FacePIN: Face Biometric Authentication

System for ATM using Deep Learning

SIVAGANESH.M 1,ABINESH.V 2,AJAY SAMSON.A 3,MOHAMAD AMJATH.S 4

COMPUTER SCIENCE AND ENGINEERING, PAAVAI ENGINEERING COLLEGE,NAMAKKAL.

banking sector. ATMs allow banking customers to avail

ABSTRACT
quick self-serviced transactions, such as cash
Automated Teller Machines also known as ATM's are
withdrawal, deposit, and fund transfers. ATMs enable
widely used nowadays by each and everyone. There is an
individuals to make banking transactions without the
urgent need for improving security in banking region.
help of an actual teller. Also, customers can avail
Due to tremendous increase in the number of criminals
banking services without having to visit a bank branch.
and their activities, the ATM has become insecure. ATM
Most ATM transactions can be availed with the use of a
systems today use no more than an access card and PIN
debit or credit card. There are some transactions that
for identity verification. The recent progress in biometric
need no debit or credit card. In 1960, an American named
identification techniques, including finger printing,
Luther George Simjian invented the Bank graph, a
retina scanning, and facial recognition has made a great
machine that allowed customers to deposit cash and
effort to rescue the unsafe situation at the ATM. This
checks into it. The first ATM was set up in June 1967 on
project proposes an automatic teller machine security
a street in Enfield, London at a branch of Barclays bank.
model that would combine a physical access card and
A British inventor named John Shepherd-Barron is
electronic facial recognition using Deep Convolutional
credited with its invention. The machine allowed
Neural Network. If this technology becomes widely
customers to withdraw a maximum of GBP10 at a time.
used, faces would be protected as well as their accounts.
Automated Teller Machines have revolutionized the
Face Verification Link will be generated and sent to user
banking sector by providing easy access to customers
to verify the identity of unauthorized user through some
and loading off the burden from bank officials. Some of
dedicated artificial intelligent agents, for remote
the uses of an ATM are The most common uses of an
certification. However, it obvious that man’s biometric
Automated Teller Machine include withdrawing money,
features cannot be replicated, this proposal will go a long
checking balance, transferring money, or changing the
way to solve the problem of Account safety making it
PIN (Personal Identification Number) Newer and
possible for the actual account owner alone have access
advanced ATMs also provide options to open/withdraw
to his accounts.
a Fixed Deposit (FD), or to apply for a personal loan.
You can also book railway tickets, pay the insurance
I . INTRODUCTION premiums, income tax & utility bills, recharge mobile,
and deposit cash. Some of these facilities require you to
Automated Teller Machines, popularly referred to as
register at the bank branch
ATMs, are one of the most useful advancements in the
• Customers can now do money transactions at their
convenience. ATMs today are installed in public spaces,

Page | 1
 highways, malls, market places, railway/airport stations, information. The end result is the same as skimming
hospitals, etc. because thieves use the stolen chip data to create
• Automated Teller Machines provide 24×7 access “cloned” versions of your debit card.
anywhere • Cash-out: This scam targets multiple accounts from the
• ATMs help to avoid the hassle of standing in long queues same financial institution. Armed with a hacked bank

at the bank even for simpler transactions like employee’s credentials, the criminal alters account
withdrawing money. It has also helped in reducing the balances and withdrawal limits. Using stolen debit card

workload of the bank officials. numbers captured from a separate skimming attack, they

Over the last two decades, automated teller can “cash out” the ATM until it’s out of money.

machines (ATMs) have become as much a part of the • Jackpotting: While there are multiple types of
landscape as the phone booths made famous by jackpotting attacks, typically, these incidents involve
Superman. As a result of their ubiquity, people casually gaining physical access to the inside of the machine. The
use these virtual cash dispensers without a second criminals may replace hardware or install malicious
thought. The notion that something could go wrong never software giving them control of the cash dispensing
crosses their minds. Most ATM scams involve criminal function. Jackpotting is similar to a cash out scam, but it
theft of debit card numbers and personal identification does not require the criminal to have any customer
numbers (PINs) from the innocent users of these account details or stolen debit card information.
machines. There are several variations of this confidence Face recognition can be used to secure ATM transaction
scheme, but all involve the unknowing cooperation of the and is used as a tool for authenticating users to confirm
cardholders themselves. the card owner.Financial fraud is a very important
ATM fraud is described as a fraudulent activity where the problem for Banks and current secure information in the
criminal uses the ATM card of another person to ATM card magnetic tape are very vulnerable to theft or
withdraw money instantly from that account. This is done loss. By using face recognition as a tool for
by using the PIN. The other type of ATM fraud is stealing authenticating users in ATMs can be confirmed as the
from the machine in the ATM by breaking in. card owner. Face Based ATM login Process the ATMs

• Skimming: This type of ATM scam involves a skimmer which are equipped with Face recognition technology

device that criminals place on top of or within the card can recognize the human face during a transaction. When

slot. To record your PIN number, the criminals may use there are “Shoulder Surfers" who try to peek over the

a hidden camera or an overlay that covers the original cardholder’s shoulder to obtain his PIN when the

PIN pad. Using the card numbers and PIN’s they record; cardholder enters it, the ATMs will automatically remind
thieves create duplicate cards to withdraw money from the cardholder to be cautious. If the user wears a mask or

consumers’ accounts. Unlike losing your debit card or sunglasses, the ATM will refuse to serve him until the

having it stolen, you won’t realize anything is amiss until covers are removed.

unauthorized transactions take place. Take a look at these • Touchless - There is no need for remembering
your passwords. Only looking at the ATM
so you know how to detect ATM skimmers.
camera will login the card holder instantly. No
• Shimming: This is the latest update to skimming. Instead physical contact is needed.
of reading your card number, criminals place a shimming • Secure - Since your face is your password, there
device deep inside the ATM to record your card’s chip is no need to worry for your password being

Page | 2
 forgotten or stolen. In addition, the face unauthorized user through some dedicated artificial
recognition engine locks access to the account intelligent agents.
and transaction pages for the card holder as the
card holder moves away from the camera of ADVANTAGES OF THE PROPOSED SYSTEM
the ATM and another face appears • The advantages can be found as that the face-id is
• Face based card holder authentication can be unique for everybody; it cannotbe used by anybody
used as primary or as a secondary other than the user.
authentication measure along with ATM PIN. • It can be used to reduce fraudulent attempts.
Face based authentication prevents ATM fraud • To prevent theft and other criminal activities.
by the use of fake card and stolen PIN or stolen • Secure facial authentication platform that users can
card itself. Face verification is embedded with trust
• Provide safe and secure lifestyle infrastructure
security features to prevent fraud, including
• Prevent unauthorized access using Face verification
liveness-detection technology that detects and
Link.
blocks the use of photographs, videos or masks
• Fast and Accurate Prediction
during the verification process.
II. IDENTIFY,RESEARCH AND COLLECT IDEA III. WRITE DOWN YOUR STUDIES AND
FINDINGS
EXISTING SYSTEM

Existing ATM authentication method is the use of ARCHITECTURE

password-PINs and OTP.QR cash withdrawals were
enabled so customers could ditch their ATM cards and
simply scan a QR-code on ATMs using the QR app to
withdraw cash.ATM security system architecture that
incorporates both the finger print and GSM technology
into the existing PIN-based authentication process.

DISADVANTAGES

• Skimming and Trapping of the ATM devices have

been designed by many Burglars hacking or even
through brute force applied to the person.

• Scanning the magnetic strip in the ATM cards will

give the complete details of the card. This causes
the duplication of the cards.
Architecture diagram

Algorithms/Techniques
PROPOSED SYSTEM
1. DCNN – Deep Convolutional Neural Network
This project proposes an automatic teller machine Face Authentication.
security model that would combine a physical access card 2. Unknown Face Verification Link Generator:
and electronic facial recognition using Deep
Face Verification Link will be
Convolutional Neural Network. When the stored image
generated and sent to user to verify the
and the captured image don't match, it means that he is
identity of unauthorized user through some
an unauthorized user. Face Verification Link will be
dedicated artificial intelligent agents, for
generate and sent to user to verify the identity of
remote certification.

Page | 3
Modules a pattern classifier, the extracted features of face image
1. ATM Simulator are compared with the ones stored in the face database.
The face image is then classified as either known or
2. Face Recognition Module unknown. If the image face is known, corresponding
Card Holder is identified and proceed further.
2.1. Face Enrollment
Unknown Face Forwarder Mechanism
2.2. Face Authentication Unknown Face Verification Link will be generate and
sent to card holder.To verify the identity of unauthorized
3. Unknown Face Forwarder Mechanism user through some dedicated artificial intelligent agents,
for remote certification, which either authorizes the
4. Transaction Module transaction appropriately or signals a security-violation
alert to the banking security system.
5. Performance Analysis
Transaction Module
ATM Simulator Enter the Withdrawal Money
ATM Simulator is a Next Generation testing
application for XFS-based ATMs (also known as In this section, you have to enter your withdrawal
Advanced Function or Open-Architecture ATMs).ATM amount and press enter.But make sure your withdrawal
Simulator is a web technology to allow ATM testing with amount does not exceed your balance in the account
a virtualized version of any ATM.ATM Simulator uses otherwise transaction will fail.
virtualization to provide with realistic ATM simulation,
coupled with automation for faster, more efficient testing Collect the Cash
for face authentication and unknown Face Forwarder
Technique. In this section, you have to collect your money from the
lower slot of the machine. Take your money before 30
Face Recognition Module seconds.
Face Enrollment
This module begins by registering a few frontal face of Performance Analysis
Bank Beneficiary templates. These templates then The performance analysis of all experiments was
become the reference for evaluating and registering the based on the most common evaluation measures used for
templates for the other poses: tilting up/down, moving statistical tests, such as accuracy, precision, recall, and
closer/further, and turning left/right.With each pose, the f_measure.
facial information including eyes, nose and mouth is
automatically extracted and is then used to calculate the SYSTEM TESTING TESTING TECHNIQUES
effects of the variation using its relation to the frontal face The purpose of testing is to discover errors. Testing is
templates. DCNN algorithms were also created to the process of trying to discover every conceivable fault
automatically detect and reject improper face images or weakness in a work product. It provides a way to check
during the enrollment process. This will ensure proper the functionality of components, sub-assemblies,
enrollment and therefore the best possible performance. assemblies and/or a finished product. It is the process of
exercising software with the intent of ensuring that the
Face Authentication Software system meets its requirements and user
After capturing the face image from the ATM Camera, expectations and does not fail in an unacceptable manner.
the image is given to face detection module. This module There are various types of test. Each test type addresses
detects the image regions which are likely to be human. a specific testing requirement.
After the face detection using Region Proposal Network BLACK BOX TESTING
(RPN), face image is given as input to the feature Black Box Testing is testing the software without any
extraction module to find the key features that will be knowledge of the inner workings, structure or language
used for classification. The module composes a very of the module being tested. Black box tests, as most other
short feature vector that is well enough to represent the kinds of tests, must be written from a definitive source
face image. Here, it is done with DCNN with the help of document, such as specification or requirements

Page | 4
document, such as specification or requirements exposing the problems that arise from the combination of
document. It is a testing in which the software under test components.
is treated, as a black box .you cannot “see” into it. The
test provides inputs and responds to outputs without
considering how the software works.
IV. CONCLUSION
WHITE BOX TESTING

White Box Testing is a testing in which in which the Biometrics as means of identifying and authenticating

software tester has knowledge of the inner workings, account owners at the Automated Teller Machines gives

structure and language of the software, or at least its the needed and much anticipated solution to the problem

purpose. It is purpose. It is used to test areas that cannot of illegal transactions. In this project, we have developed

be reached from a black box level. to proffer a solution to the much-dreaded issue of
fraudulent transactions through Automated Teller
UNIT TESTING Machine by biometrics and Unknown Face Forwarder
that can be made possible only when the account holder
Unit testing involves the design of test cases that
is physically or far present. Thus, it eliminates cases of
validate that the internal program logic is functioning
illegal transactions at the ATM points without the
properly, and that program inputs produce valid outputs.
knowledge of the authentic owner. Using a biometric
All decision branches and internal code flow should be
feature for identification is strong and it is further
validated. It is the testing of individual software units of
fortified when another is used at authentication level.
the application .it is done after the completion of an
The ATM security design incorporates the possible
individual unit before integration. This is a structural
proxy usage of the existing security tools (such as ATM
testing, that relies on knowledge of its construction and
Card) and information (such as PIN) into the existing
is invasive. Unit tests perform basic tests at component
ATM security mechanisms. It involves, on real-time
level and test a specific business process, application,
basis, the bank account owner in all the available and
and/or system configuration. Unit tests ensure that each
accessible transactions
unique path of a business process performs accurately to
the documented specifications and contains clearly
IV. FUTURE ENHANCEMENT
defined inputs and expected results.
In the future, the recognition performance should be
Integration Testing
further boosted by designing novel deep feature
Integration tests are designed to test integrated software representation schemes.
components to determine if they actually run as one
program. Testing is event driven and is more concerned
REFERENCES
with the basic outcome of screens or fields. Integration
tests demonstrate that although the components were [1] J. Liang, H. Zhao, X. Li, and H. Zhao, ``Face
individually satisfaction, as shown by successfully unit recognition system based on deep residual network,'' in
testing, the combination of components is correct and Proc. 3rd Workshop Adv. Res. Technol. Ind.
consistent. Integration testing is specifically aimed at (WARTIA), Nov. 2017, p. 5.

Page | 5
[2] I. Taleb, M. E. Amine Ouis, and M. O. Mammar, cloud,'' in Proc. Online Int. Conf. Green Eng. Technol.
``Access control using automated face recognition: (IC-GET), Nov. 2016, pp. 1-4.
Based on the PCA & LDA algorithms,'' in Proc. 4th Int.
Symp. ISKO-Maghreb, Concepts Tools Knowl. Manage.
(ISKO-Maghreb), Nov. 2014, pp. 1-5.
[3] X. Pan, ``Research and implementation of access
control system based on RFID and FNN-face
recognition,'' in Proc. 2nd Int. Conf. Intell. Syst. Design
Eng. Appl., Jan. 2012, pp. 716-719, doi:
10.1109/ISdea.2012.400.
[4] A. A. Wazwaz, A. O. Herbawi, M. J. Teeti, and S. Y.
Hmeed, ``Raspberry Pi and computers-based face
detection and recognition system,'' in Proc. 4th Int. Conf.
Comput. Technol. Appl. (ICCTA), May 2018, pp. 171-
174.
[5] A. Had, S. Benouar, M. Kedir-Talha, F. Abtahi, M.
Attari, and F. Seoane, ``Full impedance cardiography
measurement device using raspberry PI3 and system-on-
chip biomedical instrumentation solutions,'' IEEE J.
Biomed. Health Informat., vol. 22, no. 6, pp. 1883-1894,
Nov. 2018.
[6] A. Li, S. Shan, andW. Gao, ``Coupled bias-variance
tradeoff for cross-pose face recognition,'' IEEE Trans.
Image Process., vol. 21, no. 1, pp. 305-315, Jan. 2012.
[7] C. Ding, C. Xu, and D. Tao, ``Multi-task pose-
invariant face recognition,'' IEEE Trans. Image Process.,
vol. 24, no. 3, pp. 980-993, Mar. 2015.
[8] J. Yang, Z. Lei, D. Yi, and S. Li, ``Person-specific
face antispoofong with subject domain adaptation,'' IEEE
Trans. Inf. Forensics Security, vol. 10, no. 4, pp. 797-
809, Apr. 2015.
[9] H. S. Bhatt, S. Bharadwaj, R. Singh, and M.Vatsa,
``Recognizing surgically altered face images using multi
objective evolutionary algorithm,'' IEEE Trans. Inf.
Forensics Security, vol. 8, no. 1, pp. 89-100, Jan. 2013.
[10] T. Sharma and S. L. Aarthy, ``An automatic
attendance monitoring system using RFID and IOT using

Page | 6