International Engineering Journal For Research & Development

E-ISSN No: 2349-0721

Volume 1: Isuue 1
www.iejrd.in Page 1


MULTIMODAL BIOMETRICS SYSTEM- A REVIEW
Ms. Priya N. Ghotkar
1
,Prof. Vikas G. Bhowate
2

1
(PG Scholar) Department of Computer Engineering,
2
St. Vincent Pallotti College of Engg. & Tech
Nagpur,
1
vipbhowate@gmail.com,
2
priyaghotkar@gmail.com
----------------------------------------------------------------------------------------------------------------
Abstract:
In our day to day life, the automatic verification of person is a very important task. The traditional
method of establishing a persons identity include knowledge based like password or token based like
identification cards, but demonstration of these characteristics can simply be lost, stolen or shared. So for
authentication of a person some biometric characters of that person is used. For that purpose one or more
biometric characters can be used. But using one character may sometime prove less secure so more than one
characters are used. Therefore this is the survey on multimodal biometrics system.
Keywords: Biometrics, Person authentication, Unimodal, Multimodal, Fusion.
1. Introduction
In todays world where the technology is emerging rapidly, there are several person authentication
related issues that need to be handled in daily life. Biometric is the Greek word in which bios (life) and metron
(measure), and hence biological measurement is termed as biometric. It indicates to the ones physical or
biological characteristics in which physiological characters are face, speech, fingerprint, iris, etc or behavioural
characters are signature, gait or speech too. Physiological biometrics are related to the shape of the body and are
generally more stable. Behavioural biometrics are related to the behaviour of the person and are not as much of
steady. thus, a appropriate variety of biometrics is more important than building a person authentication system.
The block diagram of biometric based person recognition system is given in Figure 1.


Figure 1: Biometric-based person recognition system.[1]
International Engineering Journal For Research & Development
E-ISSN No: 2349-0721
Volume 1: Isuue 1
www.iejrd.in Page 2

For training any person validation arrangement, the biometric data are processed for attribute
extraction. The most important aspect of any biometric-based authentication system is the selection of an proper
feature set which should be reasonably invariant with different degradations. Modelling is done to make the
template in such a way that it should hold all the variations captured by that particular biometric for every
person. In the testing phase, same features are computed from the unknown test biometric template and then
compared with the models of every individual. This is proficient in the blueprint matching stage. Finally, after
blueprint matching, we get the approval or refusal of the person as the output result.
2. Comparision between UniModal Biometric System and Multimodal biometric system
2.1 Unimodal biometric system
Most real-world applications in biometric systems are unimodal, i.e., they rely on the single source of
information for authentication (e.g., single fingerprint or iris). The basic requirements of any biometric system
are input device which captures the image, biometric software for feature extraction and database for storage
and comparision.
A. Input Device
An input device such as scanner, writing pad, etc. are used to capture the images. These are then used to control
the inputs used by the software part.
B. Biometric Software
Software is used to process the input and convert it into digital form, also extracts the features and compares the
result. In terms of accuracy, the performance of a biometric system depends on the quality of the software.
C. Database
A database is used to store the information which is further used for the judgment. characteristics extracted from
the put in samples are stored since the input data will take more space and storing features saves the time for
processing samples again to extract the features.
The main processes involved are enrolment or registration, verification and identification. Fig. 2 shows
the architecture of a general biometric system.
1.Enrollment Process/Registration Process
In this process, the user need to register himself by using the input devices like scanner and writing
pads. It is necessary for identification of an individual to store his characteristics features in a record, which
been extracted as of the trustworthy samples.

International Engineering Journal For Research & Development
E-ISSN No: 2349-0721
Volume 1: Isuue 1
www.iejrd.in Page 3

Fig. 2: Basic Biometric System Architecture[2]
These features are then compared with the features extracted from the traits of the individual need to be
recognized. In command to remove characteristics, the input sample is pre-processed and feature extraction
algorithm is applied on these pre-processed samples to form feature vectors. Feature vectors are stored in the
database instead of input samples as input samples take more space on secondary memory than mathematical
data and also features are computed just once which saves a lot of processing time...
2. Verification
In the verification mode, the person is validated by comparing his captured data with the stored data or
templates in the database. In such a system, an individual who desires to be recognized claims an identity,
usually via PIN (Personal Identification number), a user name or a smart card and the system conducts a one to
one comparison to determine whether the claim is true or not [1].
3. Identification
In identification mode, the system recognizes an individual by searching the templates of all the users
in the record for a equivalent. so, the structure conducts a one to many comparisons to establish an individual
identity or fails, if the subject is enrolled in the system record. The topic does not claim any identity.
These formations have a variety of issues such as:
(a) Noise in sensed data: examples of noisy data are a fingerprint with a scar, or a voice altered by cold. It
affects the performance of the system by getting incorrectly matched with the templates in the database.
(b) Intra-class variations: These variations are typically caused by a user who is incorrectly interacting with the
sensor (e.g., incorrect facial pose), or when the characteristics of a sensor are modified during authentication
(e.g., optical versus solid-state fingerprint sensors).
(c) Inter-class similarities: In a biometric system comprising of a large number of users, there may be inter-class
similarities (overlap) in the feature space of many participants. Golfarelli et al. [2] declare that the number of
distinguishable patterns in two of the most commonly used representations of hand geometry and face are only
of the order, respectively.
(d) Non-universality: no biometric is truly universal. In fingerprint verification system, for example it says
fingerprints are universal but there are some people that does not possess fingerprint because of hand related
disabilities.
(e) Spoof attacks: An imposter may attempt to spoof the biometric trait of a legitimate enrolled user. For
example, this type of attack is especially relevant when behavioural traits such as signature or voice are used.
However, fingerprints can be constructed using lifted fingerprint impression.

2.2 Multimodal biometric systems
Some of the limitations imposed by unimodal biometric systems can be overcome by including
multiple sources of information for establishing identity [3]. Such systems, known as multimodal biometric
systems, are ordinary to be more trustworthy due to the existence of many, (moderately) autonomous pieces of
proof [4]. These systems are able to meet the performance requirements imposed by various applications. They
International Engineering Journal For Research & Development
E-ISSN No: 2349-0721
Volume 1: Isuue 1
www.iejrd.in Page 4

concentrate on the difficulty of non-universality, since several persona ensure sufficient populace reporting.
They also prevent spoofing since it would be difficult for an impostor to spoof multiple biometric traits of a
genuine user concurrently. moreover, they can make possible a challenge response type of mechanism by
requesting the user to present a random subset of biometric traits thereby ensuring that a live user is indeed
present at the point of data acquisition.

Fig 3. Multimodal Biometric System[3]
The reason to combine different modalities is to improve recognition rate. The aim of multimodal
biometrics is to reduce one or more of the following:
False accept rate (FAR)
False reject rate (FRR)
Failure to enroll rate (FTE)
Susceptibility to artefacts or mimics
Multi modal biometric systems take input from single or multiple sensors measuring two or more
different modalities of biometric description. For pattern a system with fingerprint and face recognition would
be considered multimodal even if the OR rule was mortal functional, allowing users to be confirmed by
means of either of the modalities [5].
3. Fusion In Multimodal Biometric Systems
A Mechanism that can combine the classification results from each biometric channel is called as
biometric union. We require to create this union. Multimodal biometric union together dimensions from
different biometric traits to enhance the power. union at similar score, status and conclusion level has been
extensively studied in the fiction. a variety of levels of union are: antenna level, attribute level, identical score
level and decision level.
1. Antenna level Fusion:
We merge the biometric personality taken from different sensors to form a composite biometric trait and
process.
2. Feature level Fusion:
International Engineering Journal For Research & Development
E-ISSN No: 2349-0721
Volume 1: Isuue 1
www.iejrd.in Page 5

Signal coming from different biometric channels are first previously did, and attribute vectors are separated,
using precise algorithm and we join these vectors to form a composite attribute vector. This is constructive in
arrangement.
3. Matching score level fusion:
somewhat than collecting the attribute vector, we practice them separately and individual matching score is
found, then depending on the accuracy of each biometric matching score which will be used for classification.
4. Decision level fusion:
Each modality is first pre-classified independently.This system can implement any of these fusion strategies or
combination of them to improve the performance of the system.

4. Literature Survey
Ratha et al. [6] Proposed a unimodal distortion-tolerant Fingerprint authentication technique based on
graph demonstration. with the Fingerprint details features, a subjective graph of minutiae was made for both the
query Fingerprint and the orientation Fingerprint. The planned algorithm had been experienced with excellent
results on a large private database with the use of an optical biometric sensor. relating to Iris acknowledgment
systems in the Gabor filter and 2-D wavelet filter were used for assets withdrawal. This procedure was invariant
to translation and rotation and was tolerant to enlightenment. The categorization rate on using the Gabor was
98.3% and the accuracy with wavelet was 82.51% on the Institute of Automation of the Chinese Academy of
Sciences (CASIA) record. In the proposed approach in multichannel and Gabor filters had been used to capture
local texture information of the Iris which were used to construct a fixed-length feature vector. The results
obtained were FAR = 0.01% and FRR = 2.17% on CASIA record. normally, unimodal biometric recognition
systems present different drawbacks due its dependency on the unique biometric attribute. i.e attribute
uniqueness, feature achievement, dispensation errors, and attribute that are temporally occupied can all affect
system accurateness. A multimodal biometric structure should overcome the aforementioned limits by
integrating two or more biometric features. Conti et al. [7] proposed a multimodal biometric system using two
different Fingerprint acquisitions. The matching module integrates fuzzy-logic methods for matching-score
union. investigational trials using both decision-level fusion and matching-score-level fusion were performed.
investigational result have shown an improvement of 6.7% using the matching score- level fusion rather than a
monomodal authentication system. Yang and Ma [8] worn Fingerprint, palm publish, and hand geometry to
implement personal identity verification. These three biometric features can be taken from the same image.
They perform a first fusion of the Fingerprint and palm features, and consecutively, a matching-score fusion
between the multimodal system and the palm-geometry unimodal system for comparision. The system was
tested on a database containing the features self-constructed by 98 subjects. Besbes et al. [9] proposed a
multimodal biometric system using Fingerprint and Iris type. They utilize a cross approach based on: 1)
Fingerprint minutiae extraction and 2) Iris template encoding through a mathematical representation of the
extracted Iris area. This move towards was based on two recognition modalities and every part provided its own
decision. The final decision was taken by considering the unimodal decision through an AND operative. No
investigational result have been reported for recognition performance. Aguilar et al. [10] proposed a
multibiometric method using a combination of fast Fourier transform (FFT) and Gabor filters to enhance
International Engineering Journal For Research & Development
E-ISSN No: 2349-0721
Volume 1: Isuue 1
www.iejrd.in Page 6

Fingerprint imaging. The proposed system uses the Fingerprints of mutually thumbs. every Fingerprint was
individually processed; consecutively, the unimodal results were compared in order to give the final complex
result. The tests have been performed on a Fingerprint database composed of 50 subjects obtaining FAR = 0.2%
and FRR = 1.4%. Subbarayudu and Prasad presented experimental results of the unimodal Iris system, unimodal
palm print system, and multibiometric system (Iris and palm print). The structure union utilizes a matching
score feature in which each system provides a matching score indicating the similarity of the feature vector with
the template vector.
Yeong Gon Kim et al.[11] proposed that the face recognition process is carried out according to the
following process. First, the face and eye regions are detected by AdaBoost and rapid eye detection. Second,
size normalization is conducted to eliminate variations in the detected facial region, while the illumination is
normalized using the Retinex algorithm. Third, facial features are acquired from the normalized facial image
based on principal component analysis (PCA). Finally, the matching score is calculated as the Euclidean
distance to provide an input for the SVM. During iris recognition, an iris region is segmented using
integerbased CED and with an eyelid/eyelash detection method. Iris codes are generated from the segmented
iris region. The matching score of the Hamming distance is calculated and used as the SVM input. These
procedures are performed for both the left and right iris images captured using the proposed device. The
matching scores for the face and both irises are used as SVM inputs and a final authentication is carried out
based on the outputs of the SVM.
Conclusion
This paper presents the overview about multimodal biometrics sytem, difference between single modal
and multimodal biometric. Also types of fusion that can be used in multimodal. For using the multimodal the
security of template in future should be considered before storing it in the database.

References
[1] R. Frischholz, U. Dieckmann,BiolD: A multimodal biometric identification system, Computer, Vol. 33,
No. 2, pp. 64-68,2000.
[2] M. Golfarelli, D. Maio, and D. Maltoni, On the error-reject tradeoff in biometric verification systems,
IEEE Trans. on Patt. Anal. and Mach. Intell., vol. 19, pp. 786796, July 1997.
[3] A. Ross and A. K. Jain, Information fusion in biometrics, Pattern Recognition Letters, vol. 24, pp. 2115
2125, Sep 2003.
International Engineering Journal For Research & Development
E-ISSN No: 2349-0721
Volume 1: Isuue 1
www.iejrd.in Page 7

[4] L. I. Kuncheva, C. J. Whitaker, C. A. Shipp, and R. P. W. Duin, Is independence good for combining
classifiers?, in Proc. of Intl Conf. on Pattern Recognition (ICPR), vol. 2, (Barcelona, Spain), pp. 168171,
2000.
[5] M. Indovina, U. Uludag, R. Snelick, A. Mink, and A. Jain, Multimodal Biometric Authentication Methods:
A COTS Approach,
[6] Rabiner L. R., Levinson S. E. and Sondhi M. M., "On the application of Vector Quantization and Hidden
Markov Models to Speaker-Independent, Isolated Word Recognition, " The Bell System Technical Journal,
AT&T, 1983.
[7] N. K. Ratha, R. M. Bolle, V. D. Pandit, and V. Vaish, Robust Fingerprint authentication using local
structural similarity, in Proc. 5th IEEE Workshop Appl. Comput. Vis., Dec. 46, 2000, pp. 2934.DOI
10.1109/WACV.2000.895399.
[8] V. Conti, G. Milici, P. Ribino, S. Vitabile, and F. Sorbello, Fuzzy fusion in multimodal biometric
systems, in Proc. 11th LNAI Int. Conf. Knowl.- Based Intell. Inf. Eng. Syst. (KES 2007/WIRN 2007), Part I
LNAI 4692. B. Apolloni et al., Eds. Berlin, Germany: Springer-Verlag, 2010, pp. 108115.
[9] F. Besbes, H. Trichili, and B.Solaiman, Multimodal biometric system based on Fingerprint identification
and Iris recognition, in Proc. 3rd Int. IEEE Conf. Inf. Commun. Technol.: From Theory to Applications
(ICTTA 2008), pp. 15. DOI: 10.1109/ ICTTA.2008.4530129.
[10] G. Aguilar, G. Sanchez, K. Toscano, M. Nakano, and H. Perez, Multimodal biometric system using
Fingerprint, in Proc. Int. Conf. Intell. Adv. Syst. 2007, pp. 145150.DOI: 10.1109/ ICIAS.2007.4658364
[11]Yeong Gon Kim1, Kwang Yong Shin1, Eui Chul Lee2 and Kang Ryoung Park1. Multimodal Biometric
System Based on the Recognition of Face and Both Irises. Int J Adv Robotic Sy, 2012, Vol. 9, 65:2012