2016 Online International Conference on Green Engineering and Technologies (IC-GET)

A Robust and secure authentication mechanism in

online banking
Hridya Venugopal N Viswanath
M Tech Scholar ,CSE Professor, CSE
Toc H Institute of Science and Technology Toc H Institute of Science and Technology
Cochin,India 682313 Cochin,India 682313
Email: [email protected] Email: [email protected]

Abstract—Online banking is on the up each day with for customers and transactions will address most fraud re-
a persistent rise in the number of people using this novel lated issues. Apart from incorporating strong authentication
service to carry out their financial transactions. This amplified mechanism, certain banks limit the number of online banking
interest in the use of online banking has consequently raised
the concerns over the security. This has raised the need to operations that a customer can perform each day.
protect online banking in to guard these transactions as well Biometric technology ensures the robust and safe technique
as establishing secure mechanisms for information exchange to make Secure authentications of persons. A large portion of
that prevent fraud and safeguard the personal data. With the system breaches are caused by authentication failure, either
internet now popular among all age groups, online banking has during the login process or in the transaction process which
become a necessity.Security mechanisms are, therefore a must
for the proper functioning of online banking. In addition to exist due to the limitations accompanying the existing authen-
this, all the users are required to manage multiple passwords tication methods[7]. Current authentication methods are not
and devices. Security which are provided by the extensively user oriented and are thus an endanger to users security.
used systems namely knowledge-based security and token-based In the current world, authentication of online banking users is
security can be easily breached when one reveals his password done using the following methods:[1]
and his cards are stolen. In order to overcome this, biometrics
are used. Banks have started using single biometric systems for A. KNOWLEDGE BASED
financial transactions. In order to provide further security for
online banking transactions, the proposed system introduces This method, which is the most popular and common,
the use of multiple(face and fingerprint) biometrics for online asks the users to authenticate by entering their User Id and
financial transaction where both are required for authentication password. The bank safeguards the security by ensuring that
of log-in- process and one biometric is used for transaction the users have a strong password and that are changed at a
process, thus would help overcome traditional vulnerabilities. frequent intervals which is assigned to be for few days.
Further, this proposed research further explores the matching
at the feature level, which of course is a under studied problem. B. TOKEN BASED
Here in this approach, the feature sets extracted from multiple
data sources would be fused to create a new feature set Token based method is currently used in almost all onine
to represent the individual. Since the feature set contains bank transactions. This method authenticates the users based
better-off information about the fresh biometric data compared on the knowledge based identity and something else that they
to the match score level or the final decision, combination have.This is usually done using OTP(One Time Password), or
at this level is possible to provide better authentication
token devices.
results. Initial results indicate that the planned technique can
lead to large improvement in multimodal matching performance. II. RELATED WORKS
Index Terms—Unimodal biometrics, multimodal biometrics, A. UNIMODAL BIOMETRICS
OTP. The unimodal biometric systems rely on the evidence
of a single source of information for authentication of
I. I NTRODUCTION person. Though these unimodal biometric systems have
many advantages, it has to face with variety problems like
A number of aspects, including lesser cost of network Noisy data,Intra class variation,Interclass similarities,Non
devices, larger Internet and mobile Internet penetration, avail- universality,Spoofing etc[6].
ability of devices and increased use of the smartphones have
gone into commercialising online banking around the world.
The circumstance remains that in spite of the advancements B. TYPES OF MULTIMODAL SYSTEMS
in security technology, vulnerablity still exist. Studies shows Depending on the traits, sensors and feature sets many
that many phishing and social engineering attacks take place different types of multimodal systems are there.These
around the world every month. Though there are many threats include:[2]
and vulnerabilities, a very strong authentication mechanism

978-1-5090-4556-3/16/$31.00 ©2016 IEEE

 2016 Online International Conference on Green Engineering and Technologies (IC-GET)

1) Single biometric trait, multiple sensors: Multiple the individuality corresponding to the probe is clasically
sensors are used to record the same biometric characteristic. determined by matching it against the templates of all
The raw data taken from different sensors can then be individualities in the gallery.[5]
combined at the feature level or matcher score level to
improve the performance of the system.
E. FINGERPRINT MATCHING TECHNIQUES
2) Multiple biometrics: Multiple biometric traits such as For accurate personal identification,considering all the cur-
fingerprints and face can be combined.Different sensors are rently used biometric techniques, fingerprint authentication
used for each biometric characteristic. The interdependency system is the widely used and appropriate.The existing popular
of the traits ensures a significant improvement in the fingerprint matching techniques can be broadly classified into
performance of the system. three categories depending on the types of features used:[4]
1) Minutiae-based:
3) Multiple units, single biometric traits: Two or more 2) Correlation-based:
fingers of a single user can be used as a biometric trait. It 3) Euclidean distance-based:
is inexpensive way of improving system performance, as it
doesnt require multiple sensors or incorporating additional III. P ROPOSED S YSTEM D ESIGN
feature extraction or matching modules. Iris can also be In the proposed system, the online banking system
included in this category. ensures robust and secure authentication mechanism by using
the multimodal biometrics.Multimodal system including
4) Multiple snapshots of single biometric: In this more Fingerprint and face are used for the login process. As theft
than one instance of the same biometric is used for the can occur at any point of transaction process, fingerprint
recognition. For e.g. multiple impressions of the same finger authentication is again done during transaction process.
or multiple samples of the voice. Efficient encryption and decryption methods are used for
providing the security of data transmitted and storing the data
5) Multiple matching algorithms for the same biometric: in the database. Thus the proposed system ensures improved
In it different methods can be applied to feature extraction security in online banking by using the multimodal biometric
and matching of the biometric characteristic. system.

C. FUSION LEVELS IN MULTIMODAL BIOMETRICS

There are three fusion levels in multimodal biometrics:
feature level fusion, matching score level fusion and decision
level fusion respectively. The three levels of fusion are
described as follows:[3]

1) FEATURE LEVEL FUSION: In the feature level

fusion, features from different biometric traits are initially
processed and the feature vectors are obtained are extracted
and combined to form a composite feature vector. This is
Figure 1. High level design
then combined to form a feature vector that is used for
classification.
Figure 1 describes the overall scenario in the proposed
2) MATCHING SCORE LEVEL FUSION: In matching system.The planned system consists of a client system which is
score level fusion, individual matching score is found based the user doing the online transaction. The bank server encloses
on various biometric traits and these matching scores are the database with which the details has to be compared. The
gathered to make the classification. user can login with the user id , and recognising self with
fingerprint and face . These details are compared with the
3) DECISION LEVEL FUSION: In decision level fusion, database in the server. Once the login is successful, the user
each biometric traits are captured and features are extracted can make the necessary transaction by authenticating with the
from the captured traits.The final decision of accept or reject fingerprint once again. The details are again compared with
based on the combination of the outcomes from different the server.
biometric modalities. The proposed system uses a multimodal biometric system.It
consists of two main modules namely,
D. MATCHING ALGORITHMS A. Enrolment module
Based on the pattern of the matching algorithm, the Here, the user has to register at the bank with the necessary
matching speed can vary. In a biometric recognition system, details . This includes the biometric traits as well as other
 2016 Online International Conference on Green Engineering and Technologies (IC-GET)

information needed for the authentication.

B. Authentication module
Here, the user has to authenticate him/herself using the
multi biometric traits used for the login process and unimodal
biometric , used for transaction process.The Authentication
module consist of two main processes.
1) Login Process: Here, the user has to login using the
user id followed by the recognition of face and fingerprint
for authentication .Once the user login to the system the user
can only view the account details.

2) Transaction Process: Here, the user has to again

authenticate him/her self using the fingerprint authentication.
Only when the user authenticate with the fingerprint details,
the transaction can be done.
Figure 3. Server side
The authentication mechanism includes the processes at
both the client and server side. The client side process include
capturing the finger and face image ,followed by feature various spoofing attacks that can occur while using unimodal
extraction and fusion of the feature extracted,encrypting the biometrics. Thus multimodal biometrics ensures an efficient
Euclidean distance calculated and sending it to the server.This method for authentication in online transaction.Certain threats
is depicted in Figure 2. including hacking, phishing etc can also be dispensed when
using multimodal biometrics.

R EFERENCES
[1] Available “http://www.edgeverve.com/finacle/resources/thought-
papers/Documents/what-the-future-online-banking.pdf
[2] Sheena S, Sheena Mathew,“ A STUDY OF MULTIMODAL BIOMETRIC
SYSTEM”, IJRET: International Journal of Research in Engineering and
Technology eISSN: 2319-1163 — pISSN: 2321-7308
[3] S.R.Soruba Sree , Dr. N.Radha,”A Survey on Fusion Techniques for
Multimodal Biomet- ric Identification, International Journal of Innovative
Research in Computer and Com- munication Engineering, Vol. 2, Issue
12, December 2014.
[4] Mary Lourde R, and Dushyant Khosla, ”Fingerprint Identification in
Biometric Security Systems”, International Journal of Computer and
Electrical Engineering, Vol. 2, No. 5, October, 2010 1793-8163
[5] K.Saranya,K.Baskar,”Multibiometric Secure Index Value Code Genera-
tion for Authentication and Retrieval”,Iternational Journal for Scientific
Research & Development— Vol. 1, Issue 5, 2013 — ISSN (online): 2321-
0613
Figure 2. Client side
[6] Available “http://www.airccse.org
[7] Sui, Yan, Xukai Zou, Eliza Y. Du, and Feng Li,”Design and Analysis
Figure 3 illustrates the server side process. The server of a Highly User-Friendly, Secure,Privacy-Preserving, and Revocable
Authentication Method”,IEEE Transactions on Computers, 2014.
side process include, decrypting the encrypted data, and [8] Ghayoumi, Mehdi,””A review of multimodal biometric systems: Fusion
comparing the stored data in the database. methods and their applications”,2015 IEEE/ACIS 14th International
Conference on Computer and Information Science (ICIS), 2015.

IV. C ONCLUSION
Today, the authentication mechanism in online banking
include two factor authentication which is the token based
authentication mechanism.This needs an external device to
dynamically authenticate the user. However, the chances to the
device being misplaced or loss can cause a compromise to the
bank account transactiona. There are many vulnerabilities still
concerning this area. So a robust and secure authentication
mechanism to be used in online banking is essential. This
can be achieved by using multimodal biometrics. There are