Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Mobile Based Facial Recognition Using OTP

Verification for Voting System

Ms.Ashwini Ashok Mandavkar Prof. Rohini Vijay Agawane


Computer Engineering (Computer Networks) Computer Engineering (Computer Networks)
K. J. College of Engineering & Management Research K. J. College of Engineering & Management Research
Pune, India Pune, India
ashwini.mandavkar@gmail.com agawane.rohini@gmail.com

Abstract—As we know in urban areas now a day the voting In India, elections are very expensive so common man
system is getting most complicated only because of the person’s cannot afford this. Therefore only rich people can fight
identity. They only have the voting card as a proof of elections. Also they make use of political power and money
identification. So there are lots of chances of fake voting. through illegal methods.
To avoid this we are developing this project which will store
the identity of the voters using android mobile through facial B. Misuse of official machinery:
recognition systems. This system will capture faces of the voters The election laws have to check the misuse of official
and match with the existing faces in the stored database. After machinery during the elections. But then also the ruling party
the confirmation of valid face detected, the OTP (One- Time makes use of government vehicles and servants in their favor.
Password) is generated and send to the voters registered mobile
number. Then the voter is validated and he is allowed to do the C. Use of caste/religion in election:
voting. This is very fast and helpful technique to do the
verification of the voters. This will also reduce voter’s time to
In India, many votes are cast on the basis of caste and
stand in queue for doing vote. religion. As a result of this people get divided on the basis of
caste which is very harmful for election process.
Index Terms—Eigenface, Eigen Vectors, image processing, So to avoid these inconsistencies we propose the online
OTP (One-Time Password), Principal Component Analysis (PCA), voting system through facial recognition and OTP generation.
Spontaneous facial expression. It also overcomes the above challenges.
III. LITERATURE SURVEY
I. INTRODUCTION
Polling System Using GSM Facility [1], presents the
Democracy plays a very important role in the development electronic voting system using GSM mobile technology.
countries, where the citizens choose and replace the When voter want to do the voting, then voter send a message
government by doing the elections [1]. So this whole process to GSM modem. Then the voter which sends the message that
needs lots of man power and resources too. Election provides contains the voter’s mobile id and candidate’s id and then
a chance to the people to choose leader, therefore it must be GSM modems receives the votes which are end by voter and
accurate and transparent. Existing systems contains certain then voting is done.
drawbacks as chances of the violence, damage of machineries, Smart Device Based Election Voting System Endorsed
dummy votes etc [2]. So to avoid these drawbacks this paper through Face Recognition [6], provides the online voting
implements a new technique as Android Mobile based Voting system through facial recognition using ADHAR CARD NO
System through Facial Recognition [3]. There are also so verification. Then it fetches the voter’s details using its
many other techniques such as finger print recognition system database and then matches it with ADHAR card data and if it
[12] but we are using this technique which provides bandwidth matched then only he/she can allow to vote.
utilization for face matching and reorganization [4]. Real-Time Voting Systems [8], takes different faces of
II. MOTIVATION individual persons and then it is stored in voter database by
giving appropriate ADHAR card id number. Then matching is
In 1950, 14 general elections of various State Legislative done and if matches successful then only the valid voter are
Assemblies have been held. These elections were free and fair allowed to vote. They are using Haar like features for face
but then also some weaknesses of our election system have detection and eigenface algorithm for face recognition. This
been noticed. Some of these weaknesses (challenges) are as system provides the benefits to the user that he/she can vote
under: from any place only his/her face will be matched then he can
A. Expensive elections: do the voting.

978-1-4799-8047-5/15/$31.00 2015
c IEEE 644
Highly Secured Online Voting System over Network [9] will glow that means the vote is successfully recorded [6] [7].
presents the online voting system. The people above 18 years This process repeated every time when the new voters come in
old can only give vote by online systems. Then Election for voting. Also the whole process of voting arrangement is
Commission Officer he/she verify candidates. Also here the repeated every time. Thus this process requires a lots of man
proposed software works on Ethernet and then allows online power.
voting. Here all users or voters do login by using username
and its password and then only he can do vote
Face Recognition Using Eigen face Approach [10], this
will captures the faces and then it compares with the stored
faces. Here Eigen face uses PCA algorithm for face
recognition. Also it recognizes the gender of a person
according to the frontal view, scaled frontal view, 45° view.
Face Detection and Gender Classification [11], it presents
face detection and gender classification techniques with color
images. It detects different skin regions in given image and
facial features by using pixels.
IV. NEED
Fig 1. Existing Voting System
To overcome the challenges or weaknesses of the already
existing voting system described in section II, we proposed the VI. PROPOSED WORK
online voting system through facial recognition using OTP.
The following is the need of this online voting system: A. Architecture Diagram:
A. Saved Ballot Templates:
There is no need to arrange the election process from Figure 2 shows the overall architecture of proposed mobile
scratch. Just we have to do it for one time then it will based voting process scenario. To carry out the whole
automatically save its configuration. Therefore for next procedure successfully an internet connection is essential.
elections we have to mention the candidate’s names. Also every person must have mobile phones to carry out our
application successfully.
B. Reducing costs: In the first step the registration process is done by the
It reduces costs for printing of data, mailing of documents voters through an application. Then in the second step the
and tabulating paper ballots. application will start its process. It first captures the face of the
voter using mobile front camera. Then that image is send to
C. Automatic Tallying: the existing database. Here we use already existed database or
It reduces the human efforts to count the votes. It centralized database, which contains voter’s information with
automatically tallies the votes and provides the results before photos. Then system compares the send images with already
election closes. stored images in centralized database.
Once the image is recognized successfully or matched
D. Ability to correct mistakes:
with the existing images of the database then the server sends
It provides the correction system if any mistake happens. the OTP (One-Time Password) to the voters registered mobile
Then voters can correct that mistake before final submission. number. Then voter enters that OTP, then database again
After submission it cannot be altered verifies that entered OTP by voter and if it is correct that
means it is valid user.
V. EXISTING WORK
After that the candidate’s list will appears. This list
The whole election process is divided into many parts. So contains the candidate name and in front of that name the
providing security to them is also necessary. So it needs a lot button named ‘Vote’ is provided. Voter should have to press
of man power, therefore it is quite difficult to manage it. Here that button then only the voting is done and the voting
Election commission provides the polls. These polls will be procedure is completed.
arranged in school or in halls. The people will have to know Here security is also necessary to protect the system. For
about the location before voting. So for their convenience the that session level security mechanism is being used. First
place and timing is already mentioned [5]. client need to request valid session id to server. Once server
Then voter have to reach at the polling booth for voting as validate client it will respond with session id. By handshaking
shown in Figure 1. After that the voter’s identification is done this session id communication takes place. On each request
by the duty person. Then the officer does marking on the session id will be send and validate. Once client logout from
voter’s left forefinger using inedible ink. After that voter have system it will destroy session id. Also password is encoded in
to do the signature in the register and then only he/she will MD5 format, so nobody able to see the password text.
allow going inside the compartment. Then the voter have to
press the button in front of the candidates name then red lamp

2015 IEEE International Advance Computing Conference (IACC) 645


C. Algorithm:
This method is based on information theory which will
decompose the face images. Then forms the minute set of the
characteristics features images which are called as “Eigen
faces”. This is nothing but the principal components of the
face images of the training set. For face recognition, the Eigen
face method is one of the most efficient and simplest
approaches [13]. In Eigen face method, the distance is being
measured between couples of images. If that distance is less
than a provided threshold value, then it is an identified faces
or else it is an unidentified faces [8].

Fig 2. Proposed Voting Process scenario

B. Data Flow Diagram (DFD):


In the below Figure 3, the voter name, voter date and other
information is first entered in the application. After the
successful registration, the voters sign in into the application.
Then the photo of the voters is taken and then sends it to the
centralized database. Then there face matching is done. If face
is matched then OTP is send to the voter’s registered mobile
number. Then voter enters that OTP then database again
verifies that OTP and if it is verified then voter can do the
voting. At the time of voting the candidates list will be
generated and according to that voter can select particular
candidate and then voting is done. Then it updates the vote
counter for another voters voting.

Fig 4. Data Flow Diagram of Eigen face-based facial


Fig 3. Data Flow Diagram of Proposed Voting System recognition algorithm

646 2015 IEEE International Advance Computing Conference (IACC)


In the above Figure 4, we have two set of image blocks, Here ν i is an eigenvector of L . So here M eigenvectors of
first one is training set image block and second one is test set
image block. In training set image block, firstly the Eigen face L are being used to form the M eigenvectors ui of C that
of image in the database i.e. trained image is obtained. Then will form our Eigen face basis:
the weight W1 is calculated by using the Eigen face and the M
training set. In the testing set image block, image X is the ul = ¦υlk Φ k l = 1,..., M
unknown input image which is nothing but the captured k −1
image. The weight W2 is calculated using the input image and
the Eigen face. Value of D is calculated by finding the average
Here u is Eigen faces.
of distances between W1 and W2. If the D value is less than 0,
then the face is recognized. Then the input image X and W2 2) Recognition of the faces: The recognition of new
values are stored. If the D value is greater than 0, then the face (unknown) face Γ new amongst the known faces is carried out
is not recognized. proceeds into two steps. First is the new image is transformed
into its Eigen face components and second is the resulting
1) Steps:
weights form the weight vector Ω new
T
a) Preparation of data: The faces of the training set
( Γi ) has been taken for the processing.
ωk = ukT ( Γ new − Ψ ) k = 1.....M ′
b) Subtraction of mean: Here we have to calculate the ΩTnew = [ω1ω2 .....ωM ′ ]
average matrix ( Ψ ) first, then it is being subtracted from The Euclidean distance between two weight vectors
original faces ( Γi ). Then the result of that is stored in the d ( Ωi , Ω j ) is calculated by using similarity of the images i
variable Φ i .
and j . If the minimum Euclidean distance between Γ new and
M
1 other faces exceeds - on average - some threshold value θ ,
Ψ= ¦Γ n
M n −1 one can assume that Γ new is an unknown face, Else it is
considered as a known face.
Φ i = Γi − Ψ (1)
c) Calculating covariance matrix: The covariance 3) Calculating Euclidean Distance: Let an arbitrary
matrix C is calculated as, instance x be described by the feature vector as,
x = ª¬ a1 ( x ) , a2 ( x ) ,....., ar ( x ) º¼
M
1
C= ¦Φ Φ ar ( x ) is the value of the r th attribute of instance x .
T
n n (2) Here
M n −1

Now the eigenvectors (Eigen faces) ui and its Then the distance between two instances as xi and x j is
corresponding eigenvalues λi of the vector C should be defined to be d (x ,x ) :
i j

calculated.
¦ ( a ( x ) − a ( x ))
n
d ( xi , x j ) =
2
d) Calculating eigenvectors and eigenvalues of the
r i r j
covariance matrix: The covariance matrix C in equation 2 has r =1

dimensionality as N
2
× N 2 , so it has N 2 eigenfaces and
N 2 eigenvalues. For example if we taken 256 × 256 images D. Mathematical Model:
then we have to compute total 65, 536 × 65, 536 matrixes and We are using Set theory to explain our system.
then we have to calculate total 65,536 Eigen faces. So this is Let G be a closed graph that represents our system
quite difficult and not possible. So for that we use PCA which
describes a small set of vectors with the large dimensional G = {E ,V }
space. So here PCA provides only M images and M non- Where, E represents the whole set of edges
trivial eigenvectors. Hence here for solving the eigenvectors
E = { E1 , E2 , E3 ,..., E9 }
we can take the eigenvectors of the new matrix as M × M :
L = AT A And V is a whole set of vertices
By using the below math trick we get the above equation: V = {V1 , V2 , V3 ,..., V7 }
AT Aν i = μν
i i
In the graphical representation of the system, vertices in the
AAT Aν i = μi Aν i set V represent the modules which are connected through
directed edges in the set E representing the input/output of
modules.

2015 IEEE International Advance Computing Conference (IACC) 647


VII. RESULTS
TABLE I. VERTEX DEFINITION
Vertex Modules
New user
V1
Voter & Application User
V2
Android Application
V3
Server
V4
Java Client
V5
Database
V6
Vote
V7

TABLE II. EDGES DEFINITION


Edges Input/Output
Register for new user
E1 (a) (b)
Giving user data to Java client
E2
Fetching Server
E3
Storing Data in Database
E4
Acknowledgement to User
E5
Generate New User
E6
Fetching Application
E7
Entering Vote
E8
Sending Vote to Server
E9

Let Fe be a rule of E into V such that for given edge; it (c) (d)
returns vertices.
Fe ( E ) 6 V
Thus, for our system,
1) Fe ( E1 ) = {V1} : V1 is called using E1 for
registration of new user.

2) Fe ( E2 ) = {V5 } : User data is passed to V5 using E2


to verify user documents.

3) Fe ( E3 ) = {V4 } : User data is passed to V4 using


E3 to fetch server.
(e) (f)
4) Fe ( E4 ) = {V6 } : V6 is called using E4 to store data Fig 5. The whole E-Voting procedure ((a):The Sign-In procedure for
voters, (b):If voter is not registered then voter have to fill their details
in database. and then face is captured, (c):Election details is displayed,
(d):Candidates list is displayed, (e):Voter select the candidates,

Fe ( E5 ) = {V1} : V1 is called using E5 for giving


(f):Voting is done)
5)
Figure 5(a) shows the Sign-In procedure for voters. If the
acknowledgement to the user. voter is new then register for new account and if the voter is
already registered then Sign-In using the given username and
password. In Figure 5(b) if the voter is not registered then

648 2015 IEEE International Advance Computing Conference (IACC)


voter have to fill their detailed information such as name, age, Journal of Advanced Research in Computer Science and
address, username and password. Then click on face capture Software Engineering, Volume 3, Issue 11, November 2013
button, will capture the face of the voter. Figure 5(c) provides ISSN: 2277 128X.
the detail information about the election, such as election [4] Feng-Li Lian, Yi-Chun Lin, Chien-Ting Kuo, and Jong-Hann
name, election date, and election ward and election disc. Jean, “Voting-Based Motion Estimation for Real-Time Video
Figure 5(d) provides the list of the candidates. In Figure 5(e), Transmission in Networked Mobile Camera Systems”, IEEE
the voter is free to vote for the particular candidate. Figure 5(f) TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL.
displays the message that the voting is done successfully. 9, NO. 1, FEBRUARY 2013
[5] Mr. Mayur Patil, Mr. Vijay Pimplodkar, Ms. Anuja R. Zade,
For developing the project the minimum requirement of Mr.Vinit Vibhute, Mr..Ratnakar Ghadge, “A Survey on Voting
hardware tools is Processor of Intel P4 3.0 MHz, RAM of 512 System Techniques”, International Journal of Advanced
MB and above, monitor all applicable and Hard Disk of 120 Research in Computer Science and Software Engineering,
GB only. For running the project on the android phone the Volume 3, Issue 1, January 2013 ISSN: 2277 128X.
minimum requirement is RAM of 512 MB and Processor of 1 [6] Trisha Patel, Maitri Chokshi, Nikhil Shah, “Smart Device Based
GHz only. And the software requirement to run the project is Election Voting System Endorsed through Face Recognition”,
Operating System as Windows or android, languages are Java, International Journal of Advance Research in Computer Science
and Software Engineering, Volume 3, Issue 11, November 2013
JavaScript, J2ME and PHP, Framework is MVC 2, Front End
ISSN: 2277 128X.
is Java Applet, Back End is MySQL, Building Tool is Apache [7] Hari K. Prasad, J. Alex Haldermany, Rop Gonggrijp, Scott
Ant 1.0 and Web Server is Apache Tomcat 5.5 to 6.20 only. Wolchoky, Eric Wustrowy, Arun Kankipati, Sai Krishna
By using the above configurations the online voting project is Sakhamuri, Vasavya Yagati, “Security Analysis of India’s
being developed. Electronic Voting Machines”, April 29, 2010.
[8] Ganapatikrishna P. Hegde1, M. Seetha2, “Real Time Voting
CONCLUSION System Using Face Recognition for Different Expressions and
Pose Variations”, IJRET: International Journal of Research in
This is a Novel technique that will give us best results to Engineering and Technology, eISSN: 2319-1163 | pISSN: 2321-
get better understanding of online voting systems through 7308.
facial recognition and OTP verification. [9] K. P. Kaliyamurthie, R. Udayakumar, D. Parameswari and S. N.
This voting system helps everybody to cast their votes Mugunthan, “Highly Secured Online Voting System over
without any problem. So this will increase the percentage of Network”, Indian Journal of Science and Technology.
[10] VINAY HIREMATH, ASHWINI MAYAKAR, “FACE
voting and will give us the better results. Manual counting is
RECOGNITION USING EIGENFACE APPROACH”.
not needed, so it will overcome the problems of existing
[11] S.Ravi, S.Wilson, “Face Detection with Facial Features and
system. We have used Eigen face recognition as authentication Gender Classification Based On Support Vector Machine”, 2010
in the application. This system detects the face which is Special Issue - International Journal of Imaging Science and
captured using a mobile camera. Then it compares that face Engineering.
with the database stored faces and check if the two images are [12] Sanjay Kumar, Manpreet Singh, “Design a Secure Electronic
matching or not. If it matches, then OTP is send to the Voting System Using Fingerprint Technique”, IJCSI
registered voter’s mobile number and then verification is International Journal of Computer Science Issues, Vol. 10, Issue
complete and then allows him to vote. 4, No 1, July 2013.
[13] Sheifali Gupta, O.P.Sahoo, Ajay Goel, Rupesh Gupta, “A New
ACKNOWLEDGMENT Optimized Approach to Face Recognition Using EigenFaces”,
Global Journal of Computer Science and Technology Vol. 10
Author is extremely thankful of research guide “Prof. Issue 1, April 2010.
Rohini V. Agawane”, KJ College of Engineering and
Management Research, Pune for consistent guidance,
inspiration and for valuable support. Author is also grateful of
college Principal- “Prof. Dr. S. J. Wagh”, and HOD of
Computer Department- “Prof. D. C. Mehetre”, also thankful to
PG Co-Ordinator- “Prof. M. K. Nighot” for their time to time
support and guidance.
REFERENCES
[1] Hemlata Sahu, Anupam Choudhray, “Polling System Using
GSM Facility”, International Journal of Scientific &
Engineering Research Volume 2, Issue 10, Oct-2011 1 ISSN
2229-5518.
[2] Gomathi.B, Veena priyadarshini.S, “Modernized Voting
Machine using Finger Print Recognition”, International Journal
of Scientific & Engineering Research, Volume 4, Issue 5, May-
2013 156 ISSN 2229- 5518.
[3] Dr.Aree Ali Mohammed, Ramyar Abdolrahman Timour,
“Efficient E-voting Android Based System”, International

2015 IEEE International Advance Computing Conference (IACC) 649

You might also like