BIOMETRIACALLY SECURE BANK LOCKER USING IRIS

RECOGNITION

BIOMETRICALLY SECURE BANK LOCKER USING

IRIS RECOGNITION

Page 1
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

ABSTRACT:
In today's modern world, security plays an important role. For that purpose, we proposed
advance security systems for banking locker system and the bank customers. Bank locker
security is an important thing as we people grow and achieve many things that are important
for each individual we need a proper security for such things as could be some important
papers, jewelery, personal stuffs etc . The world has changed a lot from the old and
mechanical stuffs to the modern electronics world. The proposed project deals with the
development of advanced IRIS based locker authentication using iris segmentation and
recognition. The proposed project deals with the development of the bank locker which can
be accessed by scanning the iris. The iris recognition system is developed using Dougman’s
Algorithm which uses IRIS camera interfaced to the raspberry pi to scan the iris image and
authenticate it with the image database of the authorized people. If the authentication is
successful the access to the bank locker will be granted by operating the electromechanical
lock. Thus this project proposed a novel safety and security system for bank locker access
using biometrically secure IRIS recognition and authentication.

Keywords: IRIS recognition, Segmentation, Bank Locker, Raspberry Pi, Dougmans

Algorithm, electromechanical lock, Security, Biometric etc.

Page 2
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 1.

INTRODUCTION:
The Bank which is used to place that indicates the very high level security. In day to day life
every person are involved in banking transaction. Because of high level security, we use bank
lockers to secure our important documents, expensive jewellery , or cash etc., Hence it has
become a very important part for every common human being . To suffer in this world and
for a continuous development; the banking sector needs to accommodate a very hinge rise
security. As we know new branches are opening by considering the public interest. Hence
more security for every sector is required. Because of development current system and
services becomes autonomous and banking service is not so far from that. Various researches
show that there are accountability in devices and technologies in security system . In
automatic security systems generally passwords, identification cards and PIN verification
techniques are being used but the disadvantage is that the passwords could be hacked and a
card may be stolen or lost.

Biometrics is an automatic technique that is depending upon a physical or behavioural

characteristic to authenticate an individual. Physical properties, including the face, voice,
fingerprints, even iris, behavioural properties are features that will be developed or gained,
like keystroke dynamics, dynamic signature identification, and speaker identification. For all
the particular physical characteristics, iris' biometrics is considered highly reliable and
difficult to duplicate and replicate . The iris is a vision structure, yet preserved. Iris always
does not vary over the years; therefore, it is ideal & effective for encrypting identity.

Iris recognition system is a biometric system. Today, there are many kinds of biometric
systems available in the market based on fingerprint, face, etc. But in today’s era where
demand for privacy and security has increased so much, we need a biometric system which
provides us very high security. Iris recognition system fulfils this demand, since this system
considers iris information for providing security. In various developed and developing
countries uses of iris for security purpose has replaced old technology of fingerprint scanning,
facial scanning, etc. In India, the aadhar card which has become the most important identity
proof for the citizens after the voter id uses the iris for identification of the person. Whatever

Page 3
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

service you go for from banking to education, aadhar card has been made mandatory in the
country for identification due to its employment of iris technology.

Iris feature offers a set of significant advantages over all different biometric features (like a
fingerprint, face), making it widely adopted for use in high efficiency and reliable biometric
methods. First of all, the iris feature describes the eye's circular region around the black pupil
& the white sclera; therefore, it is fully safe from different external conditions. Secondly, the
iris pattern is considered to provide a great extent of individuality, so randomness and no pair
iris patterns may remain the same, even irises for similar twins, or a person's right & left
eyes. Thirdly, each iris characteristic offers a higher level of stabilization during a human's
life from one year of age till death. Finally, this is viewed as the best reliable biometric
function toward deceitful approaches and hacking threats via an imposter. All efforts to vary
its patterns and surgery are a strong risk, whereas the fingerprint feature is considerably
simpler to manipulate. Despite these benefits, the iris identification method's implementation
is a difficult issue according to the iris acquisition process that may acquire insignificant parts
like eyelashes, eyelids, pupils, and specular reversal, which can have a significant impact on
iris segmentation & recognition results.
This project deals with the development of system for IRIS detection and Identification for
access control of bank locker.

Page 4
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 2.

LITERATURE REVIEW:
Before we start with the project a brief literature review was carried to arrive at the problem
definition. A number or research papers were studied to arrive at the scope. The problems
definition and the currently followed approaches were studied to arrive at the problem
definition. The research work of some the research scholars is given in this chapter.

[1] This system is proposed to enhance the security of the present bank locker security
system. As present system is vulnerable, for the hackers and thieves. The RFID in the system
which is presently at use where the locker can be opened by any unauthorized person if they
have the RFID tag which could be easily stolen or lost. So, to overcome this we are making
use of the face and voice to open the locker in this system. This system requires three main
components for its working, firstly we make use of the webcam which is interfaced with
raspberry pi for the facial recognition of the client and lastly, we make use of the node MCU
which is interfaced with raspberry pi for the voice recognition. Working of the proposed
system takes places, where in the first step the camera which is interfaced with raspberry pi
collects the data and stores it in the database by giving it a unique id to it, which should be
matched to open the locker. Similarly, the voice recognition is done with google assistant that
is interfaced with MCU. Hence, this system is highly secured and consumes low power for its
operation, it would be best advancements that could be implemented in the bank lockers so
that the security for the client’s valuables would get increased.

[2] The devices or the system which can be monitored without the human interface outside
the system is the main objective of the proposed security band vault system. The system has
two layers of security which makes its security strong. Vault door is secured with a numeric
and biometric authentication. The monitoring of the vault is carried out both outside and
inside its surrounding by the different sensors such as the PIR and the IR sensors. These
sensors send the continuous data of the system which are interfaced by the Arduino which is
the main controller of the proposed system. It plays the prominent role by receiving the data
from the interfaced devices and allows the rightful person to open the locker. The continuous
updates also overcome many of the human errors caused by the present system. In future

Page 5
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

days, many of the aspects could be added to the system making bank locker the most secured
and reliable one to the clients to keep their valuables.

Arvasu Chikara et.al[3] The main objective is to deal with “security” of the money in the bank
by enhancing the existing criteria by designing an advanced technology. In earlier days, only
mechanical locks were used to protect the money or the golds kept in the bank. As
technology grew electronic locks were introduced like password locking system which was
used as the verification factor. But both the techniques led to theft and unauthorized accesses.
So, in order to overcome these circumstances, we’re introducing Locker Security System
which is totally based on Face Recognition, voice detection and GSM (Global System for
Mobile) generation where, only the legal character can get admission to the treasured things
kept in the bank. The wireless digital camera is used to seize the image whose statistics is
processed by means of the MATLAB. Where in, the ARDUINO holds the manipulation over
the instructions. The mic is used for voice recognition. Motor motive force is used to open the
door lock with the use of DC motor. If an unauthorized person tries to open the locker, the
information is dispatched to the person. If the authorizer accepts-the locker opens, else the
information will be transformed to the police station and financial institution starts to alert.

Divya M N et.al[4] The main purpose of this system is to design and implement a high
security system based on fingerprint and image processing which can be organized in bank.
Here we are using automation of the banking locker with improved security features and
incorporated multiple level of security such has face recognition and fingerprint. Initially if
the person want to access, at the main door of locker his face will be captured by the camera
installed at the main gate, the image will be stored in the PC where MATLAB software will
compare the image with the image which is stored in data if it seems to be matched only then
the locker will be open otherwise it remain closed. The component used in this project
Microcontroller ATMega32, Fingerprint scanner and module, Webcam, LCD Interface,
Serial Interface. Face recognition technology. The main aim of this project is to improve the
locker management system in the bank with image processing technique which avoids the
manual interaction of person to maintain the locker for keeping it more safe. Fingerprint
authentication hardware setup which work in sink with image processing program.

Page 6
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

Sandip Dutta et. al[5] Current methods of protecting money in banks are not so secure that
they can be forged or simulated. Thieves are getting smarter. Most banks use the same
methodology lockers, but the current method are not safe. Thus, to overcome this, the new
method was proposed which does not require manual labor. Now if the bank customers like
to open a bank locker, then they can directly go for the IRIS scan and VEIN recognition
system check, after this process the customer will get their secret code with any government
identification such as aadhar number or elector identification number etc as proof of
authentication through which the verifications are done. Biometrics is used because of its
uniqueness for each and every human being. Once registration is complete, the biometrics
collected is stored in the form of a template for the future conformation. The IRIS scanner is
much more efficient as it catches the photo of the person's iris using infrared light. Infrared
light is used to separate the pupil and iris. As we all know iris is unique and veins are also
unique even for the identical twins. In vein scanner we use closed infrared lights to get vein
structure.

Nianfeng Liu et.al [12] suggested a code-level plan in heterogeneous iris recognition. The
non-linear connection betwixt binary elements codes of heterogeneous iris images was
demonstrated via an adapted Markov network. Additionally, a weight map on the
dependability of binary codes in the iris template can be gotten from the model. Broad
exploratory results of matching cross sensor, high-resolution versus low-resolution and, clear
versus obscured iris images exhibited that the code-level approach can accomplish the most
noteworthy accuracy in contrasted with the current pixel-level, feature level besides score-
level solutions.

Sheng-Hsun Hsieh et.al [13] recommended a new hardware-software hybrid strategy to build
the stand-off distance in an iris recognition framework. When designing the framework
hardware, they utilized an improved wavefront coding procedure to expand the field depth.
To recompense for the image obscuring caused by wavefront coding, on the software side,
the presented framework utilized a local patch centered super-resolution strategy to
reestablish the obscured image to its clear variant. The presented framework can expand the
catch volume of a regular iris recognition framework by three times and keep up the
framework's high recognition rate. But still, attributable to the hardware implementation

Page 7
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

Chun-Wei Tan et.al [14] researched a promising iris encoding accompanied by a matching
system for the noisy iris images obtained at a distance in the company of under less
compelled conditions utilizing visible accompanied by NIR imaging. The approach exhibited
in their work concurrently exploited the iris features computed from both the computed and
worldwide feature portrayals. The supremacy of the projected iris encoding along with
coordinating methodology has been decided by furnishing examination with numerous
contending iris encoding together with corresponding algorithms on this three freely
accessible databases: UBIRIS.v2, CASIA.v4-distance, FRGC, which recommend upgrades of
36.3%, 32.7% and29.6% in equal error rates(EER), individually, as contrasted with a quite
few contending approaches. The restriction of their work is, they don't consider the
multimodal procedure.

Page 8
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 3.

PROBLEM DEFINITION:
The current authentication system uses manual approach for dual key verification in the bank
lockers which is prone to errors and the bypass of the security is possible with little efforts.
After the literature review studied some techniques for password based authentication and
fingerprint as well as face recognition based authentication are shown or suggested which yet
lack somewhere or have their own disadvantages. It is practically difficult for the bank
security guards to determine the security breach and hence there needs to be a biometrically
secure system for bank lockers which is difficult to breach. Considering the current problems
, this motivates us to develop an IRIS recognition and authentication system which can be
used for the access to the bank lockers using biometric scanning of eyes. This can be a
suggested better solution to the problems faced.

Page 9
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 4.

OBJECTIVES:
The main objective of the project is to develop an IRIS detection and recognition system for
biometric authentication of the bank locker. The objectives of the project are:

1. To implement an IRIS recognition system which uses eye iris images captured by the
IRIS camera to detect and verify the authenticity of the person on the basis of IRIS
scanning
2. To develop an hardware based system using Raspberry pi and IRIS camera which will
scan the IRIS image and verify the identity of the person by matching it with the
stored user data
3. To develop an Bank Locker access control system which will demonstrate the use of
IRIS recognition system to authenticate and grant access to the bank locker using
IRIS scanning system

Page 10
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 5.

SYSTEM OVERVIEW

5.1 ARCHITECTURE DIAGRAM

The figure below shows the architecture diagram of the project

Page 11
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

5.2 WORKING PRINCIPLE

The figure shows above shows the architecture diagram of the system. As shown in the
architecture diagram, the system consists of development of Iris recognition algorithm for
identity authentication and implementing the same for the biometric access control of the
bank lockers. As shown in the architecture diagram, the IRIS camera is interfaced to the
raspberry pi to capture the IRIS images. The Iris images captured is then pre-processed and
fed to the IRIS Recognition system developed. Dougmans algorithm is used for the location
of the IRIS in the captured image and the feature extraction is done. The template matching is
done for the recognition and identity authentication for the captured IRIS image. If the
identity is verified then the Raspberry pi then uses the same to provide biometric access
control the bank locker. A model of bank locker with electromechanical solenoid lock is
developed. After the identity is verified, the bank locker will be granted access and unlocked
using the electromechanical solenoid lock. The LCD display is interfaced to the raspberry pi
to give the output messages regarding authentication as well as buzzer is used to indicate
system errors.

Page 12
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

The figure below shows the illustrative diagram of the system to be developed.

Page 13
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 6.
METHODOLOGY:
The brief methodology to carry the project is divided into number of different phases which
will be carried out step by step to so that the project can be accomplished with minimum
errors. The brief methodology to carry out the project is as given below.

1. Literature review and problem definition

2. Material survey and selection
3. Interfacing the IRIS camera to the Raspberry Pi for capture of the IRIS images
4. Development and implementation of Algorithms for detection, localization and
recognition of the IRIS images
5. Implementation of the Dougmans algorithm for the location of the IRIS in the
captured images
6. Feature extraction and matching to compare the features in the IRIS and match
them with the existing user data to recognize and authenticate the detected IRIS
and map to user ID
7. Development of bank locker prototype for demonstration of the biometric secure
IRIS based bank locker
8. Interfacing LCD display and buzzer to Raspberry Pi
9. Hardware Development
10. Assembly and programming
11. Conclusion

Page 14
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

The entire project is divided into following modules.

 The IRIS enrolment module:

This module consists of the enrolment system for registration of the new users in the
IRIS recognition system. The enrolment systems helps the new users to be registered
for the bank locker system. The enrolment system consist of capturing the multiple
IRIS images of the user to be registered and storing them under the unique ID which
will mapped for real-time IRIS identification
 The IRIS scanning and Verification Module:
The IRIS scanning and verification module consist of recognition and verification as
well as identification of the user on the basis of IRIS. The IRIS image is captured and
processed to map and verify against the existing users enrolled by the enrolment
module to verify the identity of the user.
 The bank locker Access control system:
This module consist of the development of the secure bank locker using IRIS
recognition and identification system. A bank locker which can be accessed using
electromechanical solenoid lock is developed in this module. The Access to the bank
locker is given post successful scanning and verification of the IRIS which will then
grant access to the valid user.

Page 15
BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

The flow diagram for the IRIS recognition is given below:

Page 16
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 7.

HARDWARE AND SOFTWARE REQUIREMENTS:

7.1 HARDWARE REQUIREMENTS:

 Windows 10/Linux OS, Core i5 or higher with 8GB of RAM

 Raspberry Pi 3b+

 IRIS Camera Module

 Buzzer

 LCD display

 I2C module

 Power supply

 Electro-mechanical Solenoid Lock

7.2 SOFTWARE REQUIREMENTS:

 Spyder

 Anaconda Navigator

 OpenCV

 Terminal Application

Page 17
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

7.3 LANGUAGES USED:

 Python 3.7+

Page 18
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 8.

ADVANTAGES AND DISADVANTAGES

8.2ADVANTAGES:
1. Lockers are authenticated using IRIS scanning hence change of fake entry is rare
2. No need of liveliness detection as in case of face recognition which is easier to cheat
3. Only the person with valid retina scan will be provided access to the locker making it
completely fail proof and secure
4. Lower response time and higher computational hardware not required

8.2 DISADVANTAGES:

1 Specialized IRIS capture Equipment needed and Normal cameras cannot be used

Page 19
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 9.

APPLICATIONS
1. Can be used in all banks
2. Can be used in lockers for high security areas
3. Can be used for IRIS recognition and verification in high security areas
4. Can be used everywhere where secure biometric is required.

Page 20
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

CHAPTER 10.

EXPECTED OUTCOMES:
The proposed project is expected to provide a better approach for IRIS detection and
Verification system for bank lockers which will make them more secure. The proposed
project is expected to provide with a standalone IRIS scanning and authentication system
implemented in the bank locker access control system which will capture the IRIS image
using the IRIS camera present on the bank locker and perform user authentication to grant
access to the bank locker using biometrically secured electro mechanical lock. Since the
system is stand alone it can be easily implemented in all bank lockers with little or no
modification.

Page 21
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

REFERENCES

1. High Protection Voice Identification based bank locker security system with live
image authentication International Journal for Research in Applied Science &
Engineering Technology(IJRASET)
2. “An IoT based Bank Locker Security System” International Journal of Engineering
Research & Technology(IJERT). ECLECTIC-2020 Conference Proceedings
3. Arvasu Chikara,” Smart Bank Locker Using Fingerprint Scanning and Image
Processing”, 6th International Conference on Advanced Computing &
Communication Systems (ICACCS),2020
4. Divya M N, “ High protection bank locker security system using live image and voice
authentication”, Department of ECE, Reva University,2019. International journal of
scientific Research and review.
5. Sandip Dutta,”Microcontroller Based Bank Locker Security System Using IRIS
Scanner and Vein Scanner”, International Conference on Inventive Research in
Computing Applications (ICIRCA),2018
6. Raj Gusain, “Enhancing bank security system using Face Recognition, Iris Scanner
and Palm Vein Technology”,2018
7. “Voice Recognition system”: speech-to text. Article in journal of applied and
fundamental sciences November 2018 Prerana Das ,Kakali Acharjee
8. N. Anusha, “Locker security system using Facial Recognition & OTP”, computer
science and engineering sathyabama university chennai, 2017.
9. “Human identification using iris recognition” Kalyani R. Rawate, Prof.P.A.Tijare of
Engineering and Technology, Maharashtra, India 2017
10. S V Tejasvi “Intellectual Bank Locker security system” ET. AL Int. Journal of
engineering and research application ISSN: 2248-9662, vol.6, issue2, February 2016
11. Avinash. D. Harale, “Iris as a Biometrics for Security System '', SKN sinhgad college
engineering korthi, Maharashtra, india, 2016.
12. Nianfeng Liu, Jing Liu, Zhenan Sun and Tieniu Tan, "A Code-level Approach to
Heterogeneous Iris Recognition", IEEE Transactions on Information Forensics and
Security, Vol. 12, No. 10, pp. 2373 – 2386, 2017.

Page 22
 BIOMETRIACALLY SECURE BANK LOCKER USING IRIS
RECOGNITION

13. Sheng-Hsun Hsieh, Yung-Hui Li, Chung-Hao Tien, and Chin-Chen Chang,
"Extending the capture volume of an iris recognition system using wavefront coding
and super-resolution", IEEE transactions on cybernetics Vol. 46, No. 12, pp. 3342-
3350, 2016.
14. Chun-Wei Tan and Ajay Kumar, "Efficient and accurate at-a-distance iris recognition
using geometric key-based iris encoding", IEEE Transactions on Information
Forensics and Security Vol. 9, No. 9, pp. 1518-1526, 2014.

Page 23