Abstract:

The evolution in banking and transaction system over the years has been enormous. The
authentication, validation and confirmation of a person are of utmost priority and should be dealt
with carefully. The first and the foremost thing that should be kept in mind is the security and
confidentiality of the consumers. ATM (Automated Teller Machine) is the most widely used
electronic equipment to carry out banking transactions. The dependence on ATM machines has
increased considerably over the last three decades or so. In this project, a system is proposed
with an aim to enhance the security of ATM machines for safer transactions. In an ATM
machine an identity test has to be passed before carrying out any transaction. The prevalent
method for this identity test is the smartcard. This method secures the customer account against
any fraud approaches by using the Personal Identification Number (PIN). But as a matter of fact
many problems are also associated with smartcard authentication system. In order to deal with
these problems, the use of voice as a biometric feature in ATM machines for secured transactions
is proposed. Security is an essential part of human life. In this era security is a huge issue that is
reliable and efficient if it is unique by any mean. Voice recognition is one of the security
measures that are used to provide protection to human’s computerized and electronic belongings
by his voice. In this paper voice sample is observed with MFCC for extracting acoustic features
and then used to trained HMM parameters through forward backward algorithm which lies under
HMM and finally the computed log likelihood from training is stored to database. It will
recognize the speaker by comparing the log value from the database against the PIN code. This
paper is attentive of providing a security by developing a voice recognition system to secure the
ATM (automatic transaction machine) using HMM with MFCC. The usage of MFCC for
extracting voice features and HMM for recognition provides a 2D security to the ATM in real
time scenario. MFCC is used to describe the acoustic features of speaker’s voice. HMM forward
backward estimation technique is used to train these features into the HMM parameters and used
to find the log likelihood of entire voice. In recognition HMM is used to compare log likelihood
to the pre-stored value and intended to recognize the speaker. If the log likelihood is matched
then it is granted otherwise failed to use ATM system.
CHAPTER 1
INTRODUCTION:

Background to the Study

The Automated Teller Machine (ATM) is a selfservice machine that dispenses cash and
performs some human teller functions like balance enquiry, bills payments, mini statements,
Fund Transfer, Cash Deposit and so on. ATM transactions are carried out through the use of a
debit/credit card which enables the card holder(s) to access and carry out banking transactions
without a teller. With ATM, customers can access their bank deposit or credit accounts in order
to make a variety of transactions mentioned earlier. If the currency being withdrawn from the
ATM is different from that in which the bank account is denominated the money will be
converted at an official exchange rate. Thus, ATMs often provide the best possible exchange
rates for foreign travelers, and ATM is widely used for this purpose .

The ATM users access their account through special type of plastic card e.g. master, visa,
verve card etc. The card is encoded with user information on a magnetic strip. The magnetic strip
contains an identification code that is transmitted to the bank’s central computer by modem. It
was developed as a result of need for self-service technology (SST) in financial service delivery
by financial institutions .

The ATM is designed to replace the manual banking transactions in which customers
walk into the bank to fill tellers, withdrawal booklets or cheque. The manual banking system are
based on paper, it is time consuming and lacks efficiency in record keeping then giving room for
manipulations. In recent time ATM has served as a device that enhances the cashless policy due
to its functionality in fund transfer between one account and other.

Apart from using ATMs other computer-based/IT based banking technologies are
available for examples internet banking and mobile banking, but the demand for cash still
remains high and bank branches are rising continually worldwide as customers demand cash to
be accessible at different locations.

The mentioned banking system/technologies such as internet, mobile and traditional

banking cannot allow customers to have access to their cash at convenient, and level of internet
access and the cost of procuring the facilities to use both the internet and mobile banking such as
computer or sophisticated mobile phone are also barrier in using such technologies.

ATM creates a paperless office, ensures security of customers’ accounts and privacy, it
grants customers 24hours access to their accounts, eliminates cash induced robbery, it reduces
cost of operation and enhances proper and effective record keeping. ATM allows you to do a
number of banking operations such as withdrawing cash from one’s account, making balance
inquiries and transferring money from one account to another using a plastic, magnetic-strip card
and personal identification number issued by the financial institution.

Currently one and major way to get access to your account in other to perform some of
the transaction mentioned earlier is through the use of ATM card. There are lot of problem
associated with the uses of ATM card, these include: Lost or Stolen Card, use of ATM card by
third party to perform transaction, ATM Card Skimming & PIN Capturing, damaging of ATM,
card expiring, charges on issuance of the ATM card and maintenances by financial institutions
etc. with all the problem mentioned, the use of ATM card has become a treat to safety of
customer funds, even though the stakeholders in financial transaction are making great efforts to
reduce ATM frauds.

A computer system that automatically identifies and verifies the person by capturing the
voice from a source like microphone is known as voice recognition. Voice recognition is one of
the terms of biometric technology. It uses to provide any authentication to any system on the
basis of acoustic features of voice instead of images. The behavioral aspect of human voice is
used for identification by converting a spoken phrase from analog to digital format, and
extracting unique vocal characteristics, such as pitch, frequency, tone and cadence to establish a
speaker model or voice sample. In voice recognition, enrollment and verification processes are
involved. Enrollment process describes the registration of speaker by training his voice features .
And verification contains to verify the speaker by comparing his current voice features to pre
stored features of voice. In real time, the verification process splits into two mechanisms. It first
compares the unknown speaker to the pre stored database of known speakers on the basis of 1:N.
and then it make decision of speaker to the exact match of 1:1. Where the one voice sample
finally matched to only 1 template stored in the database . Voice recognition has two categories
text dependent and text independent. Text dependent voice recognition identifies the speaker
against the phrase that was given to him at the time of enrollment. Text independent voice
recognition identifies the speaker irrespective of what he is saying. This method is very often use
in voice recognition as it require very little computations but need more cooperation of speakers.
In this case the text in verification phase is different than in training or enrolment phase. In Early
research Shi-Huang Chen and Yu-Ren Luo presents in the MFCC as to extract features and
trained and recognized using SVM. They defined the MFCC as the unique and reliable feature
extraction technique. That was used to find the most usable features in detailed form. In
recognition phase SVM (super Vector Machine) technique based on two class classifiers by
defining the decision in binary form was introduced. It discriminate claimed speaker and
imposter by +1 and -1 by maximizing the margins or minimizing the structural risks. It shows
results averaged to 95.1% with ERR of 0.0%. That was considered as the best results under 22nd
order of MFCC. In another research [6], Lindasalwa Muda, Mumtaj Begam and I. Elamvazuthi
also extracted the voice features using MFCC that was trained and recognized using DTW. DTW
(dynamic time Wrapping) a non linear sequence alignment is another technique that is used for
recognition process. They find it best for time sequence between two speeches. Here the optimal
wrapping path is achieved by wrapping the time distance between two signals. One other
research has done by Ibrahim Patel and Dr. Y. Srinivas Rao. They represent the voice
recognition with improvement of MFCC with frequency decomposition technique. They
introduces sub band coding in their research. The integration of MFCC with sub band coding
increases its efficincy and accurate classification as compared to MFCC separately. These two
features of MFCC and integrated sub band decomposition with MFCC are used in HMM to train
and recognize the speaker.

1.1 Voice Recognition

Voice Signal Identification transforms a speech signal into features which aid the further
processing. Quite a number of algorithms and techniques are used for the same. The ability of a
system or program to receive as well as interpret dictation, or to understand and follow the
spoken instructions, is termed as Voice Recognition. In general, it is regarded as one of the
convenient and safest recognition techniques. The basic block diagram for voice recognition is
given in Figure 1. A big challenge of proliferating accuracy and recognition speed is faced by the
real time automatic system for speech recognition. Degradation in performance of speech
recognition system occurs due to noise. Also it is affected by modifying speech data due to
dependence on speaker's gender, environmental conditions and style in which it is spoken. The
accuracy of recognition depends on the method of feature extraction and training, so one of the
core issues of speech recognition research is the recognition accuracy. Development of speech
recognition system with less recognition time and high recognition accuracy is the aim.
Moreover, system performance improves in the presence of noise with the use of noise robust
feature extraction and training technique.

A two-factor authentication of an individual can help voice to combine what masses say
and the way in which they say it. Other identifications such as fingerprints, handwriting, iris,
retina, face scans can also help, whilst voice identification is required as an authentication that is
not only secure but unique also. Voice recognition systems are economical and easily understood
by users.

1.2 Mel Frequency Cepstral Coefficient (MFCC):

In sound processing, the mel frequency cepstral (MFC) is a representation of the short
term power spectrum of a sound, based on a linear cosine transform of a log power spectrum on
a nonlinear Mel scale of frequency. MFCCs are coefficients that all in all make up a MFC. The
difference between the cepstral and the mel-frequency cepstral is that in the MFC, the frequency
groups are similarly divided on the Mel scale, which approximates the human sound-related
framework's reaction more nearly than the straightly separated frequency groups utilized as a
part of the typical cepstral. This frequency warping can allow for better representation of sound.:

A vital task in producing a better recognition system is dependent on the extraction of the
best parametric representation of the acoustic signals. MFCC is dependent on human hearing
judgement which cannot receive frequencies above 1 KHz. In other words, MFCC is dependent
on known variation of the human ear’s critical bandwidth with frequency[1]. The overall process
of MFCC is shown in Figure 2.
CHAPTER 2

2. LITERATURE REVIEW

[1]. In their work, the authors are of the opinion that Automated Teller Machine (ATM)
packages of banks in Ghana have operated for a very long time without full exploration of all
essential functions of the facility and this has been a surprised to the public and other decision
makers about the effect of ATM operations on customer demand for it. That leads to author’s
research interests to assess the operational features of the ATM and the factors that account for
customers‟ willingness to use ATM. The study modelled the operational features of ATM using
queuing modelling and a probity model was estimated to determine the factors affecting
customer usage of ATM. They use data that was collected from 160 customers of banks with
ATM facilities. It was confirmed that there was a high traffic intensity for ATMs use for most
banks in the sampled area. Also, higher educational attainment, number of ATMs per bank,
convenience, and security features, efficiency and low transaction charges have significant effect
on influencing the usage of ATM services. They recommended that management of the banks
should upsurge the number and quality of ATM services in order to increase access and usage of
ATM.

[2]. these authors investigate ATM Usage and Customers’ Satisfaction in Nigeria. He discovered
that despite the increasing number of ATM installations in Nigeria. Customers’ needs are not
satisfactorily met as customers are always seen on queue in large numbers at various ATM
designated centers as well as poor service delivery of some of this machine. He uses research
engages comparative analysis of three banks in Ogun State, Metropolis of Nigeria viz-a-viz
FirstBank, Guaranty Trust Bank and Skye Bank. Data was collected with the aid of
questionnaires to sampled population of 200 respondents and chi-square statistical tool was used
to analyze the data and the results showed a positive and significant relationship between ATM
Usage and Customers’ Satisfaction.

[3]. The study discusses the effects of the Automated Teller Machine (ATM) network market
structure on the availability of cash withdrawal ATM services and cash usage. The aim of the
study was to construct the ATM equation. The study contributes to the earlier discussion on the
effects of ATMs on cash usage. The monopolization of ATM network market structure and its
effects on the number of ATMs and on cash in circulation are analyzed in theoretically and
empirical manner. The unique annual data set on 20 countries was used in the estimations and
combined from various data sources. The observation period is 1988–2003, but the data on some
countries are available only for a shorter period. Based on our theoretical discussion, as well as
the estimation results, monopolization of the ATM network market structure is associated with a
smaller number of ATMs. It was revealed that the influence of the number of ATMs on cash in
circulation is ambiguous.

[4], this study was focused on a clear view of two different working procedure of ATM
(Automated Teller Machine) in Bangladesh. Two processes are Core Banking and Consortium of
Banks and author reached a conclusion after few surveys to the bank. The author is of the
opinion that the Major problem of using ATM is that the user do not have the information when
the ATM machine is off or on. Consequent, if users find it closed after traveling a long distance
then it is very much annoying. To support his suggestion, he built software which can be really
helpful to the situation. The author also identifies few problems like booths limitations, using
limitations and transaction limitations and suggestions was made to arrest the situation.

[5] The author appreciate the introduction of ATM in banking sector, but identify the major
setback in that innovation as customer(s) in-ability to deposit money and send it to someone at
different location. The author then came up with a framework that allows customer(s) to deposit
some amount of money into his account and later transfer the amount to another account(s) or
deposit the money directly into another person’s account over the ATM network, this can be
within a particular bank (intra bank) or involving different banks (inter-bank) in addition to the
current transactions customers make using the ATM. He proposed modification of the ATMs by
creating another input device that collects the money into the ATM system, reads its
denomination and either saves it or transfers to the required customer account. He was limited to
the software component of the ATM system. He was of opinion that the enhanced ATM will will
further decongest the banking halls thereby enhancing efficiency in the day-today running of the
banking system.

[6] Discussed recent development in ATM industry especially the improvement from the OS/2
operating systems to window based operating systems which have considerably enhanced ATM
functionality and user interfaces.
[7] studied the formal design model of the ATM system, the study included ATM design
specification using denotation mathematics known as Real Time Process Algebra (RTPA),
RTPA architectural modeling methodology, he proposed the architecture of the ATM system to
be built and refined by some set of unified data models (UDMs) which share a generic
mathematical models of tuples, he specified the static behavior of the ATM and refined it by a
set of Unified Process Models (UPMs) for the system transition processing and some supporting
processes. Based on his design model: formal design model, codes can be generated
automatically using Real Time Process Algebra Code Generator (RTPACG) or be seamlessly
transformed into programs by the programmer.

[8], studied present approach to the optimization of cash management for Automated Teller
Machine network, his approach was based on artificial neural network to forecast a daily cash
demand for every ATM in the network and on the optimization procedure to estimate the optimal
cash load for every ATM. The optimization procedure considered important factors for ATM
maintenance; cost of cash, cost of cash uploading and cost of daily services. He then present the
simulation of the ATM network’s cash management optimization system which shows a good
results, but for practical implementation of the proposed system further experimental
investigations are necessary.

[9], researched into the appraisal of the use of ATM in the banking industry in Nigeria. The
research work involves the use of diffusion innovation model, and questionnaires to collect data
in order to get the required information from the respondents. The result of the hypothesis was
stated at the end of the analysis. The results shows that there is relative advantage of using
ATMs; how hard it was to use ATMs, how compatible ATMs were with the lifestyle of the
users; how much has been registered (observed) about ATMs by the users and whether ATMs
could be tested before consistent use, were issues that influence users’ attitude towards intention
to use ATMs.

[10]. conducted an investigative research on the use of ATM card. The authors made effort to
conduct an interview with structured questions among the ATM users, data obtained was
analyzed, the result proofed that a lot of problems was associated with the use of ATM
smartcard. Some of the problems identified are: difficult to prevent another person from attaining
and using a legitimate person’s card, conventional smartcard can be lost, duplicated, stolen or
impersonated with accuracy. To address the problems, the author proposed the use of biometric
voice-based access control system in automatic teller machine. In the proposed system, access
will be authorized simply by means of an enroll user speaking into a microphone attached to the
automatic teller machine.

[11]. Yekini NA et al. (2012) presented a voice-dependent access control system for ATMs.
Figure 3 shows the block diagram of the discussed system. The system consists of 3 important
components: Voice sensor e.g. microphone, speaker verification system and access control
system for ATM machine. A low cost voice sensor is used to record the ATM user’s voice and
subsequently send it to the voice-dependent verification system to validate the authenticity of the
user.

[12]. S. Iqbal et al. (2011) made use of MFCC technique for extracting unique and reliable voice
features in the form of Mel coefficients. It is based on two security measures, PIN number as
well as voice features to provide more security to the ATM transactions. It represents the best
possible efficiency of 86.67% with the error rate of 13.33% . H. Venugopal et al. (2012) used the
concept of voice recognition in ATM machine to carry out secured transaction. They discussed a
security based implementation of Hidden markov model algorithm (HMM) to measure speech
rate, frequency and modulation pitch detection algorithm (PDA) for calculating pitch of
voiceprints and Accent Classification (AC) for the analysis of accent in voice. There was a
noticeable increase of 15% in accuracy as compared to other prevalent systems.

[13]. R. Chavan et al. (2013) tested upon own created database in different surroundings. The
recognition accuracy increases due to the combination of MFCC and HMM. The performance is
evaluated by finding word wise accuracy for five different speakers. The accuracy varies from
word to word, here the accuracy obtained for first two words is 92% and third word is 88%. A.
Patil (2014) discussed the importance of security during ATM transactions. A major concern is
the accessibility of ATM machines to differently-abled people. These concerns can be overcome
by using iris recognition for authentication and voice enabled transactions in ATM machines.
Once the authenticity of the user is verified, the user is allowed to carry out further transactions
using voice based commands by speaking into a microphone.

[14].C. Upadhya et al. (2014) observed that MFCCs for every individual user was unique. Few
variations were observed due to different surroundings of the recording area. Then the MFCCs of
the stored template and real time input are compared for every user. In programming, Euclidean
distance is used to compare the template and real time input. In this manner, MFCC algorithm is
used for voice recognition. R. Kavitha et al. (2014) discussed that voice dependent authentication
system for Tamil language has two main phases, feature excerption phase and feature matching
phase. Feature excerption phase enables the system to extract the MFCC features from the input
sample. While the feature matching module identifies the Tamil words and the user by making
use of Dynamic Time Warping algorithm that measures the warping distance between the two
sequences.

[15]. A. H. Mansour et al. (2015) discussed DTW as a flexible mathematical method and gave
high accuracy results. The signal analysis by using MFCC provides us by spectrum coefficients
which represent the accurate vocal system for the stored word, with a high level of perception for
the human voices. The spectrum coefficients has a high specification which shows their
importance depending on the speaker himself. So the back propagation algorithm has been tested
for the voice recognition, and has achieved a success rate about 95%. W. Astuti & E. Riyandwita
(2016) evaluated the effectiveness of an automated voice-dependent engine starting system. 12
Cepstral coefficients are used in order to obtain the MFCC coefficients. These filters are
simulated by integrating the Fast Fourier transform spectrum of 20 millisecond Hamming-
window voice segments in which the frame rate is 10 milliseconds. The training as well as the
testing accuracy of the system achieved 98% and 92.15%, respectively.
Existing System

Various techniques that can be used for enhancing ATM security is-

 Finger print recognition

 Iris recognition
 Retina recognition
 Face recognition
 Google verification code
 Gestures recognition
 And so many

The existing techniques for user authentication, which involve the use of either passwords and
user IDs (identifiers) or RFID (radio frequency identification which is an automatic
identification technology that uses a tag embedded in the target entity to mark it with a unique
code like ATM cards) identification cards, PINs (personal identification numbers) and tokens
suffer from several limitations.

Nguyen Van Duy et al has described that a token can be a factor to authenticate user . Tokens
can be used as an additional method to prove one’s identity who he/she claims to be. A
dynamically generated text using a hardware device or programs related to authentication can be
used as the token to authenticate user. Tokens are divided into two mechanisms: 1) software-
based 2) hardware-based physical electronic medium, such as USB-based devices for
authentication.

Sulochana Sonkamble , Dr. Ravindra Thool expresses that the finger prints of a person have
been used as person identification from long time. Finger print is the pattern of ridges and valley
on the surface of a finger-tip which is unique in every human being. The finger prints of the
identical twins are different. It is made possible to scan finger prints of a person and it can be
used through the computer for number of applications such as identifying a person etc.

Odgerel Ayurzana et al define hand geometry as a pattern of hand shape and palm contain much
information for identifying person. Hand geometry which is used to identify unique pattern of
hand, shape and also the palm lines. Handgeometry were one of the first biometrics to prove
practical in use across a variety of real-world applications.
Sulochana Sonkamble says that face recognition is the commonly used biometric technique for a
person’s recognition. The most reliable approaches to face recognition are based on shape of
facial attributes, such as eyes, eyebrows, nose, lips, chin and the relationships of these attributes
which define the human facial attire. This technique involves many facial elements; systems
which deal with all these facial elements have some sort of difficulty in matching face images.

By Anas Aloudat, Katina Michael and Roba Abbas define that iris is the colored part in the
middle of the eye, just in front of the lens. The iris is “a thin diaphragm stretching across the
anterior portion of the eye and supported by the lens”.

Disadvantages

 Vulnerable to lose or theft as user may find it difficult to keep it safe at all times which is
quite impossible.
 If CARD is lost there is no way to interact with system.
 Needs a security Guard at ATM center.
 OTP may take time to be received on mobile
 Dry or dirty skin might intend to cause errors.
 It not applicable for differently abled people.
 Xinwei Liu et al consider that a finger print sample may have lot of ambiguities because
of many properties (scenery/imaging).
 There can be drawbacks to this recognition system as this may be intrusive and require a
lot of memory to store data. This may not be cost effective.

Proposed System

Voice is the most natural way for humans to communicate .A voice recognition system is
basically developed to identify an admin voice. MATLAB software which can be used for
coding the voice recognition, the authorized person’s voice can be authenticated and he can be
identified. The key is to convert the speech waveform to a *.wav representation for further
analysis and processing. The database will be consisting of a different set of patterns which will
represent the states required for the successful completion for a given phoneme.

Devices are controlled by voice recognition process and in this system templates are stored in
database according to the sampled speech power. The access may be processed by means of an
enrolled user speaking into a microphone attached to the system which undergoes several
processes. Then will decide whether the voice matches with the stored voice.

The voice recognition is divided into two phase which are training phase and testing phase. In
training phase, the voice is recorded as long as one second. After that, silence detection will
detect the actual uttered speech. The signal is then windowed. At first transformation, is the Fast
Fourier Transform that convert voice signal from time domain into frequency domain. The
MFCC convert to Mel Frequency Cepstrum. (MEL: change the scale of frequencies;
CEPSTRUM: log followed by inverse Fourier Transform). A spectrum shows information of the
frequency components while cepstrum show information about how the frequencies change to
determine the energy within each window.

The steps training phase are exactly repeated in the testing phase. From the spectrum value get
from energy within each window is used to determine value of mean square error and average
pitch. The value is then compared with the training phase. If the result is rejected, the system will
display “You are not the same user”.

Advantages

 ATM fraud and criminal activities can be reduced or eliminate completely.

 Enhance the use of ATM by banking customers.
 It will eliminate financial burden placed on customers for issuance and maintenances of
ATM card.
 It will also reduce stress emanated from complaint related to ATM cards at the customers
care of bank on bank staff, and customers that uses ATM.
CHAPTER 4

4. PROBLEM DESCRIPTION:
ATMs enable customers to perform financial transactions like cash withdrawal, check balances
or credit mobile phones with the help of the machine and without any human intervention. A
plastic ATM card with a magnetic stripe or a plastic smart card with a chip and a unique card
number and security information is used for the transaction. Authentication is done by using the
Personal Identification Number (PIN). ATM machines also allow the conversion of one currency
into another with the best possible exchange rates. Most cash machines are connected to
interbank networks. This enables people to withdraw and deposit money from machines not
belonging to the bank where they have their accounts or in the countries where their accounts are
held enabling cash withdrawals in local currency. A voice recognition system can be used for
different purposes. Machines can identify the speaker through the voice thus enhancing the
security and privacy of the transactions. The main objective of this work is to allow illiterate
people to operate the ATMs without reading the instructions for the cash transactions. Speech
recognition is still an intensive field of research in the area of digital signal processing due to its
versatile applications. Since speech is the primary means of communication between people,
research in automatic speech recognition and speech synthesis by machine has attracted a great
deal of attention over the past five decades. Speech recognition system mostly performs two
fundamental operations: signal modelling and pattern matching. During signal modelling, speech
signal is converted into a set of parameters using different feature extraction methods. The
extracted features are compared with the parameter set from memory of the ATM machine and
the pattern which closely matches the parameter set obtained from the input speech signal is
classified during the classification stage. Better feature is good for improving recognition rate
and the performance of a speech recognition system is measured in terms of recognition
accuracy. The main advantage of this system is that the user can interact with the ATM machine
through speech that too using only spoken digits.
CHAPTER 5:
5. PROPOSED SYSTEM:
5.1 INTRODUCTION:
Proposed system description, figure 2 shows the schematic diagram of proposed intelligent
voiced-based access control for ATMs. The proposed system basically consists of 3 main
components:

1) Voice sensor
2) Speaker verification system and
3) ATM access control
A low cost microphone commonly used in computer system is used as voice sensor to record the
ATM user voice. The recorded voice is then sent to the voiced-based verification system which
will verity the authenticity of the user based on his/her voice.
Voice base verification, will enable a decision signal which will accept or reject the access that
will be sent through the parallel port of the system. The intelligent voice-based access control
will generally make possible decision as;
4) Authorized person (ATM user) accepted
5) Unauthorized person (ATM user] rejected
The qualities to measure rate of access control accuracy to reject the authorized person is called
false rejection (FRR), and that to measure rate of access control to accept the unauthorized
person is called false acceptance rate [FAR]. Mathematically, both rates are expressed as
percentage using the following simple calculation. NFR and NFA are the numbers of false
rejections and false acceptance respectively, while NFA and NIA are the number of authorized
person attempts. To generate high security of the ATM access control system it is expected that
the proposed system have both low FRR and low FAR.
A. Voiced-based verification System The use of voice for biometric measurement becomes more
popular due to the following reasons; natural signal generation, convenient to process or
distributed, and applicable for remote access. There are 2 kinds of voicedbased recognition or
speaker recognition (Campbell J.P. 1997].
1) Speaker identification
2) Speaker recognition
In speaker verification system, the system decodes that a user is who claims to be. On other hand
speaker identification decides the person among a group of person. Speaker recognition is further
divided into 2 categories which are text dependent arid text independent. Tex dependent speaker
recognition recognizes the phrase that spoken, whereas in text identification the speaker can alter
any word. Thy most appropriate method for voice-based ATM access control is based on concept
of speaker verification, since the objective in the access control is accept or reject a user to gain
access into ATM. Figure 3a and b. show the basic structure of the proposed system. There are 2
phases in the proposed system.

Training enrolment as shown in figure 2a, the authorized persons are registered and their voices
are recorded. The recorded voices are then extracted. Features extracted from the recorded voices
are used to develop models of authorized persons.
Testing or operational phase as shown in figure 3b, in this phase a person who want to access the
ATM is required to enter the claimed identity and his/her voice. The entered voice is processed
and compared with the claimed person model to verify his/her claim. At this point system
decides whether the feature extracted from the given voice matches with the model of claimed
person. Threshold is set in order to give a definite answer of access acceptance or rejection.
When degree of similarity between a given voice and model is greater than threshold the system
will accept the access, otherwise the system will reject the person to access the ATM.
After studying the prevalent systems and their inaccuracies and results, a system is proposed
which will help in creating a secured ATM access by ruling out the problems associated with the
smartcard authentication system. The system consists of user training and database of an
authorised person. This will be then compared to the real time voice input through microphone.
The comparing process is carried out by feature extraction and feature matching with that of
stored samples of authorised customer in the database. If the matching is positive then the access
to the ATM machine is enabled using an embedded design.
System Requirements
Hardware Requirements:
• Processor : Dual core processor 2.6.0 GHZ
• RAM : 4 GB
• Hard disk : 160 GB
Software Requirements:
• Operating system : Windows OS ( 2007, 2008) 64 bit
• Front End : MATLAB 15

About MATLAB:
MATLAB 15
MATLAB is a high-performance language for technical computing. It integrates
computation, visualization, and programming in an easy-to-use environment where problems and
solutions are expressed in familiar mathematical notation. Typical uses include:
• Math and computation
• Algorithm development
• Modeling, simulation, and prototyping
• Data analysis, exploration, and visualization
• Scientific and engineering graphics
• Application development, including graphical user interface building
MATLAB is an interactive system whose basic data element is an array that does not require
dimensioning. This allows you to solve many technical computing problems, especially those
with matrix and vector formulations, in a fraction of the time it would take to write a program in
a scalar noninteractive language such as C or Fortran. The name MATLAB stands for matrix
laboratory. MATLAB was originally written to provide easy access to matrix software developed
by the LINPACK and EISPACK projects. Today, MATLAB uses software developed by the
LAPACK and ARPACK projects, which together represent the state-of-the-art in software for
matrix computation. MATLAB has evolved over a period of years with input from many users.
In university environments, it is the standard instructional tool for introductory and advanced
courses in mathematics, engineering, and science. In industry, MATLAB is the tool of choice for
high-productivity research, development, and analysis.
Toolboxes
MATLAB features a family of application-specific solutions called toolboxes. Very important to
most users of MATLAB, toolboxes allow you to learn and apply specialized technology.
Toolboxes are comprehensive collections of MATLAB functions (M-files) that extend the
MATLAB environment to solve particular classes of problems. Areas in which toolboxes are
available include signal processing, control systems, neural networks, fuzzy logic, wavelets,
simulation, and many others.

The MATLAB System

The MATLAB system consists of five main parts: Development Environment. This is the
set of tools and facilities that help you use MATLAB functions and files. Many of these tools are
graphical user interfaces. It includes the MATLAB desktop and Command Window, a command
history, and browsers for viewing help, the workspace, files, and the search path. The MATLAB
Mathematical Function Library. This is a vast collection of computational algorithms ranging
from elementary functions like sum, sine, cosine, and complex arithmetic, to more sophisticated
functions like matrix inverse, matrix eigenvalues, Bessel functions, and fast Fourier transforms.
The MATLAB language. This is a high-level matrix/array language with control flow
statements, functions, data structures, input/output, and object-oriented programming features. It
allows both “programming in the small” to rapidly create quick and dirty throw-away programs,
and “programming in the large” to create complete large and complex application programs.
Handle Graphics
This is the MATLAB graphics system. It includes high-level commands for two-dimensional and
three-dimensional data visualization, image processing, animation, and presentation graphics. It
also includes low-level commands that allow you to fully customize the appearance of graphics
as well as to build complete graphical user interfaces on your
MATLAB functions:
A MATLAB “function” is a MATLAB program that performs a sequence of operations specified
in a text file (called an m-file because it must be saved with a file extension of *.m). A function
accepts one or more MATLAB variables as inputs, operates on them in some way, and then
returns one or more MATLAB variables as outputs and may also generate plots, etc.
Some functions are
Imread() – Reading the image from the graphics file.
A = imread(filename, fmt) reads a grayscale or color image from the file specified by the
stringfilename. If the file is not in the current folder, or in a folder on the MATLAB path, specify
the full pathname.
The text string fmt specifies the format of the file by its standard file extension. For example,
specify'gif' for Graphics Interchange Format files. To see a list of supported formats, with their
file extensions, use the imformats function. If imread cannot find a file named filename, it looks
for a file named filename.fmt.
The return value A is an array containing the image data. If the file contains a grayscale
image, A is an M-by-N array. If the file contains a truecolor image, A is an M-by-N-by-3 array.
For TIFF files containing color images that use the CMYK color space, A is an M-by-N-by-4
array. The class of A depends on the bits-per-sample of the image data, rounded to the next byte
boundary. For example, imread returns 24-bit color data as an array of uint8 data because the
sample size for each color component is 8 bits.
[X, map] = imread(...) reads the indexed image in filename into X and its associated colormap
into map. Colormap values in the image file are automatically rescaled into the range [0,1].
Image- write image to graphics file
imwrite(A,filename,fmt) writes the image A to the file specified by filename in the format
specified by fmt.
A can be an M-by-N (grayscale image) or M-by-N-by-3 (truecolor image) array, but it cannot be
an empty array. For TIFF files, A can be an M-by-N-by-4 array containing color data that uses
the CMYK color space. For GIF files, A can be an M-by-N-by-1-by-P array containing grayscale
or indexed images — RGB images are not supported. For information about the class of the
input array and the output image. filename is a string that specifies the name of the output file.
fmt can be any of the text strings listed. This list of supported formats is determined by the
MATLAB image file format registry. See imformats for more information about this
registry.imwrite(X,map,filename,fmt) writes the indexed image in X and its associated
colormap map tofilename in the format specified by fmt. If X is of
class uint8 or uint16, imwrite writes the actual values in the array to the file. If X is of
class double, imwrite offsets the values in the array before writing, using uint8(X–1). map must
be a valid MATLAB colormap. Note that most image file formats do not support colormaps with
more than 256 entries. When writing multiframe GIF images, X should be an 4-dimensional M-
by-N-by-1-by-P array, where P is the number of frames to write.imwrite(...,filename) writes the
image to filename, inferring the format to use from the filename's extension.
imwrite(...,Param1,Val1,Param2,Val2...) specifies parameters that control various characteristics
of the output file for HDF, JPEG, PBM, PGM, PNG, PPM, and TIFF files. For example, if you
are writing a JPEG file, you can specify the quality of the output image. For the lists of
parameters available for each format.
MATLAB applications:
The MATLAB Application Program Interface (API). This is a library that allows you to
write C and Fortran programs that interact with MATLAB. It include facilities for calling
routines from MATLAB (dynamic linking), calling MATLAB as a computational engine, and
for reading and writing MAT-files.
What Is Simulink?
Simulink, a companion program to MATLAB, is an interactive system for simulating
nonlinear dynamic systems. It is a graphical mouse-driven program that allows you to model a
system by drawing a block diagram on the screen and manipulating it dynamically. It can work
with linear, nonlinear, continuous-time, discrete-time, multirate, and hybrid systems. Blocksets
are add-ons to Simulink that provide additional libraries of blocks for specialized applications
like communications, signal processing, and power systems.
Real-Time Workshop is a program that allows you to generate C code from your block diagrams
and to run it on a variety of real-time systems.
Stateflow is an interactive design tool for modeling and simulating complex reactive
systems. Tightly integrated with Simulink and MATLAB, Stateflow provides Simulink users
with an elegant solution for designing embedded systems by giving them an efficient way to
incorporate complex control and supervisory logic within their Simulink models. With Stateflow,
you can quickly develop graphical models of event-driven systems using finite state machine
theory, statechart formalisms, and flow diagram notation. Together, Stateflow and Simulink
serve as an executable specification and virtual prototype of your system design.
Conclusion

Nowadays, ATMs have become a definite requirement for carrying out millions of transactions
that happen in day to day life. Large number of frauds can be carried out in ATM machines,
significantly because of PIN disclosure or forceful collection. This paper presents a conceptual
framework that uses voice-dependent access control system in ATMs. Recent studies have
shown that ATM users have come across many problems in day to day life which include; chip
distortion, card misplacement. Card fraud etc. These problems are mainly associated with the
smartcards. In order to overcome these problems, it is highly recommended that the banking
sector should make the use of voice biometric technique in ATMs, as this will eradicate the
loopholes that are prevalent in current system that uses smartcard to access an ATM. So, the
proposed system will not only improve the security during money transactions but also make
ATMs user friendly for the less educated. This method when fully taken into practise will not
only enhance secured and correct authentication, but will also render support in the
implementation of complex ATMs (performs deposits and money transfer), as this system
provides enhanced security.
Future Enhancement

Our future research will also address the issues of privacy through kiosk design, through the use
of highly directional speakers to create private audio zones that do not require physical barriers;
the quality of speech input through utilizing highly directional microphones. We also plan to
introduce voice commands in various languages so that even people without knowledge of
English can use ATM facility
References

[1]Yazeed Abdul Mumin1, Yazidu Ustarz1, Ibrahim Yakubu (2013) Automated Teller Machine
(ATM) Operation Features and Usage in Ghana: Implications for Managerial Decisions Journal
of Business Administration and Education ISSN 2201-2958 Volume 5, Number 2, 2014, 137-
157

[2] Odusina, Ayokunle Olumide (2014) Automated Teller Machine Usage And Customers’
Satisfaction In Nigeria Elite Research Journal of Accounting and Business Management Vol.
2(3) pp. 43 - 47, July, 2014. Available online
http://www.eliteresearchjournals.org/erjabm/index.htm

[3] Heli Snellman (2006) Automated Teller Machine network market structure and cash usage
Scientific monographs ISBN 952-462-318-8 ISSN 1238-1691(print) ISBN 952- 462-319-6 ISSN
1456-5951 (online) [4] Mosabber Hossain (2006). Understanding of ATM (Automated Teller
Machine) in Bangladesh. HTTP: dspace.bracu.ac.bd/.../Understanding+of+ATM+in+Bangl
adesh.pdf?...1

[5] Ahmed Baita Garko (2011) Enhancing The Current Automated Teller Machine (ATM) In
Nigerian Banking Sector JORIND 9(2) December, 2011. ISSN 1596 – 8308.
www.transcampus.org. www.ajol.info/journals/jorind

[6] Mary K. (2008). Breathe New Life into Your ATM Program with Progressive Marketing and
Customer Relationship Management Tools. A First Data White Paper, retrieved from
http://www.firstdata.com/en_gr/insights/atmprogramrecommendations-. Html, available on
19/7/2015.

[7] Yingxu W. (2010). The Formal Design Model of an Automatic Teller Machine, International
Journal of Software Science and Computational Intelligence, 2(1), 102-131.

[8] Rimvydas Simutis, Darius Dilijonas, Lidija Bastina, Josif Friman, & Pave Drobinov. (2007).
the optimization of Cash Management for ATM network. Information Technology and Control,
Vol.36, No.1A, retrieved from http://itc.ktu.lt/itc361/simutis361.pdf, available on 14/1/2015
[9] Yekini et al (2012) Automated Biometric Voice-Based Access Control in Automatic Teller
Machine (ATM) (IJACSA) International Journal of Advanced Computer Science and
Applications, Vol. 3, No.6, 2012.

[10] Fabumni, O.A. (2011). Appraisal of the Use Automated Teller Machine in the banking
industry in Nigeria. HTTP: https://www.unilorin.edu.ng/studproj/cis/0730gc071.pdf. Retrived
july 25th , 2015.

[11] Yekini NA, Oyeyinka IK, Itegboje AO, and Akinwole AK, “Automated Biometric Voice-
Based Access Control in Automatic Teller Machine (ATM),” International Journal of Advanced
Computer Science and Applications, vol. 3, no. 6, pp. 71–77, 2012.

[12] S. Iqbal, T. Mahboob, and M. S. H. Khiyal, “Voice Recognition using HMM with MFCC
for Secure ATM,” International Journal of Computer Science Issues, vol. 8, no. 6, pp. 297–303,
2011.

[13] A. Patil, R.D. Salagar, “Voice Enabled ATM Machine with Iris Recognition for
authentication,” Proceedings of 3rd IRF International Conference, pp. 70–75, 2014.

[14] R. Kavitha, N. Nachammai, R. Ranjani, and J. Shifali, “Speech Based Voice Recognition
System for Natural Language Processing,” International Journal of Computer Science and
Information Technologies, vol. 5, no. 4, pp. 5301–5305, 2014.

[15] A. H. Mansour, G. Zen, A. Salh, H. Hayder, and Z. Alabdeen, “Voice recognition Using
back propagation algorithm in neural networks,” International Journal of Computer Trends and
Technology, vol. 23, no. 3, pp. 132–139, 2015.

[16] W. Astuti and E. Byan Wahyu Riyandwita, “Intelligent Automatic Starting Engine Based on
Voice Recognition,” IEEE Student Conference on Research and Development (SCOReD) 2016.