International Journal of Engineering Trends and Technology (IJETT) - Volume4 Issue6- June 2013
Design For The Automation Of Inter-Networked
Banking Using Universal Sim
C. Chitra,
Professor,
Department of ECE,
PSNACET,Dindigul.
chitrasrivi@gmail.com
Contact no:97504 40704
T.Suganya Thevi,
Assistant Professor,
Department of ECE,
PSNACET,Dindigul.
suganyaathevi@gmail.com
Contact no : 97510 56856
ABSTRACT
Automated teller machines (ATMs) are
well known devices typically used by individuals
to carry out a variety of personal and business
financial transactions and/or banking functions.
ATMs have become very popular with the
general public for their availability and general
user friendliness. But the drawback in the
existing ATM system is that user should carry
the ATM card without fail. And also ATM card
can be misused by the unauthenticated person.
To overcome this problem, creation of new
generation ATM machine can be operated
without ATM card. In this proposed system the
image recognition technique can provide the
important function of security maintenance such
as identification of users and surveillance of the
ATM environment. By using this system need of
ATM card is completely eliminated. We can
operate the ATM machine by using our SIM
itself.
1.
When we insert our SIM in the reader unit of the
ATM machine it transfers the mobile to the
server. In the server, related information of the
mobile number, (i.e) users account details, their
photo etc is collected. The camera near the ATM
machine will capture the user’s image and
compare it with the user image in the server
using MATLAB. Only when the image matches
it asks the pin number and further processing
starts. Otherwise the process is terminated. So by
using this system need of ATM card is
completely eliminated. Malfunctions can also be
avoided. Our transaction will be much secured.
Mobile scanning device scans SIM
number through GSM Modem. Collected data is
given to the Teller machine for further
processing. At the same time, web camera
captures the images and compares using digital
signal processing. If images and PIN number are
same then further processing is continued.
Otherwise it gives alarm through alert module to
the user’s mobile.
INTRODUCTION
2. IMPLEMENTATION
Existing ATMs are convenient and easy
to use for most consumers. Existing ATMs
typically provide instructions on an ATM display
screen that are read by a user to provide for
interactive operation of the ATM.
Having read the display screen instructions, a
user is able to use and operate the ATM via data
and information entered on a keypad. However
the drawback in the existing system is that the
user should carry their ATM card without fail.
But in many cases we forget it. This system
helps us to use the ATM machine without the
ATM card.
By using this system, ATM machine
can be operated by our SIM in the mobile phone.
ISSN: 2231-5381
To implement this automation of
internetworked banking, microcontroller is
programmed. Now it is wondering that the nonmentioning of memory space meant for the
program storage is the most important part of
any embedded controller. Originally this 8031
architecture was introduced with on chip, ‗one
time programmable version of Program Memory
of size 4K X 8. Intel delivered all these
microcontrollers (8051) with user‘s program
fused inside the device. The memory portion was
mapped at the lower end of the Program Memory
area. But, after getting devices, customers
couldn‘t change anything in their program code,
which was already made available inside during
device fabrication. So, very soon Intel introduced
the 8031 devices (8751) with re-programmable
http://www.ijettjournal.org
Page 2417
International Journal of Engineering Trends and Technology (IJETT) - Volume4 Issue6- June 2013
type of Program Memory using built-in EPROM
of size 4K X 8. Like a regular EPROM, this
memory can be re-programmed many times.
Later on Intel started manufacturing these 8031
devices without any on chip Program Memory.
The AT89S51 is a low-power, high-performance
CMOS 8-bit microcontroller with 4K bytes of
In-System Programmable Flash memory. The
device is manufactured using Atmel‘s highdensity nonvolatile memory technology and is
compatible with the Indus-try-standard 80C51
instruction set and pin out. The on-chip Flash
allows the program memory to be reprogrammed
in-system or by a conventional nonvolatile
memory pro-grammar. By combining a versatile
8-bit CPU with In-System Programmable Flash
on a monolithic chip, the Atmel AT89S51 is a
powerful microcontroller which provides a
highly-flexible and cost-effective solution to
many embedded control applications. The
AT89S51 provides the following standard
features: 4K bytes of Flash, 128 bytes of RAM,
32 I/O lines, Watchdog timer, two data pointers,
two 16-bit timer/counters, a five-vector two-level
interrupt architecture, a full duplex serial port,
on-chip oscillator, and clock circuitry.
Since universal SIM is used for the
communication of information between the
ATM and the server, a wireless modem is
needed. Here GSM modem is used. A GSM
modem is a wireless modem that works with a
GSM wireless network. A wireless modem
behaves like a dial-up modem. It can be an
external device or a PC Card / PCMCIA Card.
Typically, an external GSM modem is connected
to a computer through a serial cable or a USB
cable. A GSM modem in the form of a PC Card /
PCMCIA Card is designed for use with a laptop
computer. It should be inserted into one of the
PC Card / PCMCIA Card slots of a laptop
computer. Like a GSM mobile phone, a GSM
modem requires a SIM card from a wireless
carrier in order to operate. As mentioned in
earlier sections of this SMS tutorial, computers
use AT commands to control modems. Both
GSM modems and dial-up modems support a
common set of standard AT commands. You can
use a GSM modem just like a dial-up modem. In
addition to the standard AT commands, GSM
modems support an extended set of AT
commands. These extended AT commands are
defined in the GSM standards. With the extended
AT commands, you can do things like
 Reading, writing and deleting SMS
messages.
ISSN: 2231-5381
 Sending SMS messages.
 Monitoring the signal strength.
 Monitoring the charging status and
charge level of the battery.
 Reading, writing and searching phone
book entries.
GSM provides recommendations, not
requirements. The GSM specifications define the
functions and interface requirements in detail but
do not address the hardware. The reason for this
is to limit the designers as little as possible but
still to make it possible for the operators to buy
equipment from different suppliers. The GSM
network is divided into three major systems: the
switching system (SS), the base station system
(BSS), and the operation and support system
(OSS). The basic GSM network elements are
shown in below figure 1.
Fig.1 GSM Network Elements
The types of software tools used here are,
 MPLAB IDE
 MATLAB
 PCA Algorithm
 ORCAD - Capture
2.1 PCA
ALGORITHM
RECOGNITION SYSTEM
-
FACE
Automatic face recognition systems try
to find the identity of a given face image
according to their memory. The memory of a
face recognizer is generally simulated by a
training set. In this project, our training set
consists of the features extracted from known
face images of different persons. Thus, the task
of the face recognizer is to find the most similar
feature vector among the training set to the
feature vector of a given test image. Here, the
identity of a person where an image of that
http://www.ijettjournal.org
Page 2418
International Journal of Engineering Trends and Technology (IJETT) - Volume4 Issue6- June 2013
person (test image) is given to the system is
recognized. PCA is itself a feature extraction
algorithm.
In the training phase, the feature vectors
for each image are extracted in the training set.
Let ΩA be a training image of person A which
has a pixel resolution of M × N (M rows, N
columns). In order to extract PCA features of
ΩA, the image is converted into a pixel vector
ΦA by concatenating each of the M rows into a
single vector. The length (or, dimensionality) of
the vector ΦA will be M ×N. Use the PCA
algorithms a dimensionality reduction technique
which transforms the vector ΦA to a vector ΩA
which has a dimensionality d where d «M ×N.
For each training image Ωi, you should calculate
and store these feature vectors Ѡi .
In the recognition phase (or, testing
phase), you will be given a test image Ωj of a
known person. Let αj be the identity (name) of
this person. As in the training phase, you should
compute the feature vector of this person using
PCA and obtain Ѡj. In order to identify Ωj, you
should compute the similarities between Ѡj and
all of the feature vectors Ѡi‘s in the training set.
The similarity between feature vectors can be
computed using Euclidean distance. The identity
of the most similar Ѡi will be the output of our
face recognizer. If i = j, it means that we have
correctly identified the person j, otherwise if i ≠
j, it means that we have misclassified the person
j. Schematic diagram of the face recognition
system that will be implemented is shown in
Figure 2.
Fig.2. Schematic diagram of a face
ISSN: 2231-5381
3.1.2 USE OF PCA
The use of PCA as a feature extractor is
explained. Assume that we have p training
images: Ωi, i = 1,2,…, p. For each training
image, you should form pixel vectors Φi where
Φi €Ŕk<k;(k = M × N). Our aim is to compute
feature vectors Ѡi where Ѡi €< Ŕd, (d « k). In
order to apply PCA to the training set, you
should first form a training data matrix A which
contains p rows: at each row Φi‘s are stored.
Thus the dimensionality of A is p × k. First, you
should compute the covariance matrix of A: CA.
Then the eigen values and their corresponding
eigenvectors of CA should be computed. There
will be k eigen value and eigenvector pairs
where each eigenvector is of dimensionality k.
Sort the eigen values in decreasing order, and
select the biggest d eigen value and eigenvector
pairs. Form the transformation matrix Ψ by
simply putting the selected eigenvectors as
columns of Ψ. You will use Ψ to compute Ѡi‘s
from Φi‘s. The computation of Ѡi is simply by
Ѡi= ѰTΦiT……………………… (1)
where ѰTand ΦiT are the transposes of Ѱ and
Φi, respectively.
Note that each column of Ѱ
corresponds to an eigenvector which is of length
k. This is equal to M × N which is the
dimensionality (resolution) of input images.
Thus, you can convert each eigenvector to an
image by reversing the concatenation operation.
These converted eigenvector images are called
eigen faces since they are similar to human faces.
Figure 2 shows 20 eigen faces that correspond to
the largest 20 eigen values of the ORL face
database. Once you obtain Ѡi‘s using the largest
d eigen vectors, you can reconstruct the image of
person i. If you use all k eigen vectors instead of
d when forming Ѱ, the reconstructed image will
the same as image Ωi.However, since our aim is
dimensionality reduction and d « k, reconstructed
image Ὡi will be an approximation of the actual
Ωi. You can reconstruct Ὡi by converting the
pixel vector: Φi = (ѰѠi) T to an image of
resolution M×N. Figure 2 shows the
reconstructed images of two persons using
different number of eigenvectors. Notice that if
you use more eigen vectors, then the
reconstructed image is more similar to the
original face image.
http://www.ijettjournal.org
Page 2419
International Journal of Engineering Trends and Technology (IJETT) - Volume4 Issue6- June 2013
The implementation details of PCA are
given below. You will use the ORL face
database which contains 10 different images of
each of the 40 subjects. For each subject five
images (instances) will be put into training set
and the rest of the images will be put into the test
set. Training and test set images will be under
train images and test images directories,
respectively. A sample MATLAB code will be
provided to you which automatically reads the
images under these directories.

The pseudo-code of
MATLAB code is as follows:


1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
the



sample
Set image resolution parameter
(im_res)
Set PCA dimensionality parameter
(PCA_DIM)
Read training images
Form
training
data
matrix
(M_train_data)
Form training class labels matrix
(M_train_labels).
Calculate PCA transformation matrix
(tmatrix).
Calculate feature vectors of all
training images using tmatrix.
Store training feature vectors in a
matrix.
Read test faces.
For each test face do step 4 to 8.
Compute the distances between test
feature vector and all training vectors.
Store the distances together with the
training class labels
Initialize error count to zero.
For each test face do step 1 to 8.
Using the distance data, determine the
person ID of the most similar training
vector.
If the found ID is not equal to the ID of
the test image increment error count.
Output
the
correct
recognition
accuracy : (1 - (error count/ total test
image count))*100 recognizer.

Fig. 3: Unauthorized person

4. RESULTS & DISCUSSIONS
The steps to be done to perform this design
are discussed below.


Open the MATLAB file and run the
program. A window will be open. Click the
‘Input’ button to give the input image.
ISSN: 2231-5381
Web cam window will open automatically.
Press enter button to take input image.
Add database image in the same manner. A
dialog box will be open. It shows the
message database successfully added to
database.
PCA algorithm compares the input and
database image. If the two images are same
then the Message ‘Authorized person’ will
be shown.
Then it will ask bank name. Select the bank
and click sign in.
Give pin no and click sign in.
In the next step, if we go with withdraw
option, it will ask to enter the amount to
withdraw. Enter the amount and click
process. Then the transaction will be
completed.
If the two images do not match it will show
the message ‘Not authorized’ and ask the
port id and gsm id number. It is shown in
figure 3.
In case of emergency, we may send our
relations or friends to take money. If it is not
authorized, a random pin no will be
automatically sent to our mobile. We can
give the pin number to them for
authentication.
Thus the transaction will be completed
successfully. The completed transaction is
shown in figure 4.
http://www.ijettjournal.org
Page 2420
International Journal of Engineering Trends and Technology (IJETT) - Volume4 Issue6- June 2013
SMC , Vol.SMC-6, No.8, Oct. 1977,
pp. 562-570.
Fig. 4: Completed transaction
2.
M. Ejiri, H. Yoda, H.Sakou and Y.
Sakamoto,
“Knowledge-Directed
Inspection for Complex Multilayered
Patterns,”
Machine
Vision
and
Applications, vol.2, no.3, 1989, pp.155166H.
3.
Jain, R. Bolle, and S. Pankati, “
Biometrics: personal identification in
networked society “, Kluwer Academic
Publishers, 1998.
H. Sako, M. Seki, N. Furukawa, H.
Ikeda and A. Imaizumi, “Form Reading
based on Form-type Identification and
Form-data Recognition,” Proc. of
International Conference on Document
Analysis and Recognition (ICDAR
2003), 2003, pp. 926-930.
4.
CONCLUSION
Today, single factor authentication, e.g.
passwords, is no longer considered secure in the
internet and banking world. Easy-to-guess
passwords, such as names and age, are easily
discovered by automated password-collecting
programs. Two factor authentication using
mobile phones has recently been introduced to
meet the demand of organizations for providing
stronger authentication options to its users. The
proposed system has two option of running,
either using a free and fast connection-less
method or a slightly more expensive SMS based
method. Both methods have been successfully
implemented and tested, and shown to be robust
and secure. The system has several factors that
make it difficult to hack.
REFERENCES
1.
S. Kashioka, M. Ejiri and Y.
Sakamoto,’A Transistor Wire-Bonding
System Utilizing Multiple Local Pattern
Matching Techniques,’ IEEE Trans. On
ISSN: 2231-5381
5.
6.
7.
S. Hoque, M.C. Fairhurst, F. Deravi,
W.G.J. Howells, “ On the Feasibility of
Generating Biometric Encryption Keys”
IEE Electronics Letters, 41(6), 309-311,
2005.
Cellan-Jones, R., London starts digital
cash trial in Technology correspondent,
BBC
News
Online,
http://news.bbc.co.uk/1/hi/tec
hnology/7117213.stm. 2007: UK.
H. Fujisawa, M. Koga, H. Ogata, T.
Kagehiro and H. Sako, “Recognition
Strategies for Japanese Mailpiece
Recipient Addresses,” Proc. of 2nd
International Conference on Multimodal
Interface (ICMI '99), 1999, pp. III6974.
http://www.ijettjournal.org
Page 2421