See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/371125630

A Project Report on Fake Currency Detection

Technical Report · May 2023

DOI: 10.13140/RG.2.2.21616.43526

CITATIONS READS
0 2,445

3 authors:

Alimul Rajee Rasel Ahmed

Comilla University Comilla University
35 PUBLICATIONS 8 CITATIONS 1 PUBLICATION 0 CITATIONS

SEE PROFILE SEE PROFILE

Shekh Humaira Sunzida

Comilla University
2 PUBLICATIONS 0 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Arduino Voice Controlled Bluetooth Car View project

GSM based Smoke & Gas Leakage Detector_lab report View project

All content following this page was uploaded by Rasel Ahmed on 29 May 2023.

The user has requested enhancement of the downloaded file.

 A Project Report

on

Fake Currency Detection

Department of Information and Communication Technology, Comilla University

Supervisor:
Alimul Rajee
Lecturer,
Department of ICT
Comilla University

Team Members:
Israt Jahan Nahin (11909004)
Shekh Humaira Sunzida (11909024)
Md. Billal Hossain (11909026)
Md. Rasel Ahmed (11909033)
Shuvra Banik (11909046)

Date of submission: March 2023

1
 Abstract

The creation and circulation of counterfeit notes are on the rise right now, as a result of advances
in color printing technology. This is a serious issue that affects practically all of the nations. The
economy is impacted. Such fake money fuels evil intentions, typically involving terrorist actions.
According to the research, this has had a highly negative effect on developing nations like India.
This research suggests a method for viewing the fake currency through its image. Pre-processing
should be used after choosing an image. The acquisition of images is completed at first during
processing. Then a conversion from RGB to GRAYSCALE is done. After conversion, the image
is segmented, its features are measured, correlation is found, and classification is completed to
determine whether the image is real or fake. Fake money has been a significant issue. Banks and
other trading places have equipment available to verify financial validity. Nevertheless, the
normal individual does not have access to such tools, which is why fake money detecting
software that is usable by regular people is required. We get at most 81% of accuracy
considering 50 items of Indian Currency notes of 2000-rupee. This project provides a thorough
explanation of a fake note detector that can be used by the average person. The suggested system
employs image processing to identify genuine currency from counterfeit money. The Python
programming language has been used to create the software in its entirety.

Keywords: Fake Currency, Image Processing, Grayscale Conversion, Segmentation, pre-

processing.

2
 Table of Contents

Title Page No
List of Figure 4
List of Table 5
Chapter One: Introduction 6
1.1 Introduction 6
1.2 Objectives 7
1.3 Project Outline 8
Chapter Two: Literature Review 9
2.1 Literature Review 9
2.2 Literature Summary 18
Chapter 3: Methodology 21
3.1 Data Description 21
3.2 Requirement Analysis 21
3.3 Features of Currency 23
3.4 Required Algorithm 27
3.5 Flowchart 30
3.6 Workflow of Proposed System 31
3.7 Implementation Steps of Proposed System 31
Chapter 4: Result and Discussion 33
4.1 Result and Discussion 33
Chapter 5: Conclusion 35
5.1 Advantages 35
5.2 Restrictions 35
5.3 Future Scope 35
Reference 36

3
 List of Figure

Figure No Figure Title Page No

Fig 3.1 All security features of Indian currency 2000 23

Fig 3.2 Portrait of Mahatma Gandhi 24
Fig 3.3 Security Thread 24
Fig 3.4 See through Register 24
Fig 3.5 Ashoka Pillar 25
Fig 3.6 Identification Mark 25
Fig 3.7 Guarantee Clause 25
Fig 3.8 Currency Numeral with the Rupees Symbol 25

Fig 3.9 Bleed Lines 26

Fig 3.10 Latent Image of Denomination Numeral 26
Fig 3.11 Micro Lettering 26
Fig 3.12 Government of India 27
Fig 3.13 Block diagram fake currency detection 30

4
 List of Table

Table No Table Title Page No

Table 3.1 Dataset 21

Table 3.2 Hardware Requirements 22
Table 3.3 Software Requirements 22

5
 Chapter One

Introduction

1.1 Introduction

Money can serve as the driving force behind any economic activity associated with
manufacturing, circulation, consumption, etc. Capital information can be used to save money and
make investments. Money is essential for everything in today's dynamic culture. There are also
other factors that are shrinking the economy as it advances. One of those things is the creation
and use of counterfeit currency. Due to the widespread use of counterfeit currency in the
economy, the typical person is the group most negatively impacted by this activity. Everyone is
afraid of accepting banknotes in the denominations of Rs. 500 and Rs. 1,000 because the bulk of
them are nearly hard to distinguish from genuine banknotes, from gas stations to the
neighborhood vegetable seller.

The issue of counterfeit money is one that is discussed and debated throughout the world. Banks
lost Rs. 16,789 crores in the most recent fiscal year due to frauds. The Reserve Bank reported
that "the amount that has been lost on account of frauds in the year 2016–17 was Rs. 16,789
crores," which was in accordance with the fraud monitoring report made by various banks and
financial institutions. According to the RBI's (Reserve Bank of India) annual report for 2021–22,
there was an increase in the number of counterfeit notes found in the denominations of Rs. 10,
Rs. 20, Rs. 200, Rs. 500 (new design), and Rs. 2,000, respectively, of 16.4%, 16.5%, 11.7%,
101.9%, and 54.6%.[51]

Inflation is the typical impact of counterfeiting on the economy. The only tool now available to
the average person to identify fake money is the Fake Note Detection Machine. The majority of
the time, this machine is only found in banks, which are not always accessible to the regular
person. In order to prove the viability of suggested solutions to a particular problem, a lot of
experimental work is required in the field of digital image processing.

6
It includes operations whose inputs and outputs are images and operations that extract properties
from photos, including the identification of specific objects. The watermark on fake currency is
created using opaque ink, white solution, and stamping with a dye that has a picture of Mahatma
Gandhi engraved on it. Visitors are the most susceptible to phony currency because they lack the
knowledge necessary to distinguish between fake and genuine currency notes.

These people will benefit from automatic currency identification using image processing
techniques. Also, it can be helpful in other workplaces. The devised system to verify the 2000-
rupee Indian currency notes. It will organize the predetermined arrangement of information and
pre-process the digital images before differentiating in monetary forms. The approach for
detecting Indian currencies suggested in this article is practical and affordable. The user can
determine whether the cash note is authentic or phony at the conclusion of the process.

1.2 Objectives

1. To examine the various security components of Indian currency notes.

2. Using a scanner or camera to gather paper money.

3. To extract characteristics from the captured image by cropping and segmenting it.

4. Creating a feature localization algorithm.

5. Designing an extraction and recognition of features.

6. To determine the right money denomination.

7. To distinguish between authentic and fake money notes

7
1.3 Project Outline

This project paper is based on Fake Currency Detection. At the first chapter it includes the
introduction and objective part which mainly describe about our project topic, why this project is
important and which methods are used in our project to make it successful. The second chapter
includes literature review which mainly represents the summary of the different project paper
related to our project topic, including of what they had achieved and by which process they had
done their project etc. The third chapter includes the methodology part which mainly represents
the information about the used dataset, software and hardware tools and tables the description of
the process with the flow-chart diagram, features of the currency by following we did our
project. The fourth chapter includes the conclusion part that includes result and discussion part,
which represents what we have got from our project with the advantages of our project and also
includes limitations of our project and the scope at which we can improve our project in the
future.

8
 Chapter Two

Literature Review

2.1 Literature Review

Colaco, Rencita Maria et al. chose to use the programming language Python and OpenCV for
their project using Canny Edge Detection algorithm. To make comparisons and determine the
outcome, a number of characteristics that define genuine currency apart from counterfeit ones
were taken into account. Identification marks, see-through registers, optical variable ink,
currency color codes, security threads, watermarks, latent images, and micro-lettering were a few
of these features. The accuracy of the proposed system was close to 80%. The goal of this
research was to create a low-cost system with quick computations so that even the average
person, who is unable to use more advanced resources, can identify counterfeit money. [1]

In this project, Vadnere, Koneri et al. used digital image processing algorithms to make currency
authentication. In essence, some features were extracted using image-based segmentation and
template matching, which worked well throughout the entire process and required little
computation time. The technique was incredibly easy & simple to use. Using this method in the
real world is highly adaptable. The folks who are unfamiliar with currencies will be benefitted
from this effort. [2]

In this experiment, Jamkhandikar, Dayanand et al.used two different currencies, and it was
discovered that the suggested method based on color and feature analysis is effective for
currencies. This project was done using image processing algorithm. The accuracy found in this
system was 70%. Doing this project allowed them to identify counterfeit money, which is
particularly helpful in preventing high-order counterfeiting that makes use of low-cost but high-
quality machinery. [3]

Sangogi, Mrs Jyoti et al. made it possible for someone who is blind to tell whether money is real
or phony. Based on the parameters of the HSV values of the currency note, the Python technique,
which was implemented in a Raspberry Pi with a scanner, could capture the currency note and
carried out the image processing techniques mandated in the project to determine whether the
currency is genuine or counterfeit. [4]

9
Shiby, Ashik et al. used image processing, this project aided in the acquisition of fake currency.
In certain ways, this can stop the propagation of fraudulent notes in the system. It can provide the
user with the opportunity to ascertain the note's legitimacy without having to visit a bank using
machine learning algorithms.The accuracy of this proposed system is approximately 80%. The
initiative covered the acquisition of Indian paper money. [5]

Warke, Kanthi et al. compared their project with other previous methodologies, image
processing techniques allow for more accurate examination of the currency image while also
saving time and money. The suggested method was designed to efficiently extract and check
features from photos of Indian cash using so many algorithms like fluorescence detection
algorithm, dimension detection algorithm, color composition algorithm etc. [6]

Aditya Sharma, Shweta Poojary et al. worked using the help of the Tensorflow and Keras
libraries, this system primarily focused on the picture categorization part of deep learning using
deep learning network algorithms, deep neural network algorithms, and convolutional neural
network algorithms. The detection of currency tooks only a few seconds, and currency
recognition was simple. The system has an excellent overall accuracy level for differentiating
between real and fraudulent cash. [7]

Kudalkar et al. proposed a technique which was appeared to be effective in determining whether
the currency is real or fake using image acquisition, noise removal, gray scale conversion, edge
detection, segmentation, feature extraction, comparison, supervised learning algorithm.
According to the project's findings, the primary security features—bleed lines, security thread,
and micro lettering—were needed to be precisely computed. [8]

Pallavi, S. et al. detect and identify banknotes of different denominations. This study presented a
system that use convolutional neural networks with deep training using CNN (Convolutional
Neural Network) model training, currency pre-processing algorithm. By acting as a feature
extractor, their ML-CNN (Machine Learning Convolutional Neural Network) based project
eliminated the need for image processing and the need for manually confirming the existence of
security characteristics in the note. The project's audio output, which will also be useful to blind

10
people, will be its final product. Further experiments with different ML-CNN (Machine Learning
Convolutional Neural Network) architectures will increase the model's accuracy. [9]

Potluri, Jahnavi et al. suggested an approach which offered a more effective way to spot a phony
banknote using data augmentation, transfer learning. This CNN (Convolutional Neural Network)
model (Mobilenetv2-FCD) was developed for a mobile application after being trained on a real-
time dataset. With smart phones, this mobile application can be used to detect counterfeit money
notes. This suggested network had an accuracy of 85% in identifying counterfeit notes. Any
currency can be used with the network after it has been trained for it. [10]

Khan, Mudassir and Mahtab Alam showed how machine learning can be used to accurately
identify fake currency. It is possible to distinguish between real and fake banknotes quickly by
using the K-Means algorithm with many clusters of size two. A third group of banknotes that
could be subject to additional analysis, such as genuine banknotes in poor condition that could be
mistaken for forgeries, can aid when there are three clusters instead of two. The model had some
flaws. In this point, an accuracy of 87% might be sufficient. Yet, it is still more precise than
human detection. Additionally, the model's accuracy can be increased any further with more data
and better analysis. [11]

Vidhate, Shah et al. examined all relevant existing architectures, and by studying how they
functioned, they found certain problems with the system that was in place at the time. With some
of the extra features in their suggested system being implemented utilizing KNN (K-Nearest
Neighbors Algorithm) algorithms or sophisticated machine learning techniques, they had
preserved all the key aspects of the existing systems as their primary focus. [12]

Kumar, Akhila et al. used three supervised machine learning (SML) methods were used in this
project to authenticate banknotes. The three methods that were used are the Decision Tree (DT),
KNN (K-Nearest Neighbors Algorithm), and SVM (Support Vector Machine) algorithms. [13]

NR, Deepak proposed an efficient method for extracting and identifying the characteristics of
Indian rupee notes was given in this study. Using digital image processing methods, the study
also included the recognition and classification of counterfeit money. [14]

11
Siddiki, Marjuk Ahmed et al. had created a system for spotting counterfeit money using a
convolutional neural network (CNN) model and certain deep learning techniques with image
processing algorithms. These were really accurate and very useful for spotting counterfeit
money. [15]

Naseem, Syeda Aimen et al. proposed an innovative method for distinguishing real banknotes
from fake ones by applying an image processing methodology that was implemented in
MATLAB. A image watermark, security thread, and serial number extraction are among the
features of the banknote. The algorithm, which is based on image processing, was used in the
proposed technique, which made use of regular RGB images to authenticate banknotes. [16]

Saxena, Ankur et al. presented a document which can give a person the ability to spot a phony
note using digital image processing algorithms, which can play a vital role in reducing
corruption. In order to detect false notes, MATLAB techniques were implemented in mobile
phones with scanners or cameras, giving average person the ability to tell if a piece of currency is
real or not. [17]

Kanojia, Mahendra, Niketa Gandhi and Amruta Rane focused on two different currencies and
used color and feature analysis to evaluate whether the given currency was real or false utilizing
image processing algorithms such image acquisition, gray scale conversion, edge detection,
image segmentation, and feature extraction etc. [18]

Shah, Raj, Mayank Champneri and Pujan Sheth employed two methods to detect fraudulent
currency notes: UV LEDs (Ultraviolet light-emitting diodes) and image analysis. To determine if
a note was false or real, a UV LED (Ultraviolet light-emitting diodes) effect (a fake note
detection unit) was positioned in front of the camera. In any country, automatic methods for
counterfeit money note detection are essential. [19]

Gaikwad, Mayadevi et al. conduct an experiment where pattern matching was used to separate
authentic notes from false notes. It was explained that various denominations have distinct
templates; if the templates match the value, the note is authentic otherwise it is counterfeit. The
input photos were preprocessed, which involved eliminating noise from the photographs and

12
performing a feature extraction technique to determine the denomination. The original note and
the phony note were then compared using pattern matching. [20]

Agasti, Tushar et al. researched a method for identifying counterfeit money began with the image
acquisition and continued with the computation of each extracted feature's intensity. Even if the
memo has scribbles on it, the algorithm can still extract features from it. This research suggested
a method for determining whether a certain frequency is real or not. [21]

Tele, Gouri Sanjay used image processing techniques were used to identify original cash using
digital image processing algorithms. Based on physical characteristics, this system provided an
effective way to detect phony cash. With this, we may be able to identify the parameters that the
fake note lacks in comparison to the authentic notes, which will help us to determine the given
note is real or not. [22]

Amirsab, Shaikh Ajig et al. proposed an example of a low-cost system that uses effective and
efficient digital image processing techniques and algorithms to detect counterfeit money. The
image processing techniques that were used in this project are- RGB to Grayscale conversion,
edge detection, sub divides the image into its constituent regions or objects, characteristic
extraction, characteristics comparison etc. [23]

Deepak, M. P. and N. B. Prajwala performed the classification of original notes and fake notes
by pattern matching. Different denominations have various templates, if the templates match the
value, then it is an original note, else it is a fake note. The images that were used as input were
preprocessed by removing of noise from the images and feature extraction technique was
performed to identify the denomination. Noise removal process was done either by linear
filtering or adaptive filtering. At the time of preprocessing, the denominations were identified.
Then, pattern matching was done to compare between the original and the fake note. [24]

This article discussed the need for accurate and efficient methods to detect counterfeit currency,
as the number of fake currency notes detected has increased in recent years. The article
highlights the limitations of traditional techniques and proposes the use of image processing
algorithms for reliable detection of counterfeit currency. Various image processing techniques
and machine learning algorithms such as artificial neural networks and support vector machines

13
have been used in the literature for this purpose. The literature suggests that image processing
techniques can provide accurate and reliable detection of counterfeit currency, but further
research is needed to improve their performance. [25]

Sawant, K. et al. used the K-means algorithm for feature clustering and SVM (support vector
machine) algorithm for data model training. Gaikwad et al. suggested a methodology based on
comparing visual features of real and fake notes. Nagpure et al. proposed a system using floral
designs on notes to distinguish between real and fake notes. Rathee et al. suggested using
supervised machine learning along with image processing to increase precision. Upadhyaya et al.
proved that logistic regression with image processing achieves accuracy above 99%. These
studies demonstrate the effectiveness of combining image processing and machine learning
techniques for detecting fake currency notes.[26]

Roy, V., Mishra et al. discussed the problem of detecting fake currency notes using image
processing techniques. The authors highlight the need for such a system in the current scenario
where counterfeiting of currency notes has become a widespread issue. They provide a brief
overview of the existing techniques for counterfeit detection and highlight their limitations,
including the need for specialized equipment and expertise. The authors then present their
proposed system, which uses image processing techniques to identify fake currency notes based
on certain features such as watermark, serial number, and texture. They also discuss the
implementation of the system and its performance evaluation, including the accuracy of the
system and its robustness against various types of attacks. Finally, the authors conclude that their
system can provide a cost-effective and efficient solution for detecting fake currency notes.[27]

Chowdhury et al. proposed an automatic model for recognizing banknotes, independent of the
orientation of banknotes, using image processing and deep learning methods. The studies utilized
various techniques such as data augmentation, RGB values as features, and ReLU nonlinearity
operations to improve the accuracy of detecting fake notes. [28]

This paper mentions five other studies related to detecting fake currency using image processing.
The first study by Ms. Monali Patil, Prof. Jayant Adhikari, and Prof. Rajesh Babu used the K-
means algorithm for feature clustering. The second study by Mayadevi A.Gaikwad, Vaijinath V.
Bhosle, and Vaibhav D Patil suggested comparing the visual features of real and fake currency.
The third study by Renuka Nagpure, Shreya Sheety, and Trupti Ghotkar proposed using floral

14
designs on currency notes to distinguish between real and fake notes. The fourth study by Neeru
Rathee, Arun Kadian, Rajat Sachdeva, Vijul Dalel, and Yatin Jaie suggested using supervised
machine learning with image processing to improve detection accuracy. The fifth study by
Akanksha Upadhyaya, Vinod Shokeen, and Garima Srivastava found that image processing with
logistic regression can achieve an accuracy of over 99%. [29]

Miss.I. Santhiya Irulappasamy included five papers on fake currency detection using image
processing techniques. Vipin Kumar Jain proposed a system for recognition of fake currency
using an Android mobile phone and hand-written character feature extraction. Amol A. Shirsat
and S.D. Bharkat proposed a currency recognition system consisting of image processing, feature
extraction, and classification using neural networks. S. S. Veling, Miss. Janhavi P. Sawal, and
Miss. Siddhi proposed a hyperspectral imaging system for fake Indian currency recognition.
Prof. Madhav Thigale and Lina Ladhane proposed a note-to-coin converter using image
processing and UV LED (Ultraviolet light-emitting diodes) and photodiode for fake note
detection. Deborah. Soniya Prathap proposed a system for fake currency detection using image
segmentation and feature extraction.[30]

Atchaya, Harani et al. focused on different techniques used for fraud detection, particularly in the
context of currency and financial transactions. The methods discussed include neural networks,
model-based reasoning, watermarking, optically variable ink, fluorescence, and image
processing techniques such as edge detection, feature extraction, and segmentation. The paper
highlights the importance of detecting fake currency as it can cause significant problems, and
various approaches are necessary to combat the issue effectively. The use of security features
specific to Indian currency is also discussed in this paper. Overall, the review presents a
comprehensive overview of different techniques and methods used for fraud detection in
financial transactions.[31]

Sruthi R. used various methods for detecting fake currency notes with machine learning and
image processing techniques. A transfer learned convolutional neural network has been used in
[3], achieving 81.5% accuracy for identifying real notes and 75% for counterfeit notes. An
ensemble of classifiers has been used in [5], with SVM (support vector machine) achieving the
highest accuracy of 82.7%. Edge detection techniques have been used in [8], achieving an
accuracy of 90.45%. Different preconfigured CNNs (convolutional neural network) have been

15
used for Indian currency recognition in [6]. KNN (K-Nearest Neighbors Algorithm), SVC
(support vector machine), and GBC (Genetic Bee Colony) have been used in [7], with KNN (K-
Nearest Neighbors Algorithm) and GBC (Genetic Bee Colony) providing higher accuracy in the
recognition task. OVD (Online Video Distributor) patches have been used in [9] to detect
counterfeit Philippine notes. [32]

Arun Anoop M et al. worked based on to remove Noise Patterns from Background a special
linear transformation function is adapted. This kinds of algorithm we typically use to solve the
problem in sequence and In this project, they actually work on the currencies. By making this
project they recognize fake currency boldly. [33]

Aakash S Patel et al. basically worked on the the Machine Learning Algorithms recognized for
Image processing and Detection. Training the machine using an already prepared dataset of
currency notes, which will contain sample images of fake and real currency notes. Analyzing the
content of the dataset, using the applied algorithm to extract required features which will help in
recognizing other input images of similar format. [34]

Vidhi Roy, Mannadiar et al. used image processing techniques to identify original cash using
digital image processing algorithms. Image Acquisition, gray scale conversion, edge detection,
image segmentation and characteristics extraction are highly required to perform this project.
[35]

Tushar Agasti, Pratik Wade and P Chitra applied here a simple algorithm which works properly.
These kinds of requirements are highly required to complete this project. Acquisition of image of
currency note under ultraviolet light by simple digital camera or scanner. Image acquired is RGB
image and now is converted to grayscale image. Edge detection of whole gray scale image. Now
characteristics features of the paper currency will be cropped and segmented. Intensity of each
feature is calculated. [36]

S.Mittal et al. basically used some algorithms such as CNN to complete the project. These kinds
of algorithms are typically SIFT (Scale-Invariant Feature Transform), SUFT, ORB (Oriented
FAST and rotated BRIEF) etc. The accuracy of the system is almost 85%. SIFT (Scale-Invariant
Feature Transform) stands for Scale Invariant Feature Transform and it is one of the most for
most image processing algorithms. Here points are detected and matched using the Euclidian

16
Distance. Speeded up Robust Features or the SURF (speeded up robust features) in an algorithm
present in computer vision that has numerous applications such as detection, recognition,
tracking, identification and extraction. The accuracy of the proposed system is close to 95%. As
the name suggests it is a fusion of the two other algorithms here. These are the FIRST algorithm
and BRIEF (Binary Robust Independent Elementary Features) algorithm. [37]

Aman Tekriwal, Prashant Mankani applied the SVM (Support Vector Machine) algorithm which
is generally used to perform the tasks. SVMs (Support Vector Machine) are supervised learning
models with associated learning algorithms that analysis data used for classification and
regression analysis. The approximate accuracy is more that 90%. Given a set of training
examples, each marked as belonging to one or the other of two categories, an SVM (Support
Vector Machine) training algorithm builds a model that assigns new examples to one category or
other. [38]

Debang S.Gujarathi, Sushii V. Nikam, Prathmesh P.gogte descrbed the technique of the
authentication of currency by applying image processing. The complete methodology works for
100,500 and 1000 currency notes. This method is very simple and easy to implement. This
project will be helpful to those people who don’t have any knowledge about currency.[39]

Chinmay Bhurke, Meghana Sirdeshmukh included two modules: One is to identify the
type/country of the currency. The second is to determine the authenticity of the currency. The
fake currency module also determines whether a currency is fake or real by comparing the
metallic black strip present on the Indian notes. Both the modules work accurately and
efficiently. [40]

17
 2.2 Literature Summary

Dataset Algorithm Result/discussion

Indian currency note Grey Scaling algorithm, By making this project they recognized fake
of 500 and 2000. Edge Detection currencies by taking advantage of cheaper and high
algorithm quality equipment.[3]
Image type dataset KNN algorithm, machine They analyzed all related existing architectures and
Source: Kaggle learning algorithms discovered some flaws in the currently existing
Sample items: 995 system to recognize fake currencies.[12]
Image type dataset of Image Scaling, This project enables a visually impaired person to
currencies binarization algorithm, detect whether the currency is original or fake.[4]
KNN classifier
Image type dataset of Machine learning This project helps to obtain false money through
currencies algorithms. image processing.[5]
Image type dataset of Canny edge detection They ended up using OpenCV along with python
currencies algorithm. programming language to detect fake currencies.[1]

Image type dataset of Digital image processing In this project, the authentication of currency is made
currencies algorithms are used. by applying image processing.[2]
Image type dataset The K-means clustering Using K-Means algorithm with several clusters equal
Source: OpenML algorithm to two can help to identify genuine or
sample items: 1,372. forged banknotes efficiently. Accuracy: 87%
Image type dataset, Data Augmentation, This CNN model (Mobilenetv2-FCD) has been
having 3 columns: Transfer Learning trained on a real time dataset and then deployed into
type, filter shape and algorithm a mobile application to determine fake currencies.
input size. Accuracy: 85%.[11]
Sample items: 500
Image type dataset. CNN model training, a system that uses convolutional neural networks
Source: Kaggle currency pre-processing using deep training to detect fake currencies.[10]
algorithm.

18
Image type dataset. Fluorescence, dimension using image processing techniques they detected the
Detection and color fake currencies.[9]
composition algorithm.
Image type dataset. SML, Decision tree, In this project, SML algorithm SVM, DT, and KNN
Sample items: 1372 KNN SVM algorithm. are applied to authenticate banknotes.[6]
Images of Indian Deep learning and neural The detection of the currency is done within seconds
currency of 10, 20, network algorithms, and the recognition is done easily.[13]
50 and 100. CNN algorithm.
Sample items: 327
Image type dataset Digital image processing The proposed system seems to be useful to detect
algorithm whether the currency is genuine or not.[7]
Image type dataset Digital image processing The study includes effective method for recognizing
algorithm the properties to detect fake notes.[8]
Image type dataset of CNN model, deep With the use of deep learning and image processing,
Bangladeshi learning technique fake currencies are detected.[14]
currencies
Image type dataset of Digital image processing suggested an approach for identifying real notes
Pakistani currencies algorithm. using the image processing technique.[15]
Srilanka Rupee SIFT 95% is achieved with note displays of 100 and 500
Notes of 100, 500, note.[16]
1000 and 2000
Total 10 different SUFT The accuracy of the system proposed is close to
data is used. 70%. Further implemented for the foreign
currencies.[17]
MATLAB, SVM Approximately accuracy is more than 90%.[18]
Ultraviolet Ray and
different notes.
UV with digital Apply here a simple The result isn’t contentment. More features are
camera or scanner, algorithm which works highly needed.[21]
RGB images properly.

19
11 different data is ORB The accuracy of the system is more than 85%.[22]
used
1000 rupees note SVM Features hence it is considered as fake note.[25]
100, 500 rupees a simple algorithm was The system compares more features for feature
notes. applied extraction than other proposed systems.[28]
Approximately 10 Canny Edge detection The accuracy of the system proposed above is close
data are used. Algorithm to 80%.[34]

20
 Chapter Three
Methodology
3.1 Data Description:

The banknote-authentication dataset is used to distinguish between genuine and counterfeit

banknotes. Images of real and fake banknote-like specimens were used to extract data from the
photos. These photos were processed and number of lines on a thin strip are measured. A
compressed version of the dataset from Kaggle was used in this experiment. There are 100
samples total. The model has been trained using 2000-rupee notes of cash from India.

To determine the dataset's input/output behavior for the system, an experiment was run. The
sample dataset utilized in the experiment is named and provided below:

Dataset Source Items Type

Indian Currency Kaggle 50 Image dataset

notes of 2000-rupee

Table 3.1: Dataset[Accessed 05 February 2023]

3.2 Requirement Analysis:

The implementation requirement details are given in this section.

Requirement Analysis method is intended in such a way that it takes fewer resources to figure
out work correctly. The minimum needs that we’d like to take care of: The system would require
a minimum of 4 GB (Gigabyte) of RAM (Random Access memory) to run all the options sleek
and unforeseen. It wants a minimum of 2 GHz (Gigahertz) processor to run the system smoothly.
The system can be operated by common people as well as commercial people.

21
Hardware Specification:

Processor 2 GHz Intel

Storage 512GB

Ram 4GB

Table 3.2: Hardware Requirements

Software Specification:

Operating System Windows 7,8,10

Programming Language Python

IDE(Integrated Development Environment) VS Code

Table 3.3: Software Requirements

Python:

Python is an interpreter, object-oriented, high-level, dynamically semantic programming

language. It is particularly desirable for Rapid Application Development as well as for usage as a
scripting or glue language to tie existing components together due to its high-level built-in data
structures, dynamic typing, and dynamic binding. Python's straightforward syntax prioritizes
readability and makes it simple to learn, which lowers the cost of program maintenance. Python's
support for modules and packages promotes the modularity and reuse of code in programs. On
all popular platforms, the Python interpreter and the comprehensive standard library are freely
distributable and available in source or binary form. [41]

Python Libraries:

OpenCV:

OpenCV is a sizable open-source library for image processing, machine learning, and computer
vision. It now plays a significant part in real-time operation, which is crucial in modern systems.
With it, one may analyze pictures and movies to find faces, objects, and even human
handwriting. To install OpenCV run the command - pip install opencv-python. Python is able to
handle the OpenCV array structure for analysis when it is integrated with different libraries, such

22
as NumPy. We use vector space and apply mathematical operations to these features to identify
visual patterns and their various features. [42]

NumPy:

Many mathematical operations can be carried out on arrays with NumPy. It provides a vast
library of high-level mathematical functions that work on these arrays and matrices, as well as
strong data structures that ensure efficient calculations with arrays and matrices. To install
NumPy run the command - pip install numpy. [43]

VS Code:

Debugging, task execution, and version control are supported by the simplified code editor
Visual Studio Code. It tries to give developers only the tools they require for a short cycle of
code-build-debugging and leaves more sophisticated processes to IDEs with more features, like
Visual Studio IDE. [44]

3.3 Features of Currency

All features of Indian currency 2000 showing in fig

Fig 3.1: All security features of Indian currency 2000[3]

23
Portrait of Mahatma Gandhi at the Center:

The intaglio printing of portrait of Mahatma Gandhi at the center of the currency.

Fig 3.2: Portrait of Mahatma Gandhi[1]

Security Thread:

When held up to the light, the security thread, which has "RBI" and "Bharat" inscribed on it
continually, can be seen at the left side of the watermark. The photo of the Mahatma has a
security thread on one side.

Fig 3.3: Security Thread[1]

See through Register:

The denomination numeral is displayed in the see-through register. Both sides of this register are
printed. One side of the two sides is hollow, and the other side is filled with material. The micro
lettering has been written horizontally along this register. The note has a latent image on the left
side. Moreover, this register is shown above the latent image. When viewed in contrast to the
light, this register appears as a single design.

Fig 3.4: See through Register[1]

24
Ashoka Pillar:

On the right side of the coin there is a picture of the Ashoka pillar.

Fig 3.5: Ashoka Pillar[1]

Identification Mark:

Just over the Ashoka's pillar symbol, there is an identification mark.

Fig 3.6: Identification Mark[1]

Guarantee Clause:

Located to the right of Mahatma Gandhi's image, the guarantee clause is signed by the governor
and includes a promise clause that is printed in intaglio.

Fig 3.7: Guarantee Clause[1]

Currency Numeral with the Rupees Symbol:

Fluorescent ink will be used for printing. When viewed from different perspectives, the
numerals change.

Fig 3.8: Currency Numeral with the Rupees Symbol[1]

25
Bleed Lines:

The oblique lines that protrude from the sides of banknotes are known as bleed lines.

Fig 3.9: Bleed Lines[1]

Latent Image of Denomination Numeral:

The right side of Mahatma Gandhi's portrait is bordered by a vertical band on the opposite side
of the denomination. A latent image of the corresponding denominational value is present in it.
Its denominational value is represented by a numerical value. The latent picture can be seen
when the coin is held horizontally, and it should also be held at eye level. While using
counterfeit money, it is not noticeable.

Fig 3.10: Latent Image of Denomination Numeral[1]

Micro Lettering:

Between the vertical band and the image of Mahatma Gandhi, micro lettering is visible. The term
"RBI" and the denominational value are written in tiny letters. The micro letters on counterfeit
money are incorrectly printed.

Fig 3.11. Micro Lettering[1]

26
Government of India:

The words "Government of India" are printed at the top of the one rupee note, directly over the
Devanagari-scripted number one. The smallest currency note now in use in India is 1 rupee, and
it is the only one that was produced by the Government of India rather than the Reserve Bank of
India like the others. Because of this, it is the only one with the Finance Secretary's signature
rather than the RBI Governor's.

Fig 3.12. Government of India[1]

3.4 Required Algorithm:

Image acquisition:

The act of obtaining an image from sources is known as image acquisition. Hardware systems
like cameras, encoders, sensors, etc. can be used to do this. It is without a doubt the most
important phase in the MV (Machine Version) workflow because a bad image would make the
workflow ineffective as a whole. As machine vision systems don't study the acquired digital
image of the object and not the object itself, acquiring an image with the proper clarity and
contrast is crucial. A set of photo-sensitive sensors turn an object's incoming light wave into an
electrical signal during the image acquisition step. These little components provide the function
of accurately describing the object to your machine vision algorithms. It's a frequent fallacy that
with an MV system, choosing the correct colors is crucial. However it's not always the case.
Colors frequently increase noise and make detection more challenging. The main objective of an
image acquisition system is to increase contrast for the important features. The ideal image is one
in which the camera can clearly see the object of interest. [45]

27
The major image acquisition components have been mentioned below:

1. Trigger
2. Camera
3. Optics
4. Illumination

RGB to Grayscale:

Taking the average of the red, green, and blue pixel values for each pixel to obtain the grayscale
value is a straightforward technique to convert a color image's 3D array to a grayscale image's
2D array. This creates an approximate gray color by combining the lightness or brightness
contributions from each color band. [46]

The Average method takes the average value of R, G, and B as the grayscale value. [47]

Grayscale = (R + G + B ) / 3

The weighted method, also called the luminosity method, weighs red, green, and blue according
to their wavelengths. The improved formula is as follows:

Grayscale = 0.299R + 0.587G + 0.114B

Image Segmentation:

Image segmentation is a technique for breaking up a digital image into smaller groupings called
image segments, which reduces the complexity of the image and makes each segment more
easily processed or analyzed. Technically, segmentation is the process of giving labels to pixels
in an image in order to distinguish between objects, persons, or other significant aspects. Object
detection is a frequent use of image segmentation. It is usual practice to first apply an image
segmentation method to discover things of interest in the image before processing the complete
image. The object detector can then work with a bounding box that the segmentation algorithm
has previously established. By stopping the detector from processing the entire image, accuracy
is increased and inference time is decreased. A crucial component of computer vision
technologies and algorithms is image segmentation. It is employed in a variety of real-world

28
contexts, including as face identification and recognition in video surveillance, medical image
analysis, computer vision for autonomous cars, and satellite image analysis.[48]

Feature Measurement:

The process of "feature detection" involves computing abstractions of image data and locally
determining whether or not each image point contains an image feature of a specific type. A
fundamental aspect of image processing is feature detection. This means that it is typically done
as the initial operation on an image and checks each pixel to see if a feature is present there. If
this is a component of a bigger algorithm, the algorithm will usually just look at the image where
the features are. The term "feature description" refers to a technique for describing the local
attributes of an image at identified key points in an image. These algorithms take advantage of
key points discovered in the image data to extract interesting information. The information
produced by these feature description techniques is frequently organized by encoding it as the
constituent parts of a single vector, or feature vector. A feature space is the collection of all
feasible feature vectors. [49]

Finding Correlation:

For finding Correlation of two images we have to follow this steps:

1. Load two images and extract their pixel-by-pixel information

2. Normalize and down sample the pixel information
3. Calculate cross-correlation using the processed pixel information
4. Generate visual summaries of cross-correlation, highlighting areas of maximum image
overlap. [50]

29
3.5 Flowchart:

Image Acquisition

RGB to Grayscale

Image Segmentation

Feature Measurement

Finding Correlation

Classification

Fig 3.13: Block Diagram of Fake Currency Detection

30
3.6 Workflow of Proposed System:

1. Image Acquisition: The model receives the image. Images should be present: the note
you're trying to identify and its real-world counterpart. Picture capture of an Indian
banknote with a basic digital camera or scanner.

2. RGB to GRAYSCALE: Acquired picture is A GRAYSCALE image is created by

converting an RGB image. The RGB image is dense and noisier. Instead of processing
three components R (Red), G (Green), and B, the image is reduced in size and gains easy-
to-process intensity information when it is converted to gray scale (Blue).

3. Segmentation: Gandhi Ji's image and a narrow strip image are cropped from the original
image. The observe and reverse aspects of the Indian paper currency will be clipped and
split. The technique of segmenting an image into separate segments and sets of pixels is
done digitally. It is also known as "picture thresholding," where a threshold is set and, if a
specific pixel's value exceeds it, the pixel turns white; otherwise, it turns black.
4. Feature Measurement: Feature measurement is done to measure the number of lines on a
thin strip. This is a really lengthy process.

5. Finding Correlation: We find correlation between Gandhi Ji’s image on the real note &
fake note using distance-weighted algorithm. If the outcome is greater than 0.5 then we
will consider it legitimate otherwise the currency is fake.

6. Classification: Finally, we will classify the image as real or fake

3.7 Implementation Steps of Proposed System:

1. At first, OpenCV was used to read the real and fake photos after importing the Modules.
2. The colorful image was then changed to black and white because a black and white
image makes it simpler to identify key aspects.
3. Then Gandhi Ji's picture was being taken from the original note. The coordinates must be
changed each time Gandhi Ji's image is extracted. The same thing also be done for fake
notes.

31
4. Then the thin strip was extracted from the real and fake notes.
5. Then the image was being converted to HSV (Hue Saturation Value). Simple terms, the
format is different. A numerical evaluation of the color in the photograph is called HSV.
The unit of measurement is degrees. Hue in RGB is the direction that the colors are
facing.
6. Then, the thin strip was extracted from the HSV images and fake images.
7. Then, the thresh value was specified in this case, and only values more than the specified
values were extracted from the cropped image. The '&' of both values will then be taken.
The goal of the entire procedure was to extract the thin strip's lines from the thin strip
image.
8. Bwareaopen is a function that will find the connectivity in a picture. It will be used to
determine how many lines there are in a short strip. It will be employed later on in the
procedure.
9. Then, the code for morphologically extracting the thin strip image can be seen. This is a
crucial step in determining the amount of lines in an image.
10. Then Bwareaopen function for the thin strip was used.
11. In next step, final counting of the number of lines in real and fake note was made.
12. A correlation function was being defined. This correlation function to our Gandhi Ji
image was applied. Simply said, we were looking to compare Gandhi Ji's image on the
real and false photos. This will accept two two-dimensional matrices as input and return a
result in the form of a number between 0 and 1. Gandhi Ji's photo on a fake note is legal
if the outcome is larger than 0.5; otherwise, it is not.
13. Lastly the correlation function was being used. Also, we were building our code to
determine whether or not the correlation value is greater than 0.5. If it is larger than 0.5,
then we were determining whether or not there are an equal number of lines.

32
 Chapter Four

Results and Discussion

4.1 Results and Discussion

Results:

There are other ways to detect if the money is phony or not, but they all follow the same basic
stages. Image capture, edge recognition, segmentation, grayscale conversion, and feature
extraction are among them. Most of the articles use MATLAB as their computation tool,
however we ultimately used OpenCV and Python as our programming language. To perform
comparisons and determine the outcome, a number of characteristics that identify genuine
currency apart from counterfeit ones are taken into account.

We are aware that these tools are used at banks and businesses to help identify counterfeit
money, but the average person who lacks these resources is susceptible to this. Our goal is to
offer a low-cost system with quick computations that can make decisions in a matter of seconds.
The entire process ought to function for Indian denomination 2000. It would be simple for the
general public to use, relatively portable, and reasonably priced. The model has some limitations.
We can get at most 81% of accuracy which may be sufficient. However, it is still more precise
than human detection. It can currently be utilized as an additional tool to lessen human mistake.
Additionally, the model's accuracy can be increased any further with more data and better
analysis.

Accuracy: The percentage of accurately classified data samples over all the data is known as
accuracy. Accuracy can be calculated by the following equation. [52]

Accuracy = (TP+TN)/(TP+FP+TN+FN)

33
 Discussion:

The comparision between accuracy that we found in our project with others project:

Name of Big Data Mobilenet Detection of A review of A review of Fake

Paper Analytics to V2-FCD: fake Fake Fake Currency
Authenticate Fake currency Currency Currency Detection
Bank Notes Currency using image Recognition Recognition
[our
Using K- Note processing Methods Methods
Proposed
Means Detection
[29] [32] [32] system]
Clustering
[10]
[11]

Algorithm Algorithm: Algorithm: Algorithm: Algorithm: Algorithm: Algorithm:

K-Means CNN K-means SVM Edge Distance-

Clustering algorithm detection weighted

Accuracy 87% 85% 99% 82.7% 90.45% 81%

34
 Chapter Five

Conclusion

5.1 Advantages

 The application will be very helpful in identifying counterfeit money.

 The software is simple to use and accessible.
 It will lessen the user's effort and save time.
 It provides the user methods that are more affordable, accurate, and give accurate
recognition of money notes.

5.2 Restriction:

 This project cannot be able to detect the currencies whether it is fake or not, of other
countries except India.
 This project is only able to detect the currencies whether it is fake or not with
denomination 2000 of Indian rupees.

5.3 Future Scope

 This project cannot be able to detect the currencies of other countries except India. So in
the future we can make this project possible to detect the currencies of other countries
also.
 This project is only able to detect the currencies whether it is fake or not with
denomination 2000 of Indian rupees. So in the future we can make it possible that it will
detect the currencies with all denomination.
 In this project, we worked using a few features of the currencies. So in the future we can
be able to work with all features of currencies to increase the accuracy of the project.

35
 Reference

1. Colaco, Rencita Maria, Rieona Fernandes, and S. Sowmya. "Efficient Image Processing
Technique for Authentication of Indian Paper Currency." 2021 International Conference
on Computer Communication and Informatics (ICCCI). IEEE, 2021.
2. Vadnere, G. V., Khilari, S. A., Sonawane, H. P., sharad Sanap, P., Koneri, A. M., &
Shinde, M. J. Indian Currency Identification Using Image Processing. Journal homepage:
www. ijrpr. com ISSN, 2582, 7421.
3. Jamkhandikar, Dayanand, Sahana Kaveri, and V. Poojashree Sudharani. "Indian Currency
Recognition System."(2021).
4. Sangogi, Mrs Jyoti, Ms Prachi Patil, and Ms Anuradha Jadhav. "FAKE CURRENCY
DETECTION USING IMAGE PROCESSING."
5. Shiby, Ashik, Fevitha Francis, Philip Jose, and Rintu Augustine. "FAKE CURRENCY
DETECTION USING IMAGE PROCESSING."
6. Warke, K. S., Kanthi, R., Makadia, D., Mogarkar, S., & Pawar, P. (2022).
COUNTERFEIT CURRENCY DETECTION USING IMAGE PROCESSING
7. Patil, Deepika P., Girija Varma, Shweta Poojary, Shraddha Sawant, and Aditya Sharma.
"Counterfeit Currency Detection based on AI."
8. Kudalkar et al., “Fake Currency Detection Using Image Processing”. AIJR Preprints, 388,
Version 1, 2022.
9. Pallavi, S., et al. "FAKE CURRENCY DETECTION." International Research Journal of
Modernization in Engineering Technology and Science (4076-4081) 4.06 (2022).
10. Potluri, T., Jahnavi, S., Motupalli, R. (2021). Mobilenet V2-FCD: Fake Currency Note
Detection. In: Luhach, A.K., Jat, D.S., Bin Ghazali, K.H., Gao, XZ., Lingras, P. (eds)
Advanced Informatics for Computing Research. ICAICR 2020. Communications in
Computer and Information Science, vol 1393. Springer, Singapore.
11. Khan, Mudassir, and Mahtab Alam. "Big Data Analytics to Authenticate Bank Notes
Using K-Means Clustering." Helix-The Scientific Explorer Peer Reviewed Bimonthly
International Journal 11.3 (2021): 1-6.
12. Vidhate, A., Shah, Y., Biyani, R., Keshri, H., & Nikhare, R. Fake Currency Detection
Application-Know Your Currency! PCE JCE_, 25.

36
13. Kumar, P. V., Akhila, V., Sushmitha, B., & Anusha, K. Detection of Fake Currency Using
KNN Algorithm.
14. NR, Deepak. "A Review on Fake currency detection using feature extraction." Available
at SSRN 4041599 (2022).
15. Siddiki, Marjuk Ahmed, et al. "Bangladeshi Currency Identification and Fraudulence
Detection Using Deep Learning and Feature Extraction." (2023).
16. Naseem, Syeda Aimen, et al. "Counterfeit Recognition of Pakistani Currency." KIET
Journal of Computing and Information Sciences 6.1 (2023): 123-147.
17. Saxena, Ankur, et al. "Fake currency detection using image processing." Int J Eng
Technology (UAE) 7 (2018): 199-205
18. Kanojia, Mahendra, Niketa Gandhi, and Amruta Rane. "Recognition and verification of
Indian currency notes using digital image processing." JOURNAL OF INFORMATION
ASSURANCE AND SECURITY 13.1 (2018): 30-37.
19. Shah, Raj, Mayank Champaneri, and Pujan Sheth. "Currency Counting Fake Note
Detection." Pujan and Gaikwad (Mohite), Vaishali, Currency Counting Fake Note
Detection (October 30, 2018).
20. Gaikwad, Mayadevi A., Vaijinath V. Bhosle, and Vaibhav D. Patil. "Automatic Indian
New Fake Currency Detection." International Journal of Engineering Research &
Technology (IJERT) 6.11 (2017).
21. Agasti, Tushar, et al. "Fake currency detection using image processing." IOP Conference
Series: Materials Science and Engineering. Vol. 263. No. 5. IOP Publishing, 2017.
22. Tele, Gouri Sanjay, et al. "Detection of fake Indian currency." International Journal of
Advance Research, Ideas and Innovations in Technology 4.2 (2018): 170-176.
23. Amirsab, Shaikh Ajij, Mohammad Mudassir, and Mohammad Ismail. "An automated
recognition of fake or destroyed Indian currency notes." International journal of advance
scientific research and engineering trends volume 2.7 (2017).
24. Deepak, M. P., and N. B. Prajwala. "Identification of fake notes and denomination
recognition." 2018 International Conference on Communication and Signal Processing
(ICCSP). IEEE, 2018.

37
25. Reserve Bank of India. Annual Report 2018-19. Available online:
https://rbidocs.rbi.org.in/rdocs/Publications/PDFs/ANRP201920F2210F9814A5B4A5B5E
5A5EF9BE9EDAB43.PDF (accessed on 1 July 2021).
26. Sawant, K. and More, C. (2016) "Currency Recognition Using Image Processing and
Minimum Distance Classifier Technique," International Journal of Advanced Engineering
Research and Science (IJAERS), Vol. 3, Issue 9, pp. 1-5. ISSN: 2349-6495.
27. Roy, V., Mishra, G., Mannadiar, R., & Patil, S. (2019). Fake Currency Detection Using
Image Processing. International Journal of Computer Science and Mobile Computing,
8(4), 88-93. ISSN 2320-088X.
28. Sonali R. Darade, Prof. G. R. Gidveer, “Automatic Recognition of Fake Indian
Currency Note”, 2016 International Conference on Electrical Power and Energy Systems
(ICEPES) Maulana Azad National Institute of Technology,Bhopal, India. Dec 15-16, 2016
29. Singh, A. (2019). Detection of fake currency using image processing. International
Journal of Engineering Research & Technology, 8(12).
30. Miss.I. Santhiya Irulappasamy. "Research on Fake Indian Currency Note Detection using
Image Processing." International Journal of Scientific Development and Research
(IJSDR), vol. 6, issue 3, March 2021, pp. 196. ISSN: 2455-2631.
31. Atchaya, S., Harini, K., Kaviarasi, G., & Swathi, B. (2016). Fake Currency Detection
Using Image Processing. International Journal of Trend in Research and Development
(IJTRD), ISSN: 2394-9333.
32. Sruthy R. (2022) "A review of Fake Currency Recognition Methods". International
Research Journal of Engineering and Technology (IRJET), Volume 09, Issue 07. ISSN:
2395-0056.
33. Arun Anoop M, “ Study in multimedia Security: Multimedia Forgery Detection and
History A survey”, International Journal of Printing Packaging and Allited Sciences, vol
4,No 4, December 2016.
34. Aakash S Patel,”Indian Paper Currency Detection” International Journal for Scientific
Research and Development(IJSRD), vol 7, issue 06, ISSN : 2321-0613 (June 2019)
35. Vidhi Roy: Gangey Mishra;Rahul Mannadiar; Sushant Patil “Fake currency detection
using image processing” (ICCSP). IEEE, 2019.

38
36. Tushar Agasti, Pratik Wade and P Chitra “Fake currency detection using image
processing” (ICCSP). IEEE, 2017.
37. S.Mittal, Indian Banknote Recogniton using Convolutional Neural Network, IEEE, 3rd
International Conference on Internet of things. (2018)
38. Aman Tekriwal, Prashant Mankani detection of fake currency." International Journal of
Engineering Research & Technology (IJERT) 6.11 (2017).
39. Devang S.Gujarathi, Sushii V. Nikam, Prathmesh P.gogte, currency recognition system
for using image processing ." International journal of advance scientific research and
engineering trends volume 2.7 (2017).
40. Bhurke, Chinmay, Meghana Sirdeshmukh, and M. S. Kanitkar. "Currency recognition
using image processing." Int. J. Innov. Res. Comput. Commun. Eng 3 (2015): 4418-4422.
41. Web link: “https://www.python.org/doc/essays/blurb/”. Accessed 28 February 2023
42. Web link: “https://www.geeksforgeeks.org/opencv-overview/”. Accessed 30 February
2023
43. Web link: “https://numpy.org/doc/stable/user/absolute_beginners.html”. Accessed 28
February 2023
44. Web link:
“https://code.visualstudio.com/docs/supporting/FAQ#:~:text=Visual%20Studio%20Code
%20is%20a,such%20as%20Visual%20Studio%20IDE”. Accessed 27 February 2023.
45. Web link: “https://qualitastech.com/image-acquisition/image-acquisition-in-machine-
vision-trigger-
mechanism/#:~:text=What%20is%20Image%20Acquisition%3F,render%20the%20entire
%20workflow%20useless”. Accessed 26 February 2023
46. Web link: “https://e2eml.school/convert_rgb_to_grayscale.html”. Accessed 28 February
2023
47. Web link: “https://www.dynamsoft.com/blog/insights/image-processing/image-rocessing-
101-color-space-conversion/”.Accessed 28 February 2023
48. Web Link: “https://datagen.tech/guides/image-annotation/image-
segmentation/#:~:text=Image%20segmentation%20is%20a%20method,analysis%20of%2
0each%20image%20segment”. Accessed 26 February 2023.

39
 49. Truong, Mai Thanh Nhat, and Sanghoon Kim. "A review on image feature detection and
description." Proceedings of the Korea Information Processing Society Conference. Korea
Information Processing Society, 2016.
50. Web link: “https://observablehq.com/@lemonnish/cross-correlation-of-2-images”.
Accessed 26 February 2023.
51. Web Link: “https://www.rbi.org.in”. Accessed 30 February 2023.
52. Samia, Moythry Manir, et al. "Aspect-based Sentiment Analysis for Bengali Text using
Bidirectional Encoder Representations from Transformers (BERT)." International Journal
of Advanced Computer Science and Applications 13.12 (2022).

40

View publication stats