Post on 04-Jun-2018
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
1/20
Salvatore Vitabile, Alessandra De Paola, Filippo SorbelloDepartment of Biopathology and Medical Biotechnology andForensics, University of Palermo, Italy
1 / 20
Journal of Ambient Intelligence and Humanized ComputingPublished on March 30, 2011
Chien-Chih(Paul) ChaoChih-Chiang(Michael) Chang
Instructor: Dr. Ann Gordon-Ross
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
2/20
An embedded monitoring system to detectsymptoms of drivers drowsiness.
2 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
3/20
MotivationRelated worksDrowsiness Monitoring System
Eye Regions SegmentationCandidate Eye Regions SelectionDrivers Eyes Detection
Drowsiness Level ComputationExperimental trialsConclusionLimitations & Future Work
3 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
4/20
10-20% of all European traffic accidents are dueto the diminished level of attention caused by
fatigue.In the trucking industry about 60% of vehicularaccidents are related to driver hypo-vigilance. [1]Automotive has gained several benefit from theAmbient Intelligent researches involving thedevelopment of sensors and hardware devices
4 / 20[1] Awake Consortium (IST 2000-28062), System for effective assessment of
driver vigilance and warning according to traffic risk estimation (AWAKE),Sep 2001 2004 [Online], available: http://www.awake-eu.org
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
5/20
The technique categories for preventing driversdrowsiness [2]
Readiness-to-perform and fitness-for-duty
technologiesMathematical models of dynamics alertnessVehicle-based performance technologies
The lateral position Steering wheel movements time-to-line crossingReal-time technologies for monitoring drivers status
Intrusive monitoring systems Non-intrusive monitoring systems
5 / 20[2] Hartley L, Horberry T, Mabbott N, Krueger G (2000) Review of fatigue detection and
prediction technologies. National Road Transport Commission report 642(54469)
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
6/20
The most accurate techniques are based onphysiological measures
Brain wavesHeart ratePulse rate
Causing annoyance due to require electrodesto be attached to the drivers
6 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
7/20
A non-intrusive, real-time drowsinessdetection system.Using FPGA instead of ASIC of DSP
Re-programmabilityPerformanceCosts
IR cameraLow light conditionsBright pupilphenomenon to detect the eyes
7 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
8/20
PERCLOS (Percentage of Eye Closure)The driver eyes are closed more than 80%
within a specified time interval is defined asdrowsiness. [3]
8 / 20
[3] W. W. Wierwille: Historical perspective on slow eyelid closure: Whence PERCLOS?,In Technical Proceedings Ocular Measures of Driver Alertness Conference, FederalHighway Admin., Office Motor Carrier Highway Safety, R. J. Carroll Ed. Washington,D.C., FHWA Tech. Rep. No. MC-99-136, 1999
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
9/20
9 / 20
Bright Pupil
Threshold Operation
Clipping & MorphologicalOperation
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
10/20
10 / 20
A list of blobsPossible Eye Pairs
Square Bounding BoxR = a
Quasi-circular shape:
R
a
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
11/20
11 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
12/20
12 / 20
Frame 1
[ (X1, Y1) ,(X2, Y2) ]t = 4
Class 1
Coordinate
At t
Class
Frame 2
[ (X1, Y1) ,(X2, Y2) ]t = 3
Class 1
Frame 3
[ (X1, Y1) ,(X2, Y2) ]t = 2
Class 1
Frame 4
[ (X1, Y1) ,(X2, Y2) ]t = 1
Class 1
Class 1
Weight 4
Class 2 Class 3 Class 4 Class 5
0 0 0 0
[ (X1, Y1) ,(X2, Y2) ]t = 5
Class 2
3 1
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
13/20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
14/20
JSP DF-402 infrared-sensitive cameraColor camera in daytime
Infrared camera under low light cond.
14 / 20http://www.es.ele.tue.nl/education/oo2/fpga/board.php
http://www2.bren.ucsb.edu/~dturney/WebResources_13/RemoteSensing/RemoteSensing.htmhttp://www2.bren.ucsb.edu/~dturney/WebResources_13/RemoteSensing/RemoteSensing.htm8/13/2019 Topic4-Realtime Drowsiness Detection System Final
15/20
Celoxica RC203EXilinX XC2V3000-4 Virtex II FPGA
Handel-C PixelStreams Library
http://www.es.ele.tue.nl/education/oo2/fpga/board.php 15 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
16/20
In light controlled environmentDrive-Camera relative distance
Not affected by driver-camera relative distance16 / 20
ID =1
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
17/20
Vertical and Horizontal of head movement
17 / 20
ID =2
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
18/20
Real operation condition(External illumination not controlled)
18 / 20
ID =3
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
19/20
An algorithm to detect and track the driverseyes has been developed by exploiting brightpupils phenomenonGood performance on rapid movements ofdrivers head. Performance not affected by driver-camera
relative distance.The drowsiness monitoring system can beused with low light conditions by usinginfrared camera
19 / 20
8/13/2019 Topic4-Realtime Drowsiness Detection System Final
20/20
Faulty operationsthe driver is wearing glasses
the drivers IR-reflecting objects such as earringDrowsiness usually happen during theevening/night hours
Light poles might be recognized as eyecandidates due to the shape and size on screen
20 / 20