Proceedings of the Fifth International Conference on Intelligent Computing and Control Systems (ICICCS 2021)

IEEE Xplore Part Number: CFP21K74-ART; ISBN: 978-0-7381-1327-2

ENHANCED SECURITY FEATURE OF

ATM’S
THROUGH FACIAL RECOGNITION
Ms. SOUNDARI D V
Department of ECE
2021 5th International Conference on Intelligent Computing and Control Systems (ICICCS) | 978-1-6654-1272-8/21/$31.00 ©2021 IEEE | DOI: 10.1109/ICICCS51141.2021.9432327

Sri Krishna College of Engineering

and Technology
Coimbatore, India
[email protected]

ARAVINDH R EDWIN RAJ K ABISHEK S

Department of ECE Department of ECE Department of ECE
Sri Krishna College of Engineering Sri Krishna College of Engineering Sri Krishna College of Engineering
and Technology and Technology and Technology
Coimbatore, India Coimbatore, India Coimbatore, India
[email protected] [email protected] [email protected]

Abstract— Automated Teller Machines also Keywords—ATM, Face-id, Eigenface algorithm,

known as ATM’s are widely used nowadays by Machine Learning.
each and everyone. The ATM machine (Automated
Teller Machine) is an electronic device that is used I. INTRODUCTION
by the banks to perform banking tasks like Nowadays, the uses of ATMs are inevitable in our
withdrawal of money, transferring of money, and day-to-day lives. Before the introduction of ATM,
many to get many information about a user's bank people used to travel with cash in their hands. This
account without the need to visit a bank. This caused the risk of getting robbed or the cash might get
System revolutionised the way of transactions. lost due to the negligence of the user. To overcome this
There were no long lines of queue in front of the issues ATM played a big role. It is like cash in a card.
bank for a simple withdrawal of money. The As the ATM technology got improved the idea of mis-
number of ATM’s a bank has can be a factor in using it also got improved. The main disadvantage of
considering the strength of a bank. As there is this system is that if the PIN is known anybody can use
increase in the number of ATM’s, there is also it to take money.
increase in the fraudulent activities in the ATM.
To address this issue, Face-id is used as a key. The
The main motivation of this project is to increase
main use of the Biometric is that it is unique for a
the security feature of the use of ATM. The current
person. The proposed innovation is not an alternative
method uses static key (PIN) for security. The
for the conventional ATM security system. The
proposed method uses Face-id as a key
proposed innovation is considered as an additional
incorporated with current method. The advantages
support for the existing method. In India, nearly 2000
can be found as that the face-id is unique for
ATM crimes are occurring every year. The proposed
everybody; it cannot be used by anybody other
method will surely reduce this rate to a greater extent.
than the user. For the implementation of the face-id
scan, the machine learning and image processing
algorithms (Eigenface algorithm) are used.

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1252

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on July 04,2021 at 12:33:32 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Fifth International Conference on Intelligent Computing and Control Systems (ICICCS 2021)
IEEE Xplore Part Number: CFP21K74-ART; ISBN: 978-0-7381-1327-2

II. RELATED WORK relation from the user's cell phone. Since no special
Mohsin Karovaliya in paper [2] proposes subscription to an SMS service provider is needed,
Eigenface based method for the face recognition. This GSM modems can be a fast and easy way to get started
system analyzes the algorithms used in the previous with SMS. A GSM modem can also be a regular GSM
systems. PCA based algorithm is more reliable, very phone with the required cable and software. A GSM
fast and the storage space takes is very less. The main modem may also be a regular GSM phone with the
drawback this method is that it can be manipulated by required cable and software driver to connect to a serial
using photos of users. This method can be improved by port or USB port.
using 3D face masks but the cost of making 3D masks
is very high. III. PROPOSED METHOD
Most people use an ATM for a lot of purposes and
The paper [3] suggests a vibration sensor which in a lot of scenarios. The most common three scenarios
senses vibrations produced from ATM machine are considered here and the solutions are found.
whenever robbery occurs. This system processes real-
x The user as himself using the ATM machine.
time data collected by the vibration sensor using an
ARM controller-based embedded system. The buzzer x If the user is threatened by somebody to get
money without the user’s willing.
will emit a beep sound once the vibration is detected.
The ATM door is closed with a DC motor. There are x If the user’s friend uses the card of the user.
also some additional security measures in place. This First of all, when a user enters the ATM, two choices
will deter theft and make it possible to arrest the are given. Whether the user comes with his card
suspect. Two software programs are used to execute the (Number 1) or if the user’s friend comes with the card
(Number 2).
software, the first of which is Keil Vision 3.0. The
Flash magic simulator is the second. Keil Vision A. User uses the ATM by himself
Debugger correctly simulates on-chip peripherals. This When a user by himself uses the ATM machine,
device aids in rapid reaction and minimization of first the card must be inserted. Now a camera captures
failure by detecting the ATM machine at real-time the image of the user. Now, the captured image is
when it has been stolen can be identified by GSM compared with the image stored in the database. If both
technology. the image is matched (Eigenface algorithm does the
In paper [4], the finger print recognition is done with comparison part) then a successfully authenticated
curvelet transform by finding the Euclidean distance message is displayed. Now the user needs to enter the
between the two corresponding finger codes. The test PIN. If the PIN entered is correct then further
finger code is compared with the entire finger codes in transactions can be done.
the database. An OTP will be sent to the matching
registered mobile number if it matches. The built-in B. User is threatened by somebody
MATLAB feature ‘imread' is used for preprocessing. Now the user is threatened by somebody to get money.
The histogram equalization approach improves an For this scenario we have a robust solution. First the
image's global contrast by shifting the intensity face of the user will match. Now for entering the right
distribution on a histogram marginally. This helps low- PIN, the user must enter a wrong PIN. This causes a
contrast areas to achieve more contrast without alert at the police portal which is at the backend. The
affecting the overall contrast. This is achieved by police portal will be at idle stage if the transactions
histogram equalization, which essentially spreads out were smooth. If the wrong PIN is entered then the alert
the most common strength values. Curvelet transform is received.
and FFT can also be used for function extraction.
C. User’s friend user the card
When the user’s friend uses the card, obviously the
The ATM security is enhanced by adding GSM face will not be matched. For this case, we proposed a
module to generate OTP in paper [5]. Where there is a clear method. The user should set a PIN and
network issue with GSM technologies, this system uses withdrawal limit in the mobile app. This must be set for
Bluetooth to link to an ATM, which produces an OTP each and every transaction. Now the user’s friend must

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1253

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on July 04,2021 at 12:33:32 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Fifth International Conference on Intelligent Computing and Control Systems (ICICCS 2021)
IEEE Xplore Part Number: CFP21K74-ART; ISBN: 978-0-7381-1327-2

enter this secret PIN. If the PIN matches, then the

amount set by the user will be automatically
withdrawn.

Result of csae1. The face is authenticated, The PIN

Flowchart is matched. Now the amount is entered. The process is
done successfully.

Result of case 2:if the face is matched and the PIN

is entered wrong then an alert is sent at the police
terminal at the back end.

Result of case 3: If the users friend uses the card,

Then the PIN (set by the user) must be entered
correctly After correct entry of the PIN amount is
withdrawn.

DRAWBACKS OF THE EXISTING SYSTEM

x Anybody who knows the PIN can use the card
for the transactions.
x Since the PIN is static it is easy for criminals to
extract it from the users, through hacking or
even through brute force applied to the person.
x Scanning the magnetic strip in the ATM cards
will give the complete details of the card. This
causes the duplication of the cards.
x Threatening people in front of ATM is done in
some places and the security feature of PIN
cannot be useful in these situations.

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1254

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on July 04,2021 at 12:33:32 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Fifth International Conference on Intelligent Computing and Control Systems (ICICCS 2021)
IEEE Xplore Part Number: CFP21K74-ART; ISBN: 978-0-7381-1327-2

ADVANTAGES OF PROPOSED SYSTEM (2) Train the Recognizer: Feed the recognizer the face
data as well as the names of each face so that it can
x The proposed method uses a biometric, this remember.
increases the security feature of the existing (3) Recognition: Feed the face recognizer with new faces
system drastically. of certain individuals to see how it knows them.
x Since we use biometric it is unique for each and OpenCV has two built-in face recognizers. The
every individual. names of those face recognizers are: Eigenfaces and
x The condition of threatening the user is well Fisher faces.
addressed here, and this can result in the safety
of the user and the user’s money.
x The withdrawal limit is set if the user’s friend
3.3.4 Eigenfaces Face Recognizer Algorithm
uses the card. This helps the user to be in control
of his account. In this algorithm, a facial image is a point from a
high-dimensional image space and a lower-dimensional
x Implementation of this method is easy, cheap, representation is found, where classification becomes
and efficient. easy. The lower-dimensional subspace is found with
Principal Component Analysis (PCA), which identifies
the axes with maximum variance. While this kind of
Biometrics Performance Accuracy Cost transformation is optimal from a reconstruction
Iris Medium Medium High standpoint, it doesn’t take any class labels into account.
Imagine a situation where the variance is generated
Fingerprint High Medium Low from external sources, let it be light. The axes with
Retina High High High maximum variance do not necessarily contain any
discriminative information at all, hence a classification
becomes impossible.
The functioning of the algorithm is that it considers
ALGORITHMS USED
that not all parts of a face are important equally to
Since we use face data as the key here. We use recognize a face. Instead, the main feature like the
image processing algorithms here. The implementation shape of nose, ears, fore-head are used and considered
is done using OpenCV in python which has built in how they vary from one another The main idea is to
image recognising algorithms. There algorithms are find the region with maximum difference. Let us
trained using machine learning. We use Eigenface consider that there will a drastic amount of variation
Algorithm. will be there when we compare the region of eyes and
nose. When multiple faces are to be compared, the
FACE RECOGNITION comparison between the faces is done looking at the
Face recognition is an easy task for humans. Seeing maximum variation among the faces, and they help to
the geometric features of the face is considered as the distinguish the faces. This is the process behind
most easy and efficient way in facial recognition. On Eigenfaces recognizer. This recognizer works by
early days the first automatic face recognition uses the looking at all the pictures that are used to train and
marker points like position of ears, position of nose etc. extracts the key components that are considered to be
These points were used in framing the feature vector relevant and ignores the other components. Those key
like distance between them or angle between them. The components are called as Principal components and
facial recognition process was done by finding the these functions as main source for the recognizer.
distance between the feature vector of the captured This face recognizer also maps the records of the
image with a reference image. Some of the latest work face to the person whose record is that. So, when a new
on geometric face recognition was, a 22-dimensional record needs to be added it follows the following
feature vector and experiments on large datasets have process. The principal component is extracted first. The
shown that geometrical features alone may not carry comparison process is done next with the images used
enough information for face recognition. to train. Find the image which matches the best. Now
Mainly Computer coding facial recognition can be map the record with the person name. In eigenface
broken down into three simple stages that are identical recognizer the light acts as one of the important needs.
to the steps used by the human brain to recognize faces. Only with the help of light the, shadowed and
These are the steps: illuminated parts of the face is found. These details are
used to find an eigenface which are used to represent a
(1) Data Collection: Gather face data (in this case, face face. The shadows and shapes of two eyes, a nose, a
images) of the individuals you want to recognize. mouth is the most used parameter in facial recognition.

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1255

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on July 04,2021 at 12:33:32 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Fifth International Conference on Intelligent Computing and Control Systems (ICICCS 2021)
IEEE Xplore Part Number: CFP21K74-ART; ISBN: 978-0-7381-1327-2

[4] [4]C. Bhosale, P. Dere, C. Jadhav, "ATM security using face and
fingerprint recognition", Interna- tional Journal of Research in
Engineering, Technology and Science, Volume VII, Special Issue,
Feb 2017.
[5] [5]Manoj V , M. Sankar R , Sasipriya S , U. Devi E, Devika T ,
"Multi Authentication ATM Theft Prevention Using iBeacon",
International Research Journal of Engineering and Technology
(IRJET).
[6] L. Wang,H. Ji, Y. Shi, " Face recognition using maximum local
fisher discriminant analysis",18th IEEE International Conference
Image showing how the Eigenface algorithm works.
on Image Processing, 2011
[7] K.Shailaja and Dr.B.Anuradha, "Effective Face Recognition using
IV. CONCLUSION Deep Learning based Linear Discriminant Classification ", IEEE
On summing up, the method what we proposed is International conference on Computational Interlligence and
Computing Research,2016.
far better in increasing the security feature. The main
[8] H. R. Babaei, O. Molalapata and A.H.Y Akbar Pandor, "Face
goal of our paper is to incorporate the facial recognition Recognition Application for Au- tomatic Teller Machines
feature along with the existing conventional method for (ATM)",International Conference on Information and Knowledge
Management(ICIKM),2012.
the betterment of the user. The Eigenface algorithm
[9] Chen, Joy Iong Zong. "Smart Security System for Suspicious
used here is used for the comparing the face of the user Activity Detection in VolatileAreas." Journal of Information
with the face in the database. Machine learning is used Technology 2, no. 01 (2020): 64-72.
to train (used in OpenCV inbuilt face recogniser) the [10] Suma, V. "Computer vision for human-machine interaction-
review." Journal of trends inComputer Science and Smart
face recogniser. The Adaboost face recognition technology (TCSST) 1, no. 02 (2019): 131-139.
algorithm has the success rate of 75% but the eigenface [11] Face Recognition using Eigenface algorithm on Laptop camera by
Algorithm produces the success rate of 80%. Rika Rosnelly, Mutiara S. Simanjuntak, Ade Clinton sitepu,
Mulkan Azhari, Sandy Kosasi in 2020 8th International Conference
The main limitation of this system is that it requires on Cyber and IT service Management.
periodic maintenance of the cameras. Twins can be an [12] Comparison of the performance of Eigen face and Fisher Face
Algorithm in the face recognition process by Delpiah
exception in this system. In rare cases, photos can be
Wahyuningsih, Chanda Kirana, Rahmat Sulaiman, Hamidah,
used to bypass the security. The future scopes of this Triwanto in 2019 7th International Conference on Cyber and IT
method are that the use of high-quality durable service Management(CITSM).

cameras. 3-d cameras can be used for the condition of [13] Face Detection based ATM security system using embedded Linux
platform Jignesh J. Patoliya, Miral M Desai in 2017 2nd
twins and photo bypassing. International Conference for Convergence in Technology(I2CT).
[14] Real Time Facial Expression and emotion recognition using
REFERENCES Eigenfaces, LBPH and Fisher Algorithms Shrayan Mukhopadhyay,
Shilpu Sharma 2020 10th International Conference on cloud
[1] J.J.Patoliya, M.M. Desai, "Face Detection based ATM Security
computing, Data Science and Engineering.
System using Embedded Linux Platform ", 2nd International
Conference for Convergence in Technology (I2CT), 2017. [15] Face recognition alert mechanism and idealness for monitoring user
motion by Web Services C. Jayaprakash, V Maheshwari 2016 10 th
[2] M.Karovaliyaa, S.Karediab, S.Ozac, Dr.D.R.Kalbande, "Enhanced
International Conference on Intelligent Systems and
security for ATM machine with OTP and Facial recognition
Control(ISCO).
features",International Conference on Advanced Computing Tech-
nologies and Applications (ICACTA), 2015.
[3] [3]Sivakumar T. 1 , G. Askok 2 , k. S. Venuprathap, "Design and .
Implementation of Security Based ATM theft Monitoring system",
International Journal of Engineering Inventions, Volume 3, Issue
1, 2013.

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1256

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on July 04,2021 at 12:33:32 UTC from IEEE Xplore. Restrictions apply.