Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

A Novel Hybrid Biometric Electronic Voting System: Integrating

Finger Print and Face Recognition
SYED SHAHRAM NAJAM*, AAMIR ZEB SHAIKH*, AND SHABBAR NAQVI**

RECEIVED ON 09.08.2016 ACCEPTED ON 21.02.2017

ABSTRACT
A novel hybrid design based electronic voting system is proposed, implemented and analyzed. The
proposed system uses two voter verification techniques to give better results in comparison to single
identification based systems. Finger print and facial recognition based methods are used for voter
identification. Cross verification of a voter during an election process provides better accuracy than
single parameter identification method. The facial recognition system uses Viola-Jones algorithm
along with rectangular Haar feature selection method for detection and extraction of features to develop
a biometric template and for feature extraction during the voting process. Cascaded machine learning
based classifiers are used for comparing the features for identity verification using GPCA (Generalized
Principle Component Analysis) and K-NN (K-Nearest Neighbor). It is accomplished through comparing
the Eigen-vectors of the extracted features with the biometric template pre-stored in the election
regulatory body database. The results of the proposed system show that the proposed cascaded design
based system performs better than the systems using other classifiers or separate schemes i.e. facial
or finger print based schemes. The proposed system will be highly useful for real time applications due
to the reason that it has 91% accuracy under nominal light in terms of facial recognition.

Key Words: Electronic Voting System, Image Processing, Finger Print Based Recognition,
Biometric Recognition

1. INTRODUCTION

E
lectoral Systems empower the citizens of a country with bags of paper votes. The central station compiles and
to elect parliament members of their choice. Paper publishes the names of winners and losers through television
based electoral system is a classical method to and radio stations. This method is useful only if the whole
accomplish the said task. In this method, printed votes are process is completed in a transparent way. However, there
submitted to various election booths of country at least one are some drawbacks to this system. These include higher
day before the election. After the election timings, sealed expenses, longer time to complete the voting process,
boxes containing votes are opened in front of all the fraudulent practices by the authorities administering
legitimate members of booth and counted. The information elections as well as malpractices by the voters [1]. These
of counted votes is submitted to a centralized station along challenges result in manipulated election results.
Corresponding Author (E-Mail: shahramnajam.neduet@gmail.com)
* Department of Electronic Engineering, NED University of Engineering & Technology, Karachi
** Department of Computer Systems Engineering, Baluchistan University of Engineering & Technology, Khuzdar.

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
59
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

Electronic Voting Systems provide efficient and reliable if the information is found to be doubtful the vote will be
technique to empower citizens of a country or members of rejected. Otherwise it can be preceded to the ballot
an organization to select a person of their choice. These counting unit. It is an effective system with proper data
systems can be classified into supervised, hybrid and remote encryption and secrecy but it lacks one feature i.e. multiple
voting styles. Supervised voting also known as offline voting votes by a single user.
is typically administered by electoral organizations. In this
Evertz [7] presents a system using WAN (Wide Area
scheme, voting machines are located at polling machines.
Network) which is connected to a server at the election
However, these machines are not connected with a
office containing the database of all the voters. First the
centralized system for cross-verification or any other
voter has to verify its identity by facial recognition, in
purpose. Hybrid voting schemes are supervised by election
which features are extracted from the face of the voter
organizing members, however, the machines are connected
and compared with pre-stored features in a database. Upon
with internet, Remote voting refers to the schemes which are
matching of the identity, a window will pop up on the
not administered by any supervising staff and the machines
screen of the computer where the voter can cast its vote.
are connected with internet [2]. Benefits of using Biometrics
But the facial recognition system used and employed is
in a voting system is to accurately recognize the voter which
quite in-effective having a success percentage of only
enables the election administrators to reduce the error rates
58% and a response time of 15 seconds. Besides, it lacks
by reducing fraudulent and bogus votes. Besides, it also
any data encryption or security for the secrecy of the
results in cost efficiency, improving physical safety and
ballot. Thus rendering it in-effective for use in real-time.
increasing convenience to the users [3].

To improve the confidentiality and privacy of the

In this regard, various authors have developed the
electronic voting systems, most of the systems use Mixnet
electronic voting systems. A smart card based voting
or homomorphic encryption techniques [8]. Additionally,
system is developed by[4]. This smart card system has
authors also claim that the homomorphic encryption is
temporary and permanent storage facilities. To address
more appropriate for the situation with several election
fraudulent practices, this card also contains biometric
candidates as well as elections with neutral votes. The
information of the end user which can be authenticated
electronic voting system is implemented extensively in
by the system. Sehr [5] present a computerized voting
developed countries such as USA. Awad and Leiss [9],
system to address issues including low attendance of
present a comprehensive study of conventional and
voters, higher administration and operation costs, longer
electronic voting systems in USA along with their
time of tabulation, and inconvenience for voters, rigid
disadvantages. Alomari and Irani [10], present e-voting
voting guidelines, and inadequate security protection.
for a developing country, hence they concluded that the
Tagawa[6], present an innovative electronic voting factors that Influence the adoption of e-voting includes
system. The proposed system encodes voting trust in internet, trust in government, attitudes, website
information. This system consists of voting unit, polling design, and compatibility including many others. Pesado
administration unit, voter list administration unit and ballot et. al. [11], have presented the challenges and solutions
or counting unit. After the vote is caste, the information of electronic voting system preferably for Argentina.
is sent to polling administration unit along with the smart Additionally, they have presented the characteristics of
card number in encrypted manner. During the comparison three different voting types, these include on site

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
60
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

electronic voting system, partially onsite and remote The Rest of the paper is organized as follows: System model
voting system. Furthermore, policy considerations are of the proposed setup is presented in Section II. Section III
also provided for the implementation of the proposed presents the performance evaluation of the proposed
system. algorithm. Section IV presents the conclusion of the paper.

Jacobs and Oostdijk [12], present a system that uses bar 2. THE PROPOSED SYSTEM
coded identifiers which are assigned either randomly or
pseudo randomly in the form of combination of numbers In this section, a brief description of various hardware
and alphabets. These encrypted codes provide security units is presented that are integrated in proposed project
from any illegal intervention.Using different identifiers to achieve the improved results for the proposed
makes this system secure in comparison to others. The electronic voting system, as shown in Fig. 1.

voter will then have to scan the bar-code and then the
Microcontroller: A microcontroller can be defined as an
system will decode and compare the code assigned with
integrated circuit that contains a core processor and
that of the database. Upon a perfect match the voter will
memory [16]. Microcontroller is also known as an
be allowed to vote. Awan et. al. [13], implement a
embedded system, capable of storing, processing and
fingerprint based electronic voting system using
transferring data and information between various
Raspberry Pi board. Vidyasree et. al. [14], fuse the
peripherals interfaced with it on some logic, i.e. like a
fingerprint and facial data to improve the identification
coordinating body of a circuit. With the advancement in
of a voter through multimodal system. The results show
the field of electronic technology especially in
a reasonable amount of improvement in comparison to
microelectronics and embedded system development,
unimodal system. Das et. al. [15], store biometric
various development boardsare available. These boards
information of the user i.e. fingerprints on RF ID tags for
include Arduino-UNO, Texas Instruments MSP 430
designing an improved electronic voting machine. The
Launchpad, Nanode, Pinguino PIC 32, Teensy 2.0,
proposed system also integrates the GSM module to
Raspberry Pi and many others. These boards not only
disseminate information from the local station to other
provide microcontroller facility to the end user but also
stations. an interfacing capability to connect different devices i.e.
Bluetooth, Zigbee, LAN and WLAN (Wireless LAN also
We develop and present an electronic voting system
called WiFi). The proposed system (in our research) uses
to eradicate fraudulent practices during public elections
Arduino-UNO board due to good processing speed as
by involving double user identification checks i.e. facial
well as memory, and capable of interfacing, controlling
recognition and finger print based identification
and monitoring of data flow [17].
methods. Facial recognition is accomplished through
a feature-extraction based machine learning algorithm,
while finger print based identification is achieved
through pattern recognition method. The facial
recognition is accomplished through cascading of
Global Principal Component Analysis and K nearest
Neighbor algorithms. The proposed method will provide
better accuracy in comparison to a single identification
FIG. 1. BLOCK DIAGRAM OF CR BASED BIOMETRIC
method. ELECTRONIC VOTING SYSTEM

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
61
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

Fingerprint Module: Unique finger impression the algorithm and its working details are provided in the
recognition or fingerprint authentication indicates the next section of this paper. A flow diagram of the proposed
mechanized strategy for checking a match between two algorithmic setup is shown in Fig. 2. The input image is
human fingerprints [18]. The examination of fingerprints processed to be utilized by trained classifiers that
for coordinating purposes requires the correlation of produce a final decision of either recognized or
components of the print design. The extracted parameters unrecognized.
of a finger pattern include edges and minutia focuses
[19]. These distinct features of a biological pattern give
3. WORKING PROCEDURE
uniqueness to a human being.
In this section, a brief working procedure of biometric
data extraction and processing is presented. Fig. 3 shows
The mechanized method for the verification of a
the registration steps to be taken for the new voter
fingerprint is done by using an electronic device called
registration into the proposed voting system. Fig. 3 shows
Fingerprint Verification Module, which captures the
the execution process of the proposed electronic voting
unique pattern of a fingerprint in the form of a
system.
computerized digital image. The digitally captured images
are then processed to prepare a biometric template. This
As shown in Fig. 3, the registration of the voter begins
biometric layout is an accumulation of extricated elements
by the start of the counter for assigning a voter number
which is stored and utilized for coordinating and matching
[20]. The proposed system uses a finger print verification
module developed by Future Electronics Egypt.

Facial Recognition System: Facial recognition system or

facial acknowledgement framework is defined as an
application capable of detecting and recognizing a person
from a digitally processed image [21]. This unit comprises
of facial recognition algorithms which includes facial
detection, facial feature extraction, formation of biometric
template by compression and formation of Eigen vectors
and their comparison. Many popular facial recognition
algorithms are available in literature that include PCA
(Principal Component Analysis) using Eigen faces, LDA
(Linear Discriminate Analysis), Fisher-face algorithm
andDynamic link matching [22].

The proposed system incorporates the facial

recognition algorithm, developed by [23]. The details of FIG. 3. SHOWING THE REGISTRATION PROCESS OF VOTER

FIG. 2. BLOCK DIAGRAM FOR FACE RECOGNITION

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
62
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

to each voter. The message is displayed on the screen to biometric layout is generated and is compared with the
place the face in front of the camera, the image is captured database in order to find a match. In case the thumb
and normalized and divided into 24X24 sub-windows. impression is not found in the database, an error is
Thus, distinct features are extracted and a vectored displayed and a message is generated for the relevant
biometric layout of the facial image is formed. The users. In case a match is found, a message is displayed
resulting biometric template can be used to train the for the voter to place the face in front of the camera. The
classifier using Adaboost trainer and then a codebook image is then normalized, 24X24 sub-windows are formed
for the Eigen vectors is formed, then the generated and features are extracted. The distinct features are
biometric layout is saved in the database against the vectored and are then compared with the biometric
encrypted ID number. This ends the first step towards layout in the database. If a match is found, the voter is
recognition of facial features. Then a message is displayed allowed to cast the vote. But in case no match is found,
to place the thumb on the scanner, the thumb sensor an error is displayed and a message is generated to the
scans and forms a biometric layout of the thumb of the relevant authority, as shown in Fig. 4.
voter and stores it against the same encrypted voter
number in the database. Now both facial and thumb print In this section, the detailed process of the individual steps
biometric templates are compared with the pre-stored is presented.
biometric layouts in the database in order to eradicate
Facial Recognition System: The facial recognition system
registration of the same voter multiple times. In case the
registered voter tries to repeat the registration step again, is the most significant feature of the proposed hybrid

the registration is rejected. biometric electronic voting system. The algorithms used
for facial recognition usually can be categorized into two
During the voting process, a message is displayed to methods firstly geometric which compare the geometry
place the thumb on the thumb sensor/scanner; a of distinct features and analyze the relative position, size

FIG. 4. EXECUTION PROCESS FOR PROPOSED ELECTRONIC VOTING SYSTEM

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
63
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

and shape of ears, eyes nose and jawbones and secondly the classifiers to recognize the relevant features. Once
photometric which is a statistical methodology to distill an integral compressed biometric template of two-
a picture into statistical values and compares the values dimensional is formed, the features stored in the layout
with the layout [24]. are converted into as set of Eigenvectors and thus an
Eigen face is formed. The formation of Eigen face is to
The algorithm used in the proposed system is based on speed up the analysis and to reduce the response time
the principle of feature extraction. Feature extraction in as shown in Fig. 6.
image processing may be defined as being a set of initial
value derived from an object in the form of a pattern which Facial recognition is implemented through a cascaded
is informative and useful for machine learning. The classifier of GPCA and KNN algorithms. KNN is a non-
algorithm can be implemented using three steps i.e. Haar parametric formula used in classification of data. It is also
feature selection, creation of an integral image and used in pattern recognition. It is one of the simplest
Adaboost training [25]. algorithms of machine learning for pattern recognition
[28]. PCAis an algorithm that convertsthecorrelated
The facial features are detected and analyzed using Haar elements to linearly uncorrelated elements through
feature selection like the positioning, distance and the orthogonal transformation. In Generalized PCA, the
geometric shape of the eyes, nose, ears and jaw bones condition of orthogonally is removed to consider an
and then using the information driven from Haar-Feature arbitrary number of spaces of unknown and different
selection, an integral image is formed [26]. The process of dimensions [29].
face detection and feature selection using Haar feature
The cascaded classifier uses the comparison of
selection can be referred in Fig. 5.
Eigenvectors of the stored bio-metric template with the
A sub-window of 24x24 pixels can exhibit a total of digital image of the voter generated at the time of voting
162,336 possible features and it would be time and then compares the nearest numbers of similarities by
consuming as well as expensive and considered to be introducing a test vector from the live scan of the voter. If
quite an impractical approach for the facial recognition the similarities is less than 90% keeping in mind the
[27]. Hence Adaboost trainer is used which eliminates environmental light and tolerated offset angles, the
the scanning of all insignificant features and also train similarities will be rejected and the voter won’t be able to
cast his/her vote.

Although many methods of face detection are present of

which the cascaded classifier method by using local PCA

FIG. 5. FACE DETECTION USING VIOLA JONES ALGORITHM FIG. 6. FEATURE REDUCTION AND FORMATION OF EIGEN-
WITH RECTANGULAR HAAR FEATURE SELECTION FACE

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
64
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

and LDA but their resulting accuracy is quite low as comparison keeping in mind the tolerated offset angle,
compared to GPCA and KNN. Fig. 7 shows the then the voter won’t be allowed to vote.
comparison of the accuracy of GPCA and KNN with LPCA
and LDA in the next section of this paper.
4. RESULTS AND DISCUSSION

In this section, the results of the proposed electronic

Apart from that the response time for the cascaded
voting system are presented along with comparison with
classifier of GPCA and KNN method is quite fast and
other systems.
responsive as compared to the other methods usually
employed for face-recognition which will be discussed in
The distinct and outstanding features of the proposed
the Results and Discussion section of this paper.
system are that the facial recognition algorithm is unique
and the accuracy yielded as compared to other cascaded
Fingerprint Processing: The method employed by the
algorithms is high. The Table 1 shows the result of the
finger print module is the optical method. Optical finger
experimental outcome of the electronic voting system for
print verification technique maybe defined as the
different K-Values:
formation of a biometric template from the digitally
computerized image for verification using visible light
The finger print module is also tested and its accuracy is
[30]. The surface for scanning the finger print is called as
shown in Table 2.
touch surface and underneath there is a light-transmitting
phosphor layer which enlightens the surface of the finger. The Fig. 9 shows a comparison between KNN and a
The light reflected from the finger goes through the cascaded classifier i.e. GPCA and KNN. The results show
phosphor layer to a variety of strong state pixels which that the cascaded system gives higher accuracy than
captures a visual picture of the finger print [31]. individual KNN classifier.

The algorithm used for the finger print verification is a

pattern-based verification in which the digital image of
the finger print is compared with the previously stored
bio-metric layout on the basis of similarities of the minutiae
features like ridge ending, bifurcation, and short ridge
[32]. The Fig. 8 shows the minutiae features of a finger.

The pre-stored template containing the features of

minutiae features are compared with the finger print of
FIG. 8. RIDGE DOT [30]
the voter and if the comparison yields less than 90%
TABLE 1. FACIAL TESTING OUTCOME

Faces tested Correct Incorrect Missed Accuracy

1 00 91 3 6 91%

TABLE 2. FINGER PRINT TESTING OUTCOME

Tested Correct Incorrect Missed Accuracy

FIG. 7. USINGK-NN METHOD BY INTRODUCTION OF TEST
VECTOR 1 00 98 0 2 98%

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
65
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

The use of cascaded classifier of KNN and GPCA rather biometric template and its effect on time response and
than just using KNN and the outcome comparison of was found to increase with the increasing number of
there accuracy with respect to changing number of K-
distinct features and founded that the algorithm had a
values compared in a single cycle can be seen in Fig. 9
response time of 4.32 seconds for a 1764 distinct features
having an accuracy of 91% for a preset value of k=1 in the
and k-value=1 (k-value is the number of features
implemented system.
compared per cycle) as preset in the system for real-time
Also from the results and comparison of outcome of other implementation.
studies and research papers, the accuracy of the outcome
of separate and paired classifiers at a constant dimension Fig. 12 shows the comparison of the accuracy of GPCA
is shown in Fig. 10 for comparison of 1764 distinct features and KNN with LPCA and LDA with respect to the distinct
and a K-value of 1 (for algorithms using K-NN). features compared with the features stored in the biometric

the experiment carried out yielded the following results template as preset in the algorithm for testing and having

which are interpolated in Fig. 11 which shows a relation an accuracy of 91% (approximately) and k-value=1 (k-
between distinct features in pixelated form stored in a value is the number of features compared per cycle) for
the 1764 distinct features.

FIG. 11. RESPONSE TIME WITH VARYING DIMENSIONS

FIG. 9. COMPARISON OF ACCURACY BETWEEN KNN AND

GPCA+KNN

FIG. 12. COMPARISON OF ACCURACY WITH DIFFERENT

FIG. 10. ACCURACY OF DIFFERENT CLASSIFIERS CASCADED CLASSIFIERS

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
66
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

5. CONCLUSION [6] Tagawa, K., “Electronic Voting System”, Google Patents,

2010.
A novel hybrid biometric voting system is proposed,
[7] Evertz, G.P., “Electronic Voting System”, Google
implemented and analyzed for fair polling process during
Patents, 2009.
general elections in developing countries including
Pakistan, Nepal, Sri Lanka and others. The investigation [8] Azougaghe, A., Hedabou, M., and Belkasmi M.,”An
results show91% accuracy of the proposed system. With Electronic Voting System Based on Homomorphic
Encryption and Prime Numbers”, 11 thInternational
the implementation of GPCA and KNN cascaded
Conference on Information Assurance and Security,
classifiers that are discussed in the previous section of
pp. 140-145, 2015.
the paper. Additionally, the proposed system also involves
the finger print based security feature to provide [9] Awad,M., and Leiss, E.L., “The Evolution of Voting:

additional authenticity of the voter. The future work will Analysis of Conventional and Electronic Voting
Systems”, International Journal of Applied Engineering
be to incorporate security features in the proposed system
Research, Volume 11, pp. 7888-7896, 2016.
by introducing encryption algorithms.
[10] Alomari, M.K., and Irani, Z., “E-Voting Adoption in a
ACKNOWLEDGMENT Developing Country”, Transforming Government:
People, Process and Policy, Volume 10, 2016.
The authors would like to thank the Administration of NED
University of Engineering & Technology, Karachi, Pakistan, [11] Pesado, P., Galdamez, N., Estrebou, C., Pousa, A.,

for providing resources to complete this research. Rodriguez, I., Eguren, S.R., Chichizola, F.,Pasini, A., and
Giusti, A.D.,”Experiences with Electronic Vote:
REFERENCES Challenges and Solutions”, Proceedings of 9 th
International Conference on Theory and Practice of
[1] Bhuiyan, M.S., Rehman, M.M., and Hossain, M.R. Electronic Governance, pp. 406-407, 2016.
“Electronic Voting System”, B.Sc. Thesis, Department
of Computer Science & Engineering, , BRAC University, [12] Jacobs, B., and Oostdijk, M., “Electronic Voting System”,

Dhaka, Bangladesh, January, 2007. Google Patents, 2002.

[2] Traoré, J., “An Introduction to Electronic Voting [13] Awan, F.G., Ijaz, M.F., Qureshi, S.A., Zahid, A., Ali, M.Z.,

Application to Single Transferable Vote”, Orange Labs, and Tariq, A., “An Automated DRE Biometric Voting

July, 2014. System with VVPAT Feature”, Science International,

Volume 28, 2016.
[3] Whither Biometrics Committee, “Biometric
Recognition: Challenges and Opportunities”, National [14] Vidyasree, P., Raju, S.V., and Madhavi, G., “Desisting

Academies Press, 2010. the Fraud in India’s Voting Process through Multi
Modalbiometrics”, IEEE 6th International Conference
[4] Drexler, J., and Dyball, C.J., “Anti-Fraud Voter on Advanced Computing, pp. 488-491, 2016.
Registration and Voting System using a Data Card”,
Google Patents, 1995. [15] Das, A., Dutta, M.P., Banerjee, S., and Bhunia,
C.T.,”Cutting Edge Multi Stratum Secured Electronic
[5] Sehr, R.P., “Computerized Voting Information System Voting Machine Design with Inclusion of Biometrics
Having Predefined Content and Voting Templates”, RFID and GSM Module”, IEEE 6th International
Google Patents, 1999. Conference on Advanced Computing, pp. 667-673, 2016.

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
67
 A Novel Hybrid Biometric Electronic Voting System: Integrating Finger Print and Face Recognition
Downloaded from search.informit.org/doi/10.3316/informit.308126203732316. on 02/18/2024 04:24 AM AEST; UTC+10:00. © Mehran University Research Journal Of Engineering & Technology , 2018.

[16] Cantrell, T.,”Microchip on the March”, Circuit Cellar, [24] Bonsor, K., and Johnson, R., “How Facial Recognition
1998. Systems Work”,HowStuffWorks.com. Np, Volume 4,
2001.
[17] D’Ausilio, A., “Arduino: A Low-Cost Multipurpose Lab
Equipment”, Behavior Research Methods, Volume 44, [25] Viola, P., and Jones, M., “Rapid Object Detection Using
pp. 305-313, 2012. a Boosted Cascade of Simple Features”, Proceedings of
IEEE Computer Society Conference on Computer Vision
[18] Kumar, R., Gowri, B.,and Kumar, K.V., “Biometric and Pattern Recognition, Volume 1, pp. I-511-I-518,
Security Based Application Development and Emulation 2001.
Framework-IEEE 802.15.4 for Intensive Care Units”,
International Conference on Recent Advances in [26] Crow, F.C., “Summed-Area Tables for Texture Mapping”,

Computing and Software Systems, pp. 228-232, 2012. ACM SIGGRAPH Computer Graphics, Volume 18,
pp. 207-212, 1984.
[19] Thornton, J., “Latent Fingerprints, Setting Standards in
[27] Szeliski, R., “Computer Vision: Algorithms and
the Comparison and Identification”, Proceedings of
Applications”, Springer Science & Business Media, 2010.
84th Annual Training Conference, California, State
Division of IAI, 2000.
[28] Peterson, L.E., “K-Nearest Neighbor”, Scholarpedia,
Volume 4, pp. 1883, 2009.
[20] Khan, M.K., “Fingerprint Biometric-Based Self-
Authentication and Deniable Authentication Schemes [29] Vidal, R., Ma, Y., and Sastry, S., “Generalized Principal
for the Electronic World”, IETE Technical Review, Component Analysis (GPCA)”, IEEE Transactions on
Volume 26, pp. 191-195, 2009. Pattern Analysis and Machine Intelligence, Volume 27,
pp. 1945-1959, 2005.
[21] Bansal, A., Mehta, K., and Arora, S., “Face Recognition
Using PCA and LDA Algorithm”,2 nd International [30] Wasserman P.D., “Solid-State Fingerprint Scanners”,
Conference on Advanced Computing & Communication Presentation, NIST, 2005.
Technologies, pp. 251-254, 2012.
[31] Setlak, D.R., “Advances in Biometric Fingerprint
[22] Brunelli, R., and Poggio, T., “Face Recognition: Features Technology are Driving Rapid Adoption”, Consumer
Versus Templates”, IEEE Transactions on Pattern Marketplace, Retrieved, Volume 13, December, 2005.
Analysis & Machine Intelligence, pp. 1042-1052, 1993.
[32] Maltoni, D., Maio, D., Jain, A., andPrabhakar, S.,
[23] Jones, M.J., and Viola, P., “Face Recognition Using “Handbook of Fingerprint Recognition”, Springer
Boosted Local Features”, 2003. Science & Business Media, 2009.

Mehran University Research Journal of Engineering & Technology, Volume 37, No. 1, January, 2018 [p-ISSN: 0254-7821, e-ISSN: 2413-7219]
68