MODEL - Project - Report (AutoRecovered)
MODEL - Project - Report (AutoRecovered)
A PROJECT REPORT
Submitted by
LINGESHWARAN.R (611218106041)
UKESH.G (611218106081)
GOWTHAM. S (611218106302)
degree of
BACHELOR OF
ENGINEERING IN
A PROJECT REPORT
Submitted by
LINGESHWARAN.R (611219106041)
UKESH.G (611219106081)
GOWTHAM. S (611219106302)
degree of
BACHELOR OF
ENGINEERING IN
NOVEMBER- 2022
BONAFIDE CERTIFICATE
Certified that this project report “INTELLIGENT VEHICLE DAMAGE
ASSESSMENT & COST ESTIMATOR FOR INSURANCE COMPANIES'' is the
bonafide work of “LINGESHWARAN R. (611219106041), UKESH G
(611219106081), GOWTHAM S. (611219106302), PRAKASH RAJ M.
(611219106305)'' who carried out the project work under my supervision.
SIGNATURE SIGNATURE
Mr. S. PREMKUMAR B.E., M.E., Mr. G. BABU B.Tech., M.Tech.,
FACULTY MENTOR FACULTY MENTOR
ASSISTANT PROFESSOR ASSISTANT PROFESSOR
Department of Electronics and Department of Computer Science
Communication Engineering, and Engineering,
Knowledge Institute of Technology, Knowledge Institute of Technology,
Kakapalayam, Kakapalayam,
Salem- 637 504. Salem- 637 504.
……………………. ……………………………………
At the outset, we express our heartfelt gratitude to GOD, who has been our strength to
bring this project to light. At this pleasing moment of having successfully completed our
project, we wish to convey our sincere thanks and gratitude to our beloved president
MR. C. BALAKRISHNAN, who has provided all the facilities to us. We would like to
conveyed our sincere thanks to our beloved Principal DR. PSS. SRINIVASAN, for
forwarding us to do our project and offering adequate duration in completing our project.
We express our sincere thanks to our Head of the Department DR. N.
SANTHIYAKUMARI, Department of Electronics and Communication Engineering for
guiding us throughout the project as a supervisor and fostering the excellent academic
environment in the Department. We express our pronounced sense of thanks with deepest
respected and gratitude to our Faculty Mentor MR. S. PREMKUMAR, Department of
Electronics and Communication Engineering, Faculty Evaluator MR. G. BABU,
Department of Computer Science and Engineering for their valuable and precious guidance
and for having amicable relation. With deep sense of gratitude, we extend our earnest and
sincere thanks to our SPOC MRS. T. DEVIKA, Assistant Professor, Department of
Electronics and Communication Engineering for her guidance and encouragement during
this project. We would also like to express our heartfelt gratitude to all the faculty
members of our department, friends and students who helped us directly and indirectly in
all aspects of the project to get completed successfully.
TABLE OF CONTENTS
1. INTRODUCTION 1
1.2 PURPOSE 2
2. LITERATURE SURVEY 3
4. REQUIREMENT ANALYSIS 12
5. PROJECT DESIGN 14
i
6. PROJECT PLANNING AND SCHEDULING 18
7.1 FEATURE 1 22
7.2 FEATURE 2 23
8. TESTING 24
9. RESULTS 26
11. CONCLUSION 30
13. APPENDIX 32
14 REFERENCES 38
ii
CHAPTER - 1
1. INTRODUCTION
1
1.2 PURPOSE
2
CHAPTER – 2
2. LITERATURE SURVEY
In this literature survey several methods have been proposed for detection of car
damage. Srimal et al. [ 4 ] proposed a solution which uses 3DComputerAided Design for
the discernment of car damage from the picture, the system only detect damage at edge
portion only. Detection of the car damage through CAD software requires some knowledge
about the software. S Gontscharovetal [ 5 ] ,the proposed system designed by using
YOLO(you only look once)algorithm to detect tha car damage, Here the multi sensor data
fusion technique is allows to locate the portion of damage more accurately and performs
detection faster compared to other algorithms which is fully automatic and doesn’t require
much human intervention. Phyu Mar Kyuet al[ 3 ], the proposed system uses deep learning
based algorithm are VGG16andVGG19 damaged car detection in the real world. This
algorithm notice the severity of the damaged car based on the location. Finally the author
concludes that L2 regularization work greater. Girish Net al [ 2], the proposed system uses
vehicle damage detection technique depends on transfer learning and mask RCNN, The
mask regional convolution neural network determines a damaged car by its position and
estimate the depth of the damage. A Neela Madheswari et al [ 1 ], the proposed system
uses convolution neural network is use to accept that image contains a car damage or not. It
take as great opportunities to attempt by classifying the car damage into different classes.
3
2.2 PROBLEM STATEMENT DEFINITION
4
CHAPTER - 3
IDEATION & PROPOSED SOLUTION
5
3.2 IDEATION & BRAINSTORMING
This step includes the formation of a team, collaborating with the team by
collecting the problems of the domain we have taken and consolidating the collected
information into a single problem statement.
6
Step 2: Brainstorm, Idea Listing and Grouping
This step of ideation includes the listing of individual ideas by teammates to help
with the problem statement framed. All the individual ideas have been valued and made
individual clusters.
Then discussed as a team and finally made an ideation Cluster A and concluded with the
most voted ideas from all the clusters together and Cluster B with the least needed ideas.
7
Step 3: Idea Prioritization
This step includes the process of listing necessary components to come up with the
working solution and making a hierarchy chart by prioritizing the components based on
importance, say from the higher being backend and lower being the user interfacing
components.
8
3.3 PROPOSED SOLUTION
Car Detection
The approach narrows down to two separate models pipelined. The first task is to
differentiate between a whole and a damaged car followed by detecting the extent of
damage and classify accordingly. Each class has at least 300 images to train upon.
Feature Extraction
An extensively comparison of the performances of many deep feature approaches
was done in terms of feature extraction and decided to use the VGG16 model with
ImageNet weights due to its simplistic model architecture and computational efficiency
9
Transfer Learning
The VGG16 architecture was selected because it has a relatively simple architecture
and Keras ships with a model that has been pre-trained on ImageNet. It is just a number of
Conv2D and MaxPooling2D layers with a dense network on top with a final softmax
activation function. Additional dense trainable layers with sigmoid function above this
model have been added.
Classification
After successfully extracting the features for the two classes, two binary
classification model for the pair of two classes were built.
Preprocessing
The RGB (Red-Green-Blue) images are Gray-scaled. The images are are resized
throughout the dataset using a predefined image size in order to change them into a
desirable format. The image data and corresponding class index are appended to training
data. The training data is randomly shuffled to ensure that each data point creates an
independent change on the model, without being biased by the same points before them.
Pickle file is generated to save the serialized format of training data to a file and load it
later to directly train the different models without repeating the hassle of data
preprocessing.
1
3.4 PROPOSED SOLUTION FIT
The Problem-Solution Fit simply means that you have found a problem with
your customer and that the solution you have realized for it solves the customer’s
problem. It helps entrepreneurs, marketers and corporate innovators identify
behavioral patterns and recognize what would work.
1
CHAPTER - 4
REQUIREMENT ANALYSIS
1
4.2 NON-FUNCTIONAL REQUIREMENTS
1
CHAPTER - 5
PROJECT DESIGN
1
5.2 SOLUTION & TECHNICAL ARCHITECTURE
The solution architecture includes the components and the flow we have
designed deliver the solution.
Here, the application is planned to be designed, where the caretaker of the
patients can feed the medicinal details to the database connected with the help of python
and API calls. By monitoring that information in the program, timely message alerts are
given to the patients to intake the medicine.
1
5.3 USER STORIES
1
1
Login and As a user, I I can log in
Customer repeated USN-5 can log into and view my High
Options usage the dashboard at
application by my demand
entering on any time
email
& password
As a user can
capture I can
Customer Web page USN-6 images of capture the High
needs to do vehicle and entire
upload web vehicle and
portal. upload
Customer Details As a user I
(Web user) about must receive a I can get the
estimated detailed report estimated High
cost based USN-7 of the insurance cost
on damage damages
present in the
vehicle and
the cost
estimated
Customer Provide As a user, I
Care friendly and need to get I can have
Executive efficient support from smooth user
customer USN-8 developers in experiences Medium
support and case of and all the
sort out the queries and issues raised
queries failure of is sorted
service
provided
Administrator Overview We need to I can finish
the entire satisfy the the work
process and customer without any
act as a USN-9 needs in an problems
bridge efficient way High
between and make
user and sure any sort
developers of errors are
fixed
1
CHAPTER - 6
Sprint-1
USN-3 As a user, they initially login High
Login using the password through
login link sent with
confirmation email.
Sprint-2 Details about USN-4 As a user, they can access the Low
insurance details about insurance
company company.
1
Sprint-3 Reports about USN-7 As a user, they receive a High
the damage and detailed report of the damages
estimated cost present in the vehicle and the
for damage cost-estimated
Rating of
Sprint-3 customers As a developer, need to know
USN-9 their service to the customer Low
needs by the customer ratings
2
6.2 SPRINT DELIVERY SCHEDULE
Velocity
2
6.3 REPORTS FROM JIRA
BURNDOWN CHART
2
CHAPTER - 7
7.1 FEATURE 1
2
7.2 FEATURE 2
<!DOCTYPE html>
<html lang="en">
<head>
<title> Vehicle Damage Assessment</title>
<link rel="stylesheet" type="text/css"
href="{{url_for('static',filename='css/prediction.css')}}">
</head>
<body>
<div class="head">
<br><h1 class="name">Vehicle Damage Detection</h1>
<nav>
<ul>
<a href="http://127.0.0.1:5000/index" >Home</a>
<a href="http://127.0.0.1:5000/logout" >Logout</a>
</ul>
</nav>
</div>
<br><br><br><br><br>
<div class="image">
<form action="">
<label for="mmm">Upload here</label><br><span><input type="file"
name="predictionimage" id="">
<input type="submit" value="Submit"></span>
</form>
</div>
<br><br><br><br><br><br><br><br>
<div class="image">
<esti>
<h6>The Estimated cost for the damage is: </h6>{{prediction}}
</esti>
</div>
<br><br><br><br><br><br><br><br><br><br><br><br><br>
<div >
<h5 class="rights" ><br>Copyrights<sup>©</sup> 2021,All Rights
Reserved<br><br></h5>
</div>
</body>
</html>
2
CHAPTER - 8
TESTING
A test case has components that describe input, action and an expected
response, in order to determine if a feature of an application is working correctly. A test
case is a set of instructions on “HOW” to validate a particular test objective/target, which
when followed will tell us if the expected behavior of the system is satisfied or not.
Characteristics of a good test case:
Accurate: Exacts the purpose.
Economical: No unnecessary steps or words.
Traceable: Capable of being traced to requirements.
Repeatable: Can be used to perform the test over and over.
Reusable: Can be reused if necessary.
Login
User login User name and Login success.
1 password
2
8.2 USER ACCEPTANCE TESTING
This sort of testing is carried out by users, clients, or other authorised bodies
to identify the requirements and operational procedures of an application or piece of
software. The most crucial stage of testing is acceptance testing since it determines
whether or not the customer will accept the application or programme. It could entail the
application's U.I., performance, usability, and usefulness. It is also referred to as end-user
testing, operational acceptance testing, and user acceptance testing (UAT).
2
CHAPTER - 9
RESULTS
Vehicle damage detection is used to reduce claims leakage during insurance processing.
Visual inception and validation are usually done. As it takes a long time, because a person
needs to come and inspect the damage.
Here we are trying to automate the procedure. Using this automation, we can avoid time
conception for the insurance claim process.
2
2
2
CHAPTER - 10
ADVANTAGES
DISADVANTAGES
Coverage failures: The primary and major disadvantage of car insurance is that
your policy does not cover the entire vehicle. Only the specific parts of the car are
under damage coverage, the policyholder needs to verify hidden clauses in the
document keenly before buying the policy.
Time taking Process: Most insurance companies take a time frame to settle the
claim amount, this is the problem most the policy holders are facing.
3
CHAPTER -11
CONCLUSION
3
CHAPTER – 12
FUTURE SCOPE
3
CHAPTER – 13
APPENDIX
INDEX
<!DOCTYPE html>
<html lang="en">
<head>
<title>Home</title>
<link rel="stylesheet" type="text/css"
href="{{url_for('static',filename='css/index.css')}}">
</head>
<body>
<div class="head">
<br><h1 class="name">Intelligent Vehicle Damage Assessment & Cost Estimator for
Insurance Companies</h1>
<nav>
<ul>
<a href="http://127.0.0.1:5000/index" >Home</a>
<a href="http://127.0.0.1:5000/login" >Login</a>
<a href="http://127.0.0.1:5000/register" >Register</a>
<a href="http://127.0.0.1:5000/prediction" >Prediction</a>
</ul>
</nav>
</div>
<div class="para">
<h2>About Our Project</h2>
<p>Vehicle damage detection is used to reduce claims leakage during insurance
processing.<br>
Visual inception and validation are usually done. As it takes a long time, because a
person needs to come and inspect the damage.<br>
Here we are trying to automate the procedure. Using this automation, we can avoid
time conception for the insurance claim process.</p>
</div>
<br>
<div>
<h5 class="rights" >Copyrights<sup>©</sup> 2021,All Rights
Reserved<br></h5>
</div>
</body>
3
</html>
LOGIN
<!DOCTYPE html>
<html lang="en">
<head>
<title>Login</title>
<link rel="stylesheet" type="text/css"
href="{{url_for('static',filename='css/login.css')}}">
</head>
<body>
<div class="head">
<br><h1 class="name">Vehicle Damage Detection</h1>
<nav>
<ul>
<a href="http://127.0.0.1:5000/index" >Home</a>
<a href="http://127.0.0.1:5000/login" >Login</a>
<a href="http://127.0.0.1:5000/register" >Register</a>
</ul>
</nav>
</div>
<div class="details">
<h3>Login</h3>
<form action="">
<div class="input">
<br>
<span><input type="email" name="email" id="email"
placeholder="Enter Email-id"></span>
</div>
<div class="input">
<br>
<span><input type="password" name="pwd" id="pwd" placeholder="Enter
Password"></span><br>
</div>
<br>
<div class="btn">
<span><button type="submit">Login</button></span>
</div>
<h6>Don't have account? <a href="Register page.html" >Register</a></h6>
</form>
</div>
<div>
3
<h5 class="rights" >Copyrights<sup>©</sup> 2021,All Rights
Reserved<br></h5>
</div>
</body>
</html>
LOGOUT
<!DOCTYPE html>
<html lang="en">
<head>
<title>Logout</title>
<link rel="stylesheet" type="text/css"
href="{{url_for('static',filename='css/logout.css')}}">
</head>
<body>
<div class="head">
<br><h1 class="name">Vehicle Damage Detection</h1>
<nav>
<ul>
<a href="http://127.0.0.1:5000/index" >Home</a>
<a href="http://127.0.0.1:5000/login" >Login</a>
</ul>
</nav>
</div>
<div class="para">
<h2>Successfully Logged Out</h2>
<p>Login for more Information...</p>
<br><br>
<div class="image">
<form action="">
<label for="mmm">Your Rating : </label><span><input type="number"
name="rating" id="">
<input type="submit" value="Submit"></span>
</form>
</div>
</div>
<div >
<h5 class="rights" >Copyrights<sup>©</sup> 2021,All Rights
Reserved<br></h5>
</div>
</body>
</html>
3
PREDICTION
<!DOCTYPE html>
<html lang="en">
<head>
<title> Vehicle Damage Assessment</title>
<link rel="stylesheet" type="text/css"
href="{{url_for('static',filename='css/prediction.css')}}">
</head>
<body>
<div class="head">
<br><h1 class="name">Vehicle Damage Detection</h1>
<nav>
<ul>
<a href="http://127.0.0.1:5000/index" >Home</a>
<a href="http://127.0.0.1:5000/logout" >Logout</a>
</ul>
</nav>
</div>
<br><br><br><br><br>
<div class="image">
<form action="">
<label for="mmm">Upload here</label><br><span><input type="file"
name="predictionimage" id="">
<input type="submit" value="Submit"></span>
</form>
</div>
<br><br><br><br><br><br><br><br>
<div class="image">
<esti>
<h6>The Estimated cost for the damage is: {{prediction}}</h6>
</esti>
</div>
<br><br><br><br><br><br><br><br><br><br><br><br><br>
<div >
<h5 class="rights" ><br>Copyrights<sup>©</sup> 2021,All Rights
Reserved<br><br></h5>
</div>
</body>
</html>
3
REGISTER
<!DOCTYPE html>
<html lang="en">
<head>
<title>Register</title>
<link rel="stylesheet" type="text/css"
href="{{url_for('static',filename='css/register.css')}}">
</head>
<body>
<div class="head">
<br><h1 class="name">Vehicle Damage Detection</h1>
<nav>
<ul>
<a href="http://127.0.0.1:5000/index" >Home</a>
<a href="http://127.0.0.1:5000/login" >Login</a>
<a href="http://127.0.0.1:5000/register" >Register</a>
</ul>
</nav>
</div>
<div class="details">
<h3>Register</h3>
<form action="/Register" method="POST" autocomplete="off">
<div class="input">
<br>
<input type="text" id="name" placeholder="Enter Name" required>
</div>
<div class="input">
<br>
<input type="email" name="email" id="email" placeholder="Enter Email-id"
required>
</div>
<div class="input">
<br>
<input type="password" name="pwd" id="pwd" placeholder="Enter Password"
required><br>
</div>
<br>
<div class="btn">
<button type="submit">Register</button>
</div>
<h6>Already have account? <a href="Login page.html" >Login</a></h6>
</form>
</div>
3
<div>
<h5 class="rights" >Copyrights<sup>©</sup> 2021,All Rights Reserved</h5>
</div>
</body>
</html>
CONTENT LINKS
GITHUB https://github.com/IBM-EPBL/IBM-Project-
45915-1660733189
3
CHAPTER - 14
REFERENCES
Pei Li, Bingyu Shen, and Weishan Dong. 2018. An anti-fraud system for
car insurance claim based on visual evidence.ar