Driver Drowsiness Detection
Driver Drowsiness Detection
A Project ReportSubmitted By
T.GURUCHARAN
200303124492
T.DHARMESH
200303124493
T.VENKATESH
200303124494
B. SESIDHAR
200303126070
Degree of
Prof.YAMINI BARGE
Assistant Professor
PARUL UNIVERSITYVADODARA
October - 2023
PARUL UNIVERSITY
C ERTIFICATE
This is to Certify that Project - 2 -Subject code 203105400 of 7th Semester entitled “HUNGERFREE -
EXTRA FOOD TRANSPORTATION TO NEEDY PEOPLE” of Group No. PIET CSE 79
has been successfully completed by
• T.DHARMESH - 200303124493
• T.VENKATESH - 200303124493
• B.SISIDHAR - 200303126070
under my guidance in partial fulfillment of the Bachelor of Technology (B.TECH) in Computer Science
and Engineering of Parul University in Academic Year 2023- 2024.
Date of Submission :-
External Examiner
ACKNOWLEDGEMENT
Behind any major work undertaken by an individual or group there lies the contribution of the people who
helped to them to cross all the hurdles to achieve their goal. It gives us the immense pleasure to express our
sense of sincere gratitude towards our respected guide YAMINI BARGE Madam for their persistent,
outstanding, invaluable co- operation and guidance. It was our achievement to be guided under them. They
are giving constant source of encouragement and momentum that any intricacy becomes simple. We gained a
lot of invaluable guidance and prompt suggestions from them during entire project work. We will be
indebted to them forever, and we take pride to work under them. we feel very privileged to have their
precious advice, guidance, and leadership. Last but not the least, our humble thanks to the Almighty God.
PLACE:VADODARA
DATE:
Project associates:
T.GURUCHARAN 200303124492
T.DHARMESH 200303124493
TVENKATESH 200303124494
B.SESIDHAR 200303126070
1
Abstract
Road accidents became a matter of concern due to the huge increase in traffic. The primary
cause of accidents is due to the drowsiness of drivers in the nighttime. Fatigue and drowsiness are
some of the leading causes of major accidents on Highways. The only solution to this problem is
detecting the drowsiness and alerting the driver.
So, in this project, we have thought of building a Driver Drowsiness Detection and Alerting
System for Drivers using Arduino Nano, Eye blink Sensor, and RF Transceiver module. The
basic purpose of this system is to track the driver’s eye movements using Eye blink Sensor and
if the driver is feeling drowsy, then the system will trigger a warning messageusing a loud buzzer
alert. Previously we have also built a Driver Drowsiness detector using Arduino UNO.
2
Contents
Certificate
Acknowledgement
Abstract
CHAPTER 1:
INTRODUCTION 2-12
1.2. Motivation 3
1.8. MEMORY 9
1.9. PINS 9
CHAPTER-2
LITERATURESURVEY 13-15
CHAPTER-3
METHODOLOGY 16-26
3.1. FLOWCHART _ 17
3.2. COMPONENTS 18
3
3.2.1. Relay 18
CHAPTER-4
4.1Conclusion 29
4.2Future Scope of the Project 30
REFERENCES 31-32
4
LIST OF FIGURES
7. 3.1 Flowchart 17
8. 3.2.1.1 Relay 19
9. 3.2.2.1 Buzzer 19
1
CHAPTER-1
2
INTRODUCTION
The goal of this project is to develop a system that can detect the sleepiness of the driver
and make alarms accordingly. There will be eye blinking sensor that will detects the driver
Driver Drowsiness
1.2. MOTIVATION
1.3. OVERVIEW
Road accidents became a matter of concern due to the huge increase in traffic. The primary
cause of accidents is due to the drowsiness of drivers in the nighttime. Fatigue and drowsiness are
some of the leading causes of major accidents on Highways. The only solution to this problem is
detecting the drowsiness and alerting the driver.
So, in this project, we have thought of building a Driver Drowsiness Detection andAlerting
System for Drivers using Arduino Nano, Eye blink Sensor, and RF Transceiver module. The basic
purpose of this system is to track the driver’s eye movements using Eye blink Sensor
3
and if the driver is feeling drowsy, then the system will trigger a warning message using a
loud buzzer alert.
Truck Accidents are most common if the driving is insufficient. The factors causing accidents are:
1) Speeding: Majorly in highways truck drivers ignore the speed limit. Speed kills and travelling
above the speed limit is an easy way to cause accident.
2) Drunk Driving: When the driver is drunk, he loses the ability to focus and function properly,
hence it is dangerous when operating a vehicle. This is one of the most contributing factors of
accident.
3) Reckless Driving: Improper driving as in speeding up the vehicle or changing lanes too
quickly or tailgating and many more can cause reckless driving. Reckless driving is when you’re
operating vehicle with deliberate disregard.
4) Night Driving: Driving in daylight can be hazardous, but driving at night nearly doubles the risk
of accident. When you can’t see what’s ahead you don’t know what to anticipate as you drive
towards it.
5) Driving Under the Influence Of Drug: Drugs, both legal and illegal can impair your ability to
fully function as a driver. Mind clearance and control over the body is need else it may cause
accident.
6) Drowsy Driving: Driver fatigue isn’t talked about a lot, but how well we can expect anyone to
drive when they are having trouble staying awake. Most of the car accidents caused by drowsy
driving occur at night.
Here we are concern about the factors which make the truck drivers fall asleep as in Drunk
Driving, Night Driving, Driving under the influence of drug and Drowsy Driving. As the result is
same for all the above mentioned factors that is,Eyes become "heavy" and blinking time is higher.
Driver tends to keep his eyes closed "at least for a few seconds. Hence here we create a buzzer and
a goggle for an adaptive driver and an application that provides driving behaviour to the company
owner. So as to warn the owner and the driver.
4
2) The implementation of Eye blinking sensor
For any vehicle accidents driver’s faults are the most accountable aspect to cause dangerous
problem to the society. Many drivers cannot control the vehicles due to different reasons it may
cause severe accidents and sometime death. For vehicle accidents various factors involved such as
drunk driving, over speeding, many distractions like texting while driving, talking with others,
playing with children etc. one of the important factor is sleeping on the wheel. People know the
dangerous of alcohol consumption and run the vehicles but they not understand the seriousness of
fatigue driving. In India, Ministry of Road Transport and Highway released a report in 2015, every
day around 1,374 accidents may happen and almost 400 people deaths occur. Every hour because of
vehicle accidents approximately 57 road accidents and 17 people dies. In that 54.1 percent of people
are in the age group of 15 to 34 years are killed in vehicle accident. The Government of India,
Ministry of Road Transport and Highway Government of India prepare a strategy to diminish the
amount of motorway accidents and losses by 50 % by 2020. Globally vehicle accidents have seemed
one of the major community health problems. In India almost 5 lakh road accidents happened in the
year 2015. A fatigue Driver those who falls asleep at the move fails to control the vehicle, not
possible to take immediate action and results in a crash so it is necessary to monitor the drowsiness
of the driver to prevent accidents.
1.3.2. COMPONENTS:
● Relay
● Piezo buzzer
● Wires:
● 9v battery:
● Gear motor:
● Wheel:
● Arduino UNO
5
● SPST switch:
CIRCUIT DIAGRAM
If found drowsy, the alarm system gets activated and the driver is alerted.
The life of the driver can be saved by alerting him using the alarm system.
Practically applicable
6
1.4. ORGANIZATION OF THESIS
Chapter 3: This chapter deals with the R-305 fingerprint module. All specifications about the
device are mentioned. Its communication protocol is also described.
Chapter 4: This chapter describes the interfacing of various devices to the Arduino Uno,
namely the LCD, the R-305 module and the four command buttons.
Chapter 5: This chapter explains the working of the software that runs our system.
How various routines and subroutines are executed is also described.
Chapter 6: This chapter gives the conclusion and the scope of future developments with regard
to this project.
Chapter 7: This chapter states the references utilized to accomplish this project work
The Uno is a MCU board based on the ATmega328P. It has been provided with 14
digital I/O pins (of these 6 can be employed as PWM O/Ps), 6 analog I/Ps, a USB connection,
a 16 MegaHz quartz crystal, an ICSP header, a power jack and a reset button. It has
7
everything required to support the microcontroller; we just got to connect it to a computer or
any such processing device with the provided USB cable or it can be powered using a 5V- 12V
AC- to-DC adapter or appropriate battery to begin operation. We can experiment with our Uno
without thinking about doing something wrong, at worst we may damage the MCU.
1.7. COMMUNICATION
The Uno contains a number of facilities and methods to communicate with the upper
computer, other Uno boards, or with other microcontrollers. The ATmega328 provides UART
TTL (5V) serial communication, which is available on digital pins 0 (RX) and 1 (TX). An
ATmega16U2 on the board channelizes this serial communication on the USB and it will
appear as a virtual com port to Arduino IDE on the upper computer
8
Serial communication on any of the Uno's digital pins is made possible using the SoftwareSerial
library provided with the Arduino IDE.
ATmega328P also supports I2C (TWI) and SPI Communication. To simplify use of the I2C
bus the Arduino Software (IDE) comes with a Wire library for the purpose.
1.8. MEMORY
The ATmega328 comes with 32 KiloByte (bootloader utilizes 0.5 KB of the flash). It
also contains 2 KiloByte of Static RAM and also 1 KiloByte of EEPROM (on which read
and write is performed with the help EEPROM library).
1.9. PINS
Arduino Uno has been provided with 14 Digital Pins which can be used as inputs or
outputs on user discretion. Using the function pinMode(), digitalRead() and digitalWrite()
the pins can be initiated. 5 volts is standard operational voltage for each of the pins. A current
of 20 mA flows though these pins under the recommended operating. A maximum current
value of 40mA mustn’t be exceeded on any Input/Output pin to avoid causing irreparable
damage to the microcontroller.
Some pins have specific purpose:
Serial Pins: 0 (RX) I/P and 1 (TX) O/P.
The use of these two pins is to receive (RX) and transmit (TX) TTL serial data. The RX and
TX LEDs blink when data is transmitted on these pins.
PWM: Pin numbers :3, 5, 6, 9, 10, and 11. Using the analogWrite() function these pins provide
8-bit PWM O/P .
LED: 13.The digital pin 13 drives a built in LED. If the pin has HIGH value, the LED glows,
when the pin has a LOW value, it's off.
Six clearly labeled pins A0 - A5 are the Arduino Uno’s 6 analog input pins, each providing 10
bits of resolution (that is 1024 different digital values). By default the upper end of their range
is 5 volts and the lower end of the range is at volts. This can be altered using the AREF pin and
the analogReference() function.
9
1.10. THE FAIL SAFE:
To protect our computer's internal parts and its USB ports from shorts and overcurrent,
the Uno contains a resettable polyfuse. It is known that most computers come with their own
system of internal protection, the fuse acts as an extra layer of safety. If a current of value more
than 500 mA is drawn to or from the USB port, the fuse would automatically break the circuit
till the short or overload is rectified
1.11. ATMEGA328/P MICROCONTROLLER:
By executing powerful instructions in a single clock cycle, the device achieves through
puts approaching 1 MIPS per MHz, balancing power consumption and processing speed.
Parameters:
Number of Comparators: 1
Pin Count: 32
11
The ATmega328P is low power Complementary MoS 8bit microcontroller, the architecture of
this is based up on the AVR enhancement Reduced instruction set computer architecture.It
will execute the instruction in a single clock cycle even the instruction s powerful. Atmega328P
can achieves high throughputs approximately 1 MIPS per MHz, so the designer have the full
flexibility to design and optimize between power consumption and processing speed.
The high performance microcontrollers integrate 32KiloBytes flash with read while
write features, 23 general purpose input output lines, 1024B EPROM, 32 general purpose
registers, 3 pliable timers with comparison modes, external and internal interrupts, a byte
directional two wire serial interfaces, serial port SPI, a six channel ten bit analog to digital
converter, the watch dog timer which can be programmable by using internal oscillator, and
power saving modes which are 5 and software selectable . The device operating range is
between 1.8-5.5 volts.
12
CHAPTER 2
Literature Survey
13
CONFERENCE-1: 2021 International Conference on Microelectronics (ICM)
TITLE: Validation of Driver Drowsiness Detection Based on Humantenna Effect Using Facial
Features
Along with the continued use of vehicles such as cars, trains, motorbikes, planes and other
transportation methods came the need to research automotive safety to avoid accidents that may be
caused by reckless, sleepy or drowsy drivers. By utilizing current technological advancements, it
could be possible to rapidly improve and create new safety measures and features in these vehicles.
This research presents validated results for a proposed driver drowsiness detectionsystem based on
the humantenna touch sensor that is simple, cost efficient, accurate and has a lot of room for
improvement. The use of this said is further validated by collecting simultaneous data obtained from
processing the images capturing the facial features of the driver through facial landmark detection
and eye aspect ratio calculation. The results obtained were based on the strong correlation between
driver drowsiness and both phenomena, the steering wheel grip pattern and eye closure. The
superposed data showed the different states possible for a driver in terms of humantenna effect and
eye aspect ratio (EAR) calculations. These results also showed the potential for more research and
improvement of the humantenna touch sensor.
CONFERENCE-2: 2020 15th Iberian Conference on Information Systems and Technologies
(CISTI)2020 15th Iberian Conference on Information Systems and Technologies (CISTI)
TITLE: Advanced Driver Assistance System for the drowsiness detection using facial landmarks
BY: Luis Darío Sinche Cueva, Jorge Cordero
This paper presents the development of a solution to detect a driver's drowsiness in real time and
issue alerts to avoid possible traffic accidents. In particular, an analysis of the methods used for the
detection of drowsiness by computer vision is performed, focusing on the use of facial
reference points. Distraction, drowsiness, tiredness, speeding and fatigue are the main causes of
accidents and, precisely, advanced driver assistance systems ADAS help reduce these serious
human errors.
14
CONFERENCE-3:2018 Tenth International Conference on Ubiquitous and Future Networks
(ICUFN)
TITLE: Real-Time Drowsiness Detection Algorithm for Driver State Monitoring Systems
BY: Jang Woon Baek, Byung-Gil Han,Kwang-Ju Kim,Yun-Su Chung,Soo-In Lee
In this paper, we proposes a novel drowsiness detection algorithm using a camera near the
dashboard. The proposed algorithm detects the driver's face in the image and estimates the
landmarks in the face region. In order to detect the face, the proposed algorithm uses an AdaBoost
classifier based on the Modified Census Transform features. And the proposed algorithm uses
regressing Local Binary Features for face landmark detection. Eye states (closed, open) is
determined by the value of Eye Aspect Ratio which is easily calculated by the landmarks in eye
region. The proposed algorithm provides realtime performance that can be run on the embedded
device. We obtained the dataset using video records from the infrared camera which is used the
real-field. The proposed algorithm tested in the target board (i.mx6q). The result shows that the
proposed algorithm outperformed in the speed and accuracy.
15
CHAPTER 3
METHODOLOGY
16
METHODOLOGY
3.1. FLOW CHART
3.2. Components:
● Relay
● Piezo buzzer
● Wire:
● 9v battery:
● Gear motor:
● Wheel:
● Arduino UNO
● SPST switch:
17
● Eye blink sensor:
3.2.1. Relay
The relay is the device that open or closes the contacts to cause the operation of the other electric
control. It detects the undesirable condition with an assigned area and gives the commands to the
circuit breaker to disconnect the affected area through ON or OFF.
1. Electromagnet
4. Spring
A piezo buzzer is a type of electronic device that’s used to produce a tone, alarm or sound. It’s
lightweight with a simple construction, and it’s typically a low-cost product. Yet at the same time,
depending on the piezo ceramic buzzer specifications, it’s also reliable and can be constructed in a
wide range of sizes that work across varying frequencies to produce different sound outputs.
18
The use of the piezo ceramic buzzer was discovered by Jacques and Pierre Curie back in 1880. They
found that electricity could be generated when a mechanical pressure was applied to particular
materials — and the inverse was true as well.
So when certain piezoelectric materials are subjected to an alternating field of electricity, the piezo
buzzer element — often a manmade piezoceramic material — stretches and compresses in sequence
with the frequency of the current. As a result, it produces an audible sound.
Unlike magnetic buzzers that have a narrow operating voltage of somewhere between one and 16
volts, piezo buzzers can typically operate anywhere between three and 250 volts. In addition,
magnetic buzzers have a higher power consumption of 30 to 100 milliamperes, while piezo buzzers
normally consume less than 30 milliamperes — even at higher rate frequencies. And although piezo
buzzers require a larger footprint than magnetic buzzers, they produce a higher sound pressure level.
3.2.3 Wires
A jumper wire is an electrical wire, or group of them in a cable, with a connector or pin at each end
(or sometimes without them – simply "tinned"), which is normally used to interconnect the
components of a breadboard or other prototype or test circuit, internally or with other equipment
or components, without soldering.
Individual jump wires are fitted by inserting their "end connectors" into the slots provided in a
breadboard, the header connector of a circuit board, or a piece of test equipment.
19
Jumper wires are simply wires that have connector pins at each end, allowing them to be used to
connect two points to each other without soldering. Jumper wires are typically used with breadboard
and other prototyping tools in order to make it easy to change a circuit as needed. Fairly simple. In
fact, it doesn’t get much more basic than jumper wires.
Though jumper wires come in a variety of colors, the colors don’t actually mean anything. This
means that a red jumper wire is technically the same as a black one. But the colors can be used to
your advantage in order to differentiate between types of connections, such as ground or power.
Jumper wires typically come in three versions: male-to-male, male-to-female and female-to-female.
The difference between each is in the end point of the wire. Male ends have a pin protruding and
can plug into things, while female ends do not and are used to plug things into. Male-to-male jumper
wires are the most common and what you likely will use most often.
3.2.4 9V Battery
The nine-volt battery, is an electric battery that supplies a nominal voltage of 9v. Actual voltage
measures 7.2 to 9.6 volts, depending on battery chemistry. Batteries of various sizes and capacities
are manufactured; a very common size is known as PP3, introduced for early transistor radios. The
PP3 has a rectangular prism shape with rounded edges and two polarized snap connectors on the
top. This type is commonly used for many applications including householduses such as smoke
and gas detector, clocks, and toys.
The 9V battery is an extremely common battery that was first used in transistor radios. It features
a rectangular prism shape that utilizes a pair of snap connectors which are located at the top of the
20
battery. A wide array of both large and small battery manufacturers produce versions of the 9V
battery. Possible chemistries of primary (non-rechargeable) 9V batteries include Alkaline, Carbon-
Zinc (Heavy Duty), Lithium. Possible chemistries of secondary (rechargeable) 9V batteries include
nickel-cadmium (NiCd ), nickel-metal hydride (NiMH ), and lithium ion. The performance and
application of the battery can vary greatly between different chemistries, meaning that some
chemistirs are better suited for some applications over other
A gearmotor, also called a gear motor or a geared motor, is a combination of a gear system or
gearbox and an electric motor. When asking yourself “what is a gear motor” or “what is a geared
motor,” you are actually probably wondering how a gearmotor can be useful for your company's
products or production facility. Geared motors are efficient because you will only have to mount
and operate one system, instead of several. Sometimes inaccurately called “gears motors,” or even
geared motors, gearmotors generally combine an efficient motor, such as an Electrically
Commutated Motor with a gear reducer or gearhead. These motor gear combinations are a great
way to minimize alignment issues with your motor and its gearbox. You don't need to be worried
about separately sourcing gears for motors; instead, you can work with a motor and gear in one: a
gear drive motor
21
Fig . 3.2.5.1 Gear motor
3.2.6 Wheel :
The Arduino UNO is a standard board of Arduino. Here UNO means 'one' in Italian. It
was named as UNO to label the first release of Arduino Software. It was also the first USB board
released by Arduino. It is considered as the powerful board used in various projects. Arduino.cc
developed the Arduino UNO board.
22
The Arduino UNO includes 6 analog pin inputs, 14 digital pins, a USB connector, a power
jack, and an ICSP (In-Circuit Serial Programming) header. It is programmed based on IDE, which
stands for Integrated Development Environment. It can run on both online a offline platforms.IDE,
which stands for Integrated Development Environment. It can run on both online and offline
platforms.
● There are 20 Input/Output pins present on the Arduino UNO board. These 20 pis include 6
PWM pins, 6 analog pins, and 8 digital I/O pins.
● The PWM pins are Pulse Width Modulation capable pins.
● The crystal oscillator present in Arduino UNO comes with a frequency of 16MHz.
● It also has a Arduino integrated WiFi module. Such Arduino UNO board is based on the
Integrated WiFi ESP8266 Module and ATmega328P microcontroller.
● The input voltage of the UNO board varies from 7V to 20V.
● Arduino UNO automatically draws power from the external power supply. It can also
draw power from the USB.
We can program the Arduino UNO using the Arduino IDE. The Arduino IDE is the Integral
Development program, which is common to all the boards.
24
We can also use Arduino Web Editor, which allows us to upload sketches and write the code from
our web browser (Google Chrome recommended) to any Arduino Board. It is an online platform.
The USB connection is essential to connect the computer with the board. After the connection, the
PWR pins will light in green.
A switch is a mechanical or controlling device that changes the flow of current direction or
interrupts the flow of current within a circuit. Switches are designed in different configurations
based on the requirements. They include various sets of contacts that are controlled through a similar
knob. These contacts may function sequentially, alternately, or simultaneously. Almost all
electronic or electrical systems include a minimum of one switch to turn ON/OFF the device. There
are different types of switches available like DPST, DPDT, SPDT, and SPST switch.
The term “SPST” in an SPST switch stands for “Single Pole Single Throw” which includes a
single input and a single output. Here, a single input is directly connected to a single output. The
main function of this switch is to control the circuit by turning ON/OFF.
Once the switch in the circuit is closed, then the circuit will be turned ON whereas the switch is not
closed or open, then the circuit will be turned off. The SPST switch example is the 25KV railway
DC voltage system & a household lamp switch. The schematic symbol of the SPST switch is shown
below.
The connections of this switch are two types like the Normally Open (NO) & the common (C).
Once the switch is enabled, then the circuit is closed. So the flow of current will be from the common
(C) terminal to the normally open (NO) terminal. When the switch is disabled, the circuit will be an
open circuit, so there is no flow of current within the circuit.
25
Fig 3.2.8.2 SPST Switch
The eye blink sensor is an infrared sensor. It contains two parts. A transmitter and a receiver. The
transmitter continuously emits infrared waves onto the eye. While the receiver continuously looks
for variations in the reflected waves which indicates that the eye has blinked.
If the eye is closed that means it will give high output. If the eye is open then it will give a low
output.
This sensor can be used in a very different variety of robotics and mechatronics projects as it
provides excellent results and is very economical.
The eye blinking system can be used for different purposes. First, let's look at the blink sensing
system. The eye blink sensor constantly sends infrared waves which are reflected and detected by
the receiver. As soon as the eye blinks, the output of the sensor goes high. This output is sent to the
Arduino board.
The Arduino sends the signal to the buzzer to make the noise. In this way, a blink is detected.The
Arduino sends the signal to the buzzer to make the noise. In this way, a blink is detectct.
26
27
28
CHAPTER 4
CONCLUSION
AND
FUTURE SCOPE OF THE
PROJECT
29
4.1Conclusion
Driver drowsiness detection is a car safety technology that helps prevent accidents caused
by the driver getting drowsy. With the advent increase in accidents on motor vehicles due
to drowsiness, driver drowsiness detection is an efficient car safety technology that helps
prevent such accidents
The future works may focus on the utilization of outer factors such as vehicle states, sleeping
hours, weather conditions, mechanical data, etc, for fatigue measurement. Driver drowsiness
pose a major threat to highway safety, and the problem is particularly severe for commercial
motor vehicle operators. Twenty-four hour operations, high annual mileage, exposure to
challenging environmental conditions, and demanding work schedules all contribute to this
serious safety issue. Monitoring the driver’s state of drowsiness and vigilance and providing
feedback on their condition so that they can take appropriate action is one crucial step in a
series of preventive measures necessary to address this problem. Currently there is not
adjustment in zoom or direction of the camera during operation. Future work may be to
automatically zoom in on the eyes once they are localized.
30
REFERENCES
1. "Road Traffic Injuries and Deaths—A Global Problem", CDC 2020, [online]
Available: https://www.cdc.gov/injury/features/global-road-
safety/index.html."Road Traffic Injuries and Deaths—A Global Problem", CDC
2020, [online] Available: https://www.cdc.gov/injury/features/global-road-
safety/index.html.
2. S. Chen, M. Kuhn, k. Prettner and & D.E. Bloom, "The global macroeconomic
burden of road injuries: estimates and projections for 166 countries", The Lancet
Planetary Health, vol. 3, no. 9, pp. 390-398, September 2019.S. Chen, M. Kuhn, k.
Prettner and & D.E. Bloom, "The global macroeconomic burden of road injuries:
estimates and projections for 166 countries", The Lancet Planetary Health, vol.
3, no. 9, pp. 390-398, September 2019.
31
8. J. Redmon, S. Divvals, R. Girshick and A. Parhadi, "You Only Look Once:
Unified Real-Time Object Detection", Proc. of International Conference on
Computer Vision and Pattern Recognition 2016, pp. 779-788, Jun. 2016.
10. S. Ren, K. He, R. Girshick and J. Sun, "Faster R-CNN: Towards Real-Time
Object Detection with Region Proposal Networks", IEEE Transaction on Pattern
Analysis and Machine Intelligence, vol. 39, pp. 1137-1149, 2016.
11. C. Kublbeck and A. Ernst, "Face detection and tracking in video sequence using
the modified census transformation", Image and Vision Computing, vol. 24, pp.
564-572, 2006.
12. D. Blome, J. Beveridge, B. Draper and Y. Lui, "Visual Object Tracking using
Adaptive Correlation Filters", Proc. of Computer Vision and Pattern Recognition
2010, pp. 2544-2550, Jun. 2010.
13. C. Sagonas, G. Tzimiropoulos, S. Zafeiriou and M. Pantic, "300 Faces in-the- Wild
Challenge: The first facial landmark localization Challenge", Proc. of IEEE Int'l
Conf. on Computer Vision, December 2013.
14. S. Ren, X. Cao, Y. Wei and J. Sun, "Face alignment at 3000fps via regressing local
binary features", Proc. Of IEEE Conf. on Computer Vision and Pattern
Recongnition, 2014.
32