Blockchain for tracking serial numbers in money exchanges

• Author: Khaled Abdel‐Hakeem,Kareem Mohamed,Amr Aziz,Mostafa Mostafa,Ayman Atia,Belal Mohamed
• Date: 01 October 2019
• DOI: http://doi.org/10.1002/isaf.1462
• Published date: 01 October 2019

RESEARCH ARTICLE

Blockchain for tracking serial numbers in money exchanges

Kareem Mohamed

1

| Amr Aziz

1

| Belal Mohamed

1

| Khaled Abdel-Hakeem

1

|

Mostafa Mostafa

1

| Ayman Atia

1,2,3

1

Computer Science, Faculty of Computer and

Information, University of Helwan, Egypt

2

Faculty of Computer Science, Misr

International University, Egypt

3

Faculty of Computers and Information HCI-

LAB, Helwan University, Egypt

Correspondence

Kareem Mohamed, Computer Science Faculty

of Computer and Information University of

Helwan. Egypt.

Email: fcih.kareem@fci.helwan.edu.eg

Summary

Money exchange is one of the most common day-to-day activities performed by

humans in the daily market. This paper presents an approach to money tracking

through a blockchain. The proposed approach consists of three main components:

serial number localization, serial number recognition, and a blockchain to store all

transactions and ownership transfers. The approach was tested with a total of

110 banknotes of different currency types and achieved an average accuracy of

91.17%. We conducted a user study in real-time with 21 users, and the mean accu-

racy across all users was 86.42%. Each user gave us feedback on the proposed

approach, and most of them welcomed the idea.

KEYWORDS

image processing, banknote serial number recognition (SNR), blockchain

1|INTRODUCTION

Payments can be represented by many forms of cash, check, debit,

and credit. Each type of payment requires a certain level of trust

(Angel & McCabe, 2015) except for cash. In a cash exchange, the

buyer and seller are directly involved in the payment process with no

need for a third party. In a sur- vey (World Cash Report., 2018), 75%

of the participating countries reported that cash is used in over 50%

of transactions. Nonetheless, cash is vulnerable to theft (Angel &

McCabe, 2015). According to the World Bank (Losses due to theft

and vandalism., 2016), money losses affect 17.0% of the annual sales

of firms in Egypt, where our case study was conducted, and tend to

exceed this ratio in some other countries. According to the statistics

that are presented in (Smith, Pieth, & Jorge, 2007), the theft of paper

money is one of the hardest issues to track; in particular, no one can

record all of the serial numbers of money that is received/sent every

day. Thus, a need for tracking ownership exists.

The serial number is the only way to distinguish one banknote

from another. Banks usually use the serial numbers on paper currency

to track it in case of theft, which leaves smaller organizations and indi-

viduals exposed, as they cannot track the money that they process on

a daily basis. The problem is the inability to track the money once it is

lost, as the stolen money becomes combined with large amounts of

other money with no clue regarding which specific banknotes were

stolen. This problem led to the invention of the money marking sys-

tem (Herndon & George, 1979), which allows money to be traced to

its source by stamping each banknote with the date of the transaction

along with a unique identifier number for the owner. The unique iden-

tifier allows the owner to notify the police with the identifier and the

date of the theft. The transaction information can be widely distrib-

uted to other merchants so that they can refuse to accept the stolen

money and contact the authorities. This idea was presented in 1979

as a USA patent; however, no one has considered its limitations in

recording all the transactions happening in the market. Today, using

the current technologies related to Big Data, cloud storage and aug-

mented reality, we can move this patent into a new dimension.

Detecting a serial number requires segmenting the serial number

without having a fixed location (region of interest) in terms of the

banknote dimensions. Figure 1 shows three samples of different

money currencies for the Egyptian pound (10 and 200 EGP). It is

remarkable that for the three types of Egyptian money, (a) 10 pounds.

200 pounds issued in 2016 and (c) 200 pounds issued in 2011,

the position of the serial number (marked with a black box) is different

for each currency category and depends on the production date as

well. The numbers, which are all written in the Hindu-Arabic numeric

system, are the target of our study.

The currency detection and tracking system (Park, 2008) also uses

money tracking to combat money laundering and other criminal

Received: 4 June 2019 Revised: 4 October 2019 Accepted: 6 October 2019

DOI: 10.1002/isaf.1462

Intell Sys Acc Fin Mgmt. 2019;26:193–201. wileyonlinelibrary.com/journal/isaf © 2020 John Wiley & Sons, Ltd. 193