International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 541

ISSN 2229-5518

Development of Online Voting System using

Minutiae based Algorithm

Talib A. Divan, Veena A. Gulhane

Abstract— Voting is one of important task which has to conduct in a country for electing a government. But being such an important part hence it has to be manage securely and accurately. But the voting percentage decreases year by year because of inflexibility. Most of ex- isting system studied suffers from high FRR rate. In this paper a minutiae based algorithm is used for designing a voting system which is implemented using combination of hardware and software. Minutiae based algorithm uses two fingerprint authentication which make sys- tem more secure. Proposed system aims to design a flexible low FRR rate system with high accuracy. The simulation result of proposed system is discussed in the result section. Also, the proposed is analysed comparing with the existing system considering the FAR and FRR rate

Index Terms— EVM, Fingerprint, FAR, FRR, Minutiae based algorithm, Template, Voting system.

——————————  ——————————

1 INTRODUCTION

lectronic voting machine is generally used now days in some countries including India for conducting election of government in a country. But the Electronic voting ma- chine has certain disadvantages like illegal voting and insecu- rity. Hence the concept of online voting system is started in some countries for conducting election. Most of the developed countries have started using online voting system but they are facing some problems in conducting it. Estonia is the only country started conducting the online voting system in na- tional election. But the percentage of voting is only 20% to
30%.
Different researchers have designed a online voting
system But the system are not so much efficient in terms of
accuracy and security. Also the voting system has high error
rate. Hence the voting system is not flexible and can be used
for specific region only. Biometric authentication is found to be more secure and accurate in certain application. Different biometric authentications like fingerprint, retina etc. can be used in designing an application to enhance the security. As fingerprint of every individual is unique it can be used for designing a voting system. Different fingerprint matching techniques has been discussed considering the FRR ratio.
Various techniques based on fingerprint matching
such as Threshold cryptography , Fuzzy Vault, Latent finger-
print matching , Minutiae based algorithm has been reviewed
depending upon their error rate. In proposed method, a voting
system has been designed using minutiae based algorithm. Minutiae based algorithm is the best techniques in fingerprint matching using the minutiae position from fingerprint tem- plate. The voting website has been designed which uses a fin- gerprint authentication to allow the user to cast his valuable vote. MATLAB software is used for processing the fingerprint image using various steps like binarization, thinning, Minutiae detection and removal of false minutiae.
GSM based EVM [1] was based on fingerprint authen- tication, but there are certain problem in these system i.e. high
error rate. Hence, the system lags with high FRR ratio. Minuti- ae based algorithm is not yet been used in designing any vot-
ing system hence proposed system implements a Fingerprint based online voting system. By using the proposed system the FRR rate is found to be less and
the voting system is found to be more effective in sense of ac- curacy and security. By using the proposed system the voting percentage in country may increase and also people who are unable cast their votes can do it easily.
To achieve the objective i.e. to calculate the FRR rate
in proposed system, fingerprint images of certain users are
taken and system is test by fingerprint authentication. As False
Rejection rate is defined as the ratio of number of user rejected which are authorised to total number of verification. The FRR is calculated which is described in result section. The web page is designed which is found to be more effective in casting votes for authorised user living anywhere in country.
The related work about the previous research is explained in section II. Section III described details of proposed method- ology and steps in minutiae based algorithm. Section IV de- scribes methodology for online voting system. Section V de- scribes the system architecture showing the proposed model. Section VI describes the results obtained by showing the snap- shot of hardware and MATLAB simulation. VII describes the result analysis of proposed system by comparing it with other existing ones. Section VIII describes the conclusion.

2 RELATED WORK

Various researches have been carried out on designing the online voting system using different techniques. Most of them were designed the online web application which just allow the user to cast their valuable votes by using the au- thentication using login id and password system. It is found to some extent that is better, but the problem is system prone to various database attacks. Hence it is not employed in some countries because of validation of the system.

In GSM based voting [1] system a single fingerprint scan-

ner is used for enrolment and authentication of particular

user. But the problem exist with this system is that it is lim-

International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 542

ISSN 2229-5518

ited to specific region and hence the system is not flexible. Also, the error rate i.e. FRR is very low. The FRR rate of this system is found to be 1.4%.

Minutiae based algorithm [2] is studied in which two fin- gerprints image are used for creating a mew template which can be stored in database. It is found with the experimental result i.e. FRR rate is very low i.e. 0.4%. Minutiae positions are extracted from the fingerprint image by considering the orientation and reference points. Reference points are select- ed from first and orientation field is selected from the second fingerprint and the template is stored in the database. Dur- ing authentication phase, the fingerprint images are matched with the template stored in the database and if shares match with it, the user gets authenticated.

Latent Fingerprint Matching [7] technique is used general- ly in some smudgy and damaged fingerprint images which are not authenticated by using different techniques. Hough transform is used which improves robustness & distortion in fingerprint image. In this process two types of methods are used manual marking and automatic marking to extract the minutiae positions. The latent fingerprint matching tech- nique is found to be more efficient in smudgy fingerprint image than normal ones. Hence, it produces very low accu- racy while using normal fingerprint images.

Moodle [3] is a software package used for creating differ- ent course and internet based websites. For registration of student to access the website a new concept i.e. fingerprint matching is used. Fingerprint image is used as the authenti- cation in login window for studying certain courses.

It is found that it is more secure by using fingerprint bio-

metric authentication than password based system.

Threshold Cryptographic [5] technique is one of

most important fingerprint matching technique. In this tech-

nique, fingerprint image is divided into two are more shares

using cryptography and then compression is applied for

compressing image in database. One part of shares is stored in database for particular user and other is given with the user. If both shares match, the user gets authenticated. The error rate i.e. FRR by using threshold cryptography is low.

Some of papers based on online voting system is also stud-

ied which describes how the online system helps in reducing time, increasing number of user and security.

Biometric-Secure e-Voting [11] is designed to replace the ex- isting system. The design of the Biometric system guarantees the community that there will be no false voting in the elec-
tion. Hence the percentage of voting user increases automati- cally. Also main focus of Biometric e-Voting is achieving the election integrity. But the system is not employed in real world using fingerprint scanner.
Online Voting System [12] based on Aadhar Id verification is proposed. In this paper a voting system is conducted based
on unique number for enrolment of particular user. A frame- work is designed which is expected to be more secure and free from accessing by any unauthorised user. But in this system user id and password based authentication is used which is not much secure.
To achieve high accuracy in proposed system, a minutiae algorithm based on two fingerprint sensor is designed. By employing two fingerprints image authentication the system will be more secure and produce accurate results.

3 METHODOLOGY FOR ONLINE VOTING SYSTEM

The following online voting system model to be con- structed aims to provide a low FRR rate. The model is divided into two parts i.e. hardware part and software part which shows simulation results in MATLAB software. Firstly, a cir- cuit diagram is designed which consist of Microcontroller (ATMEGA 16), Fingerprint sensor (R305) and Serial Commu- nication (RS232).Two Fingerprint sensors are interfaced using Relay to microcontroller ATMEGA16. Port D is used in which pin no. 2 & 3 is used for connecting serial communication i.e. RS232 for transmission and reception of data. The second stage is of software part which is again divided into two parts namely extracting minutiae position of fingerprint in MATLAB and online voting part which allow the user to cast vote by login in website. The minutiae based algorithm im- plementation is shown in MATLAB simulation. The flow chart of how the voting is carried is shown below. The user first has to enrol for only one time with fingerprint verification which is stored in the database. After enrolment the user fingerprint are applied with minutiae based algorithm in MATLAB soft- ware. After which user is authenticated with name if finger- print matches with database. Then, user has to visit the web- site. User has to create new account and then the vote is rec- orded. In this way the process of online voting system is car- ried out, the result and snapshot of proposed system is shown in further section. The flow is designed below which shows how the whole system works for casting a vote. If the user visit website without fingerprint verification then the website shows a error message showing fingerprint does not matches.

Fig. 1.Flow Chart

International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 543

ISSN 2229-5518

4 SYSTEM ARCHITECTURE

Fig. 2. Block Diagram

Fig. 3. Circuit Diagram

AVR. It has 16k Bytes of ISP Flash and 512K Bytes of EEPROM. ATMEGA16 achieves throughputs of 1 MIPS per MHz by executing instruction in single clock cycle. It consists of Two 8-bit Timer/Counters and one 16-bit Timer/Counter with separate pre scalar. In proposed system two fingerprint sensors are interfaced with microcontroller with two UART and the Serial Communication is done via a Relay as third UART port. It has 16K bytes of In-System Programmable Flash Program memory with Read-Write capabilities and 1KB SRAM. For serial communication between the DTE(Data Ter- minal Equipment) and DCE(Data Communication Equipment) RS232 is used which act as a asynchronous serial communica- tion between MAX232 IC and computer. Serial Communica- tion does not directly communicate with computer rather first it sends signals via MAX232. For sending the signals for read- ing fingerprint image from computer to microcontroller, RS232 is used. It is for reading and writing data. As ATMEGA
16 consist of only two UART, in proposed system three UART
are required i.e. for interfacing fingerprint sensor1 & 2 and
RS232 for serial communication. A relay is used which works as a additional UART in system.
The software part of proposed system composed of MATLAB and PHP server side language is used for website purpose. The Minutiae based algorithm is shown in the MATLAB using various steps involve in it like binarization, thinning, minutiae detection and false minutiae removal. The user has to keep his fingerprint in sensor and then after capturing image from both fingerprint image the algorithm is applied. The second part of software is PHP language which is used for designing the website for online voting. WAMP serv- er is used for designing the check vote, manage votes and oth- er features in the website. The flow how the proposed will work is shown below and the results obtained from the pro- posed is shown in the result section.
The system consists of fingerprint sensors, microcon- troller (ATMEGA16), RS232 (Serial Communication), PC (MATLAB Simulation). The circuit diagram is constructed as per the block diagram. Fig 2. shows the following circuit dia- gram of the system.
Sensors are the devices which are used for converting the physical quantity into electrical signals. These signals will be sensed and with the property of the data captured from sensor further processing is done. In proposed system two sensors are used which is connected to microcontroller. If two fingerprint image captured by sensor matches with data stored in database then the user gets authenticated. As two sensors are used the system will be more secure than compari- son to other ones. R305 Fingerprint sensor is used in proposed system which is a optical sensor with low power consumption, low cost, small size and performance is fine. Verification Speed and Scanning Speed of Fingerprint sensor is very low, hence it perform the operation quickly.
In proposed system, ATMEGA16 is used which is a 8- bit microcontroller with High-performance and Low-power

1. Start

2. Initialize port D and fingerprint sensor pins

3. Ask user to keep 1st fingerprint impression in sensor1
4. Ask user to keep 2st fingerprint impression in sensor2

5. If match found display the name enrolled

6. Otherwise, repeat step 3 &4 for new enrolment

7. Apply Minutiae based algorithm in both fingerprint image

8. Visit the website for casting vote.

9. Signup with same name as enrolled in database.

10. Login to cast vote.

11. If user logged in without fingerprint authentication in

http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 544

ISSN 2229-5518

MATLAB.

12. Error message as Fingerprint match not found.

13. End

5 RESULTS

Proposed system consist of both hardware and soft- ware implementation. Hardware part consist of finger- print sensor, microcontroller and RS232 for serial com- munication. Software part consist of simulation of fin- gerprint image by applying minutiae based algorithm in MATLAB software and online voting part in website. The detail of how the online voting is conducted is shown below considering various snapshots.

Fig. 6. Authenticating user through minutiae points

Fig. 4. Hardware implementation

Fig. 7. Website for online voting

Fig. 5. Applying Minutiae based Algorithm

Fig. 8. Different windows in website

International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 545

ISSN 2229-5518

The above snapshot shows how the online voting is performed using the proposed system. For analyzing proposed system from differentiating with other existing one, following analysis has been carried out. The main objective of proposed is to design a system with low error rate i.e FRR rate which is explained below.
1) FRR (False Rejection Rate)
False Rejection rate is defined as the ratio of number of times the genuine user gets rejected to total number of verifications.
It is given by

Fig. 9. Select Party to vote

Fig. 10. Voting Result Window

2) FAR (False Acceptance Rate)
False Acceptance rate is defined as the ratio of number of times an unauthorized user gets accepted to total number of verifications.
It is given by

Both the above parameters are tested in the pro- posed system and it is found that the first parameter i.e FRR(False Rejection Rate ) is very low i.e only one time the system has rejected the authorized user. The second parameter i.e. FAR (False Acceptance Rate ) is found to be zero i.e. there is no such acceptance of any unauthorized user in the system. The system is tested by enrolling more users.

TABLE 1

VALUES OF FAR & FRR

FAR	FRR
0	0.03%

TABLE 2

DETAILS OF EVALUATION

No of

Enrollment

Successful

Verification

Unsuccessful

Verification

Fig. 11. Manage voting time

International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 546

ISSN 2229-5518

TABLE 3

COMPARING WITH OTHER SYSTEM

TYPE OF SYS- TEM	FRR	FAR
ONLINE VOT- ING SYSTEM USING MINUTI- AE ALGORITHM	0.03%	0%
GSM BASED VOTING SYS- TEM	1.4%	0.02%

Fig. 12. Result Analysis

REFERENCES

[1] Sreenath Sreenath.M, Sukumar.P, Naganarasaiah Goud.K, P.Sivakalyani & V.Phani Kumar, “GSM based electronic voting ma- chine using touch screen,” IOSR Journal of Electronics and Commu- nication Engineering, June 2014.

[2] Sheng Li and Alex C. Kot “Fingerprint Combination for Privacy Protection,” IEEE Transactions on Information Forensics and Securi- ty, February 2013.

[3] Rosario Gil, Mohamed Tawfik, Alberto Pesquera Martín & Sergio Martín, “Fingerprint Verification System in Tests in Moodle,” IEEE Journal of Latin-american Learning Technologies, February 2013.

[4] Diponkar Paul & Sobuj Kumar Ray, “A Preview on Microcontroller Based Electronic Voting Machine,” International Journal of Infor- mation and Electronics Engineering, March 2013.

[5] Rajeswari Mukeshi & V.J.Subashini, “Fingerprint Based Authenti- cation System Using Threshold Visual Cryptographic Tech- nique,” IEEE-International Conference On Advances In Engineer- ing, Science And Management, March 2012 .

[6] Muhammad Umer Munir and Dr. Muhammad Younas Javed, “Fin- gerprint Matching using Gabor Filters,” National Conference on Emerging Technologies, 2004.

[7] Paulino & Jianjang Feng, “Latent Fingerprint Matching Using De- scriptor-Based Hough Transform,” IEEE Transactions on Infor- mation Forensics and Security, March 2013.

[8] Karthik Nandakumar, Anil K. Jain & Sharath Pankanti, “Finger- print-Based Fuzzy Vault: Implementation and Performance,” IEEE Transactions on Information Forensics and Security, December

2007.

[9] Sheng Li and Alex C. Kot, “A Novel System for Fingerprint Privacy Protection,” 7th International Conference on Information Assurance and Security, 2011.

[10] Xi Cheng, Sergey Tulyakov and Venu Govindaraju, “Minutiae-based Matching State Model for Combinations in Fingerprint Matching System,” IEEE Conference on Computer Vision and Pattern Recog- nition Workshops, 2013..

[11] Mohammed Khasawneh, Mohammad Malkawi,Omar Al-Jarrah, & Laith Barakat, “A Biometric-Secure e-Voting System for Election Processes,” 5th International Symposium on Mechatronics and its Applications (ISMA08), Amman, Jordan, May 27-29,2008.

[12] Himanshu Agarwal & G.N.Pandey “Online Voting System for India Based on AADHAAR ID” 2013 Eleventh International Conference on ICT and Knowledge Engineering.
The online voting system using minutiae based algorithm is compared with the existing system and it is found that the proposed system is much better than existing ones in terms FRR, FAR and accuracy. As user can cast vote more easily with very less time.The graph is plotted based on the result obtained from voting system.

CONCLUSION

A system with online voting using minutiae based algorithm with low error rate is proposed. For analyzing the system for checking for accuracy two parameters are calculate and the detailed of test results are described. Simulation re- sults of proposed system are shown in MATLAB software. Also the minutiae based algorithm is used for extracting the minutiae position of fingerprint images. The FRR rate of pro- posed is found to be 0.03% with no FAR rate.