International Journal of Computer Science Engineering

and Information Technology Research (IJCSEITR)

ISSN(P): 2249-6831; ISSN(E): 2249-7943
Vol. 4, Issue 5, Oct 2014, 49-56
TJPRC Pvt. Ltd.

A UNIQUE BIOMETRIC COMBINATION OF IRIS AND TEXTUAL

PASSWORD USING ICA
SHALLY1 & GAGANGEET SINGH AUJLA2
1
2

Research Fellow, Chandigarh Engineering College, Mohali, Punjab, India

Assistant Professor, Chandigarh Engineering College, Mohali, Punjab, India

ABSTRACT
Biometric recognition refers to an automated identification of individuals based on a feature vector(s) extracted
from their physiological or behavioral trait. A multimodal biometric authentication system can be taken as a traditional
information fusion so that we can improve the overall decision accuracy for the system. Those biometric authentication
system that use more than one physiological trait for enrolment or identification in applications such as entry/exit on the
border, ATM or access control, multi-modal biometric systems are looked reducing false acceptance and false rejection
rates, presenting an ancillary means of enrolment, and identification if adequate data cannot be obtained from a given
biometric specimen and disputing attacks to fool the biometric systems through counterfeit data sources such as synthetic
iris images. In this research paper, proposed model provides high security in authentication which protects service user
from unauthorized access. In this proposed model, user is required to authenticate himself with biometric identification
(Iris recognition) and Personal Identification Number (PIN). This model reduces complexity with authentication as
authentication is always with you with high security. It also saves time and efforts compared with card based ATMs and
also saves environmental pollution problem of excess number of plastic cards.

KEYWORDS: Biometrics, Multimodal Biometrics, Iris Recognition, ICA, ATM, High Security, Authentication
INTRODUCTION
The authentication system based on IRIS recognition is reputed to be the most accurate among all biometrics
methods because of its acceptance, reliability and accuracy. The iris is a well-protected organ that is externally visible and
whose epigenetic patterns are very unique and remain stable throughout most of a persons life. Its high uniqueness and
stability makes it a good biometrics that can be used for identifying individuals. These unique patterns can be extracted
using image processing techniques employed on a digitized image of the eye and then the results can be encoded into a
biometric template which can later be stored in a database for future comparisons. The biometric template is usually
created using some sort of mathematical operations. If an individual wants to be identified by the system, then first a
digitized image of their eye is first produced, and then a biometric template is created for their iris region. This biometric
template is compared with all the other pre-existing templates in the database using certain matching algorithms in order to
get the identification of the individual.
Strengths of Iris Biometrics:

Most accurate technology (Gregory & Simon, 2008; Deriche, 2008; Nanavati, et al., 2002).

www.tjprc.org

editor@tjprc.org

50

Shally & Gagangeet Singh Aujla

Iris biometrics is not intrusive and is hygienic (no physical contact) (Deriche, 2008; Biometrics: Newsportal.com,
2009).

Iris biometrics has low false acceptance rate (Jain, A.K. et al., 2006).

Iris biometrics has promising processing speed (Jain, A.K. et al., 2006).

The patterns of iris remain stable for all the life (Nanavati, et al., 2002; Iridian technologies, 2009).
Table 1: Comparison of Various Biometric Technologies

Limitations of Biometric Systems Using Any Single Biometric Characteristic

Noise in Sensed Data: Noisy data can also result from accumulation of dirt on a sensor or from ambient
conditions.
Intra-Class Variations: Biometric data obtained from a person during authentication may be dissimilar from the
data which was adapted to create the pattern during enrolment. This difference is usually caused by a user who is
fallaciously interacting with the sensor.
Distinctiveness: While a biometric feature is expected to diversify substantially across individuals, there may be
considerable inter-class comparisons in the feature sets used to represent these attributes. This restriction limits the
discriminability equipped by the biometric feature.
Non-Universality: While every user is awaited to have the biometric feature being obtained, in authenticity it is
possible that a group of users do not posses that particular biometric characteristic.
Spoof Attacks: An individual may attempt to forge the biometric trait. This is particularly easy when signature
and voice are used as an identifier.
Therefore, the single biometric technology i.e., IRIS recognition is not enough for high security human
identification as these recognition systems can be fooled by screening or modifying the illusion of the iris. Criminals, to
avoid detection can also wear a cover that hides their implicate iris. Identically, someone who wants to detach an individual
at a security place or point, can put on a forged iris that depicts the template on the actual trusted individual. As Dr. Flom,
the principal investigator of biometrics at the New York University School of Medicine, introduced recognition of iris in
1987 and said that we can use any digital camera as well as an iPhone, an iPad, and even Google Glasses by adding a zoom
lens which grants ingenuity of palm prints, fingerprints, faces and even Iris patterns without actually coming into contact
with the person which concludes that if the iris image clicked by any ipad, iphone, google glasses or digital camera is
stolen by unauthorized person then he can use his iphone or ipad consisting of the iris image for the false identification in
ATM.
Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0

51

A Unique Biometric Combination of Iris and Textual Password Using ICA

Multimodal Biometric Systems

A multimodal biometric authentication system can be examined as a traditional information fusion. We can be
combine evidence provided by different biometrics system so that we can improve the overall decision accuracy for the
system. Those biometric authentication system that use more than one physiological trait for enrolment or identification in
applications such as entry/exit on the border, ATM or access control, multi-modal biometric systems are looked reducing
false acceptance and false rejection rates, presenting an ancillary means of enrolment, and identification if adequate data
cannot be obtained from a given biometric specimen and disputing attacks to fool the biometric systems through
counterfeit data sources such as synthetic iris images.

PROPOSED SYSTEM
A multimodal biometric system, based on Iris characteristics and textual password, is proposed. The proposed
multimodal biometric system is composed of two main stages: the pre-processing stage and the matching stage.
Image Pre-processing
The iris is a part between the pupil which is the inner boundary and the sclera which is the outer boundary of the
eye. So, a grabbed image cannot be supposed to have only the iris part but it also carry some non-useful part, that is, pupil,
sclera, and pupil, so the iris part should be located in grabbed eye image, and normalized to polar array.
Iris Localisation
Iris localization by definition means to detach the factual iris region in a digital eye template by recognizing the
inner and outer boundary of the iris. The eyelids and eyelashes conventionally occlude the higher and lower portions of the
iris area. To detect the iris and pupil boundaries, Hough transform is used by involving canny edge detection to create an
edge map.
Iris Normalisation
Normalization refers to preparing a localised image of iris for the feature extrication process. This process of
normalization involves unbinding the image of iris and altering it into its polar equivalent. It is carried out by using
Daugmans Rubber sheet model shown in figure 1. The centre of the pupil is considered as the reference point and a
remapping formula is used to convert the points on the Cartesian scale to the polar scale. The remapping of iris image I
(x, y) from raw Cartesian coordinates to polar coordinates (r,) can be represented as
I (x(r,), y(r,)) I (r,)

(1)

Figure 1: Daugmans Rubber Sheet Model

www.tjprc.org

editor@tjprc.org

52

Shally & Gagangeet Singh Aujla

Features Extraction by ICA

Introduction
When the independence adoption is correct, blind ICA separation of a hybrid signal presents truly better results.

It is also used for signals that are not assumed to be created by a mixing for investigation purposes. A clear application of
ICA is the cocktail party problem, where the underlying speech signals are distinguished from a specimen data having
people talking concurrently in a room. Generally the problem is clarified by adopting no time delays or echoes.
A significant note to take it into consideration is that if N sources are existing, at least N examinations (e.g. microphones)
are needed to get the real signals. This constitutes the square (J = D, where D is the input dimension of the data and J is the
dimension of the model). Other cases of underresolved (J < D) and over determined (J > D) have been investigated.

Defining Component Independence

ICA finds the independent factors (latent variables/sources) by broadening the statistical independence of the

determined components. We may choose any way define independence, and this preference rules the form of the ICA
algorithms. The two broadest illustrations of independent component analysis are

Minimization of Mutual Information

Maximization of non-Gaussianity
The Non-Gaussianity family of ICA algorithms, prompted by the central limit theorem, uses kurtosis and

negentropy. The (MMI) Minimization-of-Mutual information family of ICA algorithms employs measures like KullbackLeibler Divergence and maximum-entropy. The standard algorithms for ICA use centering, whitening and dimensionality
reduction as preprocessing steps in order to clarify and diminish the complication of the problem for the real iterative
algorithm. Whitening and dimension compression can be carried out with PCA. Whitening assures that all dimensions are
handled equally a priori before the algorithm is employed. Algorithms for ICA include JADE, infomax, Fast ICA, and
many others also.
Generally, ICA cannot categorize the existent number of source signals, a solitarily accurate ordering of the
source signals, nor the appropriate scaling of the source signals. ICA is signficant to blind signal separation and has many
functional applications. It is nearly assosciated the search for a factorial code of the data, i.e., a new vector-valued
depiction of each data vector such that it gets solitarily encoded by the resulting code vector (loss-free coding), but the
code components are statistically independent.

Mathematical Definitions
Linear ICA can be split into noiseless and noisy cases, where noiseless ICA is a especial case of noisy ICA.

Nonlinear ICA should be deliberated as a different case.

General definition
The data is represented by the random vector x=(x1,xm) and the components as the random vector

s=(s1,sm). The task is to transform the observed data x, using a linear static transformation W as
S = Wx,

Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0

53

A Unique Biometric Combination of Iris and Textual Password Using ICA

into maximally independent components s measured by some function F(s1,sn) of independence.

Password Verification
In this section, password is set to the uploaded iris template values. Each password has set value which is then
stored in the database. When iris image is uploaded, its value which was set to the password is matched. If the filled
password matches the actual password of the value, then the user is authenticated. Matching is done using Sum rule-based
fusion.
Matching
The procedure for sum rule-based fusion is stated in the following. After we get a set of normalized scores
(x1,., xm) from a particular person(here the index i=1,.,m indicates the biometric matcher), the fused score fs is
evaluated using the formula fs =w1x1+...+wmxm
The notation wi stands for the weight which is assigned to the matcheri, for i=1 m. There are many choice of
how to calculate these weights based on some preliminary results, we decided to use equal weights in our experiments.
Then simplified to fs = x1+...+xm
In the next step, the fused score fs will be compared to a pre- specified threshold t. We declare a person to be
genuine if fs t, otherwise, we declare him/her an impostor.

EXPERIMENTAL RESULTS
The three algorithms have been tested on 10-20 images. Uploading iris image size is around 10-20 KB.
Table 2: Result Table
Dataset
IRIS
CASIA Data

PCA
78%
77%

ICA
81%
83%

ICA + Password
99.4%
93%

The above table in comparison of accuracy proves that the textual password fused with the ICA is more accurate
and efficient system in the Iris Recognition than the Single Biometric System ICA based Iris Recognition.

Figure 2: Graph Showing the Comparison of Iris Recognition Using PCA, ICA and ICA with Textual Password
The proposed system has been analysed by matching iris images with correct passwords and incorrect passwords
and the accuracy of the system is calculated by False Acceptance Rate and False Rejection Rate.

www.tjprc.org

editor@tjprc.org

54

Shally & Gagangeet Singh Aujla

False Acceptance Rate (FAR): It is the possibiity that the system inaccurately matches the input template to a
non-matching pattern in the database. It measures the percentage of invalid inputs which are inaccurately accepted.
False Rejection Rate (FRR): It is the posibility that the system fails to reognize a match between the input
template and a matching pattern in the database. It measures the percentage of valid inputs which are inaccurately rejected.
Table 3: Accuracy Table
Image
Name
A
B
C
D
E
F
G
H
I
J

Database
Password
ASHALLY
BSHALLY
CSHALLY
DSHALLY
ESHALLY
FSHALLY
GSHALLY
HSHALLY
ISHALLY
JSHALLY

Filled
Password
ASHALLY
ASHALLY
CSHALLY
CSHALLY
ESHALLY
ESHALLY
GSHALLY
GSHALLY
ISHALLY
JSHALLY

Matching

FAR

FRR

1
0
0
1
1
1
0
0
1
0

0
0
0
1
0
1
0
0
1
0

0
0
1
0
0
0
1
0
0
1

Accuracy=100(FAR+FRR)
100
100
99
99
100
99
99
100
99
99
99.4

Figure 3: The Graph Shows the Accuracy Rate of the Proposed System
We have filled the correct passwords and incorrect passwords to ICA data of 10 iris images to check the accuracy
of proposed system by FAR and FRR and the obtained accuracy of the proposed technique is 99.4%.

CONCLUSIONS AND FUTURE SCOPE

In the proposed system, a new multimodal biometric technique is generated at the matching-score level which is
based on the combination of matching scores after separate feature extraction and comparison between stored data and test
data for each subsystem have been compiled. Therefore, an overall matching score is generated using linear or non-linear
weighting for iris verification system to increase the accuracy of the authentication systems. In this ICA features are
extracted for iris, then password based method is used for verification after that sum rule has been applied for matching of
iris pattern. This proposed method decreased the FAR as well as FRR, & has increases the system performance on the
given data set. The accuracy of given system is 99.4%. Therefore, it will be very difficult for an attacker to gain the iris
feature and user password at the same time. Future works could go in the direction of using more robust modelling
techniques against forgeries and hybrid fusion level can be used. Multimodal modalities can be used together to make
forgeries more difficult. Also, the system should be tested on a larger database to validate the robustness of the model.

Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0

55

A Unique Biometric Combination of Iris and Textual Password Using ICA

REFERENCES
1.

Kwanghyuk Bae, Seungin Noh, and Jaihie Kim, Iris Feature Extraction Using Independent Component
Analysis, Springer 2003, Verlag Berlin Heidelberg, AVBPA 2003, LNCS 2688, pp. 838844.

2.

Seung-In Noh, Kwanghyuk Bae, Kang Ryoung Park, and Jaihie Kim, A New Feature Extraction Method Using
the ICA Filters for Iris Recognition System, Springer-Verlag Berlin Heidelberg 2005, IWBRS 2005,
LNCS 3781, pp. 142149.

3.

Jong Hyun Park and Moon Gi Kang, Iris Recognition Against Counterfeit Attack Using Gradient Based Fusion
of Multi-spectral Images, Springer-Verlag Berlin Heidelberg 2005 IWBRS 2005, LNCS 3781, pp. 150156.

4.

Jiali Cui, Tieniu Tan, Xinwen Hou, Yunhong Wang, and Zhuoshi Wei, An Iris Detection Method Based on
Structure Information, Springer-Verlag Berlin Heidelberg 2005, IWBRS 2005, LNCS 3781, pp. 157164.

5.

Topi Maenpa, An Iterative Algorithm for Fast Iris Detection, Springer-Verlag Berlin Heidelberg 2005,
IWBRS 2005, LNCS 3781, pp. 127134.

6.

Xiaoyan Yuan and Pengfei Shi, A Non-linear Normalization Model for Iris Recognition, Springer-Verlag Berlin
Heidelberg 2005, IWBRS 2005, LNCS 3781, pp. 135141.

7.

Yong Wang, Jiu-Qiang Han, Iris Recognition Using Independent Component Analysis, Proceedings of the
Fourth IEEE International Conference on Machine Learning and Cybernetics, Guangzhou, 18-21 August 2005,
pp. 4487-4492.

8.

Kousar Perveen khushk and Ahmad Ali Iqbal, An Overview of Leading Biometric Technologies Used for
Human Identity, IEEE 2005, pp. 832-835.

9.

Sanjay N Talbar, Rajesh M Bodade, Feature Extraction of Iris Images Using ICA for Person Authentication,
IEEE International Conference on Signal Processing and Communications (ICSPC 2007), 24-27 November 2007,
Dubai, United Arab Emirates, pp. 1055-1058.

10. Bo Lu, Jing-jing Wu and Yu Wang, A Iris Recognition Algorithm Based on ICA and SOM Neural Network,
3rd IEEE International Congress on Image and Signal Processing (CISP2010), pp. 2445-2448.
11. Vincenzo Conti, Carmelo Militello, Filippo Sorbello, A Frequency-based Approach for Features Fusion in
Fingerprint and Iris Multimodal Biometric Identification Systems, IEEE Transactions on Systems, Man,
and Cybernetics, Vol. 40, No. 4, July 2010, pp. 384-395.
12. Nathan D. Kalka, Jinyu Zuo, Natalia A. Schmid, Estimating and Fusing Quality Factors for Iris Biometric
Images, Transactions on Systems, Man, And Cybernetics, Vol. 40, No. 3, May 2010, pp. 509-524.
13. Imen Bouraoui and Salim Chitroub, Performance Evaluation of Independent Component Analysis in an Iris
Recognition System, IEEE 2010, pp. 1779-1784.
14. Naresh Babu N T, Vaidehi V, Fuzzy Based IRIS Recognition System (FIRS) For Person Identification,
IEEE-International Conference on Recent Trends in Information Technology, ICRTIT 2011 MIT,
Anna University, Chennai. June 3-5, 2011, pp.1005-1010.

www.tjprc.org

editor@tjprc.org

56

Shally & Gagangeet Singh Aujla

15. Kai Yang, Eliza Yingzi Du, Consent Biometrics, IEEE 2011
16. Lunman Deng and Song, Jeong-Young, A Iris Recognition based on Geometric Information, 2011 International
Conference on Network Computing and Information Security, pp. 381-384.
17. Aissa Boukhari, Salim Chitroub, Imen Bouraoui, Biometric Signature of Private Key by Reliable Iris
Recognition Based on Flexible-ICA Algorithm, Int. J. Communications, Network and System Sciences, 2011, 4,
pp. 778-789
18. Pankaj P. Chitte, J. G. Rana, Sachin Taware, Advanced Security System using RFID and IRIS Recognition
System using ICA, PCA, Daugmans Rubber Sheet Model Together, International Journal of Computer
Applications (0975 888) Volume 48 No.13, June 2012, pp. 5-11.
19. Kshamaraj Gulmire, Sanjay Ganorkar, Iris Recognition Using Independent Component Analysis,
IJETAE Volume 2, Issue 7, July 2012, pp. 433 437.
20. K. Laxmi Narshima Rao, Vikram. Kulkarni, C. Krishna Reddy, Recognition Technique for ATM based on IRIS
Technology IJERD Volume 3, Issue 11 (September 2012), pp. 39-45
21. Namit Gupta, Anu Sharma, Review of Biometric Technologies Used for ATM Security, IJEIT Volume 3, Issue
2, August 2013, pp. 460 465.
22. Thiyam Churjit Meetei, Shahin Ara Begum, Performance Evaluation of Feature Selection Methods for ANN
Based Iris Recognition IEEE 2013, pp. 208-213.
23. http://www.afcea.org/content/?q=node/11607

Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0