2. ABSTRACT
According to the previous year's statistics on road
crashes, the primary cause of such deadly road accidents is driver
irresponsibility and tiredness. This dilemma highlights the need
for a system that can detect a motorist's tiredness and provide an
alert signal to the driver before an accident occurs. As a result of
this suggested study, a drowsiness detection and accident
avoidance system based on the duration of eye blinks has been
developed. The open and closed states of the eye are initially
recognized using the eye aspect ratio (EAR). The duration or
count of blinks during the transition from open to closed eye state
is also studied. Then, when the blink duration exceeds a specified
threshold, it detects drowsiness and sounds an alarm, allowing
the driver to wake up and become awake. Our designed
technology has demonstrated excellent accuracy.
3. INTRODUCTION
Driver sleepiness is one of the leading causes of fatalities
in automobile accidents. Drivers easily become fatigued after
driving for a lengthy period of time, resulting in driver fatigue
and drowsiness. According to statistics, the majority of accidents
are caused by driver weariness. Varying countries have different
statistics on driver fatigue related accidents. According to the
report by โMinistry of Road Transport & Highwaysโ there were
4,552 accidents reported every year in India, that took lives of
thousands of people because of sleepy drivers (Road Accidents in
India 2016). Humans have always built devices and devised
tactics to make and safeguard their lives easier and safer, whether
for ordinary goals like going to work or for more intriguing ones
like flying.
4. The numbers of road Accident cause the loss of people's
lives, and the property is increasing. The dangers of drowsiness
when performing tasks that require continuous focus Sleep
deprivation and alcohol are the main causes of car accidents, both
affect the response of the human brain. Once the driver is drowsy,
the driver loses control of the vehicle. Consequently, the driver
may suddenly deviate from the road and collide with an obstacle
or car to overturn.
We then used OpenCV to draw contours around it. Now, we drew
contours around it using OpenCV. Using Scipyโs Euclidean
function, we calculated the sum of both eyesโ aspect ratio which
is the sum of 2 distinct vertical distances between the eyelids
divided by its horizontal distance. Weโll now see if the aspect
ratio is smaller than the threshold. If the value is less than that, an
alert is sounded and the user is cautioned.
5. METHODOLOGY
A. DEEP LEARNING
Deep learning allows computational models with several
processing layers to learn multiple degrees of abstraction for data
representations. These techniques have vastly enhanced the state-of-the-
art in speech recognition, visual object recognition, object detection,
and a variety of other fields like drug development and genomics. Deep
learning uses the back propagation algorithm to show how a machine
should adjust its internal parameters that are required to compute the
representation in each layer from the representation in the previous
layer, revealing intricate structure in enormous data sets. Deep
convolutional nets have revolutionized image, video, voice, and audio
processing, while recurrent nets have shed light on sequential data like
text and speech.
6. B. SUPERVISED LEARNING
Supervised learning is a method of developing artificial
intelligence (AI) that involves training a computer algorithm on input
data that has been labeled for a certain output. When provided with
never-before-seen data, the model is trained until it can discover the
underlying patterns and relationships between the input data and the
output labels, allowing it to produce accurate labeling results.
Supervised learning excels in classification and regression problems,
such as determining the category of a news article or forecasting the
volume of sales for a future date. The goal of supervised learning is to
make meaning of data in the context of a specific
7. C. CONVOLUTIONAL NEURAL NETWORK
ConvNets are built to handle data in the form of
several arrays, such as a color image made up of three 2D arrays storing
pixel intensities in each of the three color channels. Multiple arrays are
used to represent many data modalities: 1D for signals and sequences,
including language; 2D for images or audio spectrograms; and 3D for
video or volumetric imagery. Local connections, shared weights,
pooling, and the usage of several layers are the four key ideas
underpinning ConvNets, which take advantage of the features of real
signals.
8. REQUIREMENTS AND SPECIFICATION
Hardware Requirements:
โข Processor : Pentium (minimum)
โข Hard Disk : 40GB
โข RAM : 256MB (minimum)
Software Requirements:
โข Operating System : Windows or Linux
โข Technology : OpenCV, Python
9. Proposed System
The different types of methodologies have been developed to
find out drowsiness.
.Behavioral based approach: In this technique eye blinking
frequency, head pose, etc. of a person is monitored through a
camera and the person is alerted if any of these drowsiness
symptoms are detected.
10. The various technology that can be used are discussed as:
โข OpenCV: OpenCV stands for Open Source Computer Vision. It's an
Open Source BSD licensed library that includes hundreds of
advanced Computer Vision algorithms that are optimized to use
hardware acceleration. OpenCV is commonly used for machine
learning, 4 image processing, image manipulation, and much more.
OpenCV has a modular structure.
โข Kivy: Kivy is an open source Python library for developing mobile
apps and other multitouch application software with a natural user
interface (NUI). It can run on Android, iOS, Linux, OS X, and
Windows. Distributed under the terms of the MIT license, Kivy is
free and open source software. Kivy is the main framework
developed by the Kivy organization, alongside Python for Android,
Kivy iOS, and several other libraries meant to be used on all
platforms.
11. MODULE LIST
โข Open Camera
โข Face Detection
โข Eye detection
โข Recognition of Eye's State
โข Drowsiness Detection
12. MODULE LIST
Open Camera
โข START The first step in the open the camera where the user/driver
must be check the face detection and eye location
Face Detection
โข The camera will open to face detection For the face Detection it
uses Haar feature-based cascade classifiers is an effective object
detection method