Banknote Image Research Articles

The HMM-based banknote recognition system

In order to satisfy the fast, accurate and high reliability of banknote image recognition, a banknote image recognition method based on hidden Markov models (MHH) is proposed in this paper. The grid method is used to extract banknote image features and the process of image preprocessing. By analysing the experimental data, the number of Markov model states, the number of iterations and the number of Gaussian models is finally determined. This method can be used to identify any currency. In this paper, 16000 pieces of RMB were collected by the self-developed banknote cleaner as experimental samples. Experimental results show that the proposed method achieves higher recognition rate than using neural network and support vector function.

Vision Based Banknote Recognition System

Abstract: People who are visually impaired face great difficulties in recognizing banknotes for day-to-day transactions as the texture and shape of many banknotes are very similar. Also, the Indian currency system does not have a separate banknote in circulation that might help people who are visually disabled. This paper presents a real-time Indian currency detection system for visually impaired persons. The proposed work makes use of the vision-based system to facilitate the visually impaired people to prosperously recognize banknotes. In this paper, a recognition system is implemented by making use of various machine learning and deep learning approaches which can produce an accuracy of up to 90%. In this system, totally two experiments are carried out on the banknote’s image. The first one is the analysis of the different regions of the banknote, the second experiment is the examination of the banknote image which was captured using a camera in a regulated environment. Our system is extended to support the visually impaired people by executing a “Voice-based assistance system” which would help us in giving the voice output. This system can identify all the seven banknotes used in India and exhibit the result with both textual and auditory output. People who are visually impaired will be able to easily use it in daily transactions.

One- and Five-Ringgit Malaysia banknotes reader with counterfeit detection for visually impaired person using backlight mechanism and image processing techniques.

Visually impaired persons face challenges in running business activities, especially in handling banknotes. Malaysia researchers had proposed some Ringgit banknotes recognition systems to aid visually impaired persons recognize and classify Ringgit banknotes. However, these electronic banknote readers can only recognize Malaysian Banknotes' Ringgit value, they have no counterfeit detection features. The purpose of this study is to develop a banknote reader that not only can help visually impaired persons recognize the banknote value, but also to detect the counterfeit of the banknote, safeguarding their losses. This paper proposed a Malaysian banknote reader using backlight mechanism and image processing techniques to read and detect counterfeit for one Ringgit and five Ringgit Malaysian banknotes. The developed handheld banknote reader used visual type sensor to capture banknote image, passed to raspberry pi controller to perform image processing on banknote value and the extracted watermarks features. The developed image processing algorithm will trace out the region of interests: 1)see-thru windows, 2)Crescent and Star, 3)Perfect see though register and detect the watermarks features accordingly. The processed result will be passed back to the handheld banknote reader and broadcast on an attached mini speaker to aid the visually impaired understand the holding banknote, whether it is a real one Ringgit, real five Ringgit or none of them. The experimental result shown by this approach able to accomplish numerous round of banknote reading attempts with successful outcomes. Confusion matrix is further employed to study the performance of the banknote reader, in terms of true positive, true negative, false positive and false negative. Details analysis had been focused on the critical false positive cases (predicted real banknote and actually is fake banknote) and false negative cases (predicted fake banknote and it is actually real banknote).

Deep learning-based iraqi banknotes classification system for blind people

Modern systems have been focusing on improving the quality of life for people. Hence, new technologies and systems are currently utilized extensively in different sectors of our societies, such as education and medicine. One of the medical applications is using computer vision technology to help blind people in their daily endeavors and reduce their frequent dependence on their close people and also create a state of independence for visually impaired people in conducting daily financial operations. Motivated by this fact, the work concentrates on assisting the visually impaired to distinguish among Iraqi banknotes. In essence, we employ computer vision in conjunction with Deep Learning algorithms to build a multiclass classification model for classifying the banknotes. This system will produce specific vocal commands that are equivalent to the categorized banknote image, and then inform the visually impaired people of the denomination of each banknote. To classify the Iraqi banknotes, it is important to know that they have two sides: the Arabic side and the English side, which is considered one of the important issues for human-computer interaction (HCI) in constructing the classification model. In this paper, we use a database, which comprises 3,961 image samples of the seven Iraqi paper currency categories. Furthermore, a nineteen layers Convolutional Neural Network (CNN) is trained using this database in order to distinguish among the denominations of the banknotes. Finally, the developed system has exhibited an accuracy of 98.6 %, which proves the feasibility of the proposed model.

Currency Recognition System Using Image Processing: Libyan Banknote as a Case Study

Banknote recognition in the field of image processing is used to distinguish between banknotes of different countries and determine their values. One of the problems that emerged with this process is the recognization of the banknote. The recognition operations for Myanmar currency denominations comprise six stages: image capturing, preprocessing, target region extraction, number segmentation, currency value recognition, and output with screen displaying and voice. The currency recognition system is implemented by comparing the similarity of each banknote value template via image processing steps in the Matlab environment. To prove the effectiveness of the proposed system, the number of banknote images with two types (new and medium) from each of the four types of denominations were tried

A Study of Spatial Attention and Squeeze Excitation Block Fusion Improved ResNet for Identifying Bank Notes

Based on deep learning and digital image processing algorithms, we design and implement an accurate automatic recognition system for bank note text and propose an improved recognition method based on ResNet for the problems of difficult image text extraction and insufficient recognition accuracy. Firstly, a deep hyperparameterized convolution (DO-Conv) is used instead of the traditional convolution in the network to improve the recognition rate while reducing the model parameters. Then, the spatial attention model (SAM) and the squeezed excitation block (SE-Block) are fused and applied to a modified ResNet to extract detailed features of bank note images in the channel and spatial domains. Finally, the label-smoothed cross-entropy (LSCE) loss function is used to train the model to automatically calibrate the network to prevent classification errors. The experimental results demonstrate that the improved model is not easily affected by the image quality, and the model in this paper has good performance in text detection and recognition in specific business ticket scenarios.

One Ringgit and five Ringgit Malaysian banknotes reader using backlight mechanism and image processing techniques

Visually impaired persons face challenges in running business activities, especially in handling banknotes. Malaysia researchers had proposed some Ringgit banknotes recognition systems to aid visually impaired persons recognize and classify Ringgit banknotes. However, these electronic banknote readers can only recognize Malaysian Banknotes’ Ringgit value, they have no counterfeit detection features. The purpose of this study is to develop a banknote reader that not only can help visually impaired persons recognize the banknote value, but also to detect the counterfeit of the banknote, safeguarding their losses. This paper proposed a Malaysian banknote reader using backlight mechanism and image processing techniques to read and detect counterfeit for one Ringgit and five Ringgit Malaysian banknotes. The developed handheld banknote reader used visual type sensor to capture banknote image, passed to raspberry pi controller to perform image processing on banknote value and the extracted watermarks features. The developed image processing algorithm will trace out the region of interests: 1)see-thru windows, 2)Crescent and Star, 3)Perfect see though register and detect the watermarks features accordingly. The processed result will be passed back to the handheld banknote reader and broadcast on an attached mini speaker to aid the visually impaired understand the holding banknote, whether it is a real one Ringgit, real five Ringgit or none of them. The experimental result shown by this approach able to accomplish numerous round of banknote reading attempts with successful outcomes. Confusion matrix is further employed to study the performance of the banknote reader, in terms of true positive, true negative, false positive and false negative. Details analysis had been focused on the critical false positive cases (predicted real banknote and actually is fake banknote) and false negative cases (predicted fake banknote and it is actually real banknote).

Fuzzy Logic Based Perceptual Image Hashing Algorithm in Malaysian Banknotes Detection System for the Visually Impaired

Visually impaired persons have difficulty in business that dealing with banknote. This paper proposed a Malaysian banknotes detection system using image processing technology and fuzzy logic algorithm for the visually impaired. The Malaysian banknote reader will first capture the inserted banknote image, sending it to the cloud server for image processing via Wi-Fi medium. The cloud server is established to receive the banknote image sending from the banknote reader, processing them using perceptual hashing based image searching and fuzzy logic algorithm, then return the detected banknote’s value results back to the banknote reader. The banknote reader will display the results in terms of voice message played on the mini speaker attached on it, to allow visually impaired persons knowing the banknote’s value. This hardware mechanism reduces the size and costs for the banknote reader carried by the visually impaired persons. Experimental results showed that this Malaysian banknotes detection system reached an accuracy beyond 95% by running test on 600 different worn, torn and new Malaysian banknotes. After the banknote image being taken by the banknote reader’s camera, the system able to detect the banknote value in about 480 mili-seconds to 560 mili-seconds for a single sided banknote recognition. The banknotes detection speed was also comparable with human observers reading banknotes, with the response of 1.0908 second per banknote slight difference reading time. The IoT and image processing concepts were successfully blended and it provides an alternative to aid the visually impaired person their daily business transaction activities in a better way.

Banknote Classification Based on Convolutional Neural Network in Quaternion Wavelet Domain

In this paper, we propose a new framework for banknote classification based on quaternion wavelet transform (QWT) and deep convolutional neural network. Firstly, the QWT is applied to describe the phase and magnitude of different banknote images which has inherent directional sensitivity and multi-scale framework. Then we design a deep convolutional neural network which is trained on banknote images along with the magnitude and phase of quaternion wavelet coefficients. We assign the neural weights on the output probabilities of deep convolutional neural network and update these weights by utilizing the back propagation algorithm. Finally, the trained networks with decision of a weighted sum and the magnitude and the phase of quaternion wavelet networks are utilized for banknote image classification. The performance of our algorithm is experimentally verified on a variety of banknote databases. Experimental results show that the proposed algorithm achieves superior performance compared with other state-of-the-art banknote classification algorithms. The proposed algorithm can also satisfy the real-time requirements of the banknote sorting system.

Deep Learning-Based Fake-Banknote Detection for the Visually Impaired People Using Visible-Light Images Captured by Smartphone Cameras

Automatic recognition of fake banknotes is an important task in practical banknote handling. Research on this task has mostly involved methods applied to automatic sorting machines with multiple imaging sensors or that use specialized sensors for capturing banknote images in various light wavelengths. These approaches can make use of the security features on banknotes for counterfeit detection. However, they require specialized devices, which are not always available for general users or visually impaired people. Meanwhile, smartphones are becoming more popular and can be useful imaging devices. Moreover, the types of fake banknotes created by imaging devices such as smartphone cameras or scanners are sometimes cannot be recognized by especially the visually impaired people. Addressing these problems, we propose a method for classifying fake and genuine banknotes using visible-light images captured by smartphone cameras based on convolutional neural networks. Experimental results on a self-collected dataset of US dollar, Euro, Korean won, and Jordanian dinar banknotes showed that our method performs better in terms of fake detection than the state-of-the-art methods.

A hybrid convolution network for serial number recognition on banknotes

As the sole identity of banknote, serial number has played a crucial role in monitoring the circulation of currencies. Serial number recognition plays an important role in financial market, which requires fast and accurate performances in real applications. In this paper, a hybrid convolution network model has been proposed, in which a dilated-based convolution neural network is employed to improve the recognition accuracy and a quantitative neural network method is developed to speed up the identification process. In dilated-based convolution neural network, the convolution layer and the pooling layer have been replaced by dilated convolution, which can reduce the computation cost. The quantitative neural network based method quantizes the weight parameters to an integer power of two, which transforms the original multiplication operation to a shift operation and can greatly reduce the time. The proposed model was examined and tested on four different banknotes with 35,000 banknote images including RMB, HKD, USD and GBP. The experimental results show that, the proposed model can efficiently improve the recognition accuracy to 99.89% and reduce the recognition time to less than 0.1 ms, and it outperforms the other algorithms on both recognition accuracy and recognition speed.

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

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.

Deep Learning-Based Multinational Banknote Fitness Classification with a Combination of Visible-Light Reflection and Infrared-Light Transmission Images

The fitness classification of a banknote is important as it assesses the quality of banknotes in automated banknote sorting facilities, such as counting or automated teller machines. The popular approaches are primarily based on image processing, with banknote images acquired by various sensors. However, most of these methods assume that the currency type, denomination, and exposed direction of the banknote are known. In other words, not only is a pre-classification of the type of input banknote required, but in some cases, the type of currency is required to be manually selected. To address this problem, we propose a multinational banknote fitness-classification method that simultaneously determines the fitness level of a banknote from multiple countries. This is achieved without the pre-classification of input direction and denomination of the banknote, using visible-light reflection and infrared-light transmission images of banknotes, and a convolutional neural network. The experimental results on the combined banknote image database consisting of the Indian rupee and Korean won with three fitness levels, and the United States dollar with two fitness levels, show that the proposed method achieves better accuracy than other fitness classification methods.

Deep Learning-Based Banknote Fitness Classification Using the Reflection Images by a Visible-Light One-Dimensional Line Image Sensor.

In automatic paper currency sorting, fitness classification is a technique that assesses the quality of banknotes to determine whether a banknote is suitable for recirculation or should be replaced. Studies on using visible-light reflection images of banknotes for evaluating their usability have been reported. However, most of them were conducted under the assumption that the denomination and input direction of the banknote are predetermined. In other words, a pre-classification of the type of input banknote is required. To address this problem, we proposed a deep learning-based fitness-classification method that recognizes the fitness level of a banknote regardless of the denomination and input direction of the banknote to the system, using the reflection images of banknotes by visible-light one-dimensional line image sensor and a convolutional neural network (CNN). Experimental results on the banknote image databases of the Korean won (KRW) and the Indian rupee (INR) with three fitness levels, and the Unites States dollar (USD) with two fitness levels, showed that our method gives better classification accuracy than other methods.

Dominant color and texture feature extraction for banknote discrimination

Banknote discrimination with image recognition technology is significant in many applications. The traditional methods based on image recognition only recognize the banknote denomination without discriminating the counterfeit banknote. To solve this problem, we propose a systematical banknote discrimination approach with the dominant color and texture features. After capturing the visible and infrared images of the test banknote, we first implement the tilt correction based on the principal component analysis (PCA) algorithm. Second, we extract the dominant color feature of the visible banknote image to recognize the denomination. Third, we propose an adaptively weighted local binary pattern with “delta” tolerance algorithm to extract the texture features of the infrared banknote image. At last, we discriminate the genuine or counterfeit banknote by comparing the texture features between the test banknote and the benchmark banknote. The proposed approach is tested using 14,000 banknotes of six different denominations from Chinese yuan (CNY). The experimental results show 100% accuracy for denomination recognition and 99.92% accuracy for counterfeit banknote discrimination.

Multi-National Banknote Classification Based on Visible-light Line Sensor and Convolutional Neural Network

Automatic recognition of banknotes is applied in payment facilities, such as automated teller machines (ATMs) and banknote counters. Besides the popular approaches that focus on studying the methods applied to various individual types of currencies, there have been studies conducted on simultaneous classification of banknotes from multiple countries. However, their methods were conducted with limited numbers of banknote images, national currencies, and denominations. To address this issue, we propose a multi-national banknote classification method based on visible-light banknote images captured by a one-dimensional line sensor and classified by a convolutional neural network (CNN) considering the size information of each denomination. Experiments conducted on the combined banknote image database of six countries with 62 denominations gave a classification accuracy of 100%, and results show that our proposed algorithm outperforms previous methods.

Banknote recognition based on optimization of discriminative regions by genetic algorithm with one-dimensional visible-light line sensor

Banknote recognition is an important task in many automatic payment facilities and counting machines. The most popular approach is based on image processing methods in which banknote images are captured by visible light sensors and are classified by denominations and input orientations. There are regions on a banknote image that yield better recognition accuracy than the other areas. There have been few studies on optimal discriminative regions on a banknote image; therefore, we proposed a banknote recognition method to select the discriminative regions on the banknote image captured by a one-dimensional visible light sensor. The proposed method uses genetic algorithm to optimize the similarity mapping result for different classes of banknotes. Experimental results with banknote databases from various countries show that our proposed method results in better accuracies than previous methods with the average recognition accuracies of higher than 99% and small variance among five trials in each type of currency.

New banknote defect detection algorithm using quaternion wavelet transform

In order to improve the accuracy of detection of defects in banknote sorting, a new algorithm is proposed to detect cracks and scratches on banknote images. The quaternion wavelet transform and the least squares method are used for the banknote image registration. The features of the defects that are robust to gray intensity changes are constructed using edge information captured by the Kirsch operator. The banknote image is divided into several subzones of fixed size. The level of defect of the banknote image is estimated based on the defective features of each sub-zone. The experimental results show that the proposed algorithm is robust even with low quality banknote images and can obtain a high recognition rate and high stability. The proposed algorithm has already been used in a practical banknote sorting system.

Efficient Banknote Recognition Based on Selection of Discriminative Regions with One-Dimensional Visible-Light Line Sensor.

Banknote papers are automatically recognized and classified in various machines, such as vending machines, automatic teller machines (ATM), and banknote-counting machines. Previous studies on automatic classification of banknotes have been based on the optical characteristics of banknote papers. On each banknote image, there are regions more distinguishable than others in terms of banknote types, sides, and directions. However, there has been little previous research on banknote recognition that has addressed the selection of distinguishable areas. To overcome this problem, we propose a method for recognizing banknotes by selecting more discriminative regions based on similarity mapping, using images captured by a one-dimensional visible light line sensor. Experimental results with various types of banknote databases show that our proposed method outperforms previous methods.

Detection Method for Counterfeit Currency Based on Bit-Plane Slicing Technique

Counterfeiting and forging currencies is a serious threat to any economy. Even though currency exists as a variation of coins, banknotes, and electronic data, many economies remain threatened by counterfeiting which is made possible by the ongoing technological advancements in reprographic equipment available to the general public. Clearly, counterfeit currency detection is not a task that can be neglected. Digital image processing is one of the most common and effective techniques used to distinguish counterfeit banknotes from genuine ones. A new approach is presented in this paper using the bit-plane slicing technique to extract the most significant data from counterfeit banknote images with the application of an edge detector algorithm. The proposed technique consists of decomposing original images of 256 gray levels into their equivalent 8 binary images. This is useful in analyzing the relative importance contributed by each bit of the original image. Higher order bit levels are evaluated for grayscale banknote images with the application of Canny edge detection algorithm. The results are then compared with genuine banknotes and with other existing techniques used for detecting counterfeit notes. Unlike existing research, it was observed that the edges obtained using bit-plane sliced images are more accurate and can be detected faster than obtaining them from the original image without being sliced. The detection of counterfeit currency was also achieved by following the process of using Canny edge detection, image segmentation, and feature extraction.