COMPUSOFT, 3(11), 1216-1223.pdf

COMPUSOFT, An international journal of advanced computer technology, 3 (11), November-2014 (Volume-III, Issue-XI)

ISSN:2320-0790

I-Blink: Drowsiness Detection and Warning System 1

2

3

4

5

O.M.B Jayasundera , M.D.A Kavindya , T.L Guruge , D.W Senevirathne and Y. Jayaweera Faculty of computing, Sri Lanka Institute of Information Technology, Metro Campus, Colombo 03, Sri Lanaka Abstract: This research paper sees the sights of the scope which drivers could be able to warn for the cause of eye drowsiness. Vehicle accidents caused by driver inattention marks a significant increase all over the country with the development of technology and growth of human population. Drowsiness is a key factor of driver distraction and numbers of traffic accidents were caused by driver drowsiness. Even though several drowsiness detection systems can be seen currently in Sri Lanka, and the implemented system; i-Blink: Drowsiness Detection and Warning System has targeted specifically to ease the process and to reduce the difficulty for the users. As Sri Lanka is a developing country, some of the technologies which are in use internationally would be inflexible to adapt in terms of usability and the implementation cost. This research aims to introduce a system which overcomes the above mentioned constraints in Sri Lanka and a system which is affordable to almost every driver. It will also be a noteworthy and a timely asset for the drivers when considering the number of fatal accidents caused by drowsiness. The proposed system would be eligible to keep the driver alerted and conscious by delivering an effective real time alarm, saving the life of drivers as well as the lives of the pedestrians. 1, 2, 3, 4, 5

Keywords: Drowsiness, Driver distraction, Drowsiness detection systems, Real time alarm, Image processing probability o f road accidents goes high. Drowsiness INTRODUCTION “The number of human causations make happen by related accidents appear to be more severe, possibly vehicle accidents” is a major concern on today’s because of the higher speeds involved, distraction world. These happen on most factors if the driver is and the driver being unable to take any avoiding drowsy or if he is intoxicating. Therefore driver action, or even brake, prior to the accident. drowsiness is recognized as an important factor in the vehicle accidents and distraction due to drowsiness of A) Research Research problem pr oblem t o be addressed addressed drivers can be identified as the main cause for driver In Sri Lanka throughout past few years, the inattention. Drowsiness is unforeseen, unavoidable Government implemented many programs to reduce and beyond the control of the dr iver. Stress, illnesses, illnesses, road traffic accident fatalities and serious injuries and certain type of medicines, repetitive driving and of course Sri Lanka has its own set of issues to liquor can make the driver drowsy. Sleepiness address in road safety. However drowsiness can be increases reaction time which is critical in driving. taken as the unescapable and hard cause to being With the help of the advanced technology, a way of avoided when it comes to road accidents. In that case, reducing the number of accidents can be offered to the need of a drowsiness detection system is required some extent. for drivers. Many researchers have been conducted all over the word on driver drowsiness detection Road accidents in Sri Lanka cause economic losses systems on various aspects and in various angles and worth around Rs.9.34 billion each year. It can be seen in Sri Lanka, it is significantly a less amount. there are about 2,400 road deaths every year which is one death every four hours approximately. Among Though there driver inattention systems systems which comes them, 82% - 85% are males. It has been calculated along with high end luxurious vehicles, not all can around 20% of traffic accidents with driver fatalities afford that state of a vehicle as Sri Lanka is yet a are caused by driver distraction and drowsiness. It developing country. Adding to the problem, the was revealed that driving performance rapidly drop existing systems use heavy equipment such as with increased drowsiness with resulting accidents external cameras, mini computers etc. for the process creating more than 20% of all vehicle accidents. Less and they are with high cost in purchasing and attention leads the driver to being distracted and the

1216


maintenance. Convenience is a matter of fact when fixing heavy equipment to a vehicle.

face detection. It contains number of features of the face. It is constructed by using a number of positive and negative samples. First load the cascade file. Then pass the acquired frame to an edge detection function, which detects all the possible objects of different sizes in the frame. The output the edge detector is stored in an array. Now, the output of the edge detector is then compared with the cascade file to identify the face in the frame. The eye detection can be done by using the haarcascade file. By extracting the pixel value of the eye region can detect the drowsiness. If drowsiness is detected, a text message is displayed along with triggering an audio alarm [1].

LITERATURE R EVIEW EVIEW I-blink is basically focused on detecting the driver drowsiness. There are many existing systems and researches conducted under drowsiness detection all over the world world and few in Sri Lanka. In the current current era, the road accidents have been increased because of driver’s inattention and combination of physical and mental condition. It may cause regarding many reasons. Among them human drowsiness gets a vital place. There are many research projects projects done from universities and other organizations, companies in worldwide on drowsiness detection, warning and r eal time alarming. alarming. These information gathering tasks were directed at developing the best experimental research plan for i-Blink-Drowsiness Detection and Warning System. The following contains an overview of past researches conducted by several project teams.

b) Dr owsin owsin ess ess detec detection tion system ystem in Sr i L anka “Riyadisi” is another drowsiness detection system which has implemented in Sri Lanka. The system is a real time automated noninvasive system which uses computer vision and machine learning techniques to detect driver drowsiness and distraction and then alerts the driver appropriately. Riyadisi focused on driver’s eye-blinking eye-blinking rate, the yawning patterns and the movements. They used IR camera to detect the eye blinking patterns in order to analyze the blinking rate of the eye and also to overcome the challenge of different illumination conditions. They uses a video stream to analyze the drowsiness condition. They use visual cue based method to monitor the driver attention. The issue of this method is the different illumination condition to overcome this issue “Riyadisi” uses an IR camera which works both day and night [2].

a) Use of H aar Cascade Cascade Classif Classif ier J. Suryaprasad et.al explains the way of face/eye detection techniques using image processing in real time. In this research, it further explains the way of using the haarcascade samples and the differentiation of eye blink and drowsy/fatigue detection. detection. This paper introduces a vision based technique to detect the drowsiness. The major challenges are face detection, Iris detection under various conditions and developing the real time system.

There are 5 distinct modules which can be seen in their system architecture. Video acquisition is dividing in to frames, which are face detection, Eye detection, and drowsiness detection. Camera and audio alarm use as the hardware components. By making use of the camera, they get the live video of the driver and convert them into series of frames. By taking on each frame provide but the frame grabber and tries to detect the face of the driver. This is achieved through set of pre-defined haarcascade samples. The detection of the eye is making use of the set of pre-defined haarcascade samples. By taking consideration of eye states and the blinking rate detects the function of driver drowsiness. Proposed algorithm first get the video from the external camera, then convert in to frames, then in order to detect the eye and the face each frames extracted from the video. In the face detection method has 2 important functions. Identifying the region of interest, and detect the face using haarcascades. Open CV by default expects a particular resolution of the video that is being recorded. The face detection achieved by making use of the haarcascade file for

c) Use of of OpenCV i n eye bli nk detection detection

D.Jayanthi and M.Bommy et.al, have proposed a vision-based real-time driver fatigue detection system based on eye-tracking. eye-tracking. Face and eyes eyes of the driver are first contained and then marked in every frame obtained from the video source. At first, an ordinary color webcam is used t o capture the images. The first frame is used for initial face detection and eye location. The live video is captured and stored in the database. Current eye images are used as the dynamic templates for eye tracking on subsequent frames. If the process is fail, the face detection and eye location restart on the current frame. After face detection has been done the eyes are configured by giving the control points around the eyes. The eye template is then cropped and stored for the used dynamic template matching method. If the eyes of the drivers are closed at a certain time, then the driver is said to be in fatigue state and an alarm is raised. After get the eye templates, they have used gray scale correlation over eye region to find the position of the eye. However, there will be some false detection,

1217


where the results are not good when there is quick head-movement [3].

to the template image and the Template image the patch image which will be compared to the template image. To identify the matching area, have to be comparing the template image against the source image by sliding. Sliding is moving the patch one pixel at a time (left to right, up to down). At each location, a metric is calculated .so it represents how “good” or “bad” the match at that location is (or how similar the patch is to that particular area of the source image). The brightest locations indicate the highest matches. Practically we use the function min Max Loc to locate the highest value in a matrix. OpenCV implements Template matching in the function match Template. Template. There are several method for this. Some of them are, 1.method=CV_TM_SQDIFF 2.method=CV_TM_SQDIFF_NORMED 3.method=CV_TM_CCORR 4. method=CV_TM_CCORR_NORMED 5.method=CV_TM_CCOEFF 6.method=CV_TM_CCOEFF_NORMED

The proposed method was validated by tracking eye position within high and low occlusion condition. In this system use the connected component technique and centroid method to track and blinking of eyes on OpenCV platform. For that it captures the video from camera. OpenCV supports capturing images from a camera or a video file (AVI). First initialize the capture from a camera by setting zero and frame converted into Gray scale. Using OpenCV conversion, convert color image into Gray scale image. These images give as the binary image and locate it using the centroid method. To find the centroid of an image, the image first has to be binaries and then centroid program will calculate the process of the centroid. It looks like where the majority of black pixels are located. After mark the eye, it indicates with rectangle by using of the connected component technique. Those values use as parameters. If eyes are found during tracking then its X and Y coordinates are detected by finding centroid of contours. This gives the center coordinates of the eye pupil. This method is for detection of eye blinking. Based on blink detection detection it decides by that th at whether driver is drowsy or active [4].

What template matching algorithm simply does is, Loads an input input image and the image patch patch (template (template)) .Then perform a template matching procedure by using the OpenCV function match function match Template with the above mentioned 6 matching methods. After normalize the output of the matching procedure, and then localize the location with higher matching probability. probability. Finally draw a rectangle around the area corresponding to the highest match [6]. Whereas W.O.A. Basir et.al describes interfering system used to identify if the driver of the vehicle is sleepy and at risk of falling asleep at the wheel due to drowsiness. The system consist of two different sub-system and a control unit. First sub system consists of sensors of arrays placed in the vehicle seats. Second sub systems consist of heart rate monitoring sensors placed in the steering wheel. Control unit analyze the sensory data to determine the driver’s drowsiness states [7].

d) Use of M AT L AB software for i mage process processin g

M. Singh, and G. Kaur et.al, have introduced how the drowsy driver detect by using an algorithm. This is based on monitoring the changes in the eye blink duration. The system detects eye blinks via a standard webcam in real-time YUY2_640x480 resolution by using mean shift algorithm. In this they got the value of the eye closure percentage and the time for which alarm driven, this is changed according to every person, it set in the program. If the percentage increases then the alarm goes off. In the image processing approach analyze the image captured by the camera to detect the physical changes of the driver. They used MATLAB software and using image processing system of MATLAB software they measured the percentage of eyelid closure over the pupil over time. As the output first one by blowing alarm and second is graphical method to take a record. The proposed system is independent from the eye blinks duration as it works within the same frame. The proposed system detects eye blinks with 99% accuracy and a 1% false positive rate [5].

f ) Use of PCA, L DA and external external equipment equipment i n bli nk detec detection tion

J.Jo,S. J.Lee, H.G. Jung, K.R.Park, J. Kim.,propose a new driver monitoring method considering driver drowsiness and distraction. If the driver is looking ahead, drowsiness detection is performed; if not distraction detection is performed. Secondly a new eye-detection algorithm is introduced. It combines adaptive boosting, adaptive template matching, and blob detection with eye validation. Those algorithms reduce the eye-detection error and processing time significantly, by achieved the mentioned algorithms. Third they have used principal component analysis (PCA), and linear discriminate analysis (LDA) in

e) Use of of template matchin g

Template matching is a technique for finding areas of an image that match (are similar) to a template image (patch).there are two image categories the Source image the image in which we expect to find a match

1218


order to achieve accurate eye detection. Fourth, they have proposed a novel eye state – detection detection algorithm that combines appearance features obtained using PCA and LDA, with statistical features such as the sparseness and kurtosis of the histogram from the horizontal edge image of the eye. The proposed method consists of face-detection, head orientation – estimation, estimation, eye-detection, eye-detection, eye-state – detection, drowsiness-detection, and distractiondetection [8].

Research Gap A. Research Significant number of literature reviews had been reviewed by the research team of i-Blink in order to find out the gap which has not been covered. Number of developed drowsiness detection systems which comes as factory fitted for high end luxury vehicles can be seen and this technology does not come inbuilt with other normal vehicles. Existing external warning systems uses high and heavy technology where some of the drivers might not be able to afford them and the cost of maintenance is significantly high in such systems. Need of an economical, convenient and efficient system which most of the people can use or afford will be the research gap which is looked upon here in this r esearch. esearch. Usage of i-Blink is comparatively simpler in a high degree and it is with less cost than the existing systems which uses high technology and high end devises. The system is developed to detect the human eye blink and then to warn the drivers if a particular driver is detected with drowsiness with the use of a mobile application in the android platform.

METHODOLOGY Various development methodologies can be identified and used according to the specific nature of the project. Each methodology is developed considering the strengths, weaknesses and also their benefits of using the particular adaptation. Out of these methodologies, the team decided to go with Prototype development method for the development of i-Blink.

Figure 1: Flow of modules (Detailed)

The proposed system consists of two 850-nm illuminators(LED 850-66-60) a camera (EC650) having lens (ML 0614) of 6-mm focal length a laboratory-made device for controlling the illuminators, and a narrow band pass pass filter placed in in front of the camera. The system should work during both daytime and nighttime. Because visible lights can confuse drivers when driving, NIR illuminators were used to capture images.

A. Planning This was the phase where the development team identified the need of I-Blink and the value which will be gained by the use of the system. Objectives and the basic functionalities that will come use from I-Blink was clearly determined. Which are, o o o o

Face Detection Eye Blink Detection Warning the driver via an alarm Sending an SMS

Figure 2: Devices attached

A series of feasibility analysis was carried out by the team members to identify the risks, constraints and whether to proceed with the project. Technical feasibility was carried out to identify the risk associate with the familiarity of the system. Economic feasibility was performed in order to find out the costs and benefits associated with i-Blink in the long run. As a method of managing the control of the project and the direction of the project, a proper a Gantt chart was prepared.

Their eye-detection algorithm is developed by combining eye adaboost, adaptive template matching, blob detection, and eye validation. The adaboost method has been widely adopted for detecting various facial components, in this proposed system has used the standard OpenCV Haar cascade to detect the face and eyes. The template matching is a digital image processing technique. This algorithm identify the corresponding parts that match a template in an input image [9].

1219


system. As the system is a client based architecture, it would basically rely on client side where as it would consider the service of the server only for pr ocessing. When comes to the interface design, there is not much interfaces to be deal with user. So a less number of interfaces would be in use for i-Blink but, all the interfaces are designed in order to maintain user friendliness. Final stage of the design phase is the program design phase. Pieces of code were produced by each member according to the given guidelines and to achieve the derived objectives by the team. Integration of the code segments was done in this stage and a system which behaves as the system was developed. Prototype system consisted only the main functions where as it was mainly for the use to examine the failures and errors. Failures which the team found was eliminated before coming into the implementation stage.

B. Analysis Analysing process of all existing systems and past researches was carried out in this stage. Each system which is there in the use was identified and analysed separately and the team could identify their weakness and strengths. Most importantly, the research gap of the project was clearly identified in this process. A questionnaire was used to gather information which targeted the drivers in Sri Lanka. According to the results of the questionnaire, a significant number of drivers drive for a longer time in both day and night. Though they have not met with any accidents related to drowsiness driving, they prefer a system to be attached into their vehicles. Most of the drivers own a smartphone and they have android as the operating system. And in general, all the smartphones are with the needed processing power to run the developed application. This shows up a clear idea and a clear path carry on with the development development of the project. project. An entity relationship diagram was prepared to represent the data model which describes how the system represent data and how the system access them. Analysing and planning the solution was an easy task for the research team since the solution is less theoretically based but more technically weighted project.

mplementation D. I mpleme Integration of the sub modules was taken place in this stage. Failures which was detected when building the prototype was corrected by the programmers in this stage. Once the actual programming was completed and the system was nearing its completion, the team started the test plans to test the system. The results of the tests were documented in order for the future use. A complete executable program was developed once the testing was completed. The system consists with only an android application. The android application was developed using Eclipse and it uses OpenCV and basic android libraries for image processing. Few algorithms will be used to detect the eye blink and eye features. Haar Cascade classifier is being used to detect the eye and a separate classifier to detect the face was used along with other algorithms. Finally, the documentation with the final report of the complete project was prepared.

C. Design Figure 1 illustrate the high level architectural diagram of the system. Users directly interact with the mobile phone which will be the main operation and processing hub. The results will be delivered to the users on the go since the project is a real time processing system. system.

Figure 3: Architectural diagram

Conversion of the logical diagrams into physical diagram is done at the physical designing stage. This will deliver a basic idea of the blue print of i-Blink. The plan for the hardware, software will be designed in the architectural design stage. It will also help to develop the communication infrastructure of the

Figure 4: Basic Flow of the System

1220


Basic flow of the system between the modules is shown in the figure 4 above.

the application is to be used inside vehicles in Sri Lanka.

E. Testing During the test phase all aspects of the system were tested for functionality and performance. Essentially, the key elements of the testing phase was to verify that the game contains all the end user requirements laid out in the analysis phase and that it meets the quality standards of r ehabilitation ehabilitation This project was tested using two main methods.

B. Evidence This section describe about the test results of the system and there are four main testing done by the research team called software testing, unit testing, integration testing and finally system testing. Since the application was built using eclipse targeting the android users, a lot of libraries was used consisting both android and an d openCV libraries. Software Software testing was carried out to check whether these libraries and other pieces of code segments was error free. Unit testing was carried out to check the output the hardware component which is the mobile phone. Different parts of the complete project was developed accordingly. So to check these modules and to check whether all these separate code segments be able to give out a common output, Integration testing carried out. Finally a complete test was carried out to ensure that the complete system works accordingly and to check whether it gives the stated output of the project meeting the objectives. This test can be named as a System test. Objective of the a bove mentioned testing process is to give out an efficient, efficient, affective, affective, error free system to the users.

  

Integration Testing System Testing Software Testing

RESULTS AND DISUSSION This section can be divided in to three parts. Research will discussed about the findings of the research and how it works and compared the research with similar project to show how this project will unique from the others. About the tests and the interfaces are shown in the section under evidence and finally in the discussion section, research team will discuss about the reliability and accuracy level o f the system. And also discussed about the technical problems occurred in the research and finally the solutions for those t ypical problems.

Only a couple of simple interfaces were needed to carry out the while task. So the interfaces in numbers shows a less value. The figure 6 illustrates the basic view of the eye blink detecting interface which will be th e main interface out of all. As you can see, the face, eye section (both eyes) and separate eyes are detected separately. Mats are drawn around detected surfaces just for the convenience.

A. Research Before implement the system research team did a questioner to identify the research component of the system and would the project be of worth carrying out targeting the local drivers. And also to get an idea of what they really need. After getting the results, the research team identified the research component or the core of the research project which is a system to detect the drowsiness of a driver. With respect to the research component, the team found out the best possible way to deliver the solution meeting the project goals and objectives, objectives, which is developing a mobile application. Building the solution consisted of several main parts. Detecting the face of the driver, detecting the eye of the driver and detecting the eye blink which will give a blink count are the named main functionalities. For this process the team used an algorithm called template matching and cascade classifiers to detect the face and the eye. As a sub module, apart from the driver alarm, alternative SMS service was also implemented to notify the driver further. Testing was a major task to be dealt with a lot of concern since the system relies heavily on accuracy. The application was tested for different angles, different light conditions and with dynamic movement since

Figure 5: Eye blink detection interface C. Discussion

As the developed system is a warning system to warn the drivers, it heavily relies on the accuracy of the system. Drivers should be warned right on time on the correct situation in the most efficient way to get them out of the drowsiness condition. Since the

1221


captured frames are taken in the form using a red colour filter, the detection process can be carried out even in low light conditions as well. To detect drowsiness, eye blink rate was counted and it was then compared with the standard eye blink rate of a human being which is around 70-72 per minute. Calculations were based on the standard eye blink rate per minute. In addition, a simple text message is designed to be sent to a pre taken number of the user if the alarm was not r esponded on time.

come inbuilt with other normal vehicles. Existing external warning systems uses high and heavy technology where some of the drivers might not be able to afford them and the cost of maintenance is significantly high in such systems. systems. The usage of the system is comparatively simpler in a high degree and it is with less cost than the existing systems which uses high technology and high end devises. The system is developed to detect the human eye blink and then to warn the drivers if a particular driver is detected with drowsiness with the use of a mobile application via alarm alerts. This system iBlink: Drowsiness Detection and Warning System, has targeted specifically to ease the process and to reduce the difficulty for the users. As Sri Lanka is a developing country, some of the technologies which are in use internationally would be inflexible to adapt in terms of usability and the implementation cost. This research aims to introduce a system which overcomes the above mentioned constraints in Sri Lanka and a system which is affordable to almost every driver.

The main issue the team faced was gaining the technical knowledge to develop the system. As this research project was completely a whole new area of study, the language skills and th e required knowledge were two major points to consider when building the system. With the help of online tutorial and references papers of existing researches, the team could overcome the issue. The other main technical issue was to determine whether the system was reliable to be use in mobile phones. Since the processing power power of mobile phones are comparatively less, it was needed to find the best possible way to do d o the th e processing part of the system system and that too in the best efficient way not harming the accuracy and the performance of the output.

Even though the system is unique in a way of its components, functionality and has numerous benefits to its users, certain constraints were also identified in the feasibility analysis process. By considering the constraints of the system, some assumptions can be made to make the system a feasible one. One of the limitation that has been found out is the detection process cannot be done under low light situations situations and the camera should consist of a favourable resolution. Therefore the camera should be 1.3 Mega Pixel or above. Battery power of the mobile phone would be another constraint for the system and distance between the camera and the driver should not be changed as well. Since the overall objective of the system is to provide an effective drowsiness detecting and warning system to Sri Lankan vehicle drivers, the system is concerned only with drowsiness caused traffic accidents. Therefore the solution is limited to the cause of drowsiness and other types of causes are not addressing through i -Blink.

All the processing parts were designed to be implemented separately using another application which was to be developed using MATLAB. But as the research proceeded, the team find out a more effective way and moved on to stick with just a mobile application without using a third party application which will reduce the efficiency when meeting the research goals and objectives. CONCLUSION Driver drowsiness is recognized as an important factor in the vehicle accidents and distraction due to drowsiness of drivers can be identified as the main cause for driver inattention. Drowsiness is unforeseen, unavoidable and beyond the control of the driver. Stress, illnesses, certain type of medicines, repetitive driving and liquor can make the driver drowsy. Sleepiness increases reaction time which is critical in driving. With the help of the advanced technology, a way of reducing the number of accidents can be offered to some extent. Many researches have been conducted on driver drowsiness detection systems on various aspects and in various angles. Significant number of literature reviews had been reviewed by the research team of i-Blink in order to find out the gap which has not been covered. Number of developed drowsiness detection systems systems which comes as factory fitted for high class luxury vehicles can be seen and this technology does not

However the system solves issues with regard to accidents caused by drowsiness effectively and efficiently that also will help the drivers to keep in alert and be conscious. As an output, an accurate alarm at any needed moment will be provided on time regardless of the location by overcoming the issues of existing ineffective systems. I-Blink: Drowsiness Detection and Warning System will be a fruitful system which coincides with local technological standards and helpful for the researchers who are interested in carrying out further

1222


modifications to other types of causes of traffic accidents and other related fields as well. Eventually the system will be a noteworthy and a timely asset for the drivers when considering the number of fatal accidents caused by drowsiness and help to save the life of drivers as well a s the lives of the pedestrians.

http://airccj.org/CSCP/v http://airccj.org/CSCP/vol3/csit3805.pdf ol3/csit3805.pdf Aug ,18 2013].

[Accessed:

[2].M.D.Y. Fernando, S.Jayewardene, S.K.K. Wikramanayake and C.R.de Silva K.U.G.S. Darshana, "Riyadisi Automated driver attention assistance system," in National Engineering conference, Srilanka, 2013, p. 5.[Accessed:16 feb 2014].

This research project opens up pathways to conduct further more in the field of image processing for mobile applications. As this project is based in the android platform, applications targeting Windows users and IOS users would be a great asset in the asset in the future. Addition of sub modules, not only to detect drowsiness detection, but also which covers other driver inattention prevention will be another degree which can be covered by the r esearchers. Fine tuning this existing research like to suit even in the low light conditions, and a method to prevent the memory consumption of the mobile phone are some of the modification which is possible to be carried out.

[3] D.Jayanthi, M.Bommy “Vision-based “Vision -based Real-time Driver Fatigue DetectionSystemfor Efficient Vehicle Control “International Journal of Engineering and Advanced Technology (IJEAT) ,vol.2,pp.A238A242, October 2012]. [4] D.Pimplaskar1, Dr. M.S. Nagmode, A.Borkar” Real Time Eye Blinking Detection and Tracking Using Opencv” Dhaval Pimplaskar et al Int. Journal of Engineering Research and Applications, vol 3, pp.A1780-A1787, Sep-Oct Sep-Oct 2013 [5]G. K. Mandeep Singh, "Drowsy Detection On Eye Blink Duration Using Algorithm," International Journal of Emerging Technology and Advanced Engineering, vol. 2, no. 4, 2012.

ACKNOWLEDGMENT We would like to express our sincere sense of gratitude to our institution - Sri Lanka Institute of Information Technology (SLIIT). We are deeply indebted to our Lecturer in charge for the subject Comprehensive Design/Analysis Project Mrs. Gayana Fernando, whose help, stimulating suggestions, knowledge, experience and encouragement helped us in all the times of study and analysis of the project in the pre and post research period. We are also grateful grateful to Mr Iresh Bandara who helped us in many ways to a great extent in the project. Also very sp ecial thanks to our lecture panel and seniors. The completion of this undertaking could not have been possible without the participation and assistance of so many people people whose names may not all be enumerated.

[6]TemplateMatching”,Apr21,2014.[O [6]TemplateMatching”,Apr21,2014.[Online].Availab nline].Availab le.http://docs.opencv.org/doc/tuto le.http://docs.opencv.org/doc/tutorials/imgproc rials/imgproc/histog /histog rams/template_matching/templa rams/template_matching/template_matching.html te_matching.html.. [Accssed: Sep.8, 2014]. [7]W. O.A.Basir, J.P.Bhavnani, W.W.aterloo, F.Karray, and K.Desmchem, "DROWSINESS DETECTION SYSTEM," pdf US 6,822,573 B2, November 23, 2004, 2004, Accessed: Accessed: [21 Aug 2014]. [8].J.Jo ,S. J. Lee , K. R. Park , I.J. Kim, and J. Kim “Detecting driver drowsiness using feature-level feature -level fusion and user-specific user-specific classification “Jan. 2014.

R EFERENCES EFERENCES

[9].YCOE, T. Sabo, G. Kashi, “Drowsy Detection On Eye Blink Duration Using Algorithm” International Journal of Emerging Technology and Advanced Engineering,vol.2,pp.A363-A365, Engineering,vol.2,pp.A363-A365, April 2012.

[1].Dr.Suryaprasad J, Sandesh D, Saraswathi V, Swathi D, Manjunath S,” real tim e drowsy driver detection using haarcascade samples” [online].Available://

1223

COMPUSOFT, 3(11), 1216-1223.pdf

Recommend Documents