IRJET- Fake Paper Currency Recognition

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 04 | Apr 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 3717
Fake Paper Currency Recognition
Prof. Anita Lahane1, Ashwin Pandey2, Mihir Palyekar3, Ishwar Bhangare4
1Professor, Department of Computer Engineering, Rajiv Gandhi Inst. Of Technology, Mumbai, Maharashtra
2,3,4Student, Department of Computer Engineering, Rajiv Gandhi Inst. Of Technology, Mumbai, Maharashtra
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - The advancement of color printing technology
has magnified the speed ofpretend currency noteprinting and
duplicating the notes on an awfully giant scale. A few years
back, the printing could be done in a print house, but now
anyone can print a currency note with maximum accuracy
using a simple laser printer. As a result, the difficulty of
pretending notes rather than the real ones has
been magnified terribly for the most part. And counterfeit of
currency notes is additionally a giant drawback to that.
This results in the style of a system that detects
the faux currency note in very less time.
The planned system offers an associate approachtoverifythe
Indian currency notes. Verification of currency note is
finished by the ideas of the image processing. This project
includes extraction of various features of Indian currency
notes which are “security thread, serial number, latent image,
watermark, identification mark”.
The planned system has blessings like simplicity and high-
performance speed. The result can predict whether or not the
currency note is faux or not. The basic logic is developed using
Image acquisition, grayscaleconversion, edge detection, image
segmentation, feature extraction, and comparison. The
features of the note to be tested are compared with the data-
set formed from the original magnified image and finds out
whether it is fake or original. The most vital challenge is
consistently and methodologically repetition the analysis
method to scale back human error and time.
Key Words: Pre-processing, Recognition, Human Error,
Currency, Random Forest Algorithm, Neural Net
Algorithm, Application
1. INTRODUCTION
Technology is growing very fast these days.
Consequently, the banking sector is
additionally obtaining modern-day by day. This brings a
deep need for automatic fake currency detection in an
automatic teller machine and automatic goods seller
machine. Many researchers are inspired to
develop strong and economical automatic currency
detection machine. An automatic machine which
might sight banknotes is currently wide utilized
in dispensers of contemporary product like candies, soft
drinks bottle to bus or railway tickets. The technology of
currency recognition essentially aims
for characteristic and extracting visible and
invisible options of currency notes. Until
now, several techniques are projected to spot the
currency note. But the simplest approach is to use the
visible options of the note [1]. For example, color and
size. But this manner isn't useful if the note is dirty or
torn. If a note is dirty, its color characteristic is changed
widely. So it is important how we extract the features of
the image of the currency note and apply the proper
algorithm to improve accuracy to recognize the note.
2. LITERATURE REVIEW
f(x) = Fax + Fb (1)
where x is the given (input) image in grayscale, f(x) is the
resultant image; and Fa, Fb and N are selected 3, -128 and 50
respectively [1]. In this technique, the algorithm depends on
the number of paper currency denominations. Here, the
complexity of the system increases by increasingthenumber
of classes. Therefore, this technique can be used only for the
recognition of a small number of banknote denominations.
The technique discussed inthispaperisnotdependentonthe
number of paper currency classes. The features presented in
this paper are independent of the way that a paper currency
is placed in front of the sensor.
It must be famous that the mentioned technique might
not be ready to differentiate real notes from counterfeits.
Indeed, strategies like [8] that use infrared or ultraviolet
spectra is also used for discriminating between real and
counterfeits notes.
Presently, there square
measure variety of strategies for folding money recognition
[1][2][3]. Using symmetricalmaskshasbeen employedin [2]
for recognizing folding money in any direction. In this
technique, the summationofnon-maskedpixelvaluesineach
banknote is evaluated and fed to a neural network for
recognizing paper currency. In this technique, two sensors
are used for recognition of the front and back of the paper
currency, but the image of the front is the only criterion for
decision. In another study for folding money recognition [1],
initially, the edges of patterns on a paper currency are
spotted. In the next step, paper currency is divided into N
equal parts along the vertical vector. Then, for every come
near these components,thenumberofpixelsisaddedandfed
to a three-layer, back propagation neural network. In
this method, to conquer the problem of recognizing dirty
worn banknotes, the following linear function is used as a
pre-processor:

3. METHODOLOGY
The proposed methodology consists ofthree parts.Inthe
initial half, the banknotes square measure scanned and also
the info is developed. After scanning, the banknotes square
measure pre-processed for noise as second a part of the
system. In the third half, vital currency options square
measure elects and extracted. The selected options square
measure simplyextractable andhavesensiblediscrimination
power. These features are passed to the neural network for
classification in the fourth part. The fifth part shows the
experimentation results. All these elements are delineated
within the resultant sections. The procedures are as follows,
3.1 Image Acquisition
Performing image acquisition in image processing is
always the first step in the work-flow sequence because,
without an image, no processing is possible. After the image
has been obtained, varied ways of the process are often
applied to the image to perform the numerous totally
different vision tasks. There are various ways to acquire
images such as with the help of a camera or scanner. The
acquired image should retain all the features.
3.2 Pre-processing
The main goal ofthepre-processingtoreinforcethevisual
look of pictures and improve the manipulation of knowledge
sets. Image pre-processing, conjointly known as image
restoration, involves the correction of distortion,
degradation, and noise introduced during the imaging
process. Interpolation is the technique mostly used for tasks
such as zooming, rotating, shrinking, and geometric
corrections. Removing thenoiseisaveryimportantsteponce
the process is being performed. However, noise affects
segmentation and pattern matching.
3.3 Gray Image
In photography and computing, a grayscale or grayscale
digital image is an image in which the value of each pixel is a
single sample, that is, it carries only intensity information.
Images of this type, also known as black-and-white, are
composed exclusively of shadesofgray,varyingfromblackat
the weakest intensity to white at the strongest.
Grayscale pictures square measure distinct from one-bit bi-
tonal black-and-white pictures, which in the context of
computer imaging are images with onlythetwocolors,black,
and white (also called bi-level). grayscale pictures have
several reminder grays in between.
Grayscale pictures square measure usually the results of
measure the intensity of sunshine at every element during a
single band of the spectrum (e.g. Infrared, visible radiation,
ultraviolet, etc.), and in such cases, they are monochromatic
proper when only a given frequency is captured. But
conjointly they will be synthesized from a full-color image;
see the section regarding changing to grayscale.
3.4 Binarization
The image acquired is in RGB color. It is converted into
grayscale because it carries only the intensity information
which is easy to process instead of processing three
components R (Red), G (Green), B (Blue). To take the RGB
worth's for every element and build as output one value
reflective the brightness of that element. One such approach
is to require the type of contribution from every channel:
(R+B+C)/3. However, since the perceived brightness is
commonly dominated by the inexperienced part, a different,
more “human-oriented”, a method is to take a weighted
average,
e.g.: 0.3R + 0.59G + 0.11B.
3.5 Edge Detection
Edge detection is that the name for a group of
mathematical ways that aim at distinguishingpointsduringa
digital image at that the image brightnesschangessharplyor,
additional formally, has thesecontinuities. The points at that
image brightness changes sharply square measure usually
organized into a group of flexuous line segments termed
edges. Edge detection is a picture process technique for
locating the boundaries of objects at intervals pictures. It
works by detecting discontinuities in brightness. Edge
detection is employed for image segmentation and
knowledge extraction in areas like image process, computer
vision, and machine vision.
3.6 Image Segmentation
Image segmentation is themethodofpartitioningadigital
image into multiple segments (sets of pixels, also known as
super-pixels). The goal of segmentation is to modify and/or
modification the illustration of a picture into one thing that's
additional important and easier to research. Image
segmentation is often wont to find objects and limits (lines,
curves, etc.) in pictures [7].
4. FEATURES EXTRACTED
The features that are extracted to check the credibility of
the note are as follows,
1. Variance of wavelet transformed image.
2. Skewness of wavelet transformed image.
3. Curtosis of wavelet transformed image.
4. Entropy of the image.

5. SYSTEM ARCHITECTURE
Fig -1: System Architecture
6. ACCURACY COMPARISON
We have used two algorithms those are Random Forest
Classifier Algorithm and Neural Net and the after comparing
the results provided by both the algorithms following is the
accuracy provided by them,
Fig -2: Accuracy Comparison
The accuracy provided by Random Forest Classifier
Algorithm is
Accuracy - 99.12536443148689 %
And the confusion matrix generated is,
Fig -3: Confusion Matrix
7. CONCLUSION
In this paper, an efficient approach is proposed to
extract the features of Indian currency notes and recognize
it. The project also contains the fake currency detection and
authentication of Indian Rupee. Our future work will be
concentrated on the extraction of features from various
currency notes belonging to different countries as well as
recognition and classification. Currently, the system is
designed only for the computer.Also,thesystemisrestricted
to Indian paper currency only. In the near future, we will be
developing a mobile based application whichwill runon any
platform. Also, we will be adding currency detection for
various well-known countries.
ACKNOWLEDGEMENT
We would like to specific our sincere feelingtoDr. Sanjay
U. Bukade, Principal, and Dr. SatishY.Ket,H.O.DofComputer
Department of Rajiv Gandhi Institute of Technology for
providing us an opportunity to do our project work on“Fake
paper currency recognition using image processing ".
This project bears on the imprint of many peoples. We
sincerely thank our project guide Prof. Anita A. Lahane for
her guidance and encouragement in carrying out this
synopsis work.
Finally, we might prefer to impart our colleagues and
friends United NationsagencyhelpedNorthAmericannation
in finishing the Project precis work with success.
REFERENCES
[1] Yu-Ichi Ohta, Takeo Kanade, and ToshiyukiSakai,"Color
information for region segmentation," Computer
Graphics and Image Processing, vol. 13, pp. 222 - 241,
1980.
[2] J. Geronimo D, Phardin PM Assopost, “Fractal functions
and Wavelet expansions based on several Scaling
Function” Approx. Theory, pp.373-401, 1994.
[3] M. Gori, A. Frosini and P. Priami. "A neural network-
based model for paper currency recognition and
verification", IEEE Trans. Neural Networks, ppI482-
1490,Nov.1996.
[4] A Frosini, M Gori, and P Priami,"Aneural network-based
model for paper currency recognition and verification,"
IEEE Transactions on Neural Networks,vol.7,pp.1482-
1490, 1996.
[5] Sun Baiqing and Fumiaki Takeda, "Proposal of Neural
Recognition with Gaussian Function and Discussion for
Rejection Capabilities to Unknown Currencies,"
Knowledge-Based Intelligent Information and
Engineering Systems, vol. 3213, pp. 859-865, 2004.
[6] JiQian, DongpingQian, Mengjie Zhang, A Digit
Recognition System for Paper Currency Identification
Based on Virtual Instruments” 1- 4244-0555-6/06,
2006.
[7] Feature Extraction for paper currency recognition,
Department of Computer and Electrical Engineering
Noushirvani Institute of Technology, University of
Mazandaran P.O.BOX 47144,babol,Iran ,IEEE 2007.
[8] Hamid Hassanpour and Payam M Farahabadi, "Using
Hidden Markov Models for paper currencyrecognition,"
Expert Systems with Applications, vol. 36, no. 6, pp.
10105-10111, 2009.
[9] Rafael C. Gonzalez, Richard E. Woods and Steven L.
Eddins, Digital Image Processing using MATLAB.
Pearson Education, 2009.

[10] Junfang G, Yanyun Z, Anni C. A reliable method forpaper
currency recognition based on LBP, Proceedings of the
2nd IEEE International Conference on Network
Infrastructure and Digital Content, p. 359-363; 2010.
[11] Bu-Qing Cao and Jian-Xun Liu, "Currency Recognition
Modeling Research Based on BP Neural Network
Improved by Gene Algorithm," 10. Second International
Conference on Computer Modeling and Simulation, vol.
2, pp. 246 -250, 2010.
[12] Indian currency recognition based on texture analysis,
INSTITUTE OF TECHNOLOGY, NIRMA UNIVERSITY,
AHMEDABAD – 382 481, 08-10, IEEEDECEMBER,2011,
AHMEDABAD – 382481,08-10,IEEE,DECEMBER,2011.
[13] Feature extractionofcurrencynotes:Anapproachbased
on wavelet transform.Amir Rajaei, Elham Dallalzadeh,
Mohammad Imran Department of Studies in Computer
Science, University ofMysore,Manasagangothri,Mysore
- 570 006 ,Mysore, Karnataka, India IEEE 2012.
[14] Ying-Ho Liu, Anthony J.T. Lee, Fu Chang, “Object
recognition using discriminative parts,” Computer
Vision and Image Understanding, Vol. 116, no. 7, pp.
854–867, 2012.

IRJET- Fake Paper Currency Recognition

More Related Content

What's hot (19)

Similar to IRJET- Fake Paper Currency Recognition (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Fake Paper Currency Recognition