawn image (c)
this techniques system has both close and open eyes template wide open mouth larger then speaking ,its yawn (d) closing
of driver. This system can also be trained to get open and mouth completing yawn.
closed eye templates of driver. When yawn is detected by syetem then it alarm the driver.

Instead of using just one technique to detect drowsiness of

driver, some researchers [1, 2, 3] have combined various
vision based image processing techniques to have better
performance.
B. EEG Based Technique
In this technique it is compulsory to wear electrode helmet
by drivers while driving. This helmet have various electrode
sensors which placed at correct place and get data from brain.
Researchers [7] have used the characteristic of EEG signal in
drowsy driving. A method based on power spectrum analysis
and FastICA algorithm was proposed to determining the
Figure 2: Open and closed eyes template. fatigue degree. In a driving simulation system, the EEG signals
of subjects were captured by instrument NT-9200 in two
This method is simple and easy to implement because states, one state was sober, and the other was drowsy. The
templates of both open and closed eye states shown in figure 2 multi-channel signals were analyzed with FastICA algorithm,
to remove ocular electric, my electric and power frequency
are available to system. Researchers have used this technique.
interferences. Figure 4 shows how EEG based systems get data
[4]
for acquisition. Experimental results show that the method
2) Eye Blinking based Technique: In this eye blinking rate
presented in this paper can be used to determine the
and eye closure duration is measured to detect driver’s drowsiness degree of EEG signal effectually.
drowsiness. Because when driver felt sleepy at that time
his/her eye blinking and gaze between eyelids are different
from normal situations so they easily detect drowsiness. In this
system the position of irises and eye states are monitored
through time to estimate eye blinking frequency and eye
clouse duration. [5]. And in this type of syetem uses a remotly
plced camrea to aquire video and computer vision methodes
are then applied to sequentially localize face, eyes and eyelids
positions to measure ratio of closure.[11] . Using these eyes
clouser and blinking ration one can detect drowsiness of
driver.
3) Yawning Based Technique: Yawn is one of the
symptoms of fatigue. The yawn is assumed to be modeled with
a large vertical mouth opening. Mouth is wide open is larger in Figure 4: EEG data acquisition system
yawning compared to speaking. Using face tracking and then
mouth tracking one can detect yawn. In paper [6], they detect C. Artificial Neural Network Based Technique
yawning based on opening rate of mouth and the amount In this technique they use neuron to detect driver’s
changes in mouth contour area as shown in figure 3. drowsiness. Only one neuron is not much accurate and the
result of that is not good compare to more than one neurons.
Some researchers [8] are carrying out investigations in the field
of optimization of driver drowsiness detection using Artificial
Neural Network. People in fatigue exhibit certain visual
817
IJRITCC | November 2013, Available @ http://www.ijritcc.org
____________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 1 Issue: 11 816 – 819

______________________________________________________________________________
behaviors that are easily observable from changes in facial light, external illuminations interference, vibrations, changing
features such as the eyes, head, and face. Visual behaviors that background and facial orientations. This paper has very good
typically reflect a person’s level of fatigue include eyelid results of drowsiness detection but in this system, sometimes it
movement, gaze, head movement, and facial expression. To creates false alarms.
make use of these visual cues, they made artificial neural A nonintrusive prototype computer vision system for
network to detect drowsiness. They tested samples and got 96% monitoring a driver’s vigilance in real time is proposed [2]. It is
result. Figure 5 shows that flow how an artificial neural based on a hardware system for the real-time acquisition of a
network system can detect drowsiness. driver’s images using an active IR illuminator and the software
implementation for monitoring some visual behaviors that
Capture Image characterize a driver’s level of vigilance. They used Percent eye
closure (PERCLOS), eye closure duration, blink frequency,
Use Neuron nodding frequency, face position, and fixed gaze. This system
for detection consists of four major modules: 1) image acquisition; 2) pupil
detection and tracking; 3) visual behaviors; and 4) driver
vigilance. They detect drowsiness using visual behaviors and
Show Result pupil detection. These parameters are combined using a fuzzy
classifier to infer the level of inattentiveness of the driver and
Figure5: Working of ANN detect state of driver and if it detects fatigue then this system
alert the driver. This system is fully autonomous; it can
III. REVIEW OF IMAGE PROCESSING BASED DROWSINESS initialize automatically, and reinitialize when necessary. The
DETECTION performance of the system decreases during daytime, especially
General architecture for drowsiness detection using vision in bright days, and at the moment, the system does not work
based image processing techniques is shown in Figure 2. First with drivers wearing glasses.
of all in these techniques they captured video putting camera in Driver hypo vigilance (fatigue and distraction) detection
vehicle and get images from video frames. From these video based on the symptoms related to face and eye regions is
frames one can use face detection algorithms to detect face of introduced [3]. There are three main contributions in this
driver. After that eyes detection algorithms is used to detect method: (1) simple and efficient head rotation detection based
eyes. Than eyes and face tracking algorithms are used to track on face template matching, (2) adaptive symptom extraction
them. Using these various processed images one can detect from eye region without explicit eye detection, and (3)
drowsiness using various symptoms and techniques which they normalizing and personalizing the extracted symptoms using a
defined in their systems. Researches carried using this short training phase. They used eye region related symptoms
technique is reviewed here. like PERCLOS, eyelid distance changes with respect to the
Advanced Driver Assistance System (ADAS) has been normal eyelid distance (ELDC), and eye closure rate
proposed [1]. In this, a research project to develop a non- (CLOSNO). The symptom related to face region is head
intrusive driver’s drowsiness system based on Computer Vision rotation (ROT). The proposed method extracts the symptoms
and Artificial Intelligence has been presented. This system uses related to eye region using horizontal projection of top half
advanced technologies for analyzing and monitoring drivers segment without explicit eye detection; the symptom related to
eye state in real-time and in real driving conditions. They use face region is extracted based on face template matching.
different algorithms for various tasks like face tracking, eye Monitoring these symptoms, one can detect drowsiness to
tracking etc. they give separate results for face tracking, eye alarm the driver. The main disadvantage of our system is the
tracking, eye state analysis. Based on the results presented in face tracking method which is inaccurate and very
this paper, the proposed algorithm for face tracking, eye computationally complex.
detection and eye tracking is robust and accurate under varying

Figure 2: General Architecture of Drowsiness Detection

much simpler and user friendly. Driver’s spectacles make
this complex, but researches are going on to eliminate this
IV. CONCLUSION drawback. So there is a lot of scope in drowsiness detection
After reviewing various techniques used for drowsiness using Image Processing.
detection, we can concluded that, different techniques will be
suitable according to given conditions. EEG based REFERENCES
techniques are efficient but practically it is not suitable for [1] Marco Javier Flores • José María Armingol • Arturo de la
driver to wear electrodes. Artificial Neural Network based Escalera.: Real-Time Warning System for Driver Drowsiness
Using Visual Information. In: Springer Science + Business
technique is simple but if you want better result than 3 Media B.V. 2009
neurons is ideally suitable. Image Processing techniques are [2] Luis M. Bergasa, Jesús Nuevo, Miguel A. Sotelo, Rafael
one of the favorite ones for researchers. These techniques are Barea, and María Elena Lopez.:Real-Time System for
818
IJRITCC | November 2013, Available @ http://www.ijritcc.org
____________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 1 Issue: 11 816 – 819

______________________________________________________________________________
Monitoring Driver Vigilance. In: ieee transactions on Systems and Knowledge Discovery (FSKD), 2010 Seventh
intelligent transportation systems, vol. 7, no. 1, march 2006 International Conference on , 5, pp. 2164-2167, 10-12 Aug.
[3] Mohamad-Hoseyn Sigari, Mahmood Fathy, and Mohsen 2010.
Soryani.: A Driver Face Monitoring System for Fatigue and [8] Er. Manoram Vats and Er. Anil Garg.: Detection And
Distraction Detection. In: Hindawi Publishing Corporation Security System For Drowsy Driver By Using Artificial
International Journal of Vehicular Technology, Volume 2013, Neural Network Technique. In: International Journal of
Article ID 263983, 11 pages Applied Science and Advance Technology January-June
[4] D.Jayanthi, M.Bommy. : Vision-based Real-time Driver 2012, Vol. 1, No. 1, pp. 39-43
Fatigue Detection System for Efficient Vehicle Control. In: [9] Fletcher, L., Petersson, L., Zelinsky, A.: Driver assistance
International Journal of Engineering and Advanced systems based on vision in and out of vehicles. In: IEEE
Technology (IJEAT) ISSN: 2249 – 8958, Volume-2, Issue-1, Proceedings of Intelligent Vehicles Symposium, pp. 322–327
October 2012 (2003)
[5] Artem A. Lenskiy and Jong-Soo Lee.: Driver’s Eye Blinking [10] Azim Eskandarian and Ali Mortazavi, “Evaluation of a smart
Detection Using Novel Color and Texture Segmentation algorithm for commercial driver drowsiness detection”
Algorithms. In: International Journal of Control, Automation, presented at Proceeding of the 2007 IEEE Intelligent Vehicles
and Systems (2012) 10(2):317-327 DOI 10.1007/s12555-012- Symposium, June 2007
0212-0 ISSN:1598-6446 eISSN:2005-4092 [11] Amol M. Malla, Paul R. Davidson, Philip J. Bones, Richard
[6] Behnoosh Hariri, Shabnam Abtahi , Shervin Shirmohammadi, Green and Richard D. Jones, “Automated Video-based
Luc Martel.: A Yawning Measurement Method to Detect Measurement of Eye Closure for Detecting Behavioral
Driver Drowsiness. In: Distributed and Collaborative Virtual Microsleep” presented at 32nd Annual International
Environments Research Laboratory, University of Ottawa, Conference of the IEEE EMBS Buenos Aires, Argentina,
Ottawa, Canada August 31 -- September 4, 2010
[7] Ming-ai Li, Cheng Zhang, Jin-Fu Yang. :An EEG-based
Method for Detecting Drowsy Driving State. In: Fuzzy

819
IJRITCC | November 2013, Available @ http://www.ijritcc.org
____________________________________________________________________________

View publication stats