KR101232684B1 - Method for detecting counterfeits of banknotes using Bayesian approach - Google Patents

Abstract

The present invention relates to a bill authenticity discrimination method using a Bayesian approach, and more particularly, scanning the whole bill through an infrared sensor to obtain sensor data, and specifying an area of a template in which an infrared pattern exists. Dividing the specified template region into a predetermined number of unit cells, calculating a representative value representing each unit cell by using the obtained sensor data for each of the divided unit cells, and for each calculated unit cell Extracting a feature vector of a banknote having a representative value as a factor; extracting a representative feature vector by reducing the dimension of the extracted feature banknote vector using a linear feature extraction method; Bayesian, including the step of applying authenticity to Gaussian Maximum Likelihood classification By providing a method for authenticating paper money authenticity using the approach, it is possible not only to carry out the authenticity of paper money rapidly compared to the paper money authenticity discrimination method using a conventional neural network circuit, but also to accurately verify the authenticity of paper money with a low resolution paper money image. The present invention relates to a method of discriminating banknote authenticity using a discriminating Bayesian approach.

Description

Method for detecting counterfeits of banknotes using Bayesian approach

The present invention relates to a bill authenticity discrimination method using a Bayesian approach, and more particularly, a bill number is scanned by an infrared sensor to obtain sensor data, and a predetermined number of units are specified by specifying a template region in which an infrared pattern exists. After dividing into cells, a representative value representing each unit cell is calculated using sensor data measured in each unit cell, and a feature vector of a banknote whose extracted representative value for each unit cell is a factor is extracted. After reducing the dimension of the feature vectors extracted by the linear feature extraction method, the authenticity of banknotes using the conventional neural network circuit is discriminated by using the GML (Gaussian Maximum Likelihood) classification method for the feature vectors with reduced dimensions. Compared to the discrimination method, the authenticity of bills can be performed quickly, and even with a low resolution bill image, It relates to a banknote authenticity discrimination method using a Bayesian (Bayesian) that can discriminate the authenticity of the waste approaches.

The emergence of banknotes, called flowers of printing technology, is difficult to forge compared to coins, and is widely used as a popular currency in each country because it is convenient to carry and use. However, as the performance of electronic devices such as laser beam printers and digital cameras is greatly improved, the resolution of photographs taken and output is also significantly improved, and thus related to counterfeit bills using electronic devices that can obtain sophisticated images. Crime is a serious social problem.

To this end, various advanced technologies for preventing forgery and alteration have been applied in the banknote manufacturing process, and thus each banknote has a unique geological property, a background color, a surface property, an ink property, and the like. Looking at the conventional methods for determining the authenticity of the bill using the unique characteristics of the bill, the shape and geological characteristics printed on the bill, such as microlettering technology, intaglio printing technology and watermark technology, etc. Or, by analyzing the physical properties such as surface characteristics precisely to determine the authenticity, and using special ink (ultraviolet ink, infrared ink, magnetic ink, discolorable ink, etc.) to make a bill and the characteristics of the ink used in the bill There is a method of judging the authenticity of banknotes by precisely analyzing the optical or magnetic characteristics to be detected.

However, since the authenticity of the bill is determined by the expert's manual work, it takes a lot of time to determine the authenticity of the bill. Therefore, it is not applicable to a financial teller machine (ATM). For the general public, it is difficult to judge the authenticity of paper money.

Therefore, in the automated teller machine, the counterfeit is discriminated by using the optical or magnetic characteristics of the bill. Among them, the truth discrimination method using infrared data is the representative of the bill authenticity discrimination method in the automated teller machine by developing various hardware configurations and related algorithms. Has been recognized.

The authenticity determination using the infrared data of the related art is a method of determining the presence or absence of a signal compared to a reference value after measuring the IR signal strength at a specific position in a specific bill of paper using an IR sensor having a one-dimensional array structure. However, this one-dimensional array structure method has a disadvantage in that the amount of data is limited compared to the importance of IR authenticity discrimination, and thus sensitive to skew and shift environment changes occurring during banknote conveyance.

In order to solve this problem, a method of determining authenticity by acquiring 2D IR image data in the form of a contact image sensor has been introduced. The authenticity determination method using 2D IR image data is more accurate than the one-dimensional array sensor structure, but it is difficult to accurately determine the authenticity due to the low resolution due to processing speed constraints such as the number of authenticity determinations per second. Also included is a template setting process based on prior information on IR authenticity determination factors for different types of papers. Inaccurate location and area size are set due to low resolution images, which causes processing speed and authenticity discrimination performance. .

In addition, algorithm implementation becomes an important variable when using low-resolution IR images. Machine learning techniques such as neural networks and support vector machines (SVMs), which are used as conventional hand-held recognition algorithms, use kernels of nonlinear discriminant functions between layers. Due to Kernel constraints, improvement in recognition speed is required, and when a new kind is added, it is not easy to add a priesthood because it needs to go through the learning process again.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems of the prior art. That is, an object of the present invention is to divide a template region in which an infrared pattern exists among sensor data obtained by scanning a bill with an infrared sensor into a predetermined number of unit cells, and then use the sensor data measured in each unit cell. A representative value representing each unit cell is calculated, and a feature vector of a banknote that has a calculated representative value of each unit cell is extracted, the dimension of the feature vector extracted through the linear feature extraction method is reduced, and then the dimension is decreased. By using GML (Gaussian Maximum Likelihood) classification method to classify bills authenticity, it is possible to perform bills authenticity faster than bill authenticity discrimination method using conventional neural network. This paper presents a method for discriminating banknotes authenticity using the Bayesian approach, which can discriminate the authenticity of banknotes relatively accurately even with banknote images. As it has.

As a technical idea for achieving the above object, the present invention, by scanning the entire bill through the infrared sensor to obtain the sensor data, specifying the region of the template in which the infrared pattern is present, the predetermined template region predetermined predetermined Dividing into a number of unit cells, calculating a representative value representing each unit cell by using the obtained sensor data for each of the divided unit cells, and characterizing bills using the calculated representative value for each unit cell as a factor Extracting a representative feature vector by reducing a dimension of the feature vector of the extracted banknote by using a linear feature extraction method; applying a GML (Gaussian Maximum Likelihood) classification method to the extracted representative feature vector A method of discriminating banknote authenticity using a Bayesian approach comprising the step of performing authenticity determination on banknotes to provide.

The banknote authenticity discrimination method using the Bayesian approach according to the present invention primarily templates the approximate region (or entire region) of the IR authenticity image, even if there is no actual counterfeit right information, as well as the intensity of the IR signal. The similarity of the pattern's dominant shape can be measured, and since the average value calculation is applied after the block / mesh structure is applied, high-speed recognition processing is possible, and it is applicable to a low resolution CIS image.

In addition, since it is based on a statistical normal distribution estimation technique, it is possible to quantitatively predict the recognition rate by measuring the reliability of the sample distribution estimation, and the feature vector transformation is composed only of linear operations, so that it is faster than using a nonlinear discriminant function in an existing neural network. It is possible to determine the authenticity of the bill.

In addition, unlike the machine learning method, which requires a re-learning process when adding the priesthood, it is easy to determine the authenticity of the priesthood when the priesthood is added to the device. Maintenance is easy because only the comparative reference data needs to be updated without any modification.

1 is a flow chart showing a bill authenticity discrimination method using a Bayesian approach according to an embodiment of the present invention.
2 is an example of dividing a predetermined number of unit cells for each template region obtained according to an embodiment of the present invention.
3 illustrates an example of calculating a feature vector using representative values of sensor data for each unit cell according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a flow chart illustrating a method of authenticating banknote authenticity using a Bayesian approach according to an embodiment of the present invention.

In order to discriminate the authenticity of banknotes according to the present invention, first, the sensor data obtained by scanning the whole banknote for authenticity determination is obtained using an infrared sensor, and the infrared pattern is based on the pre-positional information on the IR authenticity security element. The existing template regions are specified (S110) and the specified template regions are divided into a predetermined number of unit cells (S120).

In this case, the number of unit cells for dividing the template region may vary according to the scan area of the specified template region, and typically 2 to 20 unit cells in the horizontal direction and 2 to 10 units in the vertical direction for each template region. Splitting into cells is fine. 2 illustrates an example of dividing a predetermined number of unit cells for each template region acquired according to an exemplary embodiment of the present invention.

Subsequently, a representative value representing each unit cell is calculated using sensor data measured in each of the divided unit cells (S130), and a feature vector of a banknote having a calculated representative value for each unit cell is extracted. (S140). As shown in FIG. 3, there are a plurality of pixels respectively corresponding to the IR sensor array in one unit cell, and each pixel has sensor data measured by each corresponding IR sensor. In the present invention, the sensor data is used to calculate one scalar value (representative value) representing each unit cell. In this case, the representative value representing each unit cell is an average of sensor data constituting each unit cell, Various factors such as variance or maximum may be applied, and it is preferable to use an average value of sensor data in a unit cell that can most effectively reflect the characteristics of sensor data in each unit cell.

For example, as shown in FIG. 3, each unit cell is composed of 42 unit pixels in total, and assuming that sensor data is obtained for each pixel, each unit cell exists in the unit cell A [1,4]. 9, which is an average value of the sensor data of 42 pixels, is calculated as the representative value A _1,4 of the unit cell A [1,4], and the calculated representative value is a factor of the feature vector representing the unit cell.

That is, a representative value representing each unit cell is calculated in such a manner, and a feature vector having a representative value calculated for each unit cell is extracted, and thus, a template region divided as shown in FIG. 3. A feature vector (X) having as many factors as the number of unit cells is extracted.

When the feature vector (X) is extracted by the above-described method, the authenticity of the banknote is discriminated by comparing the extracted feature vector factors. When the feature vector factor is too large, it is difficult to carry out robust and fast authenticity discrimination. A step of reducing the dimension of the feature vector is performed (S150). The reason for reducing the dimension of the extracted feature vectors is to reduce the computational object by removing unnecessary portions of the feature vectors, and to extract only representative feature vectors that are important for authenticity determination. Accordingly, the dimension of the extracted feature vector is reduced by applying the linear feature extraction method, and only a predetermined feature vector representing the characteristic of the infrared pattern is selected.

The linear feature extraction method is a method of analyzing statistical characteristics of sensor data, and the principal component analysis (PCA) and linear discriminant analysis (LDA) are representative examples. In the present invention, it is effective to reduce the dimension of the feature vector by using the principal component analysis method (PCA), which is a statistical method that can find the image features effectively. Hereinafter, the feature vector extracted using the principal component analysis method (PCA) The steps for reducing the dimension will be briefly described.

Principal component analysis (PCA) extracts a few major factor values that can represent the feature vector from several factors constituting the extracted feature vector (X), which is composed of the major factor values As a method used to extract a representative feature vector having a reduced dimension, in the present invention, a unique matrix for extracting a representative feature vector using a covariance matrix obtained using banknote feature vector data obtained from tens of thousands of bills Calculate ^T ).

The calculation process of the eigen matrix Φ ^T using the covariance matrix described above is one of widely used statistical extraction methods, and a detailed description thereof will be omitted.

That is, by applying the eigen matrix Φ ^T calculated through the above process to the extracted feature vector (X), by reducing the dimension of the extracted feature vector (X) as shown in Equation 1 below, infrared Representative feature vectors Y that better represent the characteristics of the pattern can be obtained.

The representative feature vector (Y) calculated as described above has a relatively lower dimension than the feature vector (X) extracted initially, and at the same time more effectively represents a feature for authenticity determination of banknotes.

Subsequently, authenticity determination of the corresponding bill is performed using the Gaussian Maximum Likelihood (GML) classification method on the obtained representative feature vector (S160). That is, after calculating the probability that the representative feature vector calculated through the above-described process is included in the pre-calculated pneumococcal class region, and if the probability value is lower than the preset reference, the banknote is determined as counterfeit.

One of the characteristics of the principal component analysis described above is that for a data group distributed over the same time, a vector of directions having a large degree of variance can be obtained. In other words, by calculating the matrix of the covariance matrix, the eigenvalues and the corresponding eigenvectors can be obtained.The vector with the higher eigenvalue becomes an important element of the data group, This means that the vector is less important than the first vector. Therefore, when principal component analysis is performed using tens of thousands of pneumatic features, eigenvalues for pneumococcal and corresponding eigenvectors can be obtained. do.

In this case, a method of calculating the probability that the representative feature vector is included in the corresponding pneumatic class region may be performed by Equation 2 below.

here,

ML: Maximum Likelyhood,

i: type of winding species,

Y: representative feature vector,

: 해당 권종의 분산행렬,

= Variance matrix

: 해당 권종의 평균벡터

: Mean vector

to be.

In [Equation 2], ML is a value representing the probability that the representative feature vector is included in the pneumoconiosis class of the subject species, and according to [Equation 2], it is highly likely that the representative feature vector is included in the corresponding pneumoconiosis class. The more likely it is to be a lung disease. Therefore, the ML value to be included in the authenticity class is calculated using the representative feature vector and the authenticity class average vector and the variance matrix previously stored in the database. If the calculation is less than or equal to the preset reference value, the banknote is determined as counterfeit.

As described above, the present invention determines the authenticity of banknotes using a Bayesian approach based on a representative value for each unit cell for an area in which an infrared pattern exists, thereby determining authenticity of bills using a neural network circuit. Not only can it be performed quickly, but the authenticity of banknotes can be judged relatively accurately even with a low resolution banknote image.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and it is common in the art that various substitutions, changes, and conversions are possible without departing from the technical spirit of the present invention. It will be clear to those who have knowledge of God.

Claims (5)

In banknote authenticity discrimination method,
Scanning the whole banknote through the infrared sensor to obtain sensor data, and specifying an area of the template in which the infrared pattern exists;
Dividing the specified template region into a predetermined number of unit cells;
Calculating a representative value representing each unit cell by using the obtained sensor data for each of the divided unit cells;
Extracting a feature vector of a banknote having the calculated representative value for each unit cell as a factor;
Extracting a representative feature vector by reducing the dimension of the feature vector of the extracted banknote by using a linear feature extraction method;
Performing authenticity determination on the banknote by applying a Gaussian Maximum Likelihood (GML) classification method to the extracted representative feature vector;
Banknote authenticity discrimination method using a Bayesian approach that is configured to include.
The method of claim 1,
In dividing the specified template region into unit cells,
According to the scan area of the specified template region, banknote authenticity discrimination using a Bayesian approach characterized by dividing each template region into 2 to 20 unit cells in the horizontal direction and 2 to 10 unit cells in the vertical direction. Way.
The method of claim 1,
Calculating a representative value representing each unit cell,
A method of discriminating banknote authenticity using a Bayesian approach, wherein one of average, variance, or maximum values of sensor data obtained in each unit cell is calculated as a representative value.
The method of claim 1,
The linear feature extraction method applied to the step of extracting the representative feature vector by reducing the dimension of the extracted feature vector,
Banknote authenticity discrimination method using the Bayesian approach, characterized in that the principal component analysis (Principal Component Analysis).
The method of claim 1,
Applying the GML classification method to the extracted representative feature vector to perform the authenticity discrimination for the bill,
A method of discriminating banknote authenticity using a Bayesian approach, characterized in that the counterfeit is determined as counterfeiting when the result value (ML) of the calculation formula calculated using Equation 3 below is less than a predetermined reference value.
&Quot; (3) "
,
here,
ML: Maximum Likelyhood,
i: type of winding species,
Y: representative feature vector,

= Variance matrix

: The mean vector of the volume.

Images