ABSTRACT

There is many type of security verifications present. Thus many system present there is no
proper secured system present. So actual aim of deep learning based feature extraction and
recovering for finger-vein verification using biometric is to extract and recover vein features
using limited a priori knowledge. To automatically discard the ambiguous region and to label
the pixels of clear region as foreground or background. And to recover missing finger-vein
patterns in the segmented image. Such a scheme avoids the heavy manual labeling and may
also reduce label errors, especially for ambiguous pixels. Thus this model is able to extract
the vein patterns from raw images in robust way, which leads to significant improvement in
finger-vein verification accuracy. we perform a rigorous experimental analysis that shows
that our scheme does succeed in recovering missing patterns which further improves the
verification performance.

i
 CONTENTS

CHAPTER NO TITLE PAGE NO

Acknowledgement
Abstract
List of Figures
List of Abbreviations
1 Introduction 1
1.1 Overview 1
1.2 The Role of Personal Recognition 2
Technology
1.3 Biometric based personal identification 2
system
1.4 Advantages of Biometric System 4
1.5 Disadvantage of Biometric System 7
1.6 Finger vein based authentication tool 11
1.7 Machine learning 12
2 Literature Survey 24
2.1 Introduction 24
2.2 Finger vein data acquicition 24
2.3 Human Identification using Finger Images 29
2.4 On-Line Fingerprint Verification 32
2.5 Personal Authentication Using Hand Vein 33
2.6 Analysis of Literature work 33
3 Existing system 34
3.1 Existing System 34
3.2 Disadvantages of Existing System 34
3.3 Result and Discussion 35
3.4Conclusion 36

4 Proposed system 37

ii
 4.1 Modules 37
4.2 Acquisition of an Infra-red image of the 39
finger
4.3 Normalization of the image 39
4.4 Extraction of Finger Vein patterns 39
4.5 Matching 40
5 System Requirements 41
5.1 Hardware Requirements 41
5.2 Software Requirements 41
5.3 Software Environment 42
5.4 Different modes of programming in Python 42
5.5 Flask frame work 43
6 System Study
6.1 Feasibility Study 44
6.2 System Testing 45
6.3 Types of Test 45
6.4 System Design 45
6.5 Data flow diagram 47
6.6 Use Case Diagram 48
7 Conclusion 49
Reference 50

iii
FIGURE NO TITLE PAGE NO
1.1 Categories of Biometrics 3
LIST OF
1.2 Machine Learning Algorithm 14
2.1 Finger Vein Image Acquiring Method 26 FIGURES

6.1 System Architecture 50

6.2 Data flow Diagram 51

iv
 LIST OF ABBREVATION

FVRS - Finger Vein Recognition System

FAR - False Acceptance Rate
FRR - False Rejection Rate
ROI - Region Of Interest
NIR - Near Infra-red
ML - Machine Learning
CLAHE - Contrast Limiting Adaptive Histogram Equalization
NDVI - Normalized Difference Vegetation Index
PCA - Principle Component Analysis
SVD - Singular Value Decomposition
LBP - Local Binary Pattern
LDP - Local Derivative Pattern
CN - Cross Number
LEBP - Local Extensive Binary Pattern
SVM - Support Vector Machine
CNN - Convolutional Neural Network
FCN - Fully Convolutional Network
GPL - General Public License
ICP - Increased Intracranial Pressure
EER - Energy Efficiency Ratio

v
 CHAPTER - 1
INTRODUCTION
1.1 INTRODUCTION TO FINGER-VEIN VERIFICATION
Finger veins are hidden under the skin where red blood cells are flowing. In biometrics,
the term vein does not entirely correspond to the terminology of medical science. Its network
patterns are used for authenticating the identity of a person, in which the approximately 0.3 to
1.0 mm thick vein is visible by near infrared rays. In this definition, the term finger includes
not only index, middle, ring, and little fingers, but also the thumb. Moreover, finger vein
system have some very powerful advantages. First, there is no property of latency. The vein
patterns in fingers stay where they belong, and where no one can see them- in the fingers.
This is a huge privacy consideration.

1.2 BIOMETRIC BASED PERSONAL IDENTIFICATION SYSTEM

Vein recognition is generally used in high risk deployments where space is not an issue.
Physical access control and corporate banking are verticals where the modality is more
common. But, like all biometrics, advances in mobile technology and innovations in design
and manufacturing are contributing to a broadening the areas of application for the vascular
modality. Mobile devices are beginning to surface with vein recognition capabilities. Fujitsu’s
vein scanning tablet can authenticate users via palm vein, and multiple smartphones from
ZTE are shipping with Eye print ID software, which uses a smartphone camera to capture
vein patterns in a user’s eye. There is even a rumor circulating, based on a recent patent filing
by Apple, that the Apple Watch 2 will sport some sort of vein recognition.

1.2.1 SECURE

As finger vein patterns are found internally within the body, forgery is extremely
difficult. Dryness or roughness on the surface of the skin also has no effect on the accuracy of
vein pattern authentication.

1.2.2 ACCURATE

Rates for acceptance of false users or rejection of true users are among the lowest for
biometric technologies, making finger vein authentication a reliable security solution

1.2.3 FAST

1
 Vein pattern matching is completed within the blink of an eye, affording users a
speedy authentication experience without the hassle and without the wait.

1.2.4 SMALL

Finger vein authentication devices are compact and therefore applicable as embedded
devices in a variety of applications.

1.2.5 USER-FRIENDLY

The vein patterns of each finger are unique, so each individual can register multiple
fingers as "back-up" for authentication purposes. Registration is possible even for sweaty,
oily or dirty fingers.

1.3 BIOMETRIC BASED PERSONAL IDENTIFICATION

SYSTEM

In order to distinguish a person inherently as authorized against an impostor,

biometric based personal identification systems is need of an hour. Since this system
identifies an individual based what he or she is rather than what he or she has. This
system is much less prone to forgery owing to its high level of distinctiveness, secured
and dependable (Wang 2010). With all these added advantages this system has gained
much popularity as one of most preferred personal identification measures (Huang
2017). This technology is defined as “to identify an individual based on statistical
measurement of human behavioral and unique physiological characteristics”.

The various categories of biometric systems as given below:

Figure 1.1: Categories of Biometrics

2
1.3.1 Emerging Trend in Biometric Processing

For the purpose of authentication and identification, biometric is a methodological

study of assessing and analyzing biological data. While comparing with the visual
identification methods using signatures and photographic identities, the identification
process based on biometric will be much less time consuming and more accurate.
This helps in application of bio-metric technology in diverse field of authentication
which can guarantee highly secure identification and authorization system. Hand based
biometric system using finger print or palm print is the one most commonly used in the
recent past.

However, this system is more exposed to forgery as finger prints are exposed to
other and rather sweat / dryness can hinder obtaining clear image of the same.
This leads to degradation of system performance, can be even replicated with little
effort and use of technology. To overcome this drawback, an emerging process
was identified which can use the finger vein pattern to identify individual.

1.4 ADVANTAGES AND DISADVANTAGES OF BIOMETRIC

SYSTEM

1.4.1 ADVANTAGES OF BIOMETRICS SYSTEM

Day by day with the turnaround advancement of Biometric technology all around
the world it is gaining more and more popularity. Biometric based personal identification
solutions are widely accepted by all strides of organizations, namely- government agencies,
multinational organizations, institutions, banks, hospitals, etc. The need for this technology
is budding towards almost all sectors of modern society like National Identification
Scheme, Banking and Finance Sectors, Employment and Social Security Networks, etc.
The common population of any country, is having much further confidence on the
contemporary biometric technologies relatively than on traditional biometric system as
there are numerous advantages envisaged in this system which are briefly mentioned
below:

1.4.2 SECURITY
Over the decades, the passwords are comprising of numbers, alphabets, symbols,
etc. and are easy to crack. Many hacking incidents happening every year which results in
loss of information, money, privacy, etc. constantly.

3
With the implementation of Biometric technology based identification, it will not be
possible to hack as that of passwords which is of great help, specifically for business
communities and financial sectors, who are day in and day out fighting with security
complications over a extensive period of time.

1.4.3 ACCURACY
The traditional biometric based security systems are estimate huge amount of money,
time and valuable resources.

We may also understand that the securing systems using passwords, smart cards, personal
identification number were not accurate always and tends to have problems of hacking
and misuse.

However, identification system based on biometrics which uses physical traits – Vein
patterns, retina scanning, finger print, etc. will always provide a better results which are
very much accurate anytime, anywhere.

1.4.4 ACCOUNTABILITY
With the traditional security system, the codes and passwords can be easily broken
by anybody and hack the personal information. This problem causes more risk and will
be a continuous one. In case of biometric security system, the accountability of
persons using can be tracked and accounted for any misuse or abuse, thus allowing
100% accountability for the safety of information.

1.4.5 CONVENIENT
In the case of traditional security system, individual experience in their day to day
life, of forgetting the passwords, secret codes, etc. Individuals tend to undergo nerve-
wrecking procedures to retrieve the same and the process is cumbersome. Even though,
we follow some handy methods to remember the codes and passwords, the biometric
system will always pose a handy solution, which is most convenient and reliable at times
of need as these credentials being used in biometric system are with the individual
forever and doesn’t require any hard-core remembrance or maintenance.

1.4.6 SCALABILITY
Notwithstanding with the need for which the biometric systems is implemented or
applied, it can be scaled without losing its credential, used for diversified applications
like Projects related to Government Schemes, security systems for banks, management
of workforce, etc. This scalability is not available with the traditional security systems.

4
1.4.7 ROI
As compared to traditional security systems, biometric solutions provide the best
return on investment. The capital expenditure incurred on maintenance and monitoring of
traditional systems will be much more and the gestation period also will be long as
compared to biometric security systems.

But in the case of modern biometric system, the investment is much lesser and has quicker
gestation period.

1.4.8 FLEXIBILITY
The flexibility of modern biometric system is much more as compared to traditional
security systems. The system being held and monitored will be based on our own natural
resources which definitely reduces the criticality memorizing numbers, alphabets,
symbols being used in traditional system.

1.4.9 TRUSTABLE
With the development of modern and handy technology being used in biometric
systems, makes it user friendly and trustable one. The security and reliability of access
control systems using biometric had become more trustable with financial, security and
social organizations for its operations.

1.4.10 SAVE TIME

The time consumed in biometric systems for identification process is much lesser
as compared to traditional methods. The user just need to put their finger, palm or face
to the biometric device which scans much faster than typing passwords.

1.4.11 SAVE MONEY

Due to change in government policies regarding biometric databases for
individuals for its maintenance and accessibility, the users are not required to create and
maintain their own databases, which cost money.

The existing databases available at various social organization can be enhanced for its
usage, to get accurate details about the individuals at any given point of time. This helps the
corporates and organization, to avoid spending huge capital money, for maintenance of
biometric details of their employees.

5
With the consideration of above advantages, we can determine that with the
development of biometric technology and process, there will be leaps of enhances in the
personal identification process.

Which clearly outdates the traditional method of recognition system. Many developed and
developing countries had already implemented the biometric based identification process
for the regular business and social applications, to take the advantage over traditional
methods.

1.5 DISADVANTAGES OF BIOMETRICS SYSTEM

The invention of biometric technology had brought much significant impact on our
day today life. Due to be modern biometric based identification system, our quality and
security of life had improved in many aspects. It tends to have its own advantages
and disadvantages as compared to traditional system traits.

This was well quoted by prominent researchers that “Great Power brings it along with
greater responsibilities” which is applicable to biometric identification technology also.

Even the biometric technologies have all the goodness surrounding and considered as one
of the positive development in technology, it also has its other darker side with some
disadvantages of its own. The introduction of biometric technology brings in many
benefits for individuals and society as a whole.

1.5.1 PRIVACY CONCERNS

Biometric data, such as fingerprints or facial scans, raises privacy issues as it is highly
personal. If compromised, individuals may be at risk of identity theft.

1.5.2 ACCURACY AND RELIABILITY

Biometric systems are not foolproof. Factors like environmental conditions, quality of
sensors, and variations in individuals' biometric features can lead to false positives or
negatives.

1.5.3 COST

Implementing biometric systems can be expensive due to the need for specialized
hardware and software. This cost can be a barrier for widespread adoption, especially for
smaller organizations.

6
1.5.4 INHERENT BIASES

Some biometric technologies may exhibit biases based on race, ethnicity, or gender.
This can lead to unfair outcomes, potentially discriminating against certain demographic
groups.

1.5.5 SECURITY CONCERNS

Biometric data, once compromised, cannot be changed like a password. If stolen,

individuals may face lifelong consequences, as they cannot reset or update their biometric
features.

System Vulnerabilities: Biometric systems can be vulnerable to hacking or spoofing attempts.

Techniques such as creating synthetic fingerprints or using high-quality photographs may
compromise the security of the system.

1.5.6 LACK OF STANDARDIZATION

There is a lack of universal standards for biometric data, leading to interoperability

issues. Different systems may not easily work together, hindering their widespread adoption.

1.5.7 SOCIAL ACCEPTANCE

Some people are uncomfortable with the idea of having their biometric data stored
and used, fearing misuse or unauthorized access. This lack of social acceptance can impede
the successful implementation of biometric systems.

1.5.8 LEGAL AND REGULATORY CHALLENGES

The legal landscape surrounding biometric data is complex and varies across
jurisdictions. Organizations must navigate through regulatory requirements, which can be
challenging and time-consuming.

1.5.9 LIMITED REVOCABILITY

Unlike passwords or PINs, biometric features cannot be easily revoked or changed. If

compromised, individuals may face challenges in securing their identity effectively.

1.5.10 PHYSICAL TRAITS ARE NOT CHANGEABLE

In the traditional method of security, individuals have the comfort of changing their
secret code, passwords, smart cards and other details as and when they want. But in case
of biometric based security system, this comfort is not available to the users. The present

7
biometric modalities will exertion ideally with physical personalities of human being
such as finger veins & print, iris, palm, etc. and cannot be changed for the comfort of
individuals. Cannot be changed by means due to his / her ageing, climate and
environment.

1.5.11THE EYES, FINGER PRINTS AND OF INDIVIDUALS

The passwords can be reset or cracked, but the physical traits of human beings like
fingerprints or retina are fixed and cannot be changed. Ideally the biometric data /
information collected are stored in secured Databases or with organizations designed to
provide such services. With the recent developments of sharing the data with third parties
without the consent of individuals, will impact the security of information being collected
and stored using biometric system, even if the individuals wants to change the same, it
is not possible.

1.5.12 ERROR RATE

Any biometric devices being used in the process of individual identification can
make two types of errors - False Rejection Rate (FRR) and False Acceptance Rate (FAR)
.When a unauthorized individual is accepted by the system then the same is called as
FAR and when rejection is happed then it is recognized as FRR. Due to the criticality
and complex process involved in biometric system, the chances of error is much more as
compared to traditional security system, which renders the error rate is higher. The
biometric systems can also have the effects of climate, position, physical condition, ageing
of individual and other complicated issues which caused turmoil in a large scale
process of authentication.

1.5.13 COST
The devices being used, software to be developed, hardware to be procured and
maintained, involves huge cost as compared to traditional method of security systems. A
huge of amount of money is to be spent for building and maintenance of the infrastructures
required for the fool proof biometric system.

1.5.14 DELAY

The process of scanning, processing and identifying an individual is a cumbersome

process as it depends on the individual position, weather, temperature etc. will cause
considerable delay in the identification process as compared to traditional method of
entering password, PIN number, showing digitalized cards, etc. In an organization, where

8
large number of employees needs to enter and exit every day, will cause considerable delay
in the accessing and identification process if biometric system is used .

1.5.15 COMPLEXITY
The biometric system being deployed today for identification process is much
complex and highly technical in nature which an added disadvantage.

A new entrant will always have difficulty to understand the operational process of
whole system. The organization need to employ highly technical and trained professional
for the development and implementation of the biometric based identification systems,
which adds further to its complexity

1.5.16 IDENTIFICATION

The biometric system being deployed today for identification process is much
complex and highly technical in nature which an added disadvantage. A new entrant will
always have difficulty to understand the operational process of whole system. The
organization need to employ highly technical and trained professional for the
development and implementation of the biometric based identification systems, which adds
further to its complexity.

1.5.17 DISTRIBUTION
Ideally the biometric systems being deployed to identify a person is based on
contacts with the equipment and considered as one of the unhygienic environment. Often
individuals having infectious disease, will tend to spread the same to other individuals /
employees who are using the same biometric devices for identification process. This is
most disturbing disadvantage of biometric system and needs to be addressed.

1.5.18 SCANNING DIFFICULTY

The implementation of biometric devices develops scanning difficulties especially
in the case of iris scanning. For example, eyelashes, eyelids, contact lenses being used and
reflections of the retinal causes difficulties in the scanning process and make it a difficult
process in personal identification.

1.5.19 PHYSICAL DISABILITY

Persons having physical disability, will have embarrassing situation while using the
biometric devices. For individual having lost their fingers / eyes, the use of biometric
systems will have serious psychological disturbance both physically and mentally. It will be

9
much hard time for these individuals to pass over the biometric based identification
process every time.

1.5.20 ENVIRONMENT AND USAGE MATTERS

The biometric parameters are borne to be affected by environment factors, which
cause significant impact in overall identification process. Climatic conditions in extreme
cold and hot areas create chaos in the process of biometric based identification process,
as the error rates will be high due climatic conditions which results in unreliability of the
total system.

1.5.21 ADDITIONAL HARDWARE INTEGRATION

The requirement of additional hardware integration is mandatory for the biometric
modalities which is to be used in the process of personal identification are costly,
inconvenient and complex. Even though they are much costlier as well we hard to manage.

Any new technology or development, will prone to its own criticism which is also
applicable to state of art biometric systems. Easy process / system will have its own
advantages and disadvantages and the magnitude of these will decide on the
implementation factor. With the stated advantages, we cannot conclude that
biometric based identification of fool proof and user friendly, however, considering only
the disadvantages, we should not conclude that this system is not usable. Considering its
own merits and demerits, we should evolve an ideal possibility to use of modern
biometric based identification system having hassle free and user friendly.

1.6 FINGER VEIN BASED AUTHENTICATION TOOL

Using finger vein as a personal authentication tool become an emerging trend in
the model world of technology. It is has attracted considerable attention among
researchers, engineers and professionals as an ideal tool in the process of personal
authentication. As compared among the traditional and biometric technology, the use of
finger vein based authentication process is one of the state-of-art tool which is much secure,
user friendly and cost effective. This new technology is fool proof, as the vein pattern
cannot be duplicated and misused, which can be easily captured without much effort
from individuals as well equipment (Backman 2010). In the recent epoch, finger vein
based substantiation became one of the most dependable and auspicious solution for the
identification process.

10
 Even though, finger vein based authentication process is an eye opener in the
field of personal identification system, it do have some bottlenecks of difficulties in
capturing vein images due to IR light parameters, outer skin appearance (skin decease,
dust, uncleanliness, etc.), uneven illumination, pressure, environmental conditions etc.
Moreover, the vein images being captured also tend to be affected by noise, distortion,
scaling etc., (Sprawls 1993). With the constraints and bottlenecks, the technology for
biometric authentication process using finger vein is the most promising solution in the
near future.

To overcome its bottleneck, we need to have a new approach called Region of

Interest (ROI), which is an emerging concept in the finger vein capturing process.

The successful of any authentication system depend on the robustness of ROI

identification and detection. Many researchers had published many novice methods, to
enhance the performance of the vein based authentication process, however based on ROI
identification is noteworthy to mention. Most of the research works focused on finger
vein capturing methodology, its processing techniques using mathematical and statistical
algorithm, representation methods, vein image enhancement modalities and multimodal
recognition system. Very few or less research work is carried out in the process of ROI
based finger vein authentication process, which is one of area need much more exploration,
to have robust system (Zhang 2007).

Over above the external factor, we can also infer that factors relating to collection
process of vein images like translation, orientation, rotation, finger pressure, image scale
and collection postures can also affect its features. Due to these factors, we need to
identify, robust ROI localization which is decisive for a finger vein based personal
identification system.

Finger vein based identification method as biometric tool, will have various
advantages over traditional ones, as it has unique vein patterns which cannot be forged.
It is very hard to replace the vein pattern of finger with forged ones due to its complex
internal features This tool will have high accuracy ratings and the equipment being used
for this process is much smaller as compared to equipment used for finger print or palm
print. We can acquire the finger vein patterns using Near Infra-Red (NIR) light and a
camera device. From the hygienic point of view, this method is much safer as there will not
be any physical contact between the sensor and finger. The finger vein technology had
provided significant performance development towards personal identification system over

11
a span of many years. Based on existing research works, we can infer that not only external
factors like sprinkling, patchy illumination, environment and temperature can affect its
features. Over above the external factor, we can also infer that factors relating to collection
process of vein images like translation, orientation, rotation, finger pressure, image scale
and collection postures can also affect its features. Due to these factors, we need to
identify, robust ROI localization which is decisive for a finger vein based personal
identification system.

1.7 MACHINE LEARNING

Machine Learning is a system of computer algorithms that can learn from example
through self-improvement without being explicitly coded by a programmer. Machine learning
is a part of artificial Intelligence which combines data with statistical tools to predict an
output which can be used to make actionable insights.
The breakthrough comes with the idea that a machine can singularly learn from the data (i.e.,
example) to produce accurate results. Machine learning is closely related to data mining and
Bayesian predictive modeling. The machine receives data as input and uses an algorithm to
formulate answers.
A typical machine learning tasks are to provide a recommendation. For those who have a
Netflix account, all recommendations of movies or series are based on the user's historical
data. Tech companies are using unsupervised learning to improve the user experience with
personalizing recommendation.
Machine learning is also used for a variety of tasks like fraud detection, predictive
maintenance, portfolio optimization, automatize task and so on.

1.7.1 MACHINE LEARNING VS. TRADITIONAL PROGRAMMING

Traditional programming differs significantly from machine learning. In traditional
programming, a programmer code all the rules in consultation with an expert in the industry
for which software is being developed. Each rule is based on a logical foundation; the
machine will execute an output following the logical statement. When the system grows
complex, more rules need to be written. It can quickly become unsustainable to maintain.
Traditional programming differs significantly from machine learning. In traditional
programming, a programmer code all the rules in consultation with an expert in the industry
for which software is being developed. Each rule is based on a logical foundation; the
machine will execute an output following the logical statement. When the system grows
complex, more rules need to be written. It can quickly become unsustainable to maintain.

12
1.7.2 TRADITIONAL PROGRAMMING
Machine learning is supposed to overcome this issue. The machine learns how the input
and output data are correlated and it writes a rule. The programmers do not need to write new
rules each time there is new data. The algorithms adapt in response to new data and
experiences to improve efficacy over time.

1.8 MACHINE LEARNING

1.8.1 HOW DOES MACHINE LEARNING WORK?

Machine learning is the brain where all the learning takes place. The way the machine
learns is similar to the human being. Humans learn from experience. The more we know, the
more easily we can predict. By analogy, when we face an unknown situation, the likelihood
of success is lower than the known situation. Machines are trained the same. To make an
accurate prediction, the machine sees an example. When we give the machine a similar
example, it can figure out the outcome. However, like a human, if it feed a previously unseen
example, the machine has difficulties to predict.

13
The core objective of machine learning is the learning and inference. First of all, the
machine learns through the discovery of patterns.
This discovery is made thanks to the data. One crucial part of the data scientist is to choose
carefully which data to provide to the machine. The list of attributes used to solve a problem
is called a feature vector. You can think of a feature vector as a subset of data that is used to
tackle a problem. The machine uses some fancy algorithms to simplify the reality and
transform this discovery into a model. Therefore, the learning stage is used to describe the
data and summarize it into a model.

For instance, the machine is trying to understand the relationship between the wage of an
individual and the likelihood to go to a fancy restaurant. It turns out the machine finds a
positive relationship between wage and going to a high-end restaurant: This is the model
Inferring
When the model is built, it is possible to test how powerful it is on never-seen-before data.
The new data are transformed into a features vector, go through the model and give a
prediction. This is all the beautiful part of machine learning. There is no need to update the
rules or train again the model. You can use the model previously trained to make inference on
new data.
The life of Machine Learning programs is straightforward and can be summarized in the
following points:
1. Define a question
2. Collect data
3. Visualize data
4. Train algorithm
5. Test the Algorithm
6. Collect feedback

14
 7. Refine the algorithm
8. Loop 4-7 until the results are satisfying
9. Use the model to make a prediction
Once the algorithm gets good at drawing the right conclusions, it applies that knowledge to
new sets of data.
Machine Learning Algorithms and Where they are Used?

Figure 1.2 Machine learning Algorithms

Machine learning can be grouped into two broad learning tasks: Supervised and
Unsupervised. There are many other algorithms Supervised learning.
An algorithm uses training data and feedback from humans to learn the relationship of given
inputs to a given output. For instance, a practitioner can use marketing expense and weather
forecast as input data to predict the sales of cans.
You can use supervised learning when the output data is known. The algorithm will predict
new data.
There are two categories of supervised learning:
 Classification task
 Regression task

15
1.9 CLASSIFICATION
Imagine you want to predict the gender of a customer for a commercial. You will start
gathering data on the height, weight, job, salary, purchasing basket, etc. from your customer
database. You know the gender of each of your customer, it can only be male or female. The
objective of the classifier will be to assign a probability of being a male or a female (i.e., the
label) based on the information (i.e., features you have collected).

When the model learned how to recognize male or female, you can use new data to make a
prediction. For instance, you just got new information from an unknown customer, and you
want to know if it is a male or female. If the classifier predicts male = 70%, it means the
algorithm is sure at 70% that this customer is a male, and 30% it is a female.
The label can be of two or more classes. The above Machine learning example has only two
classes, but if a classifier needs to predict object, it has dozens of classes (e.g., glass, table,
shoes, etc. each object represents a class)

1.9.1 REGRESSION
When the output is a continuous value, the task is a regression. For instance, a
financial analyst may need to forecast the value of a stock based on a range of feature like
equity, previous stock performances, macroeconomic index. The system will be trained to
estimate the price of the stocks with the lowest possible error Unsupervised learning.
In unsupervised learning, an algorithm explores input data without being given an explicit
output variable (e.g., explores customer demographic data to identify patterns)
You can use it when you do not know how to classify the data, and you want the algorithm to
find patterns and classify the data for you.

1.9.2 MACHINE LEARNING (ML) ALGORITHM:

There are plenty of machine learning algorithms. The choice of the algorithm is based
on the objective.
In the Machine learning example below, the task is to predict the type of flower among the
three varieties. The predictions are based on the length and the width of the petal. The picture
depicts the results of ten different algorithms. The picture on the top left is the dataset. The
data is classified into three categories: red, light blue and dark blue. There are some
groupings. For instance, from the second image, everything in the upper left belongs to the
red category, in the middle part, there is a mixture of uncertainty and light blue while the

16
bottom corresponds to the dark category. The other images show different algorithms and
how they try to classified the data.

1.9.3 CHALLENGES AND LIMITATIONS OF MACHINE LEARNING

The primary challenge of machine learning is the lack of data or the diversity in the
dataset. A machine cannot learn if there is no data available. Besides, a dataset with a lack of
diversity gives the machine a hard time. A machine needs to have heterogeneity to learn
meaningful insight. It is rare that an algorithm can extract information when there are no or
few variations. It is recommended to have at least 20 observations per group to help the
machine learn. This constraint leads to poor evaluation and prediction.

1.10 APPLICATION OF MACHINE LEARNING

1.10.1 AUGMENTATION
 Machine learning, which assists humans with their day-to-day tasks, personally or
commercially without having complete control of the output. Such machine learning
is used in different ways such as Virtual Assistant, Data analysis, software solutions.
The primary user is to reduce errors due to human bias.
1.10.2 AUTOMATION

17
  Machine learning, which works entirely autonomously in any field without the need
for any human intervention. For example, robots performing the essential process
steps in manufacturing plants.
1.10.3 FINANCE INDUSTRY
• Machine learning is growing in popularity in the finance industry. Banks are mainly
using ML to find patterns inside the data but also to prevent fraud.
1.10.4 GOVERNMENT ORGANIZATION
 The government makes use of ML to manage public safety and utilities. Take the
example of China with the massive face recognition. The government uses Artificial
intelligence to prevent jaywalker.
1.10.5 HEALTHCARE INDUSTRY
 Healthcare was one of the first industry to use machine learning with image detection.
1.10.6 MARKETING
 Broad use of AI is done in marketing thanks to abundant access to data. Before the
age of mass data, researchers develop advanced mathematical tools like Bayesian
analysis to estimate the value of a customer. With the boom of data, marketing
department relies on AI to optimize the customer relationship and marketing
campaign.

1.11 EXAMPLE OF APPLICATION OF MACHINE LEARNING

1.11.1 SUPPLY CHAIN
Machine learning gives terrific results for visual pattern recognition, opening up many
potential applications in physical inspection and maintenance across the entire supply chain
network.
Unsupervised learning can quickly search for comparable patterns in the diverse dataset. In
turn, the machine can perform quality inspection throughout the logistics hub, shipment with
damage and wear.
For instance, IBM's Watson platform can determine shipping container damage. Watson
combines visual and systems-based data to track, report and make recommendations in real-
time.
In past year stock manager relies extensively on the primary method to evaluate and forecast
the inventory. When combining big data and machine learning, better forecasting techniques
have been implemented (an improvement of 20 to 30 % over traditional forecasting tools). In

18
term of sales, it means an increase of 2 to 3 % due to the potential reduction in inventory
costs.
1.11.2 EXAMPLE OF MACHINE LEARNING GOOGLE CAR
For example, everybody knows the Google car. The car is full of lasers on the roof which are
telling it where it is regarding the surrounding area. It has radar in the front, which is
informing the car of the speed and motion of all the cars around it. It uses all of that data to
figure out not only how to drive the car but also to figure out and predict what potential
drivers around the car are going to do. What's impressive is that the car is processing almost a
gigabyte a second of data.
1.11.3 WHY IS MACHINE LEARNING IMPORTANT?
Machine learning is the best tool so far to analyze, understand and identify a pattern in the
data. One of the main ideas behind machine learning is that the computer can be trained to
automate tasks that would be exhaustive or impossible for a human being. The clear breach
from the traditional analysis is that machine learning can take decisions with minimal human
intervention.
Take the following example for this ML tutorial; a retail agent can estimate the price of a
house based on his own experience and his knowledge of the market.
A machine can be trained to translate the knowledge of an expert into features. The features
are all the characteristics of a house, neighborhood, economic environment, etc. that make the
price difference. For the expert, it took him probably some years to master the art of estimate
the price of a house.
His expertise is getting better and better after each sale.
For the machine, it takes millions of data, (i.e., example) to master this art. At the very
beginning of its learning, the machine makes a mistake, somehow like the junior salesman.
Once the machine sees all the example, it got enough knowledge to make its estimation. At
the same time, with incredible accuracy. The machine is also able to adjust its mistake
accordingly.
Most of the big company have understood the value of machine learning and holding data.
McKinsey have estimated that the value of analytics ranges from $9.5 trillion to $15.4 trillion
while $5 to 7 trillion can be attributed to the most advanced AI techniques.
Machine learning (ML) is the study of computer algorithms that improve automatically
through experience. It is seen as a part of artificial intelligence. Machine learning algorithms
build a model based on sample data, known as "training data", in order to make predictions or
decisions without being explicitly programmed to do so. Machine learning algorithms are

19
used in a wide variety of applications, such as email filtering and computer vision, where it is
difficult or unfeasible to develop conventional algorithms to perform the needed tasks.
A subset of machine learning is closely related to computational statistics, which focuses on
making predictions using computers; but not all machine learning is statistical learning. The
study of mathematical optimization delivers methods, theory and application domains to the
field of machine learning. Data mining is a related field of study, focusing on exploratory data
analysis through unsupervised learning. In its application across business problems, machine
learning is also referred to as predictive analytics.

1.11.4 OVERVIEW
Machine learning involves computers discovering how they can perform tasks without being
explicitly programmed to do so. It involves computers learning from data provided so that
they carry out certain tasks. For simple tasks assigned to computers, it is possible to program
algorithms telling the machine how to execute all steps required to solve the problem at hand;
on the computer's part, no learning is needed. For more advanced tasks, it can be challenging
for a human to manually create the needed algorithms. In practice, it can turn out to be more
effective to help the machine develop its own algorithm, rather than having human
programmers specify every needed step.
The discipline of machine learning employs various approaches to teach computers to
accomplish tasks where no fully satisfactory algorithm is available. In cases where vast
numbers of potential answers exist, one approach is to label some of the correct answers as
valid. This can then be used as training data for the computer to improve the algorithm(s) it
uses to determine correct answers. For example, to train a system for the task of digital
character recognition, the MNIST dataset of handwritten digits has often been used.

1.11.5 MACHINE LEARNING APPROACHES:

Machine learning approaches are traditionally divided into three broad categories, depending
on the nature of the "signal" or "feedback" available to the learning system:
Supervised learning: The computer is presented with example inputs and their desired
outputs, given by a "teacher", and the goal is to learn a general rule that maps inputs to
outputs.

20
Unsupervised learning: No labels are given to the learning algorithm, leaving it on its own
to find structure in its input. Unsupervised learning can be a goal in itself (discovering hidden
patterns in data) or a means towards an end (feature learning).
Reinforcement learning: A computer program interacts with a dynamic environment in
which it must perform a certain goal (such as driving a vehicle or playing a game against an
opponent). As it navigates its problem space, the program is provided feedback that's
analogous to rewards, which it tries to maximize.
Other approaches have been developed which don't fit neatly into this three-fold
categorization, and sometimes more than one is used by the same machine learning system.
For example topic modeling, dimensionality reduction or meta learning.
As of 2020, deep learning has become the dominant approach for much ongoing work in the
field of machine learning.

1.12 HISTORY AND RELATIONSHIPS TO OTHER FIELDS

The term machine learning was coined in 1959 by Arthur Samuel, an American IBMER and
pioneer in the field of computer gaming and artificial intelligence. A representative book of
the machine learning research during the 1960s was the Nilsson's book on Learning
Machines, dealing mostly with machine learning for pattern classification.
Interest related to pattern recognition continued into the 1970s, as described by Duda and
Hart in 1973. In 1981 a report was given on using teaching strategies so that a neural network
learns to recognize 40 characters (26 letters, 10 digits, and 4 special symbols) from a
computer terminal.

Tom M. Mitchell provided a widely quoted, more formal definition of the algorithms studied
in the machine learning field: "A computer program is said to learn from experience E with
respect to some class of tasks T and performance measure P if its performance at tasks in T,
as measured by P, improves with experience E. "This definition of the tasks in which machine
learning is concerned offers a fundamentally operational definition rather than defining the
field in cognitive terms. This follows Alan Turing's proposal in his paper "Computing
Machinery and Intelligence", in which the question "Can machines think?" is replaced with
the question "Can machines do what we (as thinking entities) can do?.

Modern day machine learning has two objectives, one is to classify data based on models
which have been developed, the other purpose is to make predictions for future outcomes

21
based on these models. A hypothetical algorithm specific to classifying data may use
computer vision of moles coupled with supervised learning in order to train it to classify the
cancerous moles. Whereas, a machine learning algorithm for stock trading may inform the
trader of future potential predictions.

1.13 ARTIFICIAL INTELLIGENCE

Machine Learning as subfield of AI
Part of Machine Learning as subfield of AI or part of AI as subfield of Machine Learning
As a scientific endeavor, machine learning grew out of the quest for artificial intelligence. In
the early days of AI as an academic discipline, some researchers were interested in having
machines learn from data. They attempted to approach the problem with various symbolic
methods, as well as what was then termed "neural networks"; these were mostly perceptronz
and other models that were later found to be reinventions of the generalized linear models of
statistics. Probabilistic reasoning was also employed, especially in automated medical
diagnosis.
However, an increasing emphasis on the logical, knowledge-based approach caused a rift
between AI and machine learning. Probabilistic systems were plagued by theoretical and
practical problems of data acquisition and representation. By 1980, expert systems had come
to dominate AI, and statistics was out of favor. Work on symbolic/knowledge-based learning
did continue within AI, leading to inductive logic programming, but the more statistical line
of research was now outside the field of AI proper, in pattern recognition and information
retrieval. Neural networks research had been abandoned by AI and computer science around
the same time. This line, too, was continued outside the AI/CS field, as "connectionism", by
researchers from other disciplines including Hopfield, Rumel hart and Hinton. Their main
success came in the mid-1980s with the reinvention of back propagation.
Machine learning (ML), reorganized as a separate field, started to flourish in the 1990s. The
field changed its goal from achieving artificial intelligence to tackling solvable problems of a
practical nature. It shifted focus away from the symbolic approaches it had inherited from AI,
and toward methods and models borrowed from statistics and probability theory.
As of 2020, many sources continue to assert that machine learning remains a subfield of AI.
The main disagreement is whether all of ML is part of AI, as this would mean that anyone
using ML could claim they are using AI. Others have the view that not all of ML is part of AI
where only an 'intelligent' subset of ML is part of AI.

22
The question to what is the difference between ML and AI is answered by Judea Pearl in The
Book of Why. Accordingly ML learns and predicts based on passive observations, whereas
AI implies an agent interacting with the environment to learn and take actions that maximize
its chance of successfully achieving its goals.

23
 CHAPTER 2

LITERATURE SURVEY

2.1 LITERATURE SURVEY INTRODUCTION

In this chapter, an attempt is made to illustrate methodically, the analysis and review
being done on various techniques applied for finger vein recognition system by eminent
personalities. A brief analysis of each method was carried out to identify its own merits
and limitation. The literature survey serves as stare of the art tool to develop an efficient
method which can be used as authentication tool based on finger vein. The survey helps
to gather novel ideas from the existing methods such as capturing of finer image without
noise, enhancing the image captured, segmentation technique and matching process.

2.2 FINGER VEIN DATA ACQUISITION

David Mulyono proposes an ideology for finger vein image off-line and online acquisition.
The images which are acquired during the course of real time is called On-line images.
Images which are acquired from already existing source such as database, historical
information, etc are called as Off-line images. By using equipment like Web Camera or any
device so designed using light transmission technology are the distinct methods for
capturing on-line images. Light reflection and transmission are the two prominent
methods used for acquiring on-line images and the prominent difference between the two is
the location where in the near infrared light is positioned. In the case of light reflection
method, the pattern of finger vein is acquired based on near infrared light reflection
from palmer surface but in the method of light transmission acquiring the reflected light
from finger palmer surface, which penetrates to capture finger vein pattern. While
comparing between light reflection and transmission methods, the later one can acquire
vein images with high contrast, and most of the devices apply this methodology.

24
 Figure 2.1: Finger Vein image acquiring methods

Ton et.al (2013) presented a distinctive device using transmission of light method for
acquiring finger vein images (Huang 2017). This device normally designed, by using
infrared light. This method can help to capture the vein patterns more distinctly and
accurately.

Yin et al (2011) proposed homologous multimodal database, the first one of its kind
named SDUMLA-FV and was constructed by Shandong University (Wang 2012). As part
of this database, Ajay and Zhou, published another finger vein database which was
considered as an extended arm and called as HKPU-FV database. Subsequently another
database was formed by University of Twenty, named as UTFV database. For all the
practical purpose of research work, the two finger vein databases which were available, by
Tsinghua University and Chonbuk Nation University are used. The first one was called as
THU- FV database which is part of Homologous multimodal database and second one
is termed MMCBNU_6000 database.

The images acquired based on, light transmission-technology was used on all databases
and are different in sizes, contrast, background and quality.

In finger vein based biometric systems, various pre-processing task are to be performed –
retrieving information regarding edges, enhancement of contrast and brightness, noise
removal, sharpening of images, etc. in order to augment the excellence of image captured.
These pre-processes helps to improve image quality, which can be used as an input
during later stage of process for obtaining more relevant information and authentication

25
tool. Indeed, if the quality of the image is better, then better accuracy can be gained
which helps to advance upon the authenticity of the biometric system. Primarily the
pre-processing activity involves segmentation of finger vein images and alignment
denoising of images, detection of Region of Interest (ROI), normalization of image size,
and image augmentation.

M.Khalil-Hani et al proposed an essential thresholding operation called the process of

binarization which can help to select single threshold value, peak values as considered
for the image in the process of identification of finger vein pattern.

Zhi Liu (2012) proposed based of using mobile devise, finger vein pattern can be
identified in the process of alignment and segmentation of vein image is carried out by:

By using Bi-cubic interpolation for resizing and enhancement of the image is carried after
image segmentation and alignment. To enhance the image grey level contrast, the principle
of histogram equalization is applied. Three processes are involved in the image
segmentation function – (a) by applying canny edge detection algorithm the finger
edge is detected (b) By applying morphological dilation algorithm, the broken edges of
the finger are joined to achieve edge smoothening and (c) By applying Histogram
equalization algorithm, inside region of the finger area are filled with white pixels.

One of the methods used in image enhancement process is Histogram equalization

which usually help to increase the image contrast in those cases were the close contrast
values represent the image usable data. In order to generalize this method, multiple
histograms are applied which emphasizes local contrast, rather than overall contrast.
For the process of image portioning into image regions, the principles of histogram
equalization and Contrast Limiting Adaptive Histogram Equalization (CLAHE) are
pragmatic for performing each region equalization separately. The finger vein images will
poise to have noises after application of thresholding functions, and these unwanted
noises are to be removed based on binary median filtering technique. For denoizing the
finer vein image the conventional methods – median filter, sorting of pixel value etc. are
to be performed to denoised. By sorting of all pixels presents in the window we can obtain
the medial values, considering the assumed pixel as centre. We can also applied another
novel method called Gaussian Blur algorithm, for the removal of noise from finger vein
image.

26
2.3 HUMAN IDENTIFICATION USING FINGER VEIN

Fernando C. Monteiro (2015) projected a novel technique for segmenting finger vein
image based on edge information which can be obtained by watershed of morphological
algorithm and spectral method. R.V.Patil (2010) prerogative that better result can be
provided by K-means image segmentation, if the estimation of cluster numbers are
measure accurately. The process of edge detection is considered as major phenomenon
for estimating the cluster numbers accurately. To detect the edges and find cluster, phase
congruency was proposed by the author. The clusters are made and identified based on
Threshold and Euclidean distance computation and to find the segmentation of the image
K-means is use. Results obtained based on experiments on nine different images proves
that the identification of clusters based on the proposed method was accurate and optimal.

Weihong Cui Yi Zhang (2010) projected a new method for generating segmentation of
multi-scale based on method by selection edge based auto threshold. For calculation of
edge weight to prominent method named Band weight and Normalized Difference
Vegetation Index (NDVI) is use for the performance of segmentation of images. Over
and above edge based Threshold technique is also used for image segmentation. Based on
experiments using on multi-scale determination images, this proposed methodology proves
to maintain the boundaries and information of objects intact, while segmenting the same.
Anna Fabijańska (2011) presented a state of art technique for image segmentation process
by using Variance Filter technique, which helps to identify edge position of the image. In
this proposed technique, the edge information is extracted and compared using Sobel
Gradient filter with K-means. This method is appropriate, if 9 x 9 matrix window to be
used to extract the edge information, which helps to match accurately with the shape of
image object. We can also use small filtering window for the larger detailed images.

Mohammed J. Islam (2011) initiate that in case of pharmaceutical industries, for the
real time inspection of images, the best method is application of Computer Vision. The
Author had proposed a new system for inspection of images with respect to quality by
applying image segmentation edge based technique. Edges with noise-suppression property
was detected based on Sobel Edge Detector technique.

27
2.3.1 STATEMENT OF MD KHALED HASAN 2021

Present detailed comparison amoung the algorithm in all-inclusive and under sub-branches.
we provide the strengths and limitations of these algorithms and a novel literature survey.

After detection of these edges, in order to localization of foreground and background

pixels, Otsu Thresholding technique is applied. The experiments results obtained using this
method, was associated with NN based segmentation technique, by using Visual C++
software. As compared to NN method, on the basis of accuracy, the proposed method
outperform other version with a margin of difference by 10ms.

D. Barbosa (2012) suggested a new technique using Morphological Watershed Algorithm,

which joins region and edge based information for image segmentation. During pre-
processing stage, the noise filter is used with Magnitude and region merging is done based
on pre-segmentation process. By using Multi Class Normalized Cut the resemblance
graph is generated to perform image segmentation process. While comparing this proposed
technique with MNCUT, Mean Shift and JSEG on natural images, it significantly
overwhelms the results produced by Spectral Clustering Methods with thin margin of
improvement.

Gang Chen (2009) observed that, in the course of real time image processing, extraction
of image information as fast as required is a problem for a given image. Effective image
segmentation and the process is much time consuming, if region based methods are used.
To overcome the bottleneck, an application which can use new region based information
methods like Least Square Method which can detect the objects very fast as well
accurately by use of weight matrix and taking care of image local information into
account. This proposed technique provides very optimal and fast segmentation results,
which can be compared with other traditional techniques. Comparatively the projected
technique can extract image features more effectively and accurately at a given situation.

Zhen Hua, Yewei Li (2010) proposed an innovative method for image segmentation
created on two approaches – Region increasing and Watershed Method to improve
visual attention. With the use of Guass-Laplace and Gabor Filters, the edges and grey
values of image are extracted and subsequently ANN method is used for ROI extraction.
Experiments were conducted on natural images and the results obtained were associated
with other methods, proves that the newly developed algorithm helps perfectly for image
segmentation and also improves to retain the salience edges of the images.

28
ancan Mei (2011) proclaims that in the process of image segmentation, due to large
range of interactions, the Markov Random Field (MRF) is suffered. A new improved
version was proposed for image segmentation which can over this bottleneck, based on
Region Based Multi Scale Segmentation method.Tiancan Mei (2011) proclaims that in
the process of image segmentation, due to large range of interactions, the Markov
Random Field (MRF) is suffered. A new improved version was proposed for image
segmentation which can over this bottleneck, based on Region Based Multi Scale
Segmentation method.

After detection of these edges, in order to localization of foreground and background

pixels, Otsu Thresholding technique is applied. The experiments results obtained using this
method, was associated with NN based segmentation technique, by using Visual C++
software. As compared to NN method, on the basis of accuracy, the proposed method
outperform other version with a margin of difference by 10ms.

D. Barbosa (2012) suggested a new technique using Morphological Watershed Algorithm,

which joins region and edge based information for image segmentation. During pre-
processing stage, the noise filter is used with Magnitude and region merging is done based
on pre-segmentation process. By using Multi Class Normalized Cut the resemblance
graph is generated to perform image segmentation process. While comparing this proposed
technique with MNCUT, Mean Shift and JSEG on natural images, it significantly
overwhelms the results produced by Spectral Clustering Methods with thin margin of
improvement.

Gang Chen (2009) observed that, in the course of real time image processing, extraction
of image information as fast as required is a problem for a given image. Effective image
segmentation and the process is much time consuming, if region based methods are used.
To overcome the bottleneck, an application which can use new region based information
methods like Least Square Method which can detect the objects very fast as well
accurately by use of weight matrix and taking care of image local information into
account. This proposed technique provides very optimal and fast segmentation results,
which can be compared with other traditional techniques. Comparatively the projected
technique can extract image features more effectively and accurately at a given situation.

Zhen Hua, Yewei Li (2010) proposed an innovative method for image segmentation
created on two approaches – Region increasing and Watershed Method to improve
visual attention. With the use of Guass-Laplace and Gabor Filters, the edges and grey
values of image are extracted and subsequently ANN method is used for ROI extraction.
Experiments were conducted on natural images and the results obtained were associated

29
with other methods, proves that the newly developed algorithm helps perfectly for image
segmentation and also improves to retain the salience edges of the images.

Tiancan Mei (2011) proclaims that in the process of image segmentation, due to large
range of interactions, the Markov Random Field (MRF) is suffered. A new improved
version was proposed for image segmentation which can over this bottleneck, based on
Region Based Multi Scale Segmentation method. The data set of natural scenario images
are used, and by using multi-scale MRF model for regions as parameter, this algorithm
is considered as better option than other techniques towards image segmentation process.
From the results derived it was concluded that, as compared to MSAP, RSMAP
algorithm provides improvised results for image segmentation.

2.4 ON-LINE FINGER VEIN VERIFICATION

In order to extract finger vein patterns from non-uniform images, by a robust system, the
method proposed should consider tracking dark lines presents in an image repeatedly.
Established on the quantum of times, wherein the tracking lines pass through the points, the
extracting process is defined. A few literatures published by eminent personalities,
describing the process and modalities for extracting finger vein images are discussed in this
thesis work.

N. Miura, A. Nagasaka, and T. Miyatake (2004) proposed a new method for future
extraction based on repeated line tracking technique. This method adopts the pattern of
extraction by applying the concept of number of times the tracking lines pass through
the points[63]. This method helps to identify local dark lines, which starts at various
positions for line tracking and the same is executed by moving along with the lines by
pixel by pixel.. During this process if a dark line is not detectable, then start a new
tracking operation from another position until all the dark lines present in an image is
identified. This process is carried out repeatedly, by executing local line tracking
operations and based on this operations the location of the lines overlap.

2.5 PERSONAL AUTHENTICATION USING HAND VEIN

Classification and Matching process is one of the important stage in finger

vein identification. This process measures the similarity or dissimilarity among the real
time input finger vein images and historical images previously stored in the database
(Jian-Da Wu 2011). This is the process which is performed immediately once the finger

30
vein features are extracted. Once this process completed, then the next stage is minutia
matching which involves three stage Literature survey is carried on based on the papers
published by various authors on their contribution towards these process are evaluated and
submitted in this thesis work.

Fei Liu et al (2014) projected a new technique for minutiae pairing which is called
Singular Value Decomposition (SVM). The process of False Removing is performed
based on Local Extensive Binary Pattern (LEBP) by applying the local characteristics
which are rich in nature for minutiae representation. LEBP is the combination of Local
Directional and Local Multilayer binary pattern. For image classification Support Vector
Machine (SVM) is widely used for image classification and matching.

Jian-Da Wu et al (2011) proposed a new method to find hyper plane classifier which helps
for classifying Support Vector Machine. The SVM usage was proposed due to its wide
application usage and its capacity to handle data of nonlinearly separable. This classified
is robust in nature and can operate on much lesser time.

Pandillapalli Gowtham et al (2013) developed a new methodology of using Score

Level Fusion, to recognize the finger vein images which involve the process of matching
the removed vein image with the available database which contains all the features of vein
images. The features extracted from the given input vein images are harmonized with
the images available in database. Based on matching of input image with the stored image,
then the user will be declared as authorised.

A Kumar et al(2012) proposed a new technique for finger vein matching based
on finger vein and its dorsal texture. By combing these two properties the score level
combination is calculated, which is a better decision level as compared to feature level
combination. The user will be declared as imposter, one the input image is not matching
with the combination of finger vein and its dorsal texture. In order to calculate the score
level, the holistic and nonlinear fusions are considered. The prior knowledge existing in
the dynamic combination process of matched scores will be utilized in the process of
Holistic fusion. Subsequently, based on the degree of uniformity between the two
identical scores the non-linear fusion is adjusted with the combined score.

Cheng-Bo Yu et al (2009) suggested a new method to abstract the critical points of

the extracted image such as bifurcation point and ending point which can be used to
compute the similarities between the input and template image available in the given
database .

31
 Wonseok Song et al (2011) proposed a novel methodology based on Mean
Curvature for fing er vein verification system. The two dimensional point of the finger
vein pattern can be measure using Hausdorff distance. The Multiple Pixel Ratios (MPR) is
defined as the ratio of matching the quantum of pixels and total number of pixels
available in finger vein image patterns. This new method can achieves EER value of
0.761% [75]. We cannot achieve the process of matching the pixels, if existence of any
rotation or translation then the EER of mean curvature value will be 0.0025%.

2.6 ANALYSIS OF LITERATURE WORK

With the extensive analysis of literature works, we can envisage the horizon to overcome
the limitations understood during the course of literature work, that the proposed methods
should built upon. Based on the understanding, we proposed to develop a robust finger
vein based identification system, which has assumption that the unique feature of vein
pattern will be a milestone towards the personal identification system using biometrics.

As compared to the traditional biometric based identification system, the proposed

methodology will have much more advantages, which can be briefly enumerated as given
below.

32
 CHAPTER-3

EXISTING SYSTEM

3.1 EXISTING SYSTEM:

In this approaches extract vein patterns by assuming that they generate distributions such as
valleys and line segments. They can be broadly classified into two categories:

 Finger-vein extraction approaches based on detecting valley features

 Finger-vein extraction approaches based on detecting line-like features

3.2 DISADVANTAGES OF EXISTING SYSTEM:

 It completely depend on domain knowledge.

 Lack of robustness to extract finger-vein features from raw images.
 Verification accuracy is low.
 Imbalanced and poor local illumination problem.
 Privacy and security problem occur.
 Not always effective to detect the finger vein patterns.
 Not easily observed in visible light.

In this, a Convolutional Neural Network (CNN) is trained on the resulting dataset to predict
the probability of each pixel of being foreground given a patch centered on it. The CNN
learns what a finger-vein pattern is by learning the difference between vein patterns and
background ones. We propose another new and original contribution by developing and
investigating a Fully Convolutional Network (FCN) to recover missing finger-vein patterns in
the segmented image. We propose an automatic scheme to label pixels in vein regions and
background regions, given very limited human knowledge.

We employ several baselines approaches to extract (segment) the vein network from an image
and use their combined output automatically to assign a label for each pixel.

Such a scheme avoids the heavy manual labeling and may also reduce label errors, especially
for ambiguous pixels. As a finger-vein consists of clear regions and ambiguous regions,
several baselines are employed to automatically label pixels as vein or background in the
image clear regions, thus avoiding the tedious and prone-to-error manual labeling.

28
 A CNN-based scheme is employed to automatically learn features from raw pixels for
finger-vein verification. First, a dataset is constructed based on patches centered on the
labeled pixels, and we take the patches as input for CNN training. Secondly, in the test phase,
the patch of each pixel is input into CNN the output of which is taken as the probability of the
pixel to belong to a vein pattern. Then, the vein patterns are segmented using a probability
threshold of 0.5. Compared to existing approaches, our CNN automatically learns robust
attributes for finger-vein representation. This paper investigates a new approach for
recovering vein patterns in the extracted finger-vein image.

As finger-vein patterns may be missing by corruption during the imaging stage and the
inaccurate estimation of parameters during the preprocessing stage (i.e. alignment and feature
extraction), we develop a robust finger-vein feature recovering scheme based on a Fully
Convolutional Network (FCN).

In this context, we perform a rigorous experimental analysis that shows that our scheme does
succeed in recovering missing patterns which further improves the verification performance.

3.4 ADVANTAGES OF EXISTING SYSTEM:

FINGER vein authentication can be done using the vascular pattern on the back of a hand or a
finger. However, the FINGER vein pattern is the most complex and covers the widest area,
because the FINGER has no hair, it is easier to photograph its vascular pattern.

 1. The FINGER also has no significant variations in skin colour compared with
fingers or back of the hand, where the colour can darken in certain areas. Also we
can use fusion of two technologies, FINGER vein and FINGER print, which
will be more complex and more reliable but costly.
 2. FRR and FAR are very low in comparison to other biometric technologies. So it is
more secure and reliable.
 3. FINGER vein pattern of any individual cannot be theft. Also since it is contactless,
privacy cannot be invaded.
 4. The completely contactless feature of this device makes it suitable for use
when high levels of hygiene are required. It also eliminates any hesitation people
might have about coming into contact with something that other people have already
touched.

29
 CHAPTER 4

PROPOSED SYSTEM

4.1 MODULES
Based on the exploration of various literatures published in the field of finger vein
authentication system, we could estimate the truthfulness of these methods to extract
the finger vein patterns. We need to regulate whether the projected method over
comes the drawbacks of the various existing methods, with respect to its robustness,
repeatability, accountability and effectiveness. The experimental results thus arrived based
on conventional methods of biometric system as that of new methods are validated and
compares for its effectiveness of qualitatively and cost effective. The traditional methods
uses matched filtering technique for the comparison of biometric parameter to identify a
person where in the profile matching technique is being used in the finger vein based
identification system.

After careful analysis of these literatures, we could able to propose the finger vein
based recognition framework, which is based on the principles of Acquisition,
Normalization, Extraction, Matching and output identification of captured image.

4.2 ACQUISITION OF AN INFRARED IMAGE OF THE FINGER:

We can obtain the NIR image of the finger using special imaging device, which is having
the provision to capture the image of vein patterns using NIR. This NIR light is captured
by the camera located at the opposite side of the finger and the concentration of the
light is accustomed with reference to the brightness of the image required.

4.3 NORMALIZATION OF THE IMAGE:

The captured image requires some normalization approach as required due to the location
and angle of the finger. These qualities are often vary from person to person and much
depends on individual traits. Ideally using the outline of the finger, we apply two
dimensional normalization method, based finger on the assumption that angle of the finger
and it three dimensional location are constant.

30
4.4 EXTRACTION OF FINGER-VEIN PATTERNS:

From the normalized image of finger vein, we extract its pattern for matching the
same. The finger-vein pattern is extracted from the normalized infrared image of the finger.

4.5 MATCHING:

The input and registered image patterns are compared for correlations as computed
and based on the same the identification process if carried out for further processing of the
system.

31
 CHAPTER -5
SYSTEM REQUIREMENTS

5.1 HARDWARE REQUIREMENTS

 System : Pentium IV 2.4 GHz.
 Hard Disk : 40 GB.
 Floppy Drive : 1.44 Mb.
 Monitor : 15 VGA Colour.
 Mouse : Logitech.
 Ram : 512 Mb.

5.2 SOFTWARE REQUIREMENTS

 Operating system : Windows
 Coding Language : Python
 Database : MYSQL

5.3 SOFTWARE ENVIRONMENT

5.3.1 PYTHON:

Python is a high-level, interpreted, interactive and object-oriented scripting language. Python is

designed to be highly readable. It uses English keywords frequently where as other languages
use punctuation, and it has fewer syntactical constructions than other languages.

 Python is Interpreted − Python is processed at runtime by the interpreter. You do not

need to compile your program before executing it. This is similar to PERL and PHP.

 Python is Interactive − You can actually sit at a Python prompt and interact with the
interpreter directly to write your programs.

 Python is Object-Oriented − Python supports Object-Oriented style or technique of

programming that encapsulates code within objects.

32
  Python is a Beginner's Language − Python is a great language for the beginner-level
programmers and supports the development of a wide range of applications from simple
text processing to WWW browsers to games.

5.3.2 HISTORY OF PYTHON

Python was developed by Guido van Rossum in the late eighties and early nineties at the
National Research Institute for Mathematics and Computer Science in the Netherlands.

Python is derived from many other languages, including ABC, Modula-3, C, C++, Algol-68,
SmallTalk, and Unix shell and other scripting languages.

Python is copyrighted. Like Perl, Python source code is now available under the GNU General
Public License (GPL).

Python is now maintained by a core development team at the institute, although Guido van
Rossum still holds a vital role in directing its progress.

5.3.3 PYTHON FEATURES

Python's features include

 Easy-to-learn − Python has few keywords, simple structure, and a clearly defined
syntax. This allows the student to pick up the language quickly.
 Easy-to-read − Python code is more clearly defined and visible to the eyes.
 Easy-to-maintain − Python's source code is fairly easy-to-maintain.
 A broad standard library − Python's bulk of the library is very portable and
cross-platform compatible on UNIX, Windows, and Macintosh.
 Interactive Mode − Python has support for an interactive mode which allows
interactive testing and debugging of snippets of code.
 Portable − Python can run on a wide variety of hardware platforms and has the
same interface on all platforms.
 Extendable − You can add low-level modules to the Python interpreter. These
modules enable programmers to add to or customize their tools to be more
efficient.

33
  Databases − Python provides interfaces to all major commercial databases.
 GUI Programming − Python supports GUI applications that can be created and
ported to many system calls, libraries and windows systems, such as Windows
MFC, Macintosh, and the X Window system of Unix.
 Scalable − Python provides a better structure and support for large programs than
shell scripting.

Apart from the above-mentioned features, Python has a big list of good features, few are listed
below

 It supports functional and structured programming methods as well as OOP.

 It can be used as a scripting language or can be compiled to byte-code for
building large applications.
 It provides very high-level dynamic data types and supports dynamic type
checking.
 It supports automatic garbage collection.
 It can be easily integrated with C, C++, COM, ActiveX, CORBA, and Java.

Python is available on a wide variety of platforms including Linux and Mac OS X. Let's
understand how to set up our Python environment.

34
 CHAPTER - 6
SYSTEM STUDY

6.1 FEASIBILITY STUDY

The feasibility of the project is analyzed in this phase and business proposal is put forth with a
very general plan for the project and some cost estimates. During system analysis the feasibility
study of the proposed system is to be carried out. This is to ensure that the proposed system is
not a burden to the company. For feasibility analysis, some understanding of the major
requirements for the system is essential.

Three key considerations involved in the feasibility analysis are

 ECONOMICAL FEASIBILITY
 TECHNICAL FEASIBILITY
 SOCIAL FEASIBILITY

6.1.1 ECONOMICAL FEASIBILITY

This study is carried out to check the economic impact that the system will have on the
organization. The amount of fund that the company can pour into the research and development
of the system is limited. The expenditures must be justified. Thus the developed system as well
within the budget and this was achieved because most of the technologies used are freely
available. Only the customized products had to be purchased.

6.1.2 TECHNICAL FEASIBILITY

This study is carried out to check the technical feasibility, that is, the technical requirements of
the system. Any system developed must not have a high demand on the available technical
resources. This will lead to high demands on the available technical resources. This will lead to
high demands being placed on the client. The developed system must have a modest
requirement, as only minimal or null changes are required for implementing this system.

35
6.1.3 SOCIAL FEASIBILITY
The aspect of study is to check the level of acceptance of the system by the user. This
includes the process of training the user to use the system efficiently. The user must not feel
threatened by the system, instead must accept it as a necessity. The level of acceptance by the
users solely depends on the methods that are employed to educate the user about the system and
to make him familiar with it. His level of confidence must be raised so that he is also able to
make some constructive criticism, which is welcomed, as he is the final user of the system.

6.1.4 OPERATIONAL FEASIBILITY

Operational feasibility is the measure of how well a proposed system solves the problems, and
takes advantage of the opportunities identified during scope definition and how it satisfies the
requirements identified in the requirements analysis phase of system development.

The operational feasibility assessment focuses on the degree to which the proposed development
project fits in with the existing business environment and objectives with regard to development
schedule, delivery date, corporate culture and existing business processes.

To ensure success, desired operational outcomes must be imparted during design and
development. These include such design-dependent parameters as reliability, maintainability,
supportability, usability, producibility, disposability, sustainability, affordability and others.
These parameters are required to be considered at the early stages of design if desired operational
behaviours are to be realised. A system design and development requires appropriate and timely
application of engineering and management efforts to meet the previously mentioned parameters.
A system may serve its intended purpose most effectively when its technical and operating
characteristics are engineered into the design. Therefore, operational feasibility is a critical
aspect of systems engineering that needs to be an integral part of the early design phases.

36
6.2 SYSTEM TESTING
The purpose of testing is to discover errors. Testing is the process of trying to discover every
conceivable fault or weakness in a work product. It provides a way to check the functionality of
components, sub-assemblies, assemblies and/or a finished product It is the process of exercising
software with the intent of ensuring that the Software system meets its requirements and user
expectations and does not fail in an unacceptable manner. There are various types of test. Each
test type addresses a specific testing requirement.
6.3 TYPES OF TESTS
6.3.1 UNIT TESTING
Unit testing involves the design of test cases that validate that the internal program logic is
functioning properly, and that program inputs produce valid outputs. All decision branches and
internal code flow should be validated. It is the testing of individual software units of the
application .it is done after the completion of an individual unit before integration. This is a
structural testing, that relies on knowledge of its construction and is invasive. Unit tests perform
basic tests at component level and test a specific business process, application, and/or system
configuration. Unit tests ensure that each unique path of a business process performs accurately
to the documented specifications and contains clearly defined inputs and expected results.

Unit testing is usually conducted as part of a combined code and unit test phase of the software
lifecycle, although it is not uncommon for coding and unit testing to be conducted as two distinct
phases.

TEST STRATEGY AND APPROACH

Field testing will be performed manually and functional tests will be written in detail.

TEST OBJECTIVES

 All field entries must work properly.

 Pages must be activated from the identified link.
 The entry screen, messages and responses must not be delayed.

37
FEATURES TO BE TESTED

 Verify that the entries are of the correct format

 No duplicate entries should be allowed
 All links should take the user to the correct page.

6.3.2 INTEGRATION TESTING

Integration tests are designed to test integrated software components to determine if they
actually run as one program. Testing is event driven and is more concerned with the basic
outcome of screens or fields. Integration tests demonstrate that although the components were
individually satisfaction, as shown by successfully unit testing, the combination of components is
correct and consistent. Integration testing is specifically aimed at exposing the problems that
arise from the combination of components.

6.3.3 FUNCTIONAL TEST

Functional tests provide systematic demonstrations that functions tested are available as
specified by the business and technical requirements, system documentation, and user manuals.

Functional testing is centered on the following items:

 Valid Input : identified classes of valid input must be accepted.

 Invalid Input : identified classes of invalid input must be rejected.
 Functions : identified functions must be exercised.
 Output : identified classes of application outputs must be exercised.
 Systems/Procedures : interfacing systems or procedures must be invoked.
Organization and preparation of functional tests is focused on requirements, key functions, or
special test cases. In addition, systematic coverage pertaining to identify Business process flows;
data fields, predefined processes, and successive processes must be considered for testing.
Before functional testing is complete, additional tests are identified and the effective value of
current tests is determined.

38
6.3.4 SYSTEM TEST
System testing ensures that the entire integrated software system meets requirements. It tests a
configuration to ensure known and predictable results. An example of system testing is the
configuration oriented system integration test. System testing is based on process descriptions
and flows, emphasizing pre-driven process links and integration points.

6.3.5 WHITE BOX TESTING

White Box Testing is a testing in which in which the software tester has knowledge of the
inner workings, structure and language of the software, or at least its purpose. It is purpose. It is
used to test areas that cannot be reached from a black box level.

6.3.6 BLACK BOX TESTING

Black Box Testing is testing the software without any knowledge of the inner workings,
structure or language of the module being tested. Black box tests, as most other kinds of tests,
must be written from a definitive source document, such as specification or requirements
document, such as specification or requirements document. It is a testing in which the software
under test 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.

6.4 SYSTEM DESIGN

It is a testing in which the software under test 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.

6.5 DATA FLOW DIAGRAM

 The DFD is also called as bubble chart. It is a simple graphical formalism that can be
used to represent a system in terms of input data to the system, various processing carried
out on this data, and the output data is generated by this system.
 The data flow diagram (DFD) is one of the most important modeling tools. It is used to
model the system components. These components are the system process, the data used
by the process, an external entity that interacts with the system and the information flows
in the system.

39
  DFD shows how the information moves through the system and how it is modified by a
series of transformations. It is a graphical technique that depicts information flow and the
transformations that are applied as data moves from input to output.
 DFD is also known as bubble chart. A DFD may be used to represent a system at any
level of abstraction. DFD may be partitioned into levels that represent increasing
information flow and functional detail.

Fig .6.2 Data Flow Diagram

6.6 USE CASE DIAGRAM:

A use case diagram in the Unified Modeling Language (UML) is a type of behavioral
diagram defined by and created from a Use-case analysis. Its purpose is to present a graphical
overview of the functionality provided by a system in terms of actors, their goals (represented as
use cases), and any dependencies between those use cases. The main purpose of a use case
diagram is to show what system functions are performed for which actor. Roles of the actors in

40
the system can be depicted.
6.6.1 CLASS DIAGRAM:
In software engineering, a class diagram in the Unified Modeling Language (UML) is a type of
static structure diagram that describes the structure of a system by showing the system's classes,
their attributes, operations (or methods), and the relationships among the classes. It explains
which class contains information.

Input Output

Features extraction

Input data Regression

Preprocessing ( ) Finally get Regression &

Display Result: Predicted
Vein

6.6.2 SEQUENCE DIAGRAM:

A sequence diagram in Unified Modeling Language (UML) is a kind of interaction diagram that
shows how processes operate with one another and in what order. It is a construct of a Message
Sequence Chart. Sequence diagrams are sometimes called event diagrams, event scenarios, and
timing diagrams.

41
6.6.3 ACTIVITY DIAGRAM:

Activity diagrams are graphical representations of workflows of stepwise activities and actions
with support for choice, iteration and concurrency. In the Unified Modeling Language, activity
diagrams can be used to describe the business and operational step-by-step workflows of
components in a system. An activity diagram shows the overall flow of control.

42
 CHAPTER 7

CONCLUSION

In conclusion, finger vein verification based on deep learning offers a robust and secure
biometric authentication method. Its ability to leverage intricate patterns within the veins ensures
a high level of accuracy and resistance to fraudulent attempts. As technology advances, further
refinements in deep learning models are likely to enhance the efficiency and reliability of finger
vein verification, making it a promising solution for secure identity authentication in various
applications. However, challenges and considerations persist. Privacy concerns associated with
biometric data storage and potential vulnerabilities, such as spoofing attempts, demand careful
implementation and adherence to stringent security measures. Additionally, the system's success
relies on overcoming issues like environmental variations and ensuring social acceptance.

43
 REFERENCE
• M. A. Turk and A. P. Pentland, “Face recognition using Eigen faces,” CVPR, pp.
586–591, 2021.
• A. Jain, L. Hong, and R. Bolle, “On-line fingerprint verification,” IEEE Transactions
on Pattern Analysis and Machine Intelligence, vol. 19, no. 4, pp. 302–314, 2017.
• J. Daugman, “How iris recognition works,” IEEE Transactions on Circuits and
Systems for Video Technology, vol. 14, no. 1, pp. 21–30, 2014.
• A. Kumar and Y. Zhou, “Human identification using finger images,” IEEE
Transactions on Image Processing, vol. 21, no. 4, pp. 2228–2244, 2012.
• A. Kumar and K. V. Prathyusha, “Personal authentication using hand vein
triangulation and knuckle shape,” IEEE Transactions on Image Processing, vol. 18,
no. 9, pp. 2127–2136, 2009.

44