AUTOMATED COUNTERFEIT CURRENCY RECOGNITION

AND FRAUD DETECTION SYSTEM

Main Project Report

Submitted to the APJ Abdul Kalam Technological University

in partial fulfillment of requirements for the award of degree

Bachelor of Technology

in

Computer Science and Engineering

by

ALEENA SARA VARGHESE (LBT21CS013)

ARYA B NAIR (LBT21CS021​)
BIYA MERIN BINU (LBT21CS030)
CHANCHAL MANJU RAJEEV (LBT21CS032)

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

LBS INSTITUTE OF TECHNOLOGY FOR WOMEN

TRIVANDRUM, KERALA
 TABLE OF CONTENTS

TITLE​ ​ ​ ​ ​ ​ ​ ​ ​ PAGE NO:

INTRODUCTION​ ​ ​ ​ ​ ​ 1

OBJECTIVES​​ ​ ​ ​ ​ 2​ ​

METHODOLOGY​ ​ ​ ​ ​ 3

IMPLEMENTATION​​ ​ ​ ​ 4

PROJECT TIMELINE​ ​ ​ ​ ​ 18

CONCLUSION​ ​ ​ ​ ​ ​ 19

REFERENCES​ ​ ​ ​ ​ ​ ​ ​ 20
 INTRODUCTION

Currency notes hold immense value in modern economies as a physical representation of

monetary transactions. Their importance goes far beyond their physical attributes, encompassing
their role as a medium of exchange, a store of value, and a standard for deferred payments. The
widespread use of currency notes makes them essential for facilitating both small-scale
day-to-day transactions and large-scale commercial operations, acting as a bridge between
individuals, businesses, and governments in the economic system. The intrinsic value of a
currency note is the cost of its production, which includes the materials, printing, and security
features. This value is relatively low compared to the face value of the note, which is the
denomination it represents and the purchasing power it holds in the economy. For example, the
production cost of a ₹100 note is significantly less than its face value, but the trust and
confidence of the public in the monetary system allow it to function effectively as legal tender.
The difference between intrinsic and face value underscores the importance of trust in currency.
This trust is built and maintained through the central bank's credibility, the use of advanced
security features, and the consistent enforcement of anti-counterfeiting measures. Without this
trust, the economic system would face disruptions, as individuals and businesses might become
reluctant to accept currency for transactions.

As a medium of exchange, currency eliminates the inefficiencies of the barter system by

providing a universally accepted means of exchange, enabling smoother and quicker
transactions. Currency serves as a measure of value, providing a standard for pricing goods and
services, facilitating economic comparisons, and supporting informed decision-making.
Currency also acts as a standard for deferred payments, enabling its use in credit systems and
long-term financial contracts, which supports economic stability and fosters growth. These roles
make currency indispensable, and their reliability is a cornerstone of public confidence in the
economy.

To safeguard the value of currency notes, central banks incorporate a range of security
features. Watermarks, for instance, are embedded designs visible when held up to light,
enhancing the note’s [Link] are reflective images that change when viewed from
different angles, making replication difficult. Microprinting and fine line patterns are intricate
details that standard printing technologies cannot replicate, serving as a deterrent to
counterfeiting. Color-changing ink, which shifts visibly when the note is tilted, is another
advanced feature that adds an additional layer of security. These features are intended to ensure
the authenticity of notes and protect the monetary system from counterfeiting. However,
advancements in technology, such as high-resolution printers and scanners, have made it easier
for counterfeiters to mimic these features, posing a significant threat to the integrity of currency.
Counterfeit currency diminishes the trust people place in physical money, leading to broader
economic consequences. Fake notes reduce the purchasing power of individuals and lead to
direct financial losses for businesses and banks. An influx of fake money can disrupt the
supply-demand balance, contributing to inflation and undermining economic stability. Public and
institutional confidence in the currency and the economy as a whole may deteriorate, leading to
reduced participation in financial activities and a weakening of the monetary system. The critical
value of currency notes necessitates continuous efforts to preserve their authenticity and
functionality. This includes not only refining security features but also developing automated
systems to detect counterfeit notes efficiently. As technology evolves, so must the mechanisms to
protect the integrity of currency notes, making this an ongoing and vital focus for governments,
financial institutions, and technologists worldwide.\

OBJECTIVES
1.​ Develop an efficient system for recognizing and classifying currency denominations using
advanced image processing and deep learning techniques.
2.​ Develop a robust approach for verifying the authenticity of banknotes by examining distinct
features and patterns to detect counterfeit currency.
 METHODOLOGY
SIFT-FLANN

Sift and Flann is used to identify the counterfeit notes by matching the keypoints in the input
image and reference [Link] identifies the keypoints in the notes and flann matches the
keypoints of both the images.
The figure below gives the working of counterfeit detection using Sift and Flann.

ResNet -18
ResNet-18 is a deep learning technique that can be used to identify the denomination of
each class in our [Link] below shows the working of denomination identification
using ResNet-18
YOLOv11

YOLOv11 can be used to classify currency notes as real or fake while also identifying their
denominations.
Figure below shows the working of YOLOv11 used for currency denomination and classification
and counterfeit detection.

IMPLEMENTATION
SIFT-FLANN

Sift and flann is tested with taking 3 features,8 features and full features of the currency note.

3 FEATURES
OUTPUT: 3 FEATURES
8 FEATURES
OUTPUT: 8 FEATURES
FULL FEATURES
OUTPUT: FULL FEATURES
The 3 features system is more accurate than taking 8 features and full [Link],it is not
advisable to take only 3 features of the currency note for counterfeit detection.

ResNet-18
OUTPUT
YOLOv11

Figure below shows the code used for training the model

Figure below shows the code for detecting the denomination as well as authenticity of the note:

OUTPUT
TIMELINE
 CONCLUSION
In conclusion, our system presented an integrated approach for counterfeit currency detection
and denomination classification, using the strengths of SIFT and FLANN for feature matching
and YOLOv11 for classification. Our proposed system addresses the dual challenge of
identifying counterfeit notes and accurately determining their denomination. The combination of
SIFT and FLANN allowed for reliable counterfeit detection by comparing key features of
suspect notes against genuine currency templates. Simultaneously, the YOLOv11 model
demonstrated its efficacy in accurately classifying the denomination of genuine currency,
contributing to the overall system's accuracy and efficiency. The experimental results
demonstrate the potential of our approach for real-world applications, offering a solution for
automated currency verification and reducing the risks associated with counterfeit money. Future
work may explore further optimization of the algorithms, the integration of additional features
for enhanced counterfeit detection, and the development of a more user-friendly and deployable
system. Additionally, exploring the impact of different lighting conditions and image quality on
the system's performance would be a valuable direction for future research.

REFERENCES

[1] Mr. K Bhushanm, [Link], [Link] Sultana , [Link] Kumar , S. Mahesh Babu \”Fake Currency
Detection using Deep Learning \”, 2024 International Journal of Advanced Research in Computer and
Communication Engineering (IJARCCE ).

[2] Kiran Kamble, Anuthi Bhansali, Pranali Satalgaonkar, Shruti Alagundgi \”Counterfeit Currency
Detection using Deep Convolutional Neural Network\”,2019 IEEE Pune Section International Conference
(PuneCon).

[3] Ms. Megha Jadhav, Dr. Yogesh kumar Sharma, Dr. G. M. Bhandari \”Currency Identification and
Forged Banknote Detection using Deep Learning\”,2019 International Conference on Innovative Trends
and Advances in Engineering and Technology (ICITAET).

[4] Rohith Pokala, Varun Teja \”Indian Currency Recognition For Blind People\”,2020 International
Research Journal of Engineering and Technology (IRJET).
[5] [Link] , N. Sailaja, [Link], [Link], M. Sai Reethika \”Counterfeit Currency Detection
using Machine Learning \” ,2023 IJRTI Volume 8 Issue 4

[6] Rahima Khanam and MuhammadHussain \”YOLOV11: AN OVERVIEW OF THE KEY

ARCHITECTURAL ENHANCEMENTS “ 2024

[7] Ms. Agrima Gupta ,Ms. Rashleen Kour \”Fake Currency Detection Using ORB Algorithm \”, 2022
International Journal for Research in Applied Science & Engineering Technology (IJRASET).

[8] Rakesh Chandra Joshi ,Saumya Yadav, Malay Kishore Dutta \” YOLO-v3 Based Currency Detection
and Recognition System for Visually Impaired Persons\”2020 International Conference on Contemporary
Computing and Applications (IC3A) .

[9] Mr. K Bhushanm, [Link], [Link] Sultana, [Link] Kumar, [Link] Babu ,\”Fake Currency
Detection using Deep Learning \”, 2024 International Journal of Advanced Research in Computer and
Communication Engineering

[10][Link] Kumar, Gaurav Singal, Shwetha Sirikonda, R. Nethravati\”A Novel Approach for Detection
of Counterfeit Indian Currency Notes Using Convolutional Neural Networks \”2020 IOP Conference
Series : Materials Science and Engineering (ICRAEM 2020)