Biometric-based Voters Identity Management to Combat Impersonation Vulnerability for A Credible E-Voting Process in Nigeria

(1)

Biometric-based Voters Identity

Management to Combat Impersonation

Vulnerability for A Credible E-Voting

Process in Nigeria

Duke, Etim Bassey

25831127

Thesis submitted in partial fulfillment of the

requirements

for the Award of Master of Science

(M.Sc) Degree in Computer Science at the

Mafikeng Campus of the North-West University

Supervisor:

Prof. B. M. Esief arienhre

Co-Supervisor:

Dr. N. Gasela

November, 2015

(2)

DECLARATION

I declare that this research project on

the

Biometric-based Voters Identity

Management to Combat Impersonation

Vulnerability

for A Credible

E-V

oting

Process

in Nigeria is my

work, and

has never been presented for the

award of any degree in any University. All the information used has been dully

acknowledged both in text and in the references.

Signature ....

z

....

.

Date

...

.

...

~;q~o/.'~

..

.

Duke, Etim Bassey

Approval

Signature...

···";1

Date ...

9. . . ...

1-~~··t<vdb

Supervisor:

Prof. B. M. Esiefarienhre

Department of Computer Science

North-West University

Mafikeng Campus

South Africa.

Signature .. ~ · · ·

Date . .

fl

3f:1/

?:'?.(k? ...

.

Co-Supervisor:

Dr. N.

Gasela

Department of Computer Science

North-West University

Mafikeng Campus

South Africa.

(3)

DEDICATION

(4)

ACKNOWLEDGEMENT

Firstly,

I

would like to thank the

Almighty

GOD for

allowing me to finish this

research in good health and mind.

I am

truly indebted and thankful to my supervisor - Prof. B. M.

Esiefarienhre

and

co-supervisor - Dr.

N. Gasela. Thank you

for the

support,

the

wisdom you

shared and encouragement throughout the course of this research.

Many thanks, to the members of

staff in my department

especially my lecturers

Prof.

0 . 0

. Ekabua,

Dr. Bassey Isong,

Ms Ohaeri and Ms Nochipo for offering

valuable

advice.

Without you, it would not have been possible for me to finish

this research project.

I

owe sincere thanks to my mother, brothers and sisters for their

support

and

love throughout the course of this year.

Lastly, I would like

to thank

my

wife for the undying love and

support

throughout the course of this project.

(5)

LIST OF FIGURES

Figure 2.1 INEC Voters Registration _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 14

Figure 2.2 INEC Voting Process 15

Figure 2.3 Human physiological and behavioral traits 18

Figure 2.4 Biometric System 20

Figure 2.5 Minutia template 21

Figure 2.6 Minutia Based Algorithm [44] 22

Figure 2.7 Pattern Based Algorithm [44] 23

Figure 3 .1 System Actors 31

Figure 3.2 System use Case Model 31

Figure 3.3 INEC Staff Sequence Diagram 35

Figure 3.4 INEC central database schema 37

Figure 3.5 INEC Biometric-based election framework design 38 Figure 3.6 Voter fingerprint authentication architecture 39

Figure 4.1 INEC Election process 40

Figure 4.2 INEC system front page 42

Figure 4.3 Unsuccessful logon page 42

Figure 4.4 Successful logon page 42

Figure 4.5 System admin page 43

Figure 4.6 INEC Staff page 43

Figure 4.7 Voter registration page 44

Figure 4.8 Voter identification page 45

Figure 4.9A Electronic ballot paper 46

Figure 4.9B Electronic ballot paper 46

Figure 4.10 Successful voting 47

Figure 4.11 INEC information printing page 4 7

Figure 4.12 INEC voter register printing page 48

Figure 4.13 INEC Voters Register 48

(9)

LIST OF TABLES

Table2.l Key Terminologies _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 10

Table 3.1 Biometric Systems Evaluation 27

Table 3.2 System FR and NFR 29

Table 3.3 System login 33

Table 3.4 Voters registration 33

Table 3.5 Voters Identification and authentication 34

(10)

ABSTRACT

Today, democracy has become a de

facto

standard of every form of government

in the world, nations and organizations. People are given the power to vote for

candidates of their choice to represent them in offices. These require that

elections are organized in a manner that is unbiased, fair and free of

manipulation

s.

This is because confidence can only be gained and the results

generally accepted if the electorate sees the electoral process a

s

being

transparent, ac

cura

te and fair. Although Information and Communication

T

echnology is being applied during the Ni

g

erian ele

c

toral process, the process

i

s

yet to be efficient

,

transparent and trust

w

orthy

.

Nigerian elections are marred

w

ith violence

, v

oters' fraud such a

s

rig

g

ing,

impersonation

,

multiple voting and

v

ote count falsification

.

These are seen as threat

s

to the integrity of the election

proce

ss

. The major cause is the lac

k

of strong security measures in place to keep

the

s

e acti

v

ities under check. Therefore, in this research

,

a possible solution is

offer to t

h

e challenge

s

faced by the Independent National Electoral Commission

(INEC)

t

oday in conducting a credible e-

v

oting process in Nigeria. In this

r

e

s

ear

c

h wo

r

k

, a

framework was designed for managing voters' identity for

credibl

e

elections and electronic voting in Nigeria. The objective was to protect

the electoral system integrity by ensuring that an eligible voter has one

registration and vote only once for a candidate in an election. The approach was

by collecting potential voters' information during the voter registration exercise

and the actual identification of voters on Election Day in a real-time mode using

biometric as authentication factor. In addition and as a proof of concept, this

research implemented a web-based system prototype for INEC that demonstrate

(11)

the effective operation of our idea. The system was tested and the results

obtained

s

howed an improved and credible electoral process through ICT

(12)

Acronym

ICT

INEC

FIS

ID

IDm

PVC

WLAN

DREs

LIST

OF

ABBREV

ATIONS

Meaning

Information

and

Communication Technology

Independent

Na

tio

nal E

l

ectoral Commission

Fingerprint

Identification

System

Identity

Identity Management

Permanent Voters Card

/uUt¾

J

Wireless

Local

Area Network

(13)

Chapter 1: Introduction and

Background

1.1 Introduction

In recent years, democracy has become a de facto standard of every form of government in the world, nations, and local governments and even in organizations. It has been described as the healthiest, participatory and most acceptable form of government that encourages coordination and collaboration between the private and public sectors as well as the citizens and government. Democracy promotes electoral processes and within the sphere of a democratic government, elections occur frequently. The function of elections is chiefly to elect representatives and more infrequently to take decisions about representatives regarding issues that are specific [ 1]. In other words, elections are formal decision-making processes by which a population chooses individuals to hold public office [2]. With the proliferation in infonnation and communication technology (ICT) in the e-world today; elections have taken a new dimension as they can be electronically monitored where a number of possible technological applications can be used. One such application is electronic voting (hereafter referred to as e-voting). E-voting is used to describe several types of voting approaches which involve both casting of a vote and counting a vote electronically [3]. The approach in its mode of operations does not involve a new electoral or voting process but rather, a modification of the existing voting process [3]. In this case, it is geared towards re-engineering the existing manual or traditional electoral process to make it more advanced. In electoral systems today, the introduction of e-voting has been to enhance the existing approach with respect to transparency and efficiency especially, when it comes to national elections [ 4]. This goes with the critical need to organize elections in such a way that is unbiased, fair and free of manipulations [ 4]. The rationale is that the public can only gain confidence in the electoral system if it is based upon transparency, accuracy and fairness [5, 6, 7]. In many developed nations of the world today such as the US, the UK, France, Brazil, Israel and Russia, they have witnessed the advantages of e-voting. Though the benefits are numerous, e-voting also comes with pitfalls. The electoral technology is viewed by many as a threat to the integrity of elections where new sets of risks, doubts and opportunities for fraud and failure are created [8]. In other words, they are not as safe, accurate and trustworthy as they were promised to be [8]. Most of these problems have been experienced in the US

(14)

elections, especially the year 2000 elections where several security issues, design and certification flaws, difficulty in auditing or recounting [9, 10, 11, 12], were reported. Though most of the reported problems have been fixed, it is still considered a serious and complex problem in a continent like Africa.

In the Nigeria of today, though some elements of ICTs have been incorporated into the electoral system, the manual process is still in use. However, the greatest challenges are the issues of frauds and malpractices stemming from rigging and impersonation. These forms of practices have been seen as undermining the integrity of the elections, often creating an atmosphere of instability. Though impersonation gives way to rigging, the cause is due to the inability of the process to effectively manage the identity of voters. These lapses will not only give room for impersonation, but also threaten the credibility and integrity of the election as has always being. Given the critical context, it becomes imperative that the identities of potential voters are effectively managed using a biometrics-based approach to uphold elections integrity in Nigeria.

1.2 Background Information

With the exponential growth of the Internet, activities on the Internet have expanded to various areas and sectors and have reached even hard to reach areas of the world. In life, a person is known to have one name and identity, but the case is not the same on the Internet as

one person can have several identities [13]. Consequently the different identities can be used by the person for different purposes or services. When people have multiple identities as this, it becomes a problem as human can control several identities. Thus, a scheme that differentiates the person for an online identity is indispensable [ 13]. This is the challenge faced by elections in the African continent today, Nigeria in particular which has been marred with impersonation which in turn gives room for rigging.

In order to solve the problem and enhance the existing electoral process, effective identity management of voters is required. Identity Management (IDm) covers the processes and systems for management and control of identities of individual principals, their authentication, authorization [ 14, 15] with the objective of increasing security and productivity while minimising cost and repetitive tasks. IDm constitutes a practice where features that can uniquely identify individuals in a given setting are managed [ 16]. Today, there are several ways to identify an individual which are broadly classified (1) by something

(15)

he knows or memonzes, (2) by something he cames and, (3) by a personal physical characteristic (17]. All these approaches have their strengths and weaknesses which are some of the determinant factors for their choice of application.

In particular, the approach which uses personal physical characteristic of humans called biometrics has been one effective approach of IDm. Today, biometrics has gained considerable interest and attracted widespread usage such as in immigration control system, crime investigation by Police etc. Biometrics offers a natural identity management tool that is characterized by greater security and convenience than traditional methods of personal recognition (17) which include, name, passwords, PINs, ID Cards, etc. In biometrics, identification or identity verification of a person is based on the physiological and behavioural characteristics of the person [ 18, 19]. This is made possible due to the permanence, uniqueness, and universality of human characteristics (17]. In this case, biometrics is used to overcome the danger of loss and failure to recall by other methods of identifications. Biometrics assures that each person has only one identity in the system and that only one person can access each identity (20]. Common individual biometrics used for recognition includes finger and palm friction ridges, iris, face, voice, handwriting, hand shape, and hand vein patterns [ 1 7).

As one of the core objectives of electoral system is to ensure free and fair elections, where one voter is allowed to register and vote once, it is imperative that employing biometrics into the system would definitely help protect its integrity. Using biometrics such as fingerprint instead of identification with names, passwords, PINs, ID Cards, PVCs during the process of registration and voting in Nigerian elections will go a long way to c3reate an electoral system that is credible, trustworthy and transparent.

1.3 Problem Statement

Today, ICT has played a critical role in every walk of life and electoral process is not an exception. The application of ICT in the form of e-voting has been valuable as experienced by the majority of the developed nations of the world. In the same way, they come with several challenges which we believe cannot be contained in the Nigerian elections context especially when it comes to issues such as registration mix-ups, system failures and audit prevention due to software glitches (9, 10, 12]. In addition, the existing manual electoral process in Nigeria is marred by malpractices such as voter impersonation which provide

(16)

room for rigging votes and vote count falsification. In the event of these problems elections can be disastrous. Consequently, this could lead to situations ranging from national instability, endless wars resulting to deaths of citizens. Also, in Nigeria, biometrics is collected during registration locally and not in a real-time mode [27, 46]. In addition, ID cards or permanent voter cards (PVC) are used as a means of identification on Election Day while the collected biometrics are sometimes not used.

Given the critical context it becomes imperative that the existing electoral system in Nigeria has to be reengineered for fitness of purpose such as ensuring the privacy of individual voter as well as one person having one registration and vote. In addition, other benefits ( efficiency, speed, and accuracy) can also be derived from the system. This therefore forms the motivation for this research study.

1.4 Research Questions

In consideration of the above stated problem, this research would be addressing the following research questions (RQs).

RQl: What existing biometric technology can be applied to manage a credible electoral process in Nigeria?

This research question is answered in Chapter 3. The basis to explore the different biometric systems to identify which one is convenient and suitable for use in this research as the authentication factor.

RQ2: How can we manage the identity of each voter to ensure its privacy?

This research question is also answered in Chapter 3 and Chapter 4. The basis to identify different information to be collected from voters and how to access them using biometrics.

RQ3: How can we develop a generic framework for a credible electoral process in Nigeria that can identify an intending impersonator?

This research question is answered in Chapter 3 and the implementation prototype is in Chapter 4. The basis to design the overall idea as a blue-print for implementation.

1.5 Research Goal and Objectives

(17)

1.5.1 Research Goal

The main goal of this research is to develop a novel framework for managing voters' identity

using a biometric-based approach to combat impersonation in Africa's elections.

1.5.2 Research Objectives

In order to achieve the main goal of this research, the following four objectives will be carried out:

1) To analyse current or existing biometric-based identity management technology to see which one is suitable in this context and how it can be applied to ensure voters'

pnvacy;

2) To evaluate existing electoral processes to find out how identity management can be incorporated effectively and efficiently;

3) To develop a framework for electoral system in Nigeria that incorporate biometric identification both at registration and actual voting; and

4) To analyse the benefits that can be derived from the systems. 1.6 Rationale of Study

The credibility of any election depends solely on the processes involved, how transparent it is. Having substantial public confidence in the electoral process and its results is a major

issue [ 4]. In Nigeria today, most people have lost confidence in the traditional election process due to malpractices such as fraud, ghosts voting or impersonation and chain voting

and multiple voting [12]. The above situations have brought about lack of transparency and

accuracy and have also reduced public confidence and trust in the democratic process. In

addition, the reported flaws and problems associated with e-voting are also the rationale why aii developing countries are currently rejecting and criticising the voting technology. With the

problems of e-voting and the current challenges facing the traditional electoral system in Nigeria, there is therefore the need to restructure the traditional electoral system.

This research is therefore seeking to provide lasting solutions to the elections system in Nigeria. The research work aims to eliminate to the barest minimum, the case of impersonation and rigging practices, whether acted directly by officials or indirectly by

voters. In achieving this, a biometric-based IDm approach is implemented which can operate

(18)

impersonation or rigging activities. The objective is to ensure that Nigerian elections can be run with one man having one registration and one vote.

1.7 Research Methodology

Research method is necessary to provide clarity and transparency in terms of research

reporting methods, procedures in order to responsibly show how data have been collected, synthesized, analysed and discussed [21]. Therefore, in this research, the following research methodologies will be used:

1. 7.1 Literature Survey

In this study we carried out a comprehensive literature survey and review of some closely

related literature in the area of biometrics identification and privacy, election processes and their challenges. The knowledge gained from this review will then provide the current research with what has been and has not been researched. The researched material would be acknowledged to provide easy reference for future research.

1. 7. 2 Framework Analysis and Design

The knowledge and information gained from the literature review will be used to formulate a generic framework for credible elections in Nigeria that will ensure privacy of voters and ensure that one individual may register and vote once. This is indeed necessary to ensure that the integrity of the elections is upheld always.

1. 7. 3 Framework Implementation

As a proof of concept for this research work, we shall implement the design framework to ascertain its suitability in combating electoral security vulnerabilities.

1.8 Research Scope

The scope of this research will be centred on the design. However, implementation shall be

limited to a prototype system. This is due to the time constraint associated with this research.

1.9 Chapter Organization

The remaining part of this research work is organized as follows:

Chapter two is on review of related literature and it looks at what has been done in the

literature in the areas of elections integrity, Nigerian election systems and challenges as well

(19)

Chapter three is the biometrics-based framework analysis and design which describes

the system requirements and design components.

Chapter four presents the system evaluation and implementation by g1vmg an

explanation of the system functionalities and their working, while

(20)

Chapter 2: Literature Review

2.0 Chapter

Overview

This Chapter focuses on what has been done so far with specific reference to the areas of the use of ICT in electoral systems. It starts with an introduction and progresses to biometrics system that will be deployed to enhance the authenticity and integrity of the elections in Nigeria. In addition, we discuss the different terminologies and concepts that are of essences in this research study.

In today's modern society, for people to be served fairly, democracy has become the most reliable process used by many countries in the world to elect persons into government. With democracy, people are given the power to choose, elect or vote for candidates of their choice they consider capable to represent them in public office. This is carried out through the process of elections which can be manually done in a traditional way or through the use of ICT tools in the form of e-voting.

In today's e-society, the exponential growth of ICT especially the Internet has been valuable and produced positive results. The impact of these ICTs has been felt in all walks of life through e-services and e-practices [5, 22]. Withe-services, e-government which is considered more democratic, transparent, trustworthy and efficient has emerged [23, 24]. E-government has been well-known in the enhancement of e-democratic process that has been known for effectiveness in engaging citizens' support for democratic decision-making processes. It has also strengthened representative democracy via ICTs tools such as e-voting. The application of these tools in electoral processes has been commendable as experienced in some developed countries of the world like the US, the UK, Japan, Brazil, Israel, India, etc.[9, 25]. In order for democracy to function effectively, the issues of building trust and confidence in the voters is critical to the success and failure of the process. This ensures that the electoral process is transparent and all participants are allowed to exercise their rights as stipulated in the constitution [23]. A study shows that there is a decline in levels of voters' participation if there is a lack of confidence in the freedom and fairness of the elections [ 10, 24]. Hence, elections which are at the core of a democratic government have to be conducted in an orderly manner which is free of fraud and malpractices [5]. This research is therefore of the

(21)

opinion that voters' confidence in electoral processes in Nigeria should be improved and the

integrity upheld. This can be achieved through the use of biometrics technology in both

voters' registration and actual voting on Election Day.

2.2 E-democracy and Voting

Today, most nations in the world that is considered legitimate is one that is democratic.

Democracy is a system of government by all the eligible members of a state through elected

representatives [3]. Due to advances in ICT, today elections can be improved using a number

of possible technological applications that allows it to be done electronically. In this case, it is

called electronic voting or simply e-voting. This has however given birth to electronic

democracy which a new form of democracy that is boosted by e-government [25].

E-democracy is simply the use of ICT tool, network and other related technologies like the

Internet to improve democratic processes within a democratic nation [23, 25]. It involves

activities such as voting, registration, electoral campaigns, communication exchange,

participation, consultation and public opinion polling [3). The application of ICT in electoral

processes has been quite commendable as it is used to strengthen the democratic processes

and institutions, promote participations, build trust and confidence on the election process

and so on [23, 24).

In the perspective of the e-voting, it is not a new form of electoral process as it is viewed by

some people. It is rather a process that modifies or re-engineers the existing voting process

making it possible to be conducted electronically. In this case, voters can cast their votes

electronically using technologies such as the internet, telephones, interactive digital

television, SMS text messaging and Direct Recording Electronic Voting Machine (DREs) [3].

By e-voting, the ballot paper which is a hardcopy document is replaced with digital ballot

which is electronic thereby making the process paperless. In e-voting, for example in the UK

and US, some of the processes that can be automated electronically are the Voter registration,

ballot boxes, smartcard for voter identification and ballot paper counting [3]:

l

NWU

· '

·

1 LJBRARY_

The application of technology in the electoral process especially in developed countries has

been advantageous and has strengthened their democratic system. However, the technology

has also introduced risks and opportunities for fraud and manipulation [26]. Cases such as

vote counting inegularities, software bugs that prevents auditing, design flaws, confusing

(22)

the trust and confidence people have in the system. Therefore, in this research, we believe that the Nigerian INEC will meet some challenges in handling these types of issues. Instead, the current system has to be reengineered especially in the identity management of voters using real-time biometrics-based registration and voters' identification.

2.3 Terminologies

In this section, we shall discuss the various terminologies that will be used throughout this study.

Table 2.1: Key Terminologies

Term Definition

Authentication It is the establishment of whether a claim is true or not.

Biometrics Technology that measures and analyses human body characteristics, such as DNA, fingerprints, eye retinas and irises, voice patterns, facial patterns and hand measurements, for authentication purposes.

Ballot paper A paper used for voting where names, party and other information about the candidate is written.

Biometric One or more stored biometric samples, biometric templates or biometric

reference models attributed to a subject and used for comparison.

Biometric sample Analogue or digital representation of biometric characteristics prior to the feature extraction process, obtained from a biometric.

Biometric Stored biometrics features which are compared directly to biometrics template features of a recognition biometric sample.

Confidentiality Ability to protect our data from those who are not authorized to view it.

Democracy This simple government of the people, by the people and for the good and benefit of the people.

E-voting The application of electronic system to help in casting and counting of

votes in an election.

Elections A process by which citizens in a democratic nation, organization or

society choose candidates that they want to represent them in public

offices. It includes registration, verification, voting and declaration of

results

Fingerprint An impression left by the friction ridges of a human finger and extracted

with the aid of an electronic device.

Identification It is the claim of what someone or something is.

Integrity ability to prevent our data from being changed in an unauthorized or undesirable manner.

Manual system A system that is manipulated by a human operator and is not done automatically or electronically.

Polling Station A place set aside for voting during elections.

Security The degree of resistance to, or protection from potential harm or vulnerabilities.

Vote The process of choosing candidates in an election. Voter A person who is eligible to vote in an election.

(23)

2.4 Elections in Nigeria and Challenges

Nigeria is currently a democratic nation in Africa where candidates are voted into offices to represent their people. Nigeria operates a-three-tiers of government: the Local, State and the Federal [ 45]. Elections take place in each tier of government to elect councillors, chairmen, and members of the state house of assembly, state governors, and members of the federal

house of representative, the senates and the president [ 45]. All these divisions are based on

effort to ensure that democracy or government is felt at all levels and to assist in managing the nation effectively. All elections are conducted and monitored by a body called the INEC which was established by the 1999 constitution [ 45]. The values, mission and vision statements of INEC can be found at [ 45]. The mission of INEC is to serve as an independent and effective Election Management Bodies (EMB) committed to the conduct of free, fair and credible elections for sustainable democracy in Nigeria. The vision is to be one of the best EMB in the world that meets the aspirations of the Nigerian people and they are guided by

the following values:

Autonomy: INEC shall carry out all its functions independently, free from external control and influence.

Transparency: INEC shall display openness and transparency in all its activities and in its

relationship with all stakeholders.

Integrity: INEC shall maintain truthfulness and honesty in all its dealings at all times

Credibility: INEC shall ensure that no action or activity is taken in support of any candidate

or political party.

Impartiality: INEC shall ensure the creation of a level playing field for all political actors.

Dedication: INEC shall be committed to providing quality electoral services efficiently and

effectively, guided by best international practice and standards

Equity: INEC shall ensure fairness and justice in dealing with all stakeholders.

Excellence: INEC shall be committed to the promotion of merit and professionalism as the

basis for all its actions.

Team work: INEC shall create a conducive environment that promotes teamwork among its

(24)

INEC have the sole responsibilities of registering voters, political parties, and conduction of elections and so on.

The Challenges of Electoral Processes in Nigeria according to [50] includes the following: ethnicity and communal tensions, ethnic and religious divides, injustice and culture of impunity, economic vulnerabilities, confidence in electoral tribunal, no internal party democracy, electoral malpractice, media inflammatory statement as well as power sharing problem. These, for some time now have continued to affect the electoral processes in Nigeria and currently, are still eating deep into the electoral system. Most Nigerians today, vote for candidates from their ethnic group while some look at religious affiliations. There northerners will prefer to vote for a candidate from the northern part of the country while southerners will equally bring in such sentiments. Whenever candidates and any political parties are identified with a particular region, state or religious group, sentiment will be brought into play. This was found in the case of 2011 elections in Nigeria where all voters from the south rallied round to ensure that Goodluck Jonathan was voted in as the next president of Nigeria. Thus, electoral violence is triggered by individuals and political groups to use all available means, like the use of violence to defend their communal honour

Electoral malpractices have also been seen as a serious challenge to Nigeria electoral processes.

Insecurity in Nigeria cannot be left out as one of the challenges in the electoral processes. According to [ 50] political assassination, intimidation, ballot snatching, arson, thuggery, harassment of election observers, militancy and insurgency has always been responsible for why a very successful electoral processes cannot be completed in Nigeria.

According to [50], cases of inter-communal clashes, political upnsmgs, sharia-induced unrests and economic sabotage, ethno-based organizations from Afenifere, Arewa, Ohaneze, Ijaw union became pronounced and that the 2007 election registered repeated violence. The level of killing and intimidation was high. Candidate for any political office should be ready for assassination and other unprecedented violence, since Nigeria's April 2007elections, every other election held in Nigeria records violence, assassination, intimidation as well as kidnapping.

(25)

For some years now, the democratic governments in Nigeria have been in principle considered as stable. A number of people in all the tiers of government have voted and been

voted into offices and replaced as well. However, the process involved in the elections has

been questionable and undemocratic. Elections are often marred by insecurity mostly at the

North East and South South zones of Nigeria, violence, malpractices and voter fraud such as

impersonation, rigging and hijacking. These threaten the democratic system, the election

integrity and the voters' confidence are being degraded. This is made possible because there

are no strong security measures in place to ensure that elections are free and fair. In this case, it is important that strong security measures supported by ICT tool are in place. The section that follows discusses the election process in Nigeria.

2.4.1 Election Process in Nigeria

In Nigeria, the election system assumed the traditional-based approach enhanced with little

ICT tools. E-democracy and ICTs application is limited to the communication of information

and services, participation promotion and voters registration. In this case, much has not been

done to protect the integrity of the elections especially during voting. In the Nigerian

elections point of view, though it employs some ICT in its electoral process of voter

registration, relevant tools such as dedicated bandwidth for high speed internet access, very fast and accurate verification with latent print image removal that can encrypt fingerprint

templates with a multi device connection reader and has high performance with optical

fingerprint sensor. A very high configuration computer system with not less than 4 gigabytes

RAM (Random Access Memory) with Hard Disk Driver (HDD) in terabytes and not less than

3.0GHz speed are not effectively employed during the voters registration and voting which aim at ensuring the integrity of the process. This makes it cumbersome to keep voter fraud and malpractice under control. in essence, protecting elections' integrity is to ensure one registration and one vote for every eligible citizen. The process involved in elections starts

from voter registration, voting on Election Day, collation and counting of results, decision on

the out is made and finally, a winner is declared or announced.

2.4.1.1 Voter Registration

Voter registration is the first step towards achieving successful elections. In this process,

eligible voters' information is captured for statistical purposes and for identification during

(26)

the time of registration and must have registered in the constituency where he intends to vote as well as possesses a registration card obtained during voters' registration.

To enhance the authenticity of voters in an attempt to promote one man one vote, recently, the biometric-based registration has been introduced by INEC. Though, the fingerprints of potential voters are captured, it still does not protect the integrity of the election for the reason that the data bases where the finger prints are held are fragmented and are not used on a real-time mode, thereby giving some people the opportunity to register more than once in different places [27, 46]. We therefore consider this approach a weak security measure to protect the integrity of the elections. Given this critical context, it is important that strong security measures be in place in the form of real-time biometric-based identity management in order to ensure that each potential voter registers only once irrespective of their location.

Name Date_of_Birth id_number FINGERPRINT Sex Town Ward Local_ Government State_ Origin Database

INEC REGISTRATION CENTER

...

~,

~

INE~ration Staff Electorates

Figure 2.1 INEC Voters registration

As shown in Figure 2.1, as the electorate enters the registration centre, his/her information is captured by the registration officer and stored in the database that is local and not central. The information collected from each voter is shown in the database in Figure 2.1.

2.4.J.2 Actual Voting

Voting is what is done on the Election Day aimed at choosing candidates the electorate wants to represent them. However, for the elections to be transparent and results generally accepted,

(27)

every voter has to be identified as a valid voter. One effective approach is the

biometric-based approach. Unfortunately, in Nigeria today, though biometrics is collected during voter

registration, the biometrics information is not used for identification and authentication on Election Day [27]. Instead, the manual approach is used where a voter is identified against their names on the voters' roll. In this case, a voter is allowed to vote if his/her name is found,

otherwise, not allowed. In this study, we consider this approach a weak security measure as it is difficult to ensure one man one vote. The process involve in the actual voting during elections is captured in Figure 2.1.

As shown in Figure 2.2, the diagram depicts the voting process during elections in Nigeria.

As the electorate enters the voting station, he/she is identified first. If he is a registered voter,

he/she is given a ballot paper and allowed to vote, otherwise, sent away or apprehended by security bodies.

In this study, we aim to curb voter fraud in the form of multiple registrations, multiple voting,

vote count falsification and impersonation as they are common-place in Nigerian elections.

The trustworthiness and the integrity of the elections will continue to be compromised if

electoral system is surrounded by irregularities and democracy will fail [3, 12, 41). Thus, an approach that increases the trustworthiness, is cost-effective and is reusable is needed to protect the integrity of Nigerian elections.

Bal.ot Paper Voter ld•ntifi..ci V rs' Id ritlfic;;, ion Dosk (Voter's Card) Vourtob• ld•ntified C VOTI G STATI Voter coi n,g to 11<ot.•

Figure 2.2 INEC Voting process

,

1'4

Vlt

tJ _

I

l

LJBRARYJ

~ - - - ~

Br,llol Box After Votin,g INEC Auditoi-s

(28)

The new approach is not the introduction of e-voting into the system but the enhancement or the re-engineering of the traditional approach using biometric-based identification. The justification for not e-voting is due to the series of irregularities that have been reported about

e-voting such as software flaws, problematic voting machines and so on. We believe that

Nigeria is technologically less advanced and will have difficulty to manage such risks

associated withe-voting. 2.5 Identity Management

With the growth in economic importance of the Internet, activities that are performed online

are increasing in prominence just as is the case with other online activities [13, 28]. In recent

years, online identity management issues have gained considerable prominence. While in real-life or physical activity a person can have one name to identify him/her, online. An individual may assume a variety of identities. In this case, different identities can be assumed

by a single person for different services [13]. This however becomes a problem in identifying uniquely a person who is responsible for certain actions. That is, several identities can be

controlled by a single human being. One effective solution is the proper management of people's identities.

The uniqueness and distinctness of an individual can be established using individual identity

feature [30, 31]. This is determined by the features owned by the individual which are temporary or permanent [31]. The concept of identity management (IDm) is basically the

practice of managing the essential features that individuals can be identified uniquely within

a given context [28]. IDm comprises of all the processes, systems and technologies that are

employed in the management and control of entities' identities either in computer or related devices [31]. IDm can take different forms such as access control authorization, through

authentication, entity genuineness verification, and so on [29]. Today several tokens exist that can be used for individual identity such addresses, mobile phone numbers, date of births, driving license, social security numbers and so on [29]. The tokens used for identity can be used for verification activities in traveling, receiving, sending, banking, insurance etc.

(29)

Though the management of identities on the Internet has attracted considerable attention in

recent years, there are still issues of identity in the real-life situation. One area is the

management of identities in the electoral system practiced by developing countries like

Nigeria. This is a serious issue which has threatened the nation's democracy, reducing people's confidence and trust on elections and their results. Electorates are not effectively identified thereby providing avenue for frauds, malpractices such as rigging, impersonation and ghost voting [29]. To solve this problem, biometric-based IDm is proposed. The justification for this choice is that biometrics has been known for its security and strength and can help to eliminate the threats and vulnerability posed by other forms of tokens like the

permanent voters card (PVC) which can be lost, stolen, forged or forgotten as they originates from authentication type dependent on what you know and what you have respectively.

2.6-Biometric System

Today, biometric-based IDm is increasingly being used by organizations, governments and other agencies for the purposes of verifying identity and enforcement of the law [32]. Biometrics offers a natural identity management tool that is characterized by greater security and convenience than traditional methods of personal recognition [33]. In biometrics,

identification or identity verification of a person is based on the physiological and behavioural characteristics of the person [19, 34]. This approach is found advantageous as every individual has unique traits that cannot be forged, stolen or lost. Biometrics focuses on unique characteristics that are measurable, hence consisting of properties including universality, uniqueness, permanence, collectability even though they all may be fit and convenient for a specific environment [32, 33]. Biometrics assures that each person has only one identity in the system and that only one person can access each identity [20]. Common features for individual biometrics used for recognition includes finger and palm friction ridges, iris, face, voice, handwriting, hand shape, DNA and hand vein patterns [33]. (see Figure 2.3)

(30)

Biometrics

Behavioral

Physiological

Signature

~

ft

_'

_{...___}

_D

_{_}

_NA

_{_}

_{_____,}

Voice

Hand

Keystroke

-

_Iris

Fingerprint

Face

Figure 2.3. Human physiological and behavioural traits

Biometrics technology exists today that is used to automate the recognition of individuals

using biological and behavioural traits as shown in Figure 2.3. The technology is utilized

today because of its ability to minimise errors made by users during the identification

process. It is also able to increase both user satisfaction and automatic control for security.

However, the main concern when using biometric technology for authenticating users is to provide storage for the captured biometric feature in a secured manner [34]. In addition,

while biometrics usage has been praised in tenns of efficiency and scalability, several

challenges in the form of loss of privacy, and the potential for fraud and abuse have been reported [32].

2.6.1 Properties of Biometrics

Desirable properties of human physiological and behavioural features chosen for authentication as biometrics are universality, uniqueness, permanence and collectability [32, 35, 48]. Other characteristics such as performance, acceptability and circumvention are regarded as the qualities attributed to biometric systems [35, 48]. The characteristics are discussed as follows:

(31)

1) Universality means that the characteristics should be owned by every individual. 2) Uniqueness means that every individual should have a different characteristic.

3) Permanence means that the characteristics should not show any variation that may be

caused by the ageing process.

4) Collectability means that the characteristic should be able to be captured and stored. 5) Performance means achievable and acceptable identification accuracy, resource

requirements and environmental factor that affect identification accuracy.

6) Acceptability means the degree to which people are willing to accept it as a

biomehics.

7) Circumvention means how easy it is to carry out fraudulent act on the system.

As shown in Figure 2.3, fingerprints have been one of the biomehics used for identification

and authentication for quite some time now. They are used for verification purposes wherein they are matched with the identity that the user is claiming to be. The factors that make fingerprint based systems to be more considerable for use is the fact that the devices used to capture the fingerprints are not expensive and are not frustrating since they are not huge, the computing hardware they use are fast and the rate and speed it takes to recognize is not too long in a way that a user can find it not usable. Based on this factor, in this study, fingerprints

will be used as an authentication factor.

2.6.2 Biometric Working Process

The process involved in the application of biometrics 1s twofold: enrolment and

authentication or verification as shown in Figure 2.4 [33, 35]. In the enrolment stage,

biomehics data is collected from persons or individuals and stored in the database together with their identity. The captured biomehics data is subject to processing in order to extract only the significant and distinctive features which are the only features to be stored and the

raw data are cast-off [33, 35]. In using biometrics systems, what is needed is a scanner, the

software that will be used in encoding the information about a user into digital data for comparison purposes and also an encrypted database that will be used for storing captured data. Identification process using this method is quick, harmless and eliminates identity theft

(32)

ENROLLMENT

Present

B1ome nc Capture Process

NO MATCH

Compare

AUTHENTICATION

MATCH

Capture Process

Figure 2.4 Biometric system

On the other hand, during the course of the authentication or verification, currently captured

biometric data collected from individuals are compared against the ones stored in the

database for authentication purposes. This is then captured as shown in Figure 2.4.

In general, biometrics has two applications: identification and authentication. In addition,

biometrics systems have heen used for managing attendance as well as identification. These

two forms of application directly corresponds to the security issues facing INEC today. In

this study, we aim at utilizing these capabilities to protect the integrity of our elections in

Nigeria by ensuring that one man has one registration and vote.

2.7 Fingerprint Identification System

In this study, we have adopted the use of fingerprints as an authentication factor to be used in

the process of curbing election malpractices such as impersonation and other forms of fraud.

Law enforcement agencies and others have successfully used fingerprint matching

technology for over a century now [19, 32, 42]. The scientific foundation for using

fingerprints as a method for identification for a person was documented by Henry Faulds,

Francis Galton, and Edward Henry, among others [ 42]. For the reason that all other biometric

traits as shown in Figure 2.3 do not exhibit the characteristic of extreme uniqueness

(33)

A fingerprint recognition system provides the two basic functionalities of verification and identification. In verification, an input fingerprint is compared to the "enrolled" fingerprint of a specific user to ascertain a match or mismatch (this is referred to as 1:1 match). In

identification, the system first maintains a database of fingerprints against which it compares an input fingerprint to determine a match (this is referred to as 1: N match). The process is

captured in Figure 2.4. The operational principle of an automated Fingerprint Identification System (FIS) is divided into three functional modules namely: sensor, feature extractor, and matcher.

Enrolment: As shown in Figure 2.5, during the enrolment phase, the sensor scans the user's

fingerprint as a user places his or her finger on the scanner and converts it into a digital rmage.

Minutia extractor: The image taken in the enrolment phase is processed by the minutiae

extractor to identify specific details known as minutia points that are used to distinguish

aifferent users. Locations where friction ridges end abruptly or where a ridge branches into

two or more ridges as shown in Figure 2.5 are known as minutia points [ 42, 43]. This minutiae information is stored along with the user's demographic information as a template

in the enrolment database.

Figure 2.5 Minutia template [43)

Matching: During the identification phase (matching), the user touches the same sensor,

generating a new fingerprint image called a query print. Minutiae are extracted as explained

earlier and the results are compared by the matcher module to the stored minutia templates in the enrolment database to find the number of common minutia points. In the process, the matcher is able to determine the number of pairs of matching minutiae that is two minutia points that have similar location and directions.

(34)

2.7.1 Fingerprint Matching Algorithms

In this section we discuss two main algorithms commonly used to recognize fingerprints:

minutia based and pattern based matching. The difference between these algorithms is based

on their method of evaluating fingerprint images [ 44].

Minutia Matching: In a fingerprint, there are a number of ridges and valleys - ridges are the

upper skin layer segments of the finger while valleys are the lower segments. The ridges

constitute what is called minutia points [ 43, 44]. The minutia matching algorithm simply

analyses the points at which the fingerprints split, intersect or end. Figure 2.4 represents the

three phases involved in the fingerprint identification process as discussed earlier using the

minutia extraction algorithm.

Pre-processing Minutia Extraction Template

(33, 11, 4, 1) (10011001)

(25, 18, 2, 0) (10101001)

..

(18, 13, 0, 1) (10010011)

(13, 27, 2, 0) 1111 (11010100)

(75, 4, 8, 0) (10100110)

Figure 2.6 Minutia Based Algorithm [ 44]

Pattern Matching: Pattern matching algorithm in contrast to minutia matching takes into

account the overall fingerprint characteristics, not just individual points. This makes it less

prone to difficulties of recognizing a finger with varying fingerprint quality [ 44]. Basically,

this algorithm identifies fingerprints through the graphical comparison of fingerprint image,

as shown in Figure 2.7.

(35)

Pre-processing Minutia Extraction Template (10011001)

..

(10101001) (10010011) (11010100) (10100110)

Figure 2.7 Pattern Based Algorithm [44]

2.7.2 Performance

A fingerprint matcher can make two types of errors: a false match and a false non-match [ 43, 44]. A false match is one in which the matcher declares a match between images from two different fingers. This type of error is sometimes statistically referred to as false acceptance rate (FAR) and a false non-match, in which it does not identify images from the same finger

as a match. This is also called false rejection rate (FRR).

The classification strength of a biometric algorithm is determined by these two statistical

measures. "A system's false match rate (FMR) and false non-match rate (FNMR) depend on

the operating threshold and are dependent operational threshold set [ 42]. In summarizing the

above challenge, [44] further maintained that FAR and FRR are diametrically opposed, suggesting that increasing FAR will lower FRR and vice-versa. FIS performance is measured

in terms of its false positive identification rate (FPIR) and false negative identification rate

(FNIR). A false positive identification occurs when the system finds a hit for a query

fin-gerprint that is not enrolled in the system. A false negative identification occurs when it finds

no hit or a wrong hit for a query fingerprint enrolled in the system. The relationship between

these rates is defined by FPIR = 1 - (1 - FMRt, where N is the number of users enrolled in

the system. Hence, as the number of enrolled users grows, the fingerprint matcher's FMR needs to be extremely low for the identification system to be effective [ 42].

(36)

2.8 Related Works

In this section, we provide several works that have been carried out which concerns the

integrity of election system and the applications of biometric system.

A study carried out by Jones [36] shows that the voting system provided by e-voting does not pledge that the vote is cast by an authentic person or one person, one vote. In addition, this type of situation can give way to foul practices in the electoral process like chain voting,

multiple voting, ghosts voting or impersonation. In another study by Simons [ 1 O], it was stressed that having a voting system like an e-voting system cannot bring about trustworthiness and the integrity of the elections can be threatened and compromised [26].

Also in a press release on e-voting [24], it was emphasized that the voting system can bring

about failures or corruption of democracy. Furthermore, it was stressed that if the democracy

is compromised by corrupt practices, the people inversely will also be corrupt. The study by

[36, 37, 38, 39] also reiterated that for any election to be completely conducted in a manner

that no room is made for doubt and opportunities for fraud and failures the entire systems has

to be made more transparent, trustworthy, accurate and reusable.

lL,:~i,J

In an attempt to reengineer the traditional electoral systems in Africa since e-voting cannot be practiced due to the risks associated with it, several approaches have been proposed. One of

such approach is the proposal by Isong et al [ 12, 41]. In their studies they proposed an

approach to ensure that one man has one registration and vote through real-time registration and identification of voters on Election Day. [12, 41] Proposed the use ofRFID-based Voters Card and biometrics in the South African elections.

In the area of biometrics, large scale biometric-based IDm system, the fingerprint in paiiicular, have found application in immigration control of several countries, vehicle licensing offices for vehicles and driver's license, ministry of transportation, social security

departments, health ministry, [19, 32].

However, this research aims at providing a solution to the elections problems in Nigeria,

especially in the aspect of registration and identification. In this case, the major goal is to design and implement system prototype that will assist in automating voters registration and

(37)

identification in a real-time mode to ensure that one person votes only once in an election and

is the authentic person without impersonation.

2.9 Chapter Summary

In this Chapter, we have discussed the various concepts and what has been done so far in the literature. We have presented the election system in Nigeria and the challenges thereof. We have also discussed biometrics system and have chosen fingerprints to be used as an authentication factor. Also presented are related studies, key terminologies and identity management issues. In the next Chapter, the analysis and design of the system will be discussed.

(38)

Chapter 3:

Biometric-based

Framework

Analysis and Design

3.0 Chapter Overview

This Chapter focuses on the framework analysis and design of the system which is based on biometric system. It starts with an introduction and progresses to a justification of using

fingerprint as authentication factor, requirements analysis and the design which includes the

components identification and architecture. The objective here is to analyse and design an approach that will be used to meet the overall goal of this research.

An election is considered free and fair if the electorate has confidence in it, the process is

transparent, trustworthy and the result is generally acceptable as the people's will. This is the

situation we want the election system in Nigeria to be. As discussed in Chapter 2, currently the election process in Nigeria does not pledge that the person who cast the vote is an authentic voter. This is due to the weak security measures that are in place which in turns have given room for impersonation, rigging and other forms of malpractice.

Therefore, this Chapter 3 will specify and design an approach to assist INEC in combating these threats that are gradually eating up Nigeria nascent democracy. In this case, we have

chosen the use of biometrics, fingerprints in particular, to help protect the integrity of the

system ranging from voter registration to actual voting in the capacity of identification and

authentication. The intuitive belief is that if voters are in real-time identified in both

registration and actual voting, it will go a long way to uphold elections integrity, not only in Nigeria but in all nations where it is adopted and as well boost citizens confidence in the election system.

3.2 Justification for Fingerprint Identification System

In today's e-society, identification and authentication problems have been considered

challenging [ 47]. As we have witnessed in a world today that is complex, increasingly interconnected electronically and geographically mobile, identifying individuals accurately is

increasingly becoming important in the same way it becomes difficult to identify an individual [ 47]. However, the importance associated with accurate identification is that it

(39)

could work as a deterrence in crime commission and fraud, protect resource that are critical and so on [ 47].

Of the several approaches to curb this situation the biometrics-based approach is considered the best. There are several biometric systems with each having its own strengths and weaknesses in terms of security, cost, usage, etc. [ 4 7, 48]. Among the different biometric traits captured in Figure 2.3 of Chapter 2, this study has adopted the fingerprint as a feature to be used in identifying individual as authentic voters during elections. The justification for this

choice is its desirable properties and mode of operation.

3.2.1 Desirable properties of Biometrics

We chose the fingerprint to be used in this research based on the satisfaction of the essential properties of any human physiological and behavioural features to be consider as biometric system as discussed in Section 2.6.1 of Chapter 2. We used the assessment scale of low,

medium and high to evaluate each known biometrics system.

Table 3.1: Biometric systems evaluation

Biometric Univer- Unique- Perma- Collec- Perfor- Accep- Circum-sality ness nence tability mance tability vention Human Medium Low Low medium Low High Low

Voice

Finger Medium High High Medium High Medium High

print

Face High Low Medium High Low High Low

Iris High High High Medium High Low High

DNA High High High Low High Low Low

Signature Low Low Low High Low High Low

Hand Medium Medium Medium High Medium Medium Medium

Geometry

As shown on Table 3.1, we have chosen fingerprint based on its strengths in satisfying the desirable properties. This is in line with the evaluation studies carried out by [ 4 7, 49]. With these strengths, we believed it will be advantageous to employ it in the real-time identification of voters.

3.2.2 Mode of Operations

Another factor that prompts our choice of fingerprint is due to the mode of its operations

which is also applicable to other biometric systems. These are the enrolment and the verification or authentication process. Firstly, the process of enrolment corresponds to the