Deep Learning-Based Multinational Banknote Type and Fitness Classification with the Combined Images by Visible-Light Reflection and Infrared-Light Transmission Image Sensors.

Tuyen Pham,Chanhum Park,Kang Park,Dat Nguyen

doi:10.3390/s19040792

Abstract

Automatic sorting of banknotes in payment facilities, such as automated payment machines or vending machines, consists of many tasks such as recognition of banknote type, classification of fitness for recirculation, and counterfeit detection. Previous studies addressing these problems have mostly reported separately on each of these classification tasks and for a specific type of currency only. In other words, there has been little research conducted considering a combination of these multiple tasks, such as classification of banknote denomination and fitness of banknotes, as well as considering a multinational currency condition of the method. To overcome this issue, we propose a multinational banknote type and fitness classification method that both recognizes the denomination and input direction of banknotes and determines whether the banknote is suitable for reuse or should be replaced by a new one. We also propose a method for estimating the fitness value of banknotes and the consistency of the estimation results among input trials of a banknote. Our method is based on a combination of infrared-light transmission and visible-light reflection images of the input banknote and uses deep-learning techniques with a convolutional neural network. The experimental results on a dataset composed of Indian rupee (INR), Korean won (KRW), and United States dollar (USD) banknote images with mixture of two and three fitness levels showed that the proposed method gives good performance in the combination condition of currency types and classification tasks.

Highlights

Despite the growth of electronic financial transactions that have caused a decrease in the use of physical currency, transactions involving banknotes are still playing an important role in daily life as well as large-scale commercial exchanges
Regarding the functionalities of the automated banknote sorting system, one popular approach is based on image processing, in which input banknotes are captured by various imaging sensors and their optical characteristics are used for classification tasks [1,2]
We proposed a banknote classification method that simultaneously classifies banknotes from multiple national currencies in both types and fitness levels using the combination of infrared transmission (IRT) and visible-light reflection (VR) images of the input banknote and the convolutional neural network (CNN)

Summary

Introduction

Despite the growth of electronic financial transactions that have caused a decrease in the use of physical currency, transactions involving banknotes are still playing an important role in daily life as well as large-scale commercial exchanges. Automated machines involve many processes in these transactions and have the ability to handle multiple tasks, which are banknote recognition, fitness classification, counterfeit detection, and serial number recognition. Banknote recognition determines the denomination of the input currency paper, and fitness classification evaluates the physical condition of the banknote and decides whether it is suitable for recirculation or if it should be replaced by a new one. The determination of a banknote’s denomination is the primary function of the counting system; fitness classification helps to prevent problems that might occur due to a low quality or damaged banknote being inserted into the system, such as jams or incorrect recognition. Regarding the functionalities of the automated banknote sorting system, one popular approach is based on image processing, in which input banknotes are captured by various imaging sensors and their optical characteristics are used for classification tasks [1,2]. Previous studies about banknote recognition and banknote fitness classification are mostly reported separately for each of the problems; that is, when dealing with banknote recognition, little research has considered the fitness of the banknotes for recirculation, fitness classification studies have mostly been conducted under the assumption that the type and denomination of banknotes have been correctly pre-classified

Objectives

Methods

Findings

Conclusion