WO2017183867A1

WO2017183867A1 - Forged-fingerprint distinguishing apparatus and method capable of distinguishing forged fingerprint on basis of slight change in brightness of fingerprint image, caused by living body's heartbeat

Info

Publication number: WO2017183867A1
Application number: PCT/KR2017/004093
Authority: WO
Inventors: 백영현; 신요식
Original assignee: 주식회사 유니온커뮤니티
Priority date: 2016-04-18
Filing date: 2017-04-17
Publication date: 2017-10-26
Also published as: KR20170119245A; KR101798395B1

Abstract

Disclosed is a forged-fingerprint distinguishing apparatus and method capable of distinguishing a forged fingerprint on the basis of a slight change in the brightness of a fingerprint image, caused by a living body's heartbeat. On the basis of a slight change in a biometric fingerprint, caused by a living body's heartbeat, the forged-fingerprint distinguishing apparatus according to the present invention can determine a fingerprint not having such a change as a forged fingerprint. To this end, at least two fingerprint images are acquired from a relevant fingerprint, the difference between the brightness values of each two pixels at identical coordinates in the two fingerprint images is extracted, the number of pixels having the same difference is derived in the form of a cumulative histogram, and then, the number of peak points in the corresponding cumulative histogram is calculated to determine whether a fingerprint is a forged fingerprint.

Description

Counterfeit fingerprint discrimination apparatus and method for discriminating forged fingerprint based on minute change of brightness of fingerprint image according to heartbeat of living body

The present invention relates to a method for determining a forgery fingerprint in the optical fingerprint image acquisition method, and forgery that judges a fingerprint without such a change as a forgery fingerprint based on a minute change in a biometric fingerprint caused by a heartbeat of a living body. A fingerprint identification device and a method thereof are provided.

The use of user's biometric information with invariability and uniqueness for personal authentication using information devices has already been generalized. Among them, fingerprint recognition has a very simple structure, and its performance is very excellent. It is becoming.

Normal personal authentication is mainly used in areas where security is important, such as access control, e-commerce, financial transactions, security of personal computers (PCs), and office payment systems. The most important thing is to effectively distinguish fingerprints (hereinafter referred to as 'false fingerprints').

Disclosure of Invention An object of the present invention is to provide a counterfeit fingerprint discrimination apparatus and method for determining a fingerprint without such a change as a fake fingerprint based on a minute change in a biometric fingerprint generated according to the heartbeat of a living body.

In accordance with another aspect of the present invention, there is provided a method for determining a fake fingerprint, the method including: obtaining first and second fingerprint images within a unit time from a fingerprint in contact with an optical fingerprint sensor unit; Calculating a difference in brightness between pixels of the same coordinates of the first and second fingerprint images, and obtaining a cumulative histogram of the number of pixels having the same difference; And determining the fingerprint as a biofingerprint when there are at least two peak points in the cumulative histogram.

According to an embodiment, the step of obtaining a cumulative histogram may include extracting an object region having a fingerprint region from the first and second fingerprint images; And extracting only LL frequency components by performing discrete wavelet transformation on the object region to obtain a first LL fingerprint image and a second LL fingerprint image. In this case, the difference in brightness between the pixels is calculated based on the first and second LL fingerprint images.

According to another exemplary embodiment, a method of determining a fake fingerprint includes: obtaining an average brightness value of each of the first and second fingerprint images; And performing first-order filtering to determine the fingerprint as a fake fingerprint when the difference between the average brightness values is smaller than the first reference value, so that there is no change in the brightness value without using the above method. Fingerprints can be determined. Naturally, the step of obtaining the cumulative histogram is not performed on the first filtered fingerprint.

Counterfeit fingerprint discrimination apparatus according to another embodiment of the present invention includes a fingerprint sensor unit and the color distribution analysis unit. The fingerprint sensor unit includes a photorefractor to obtain first and second fingerprint images within a unit time from a fingerprint optically contacting the fingerprint contact surface of the photorefractor. The color distribution analyzer calculates a difference of brightness values between pixels of the same coordinates of the first and second fingerprint images, and calculates the number of pixels having the same difference as a cumulative histogram. Subsequently, the color distribution analyzer determines the fingerprint as a bioprint when the cumulative histogram has at least two peak points, and determines that the fingerprint is a fake fingerprint when the peak point is less than two.

According to an embodiment, the counterfeit fingerprint discrimination apparatus may extract an object region having a fingerprint region from the first and second fingerprint images, and perform discrete wavelet transform on the object region to extract only LL frequency components. The apparatus may further include a wavelet converter configured to obtain a first LL fingerprint image and a second LL fingerprint image. In this case, the color distribution analyzer calculates a difference in brightness values between the pixels based on the first and second LL fingerprint images.

According to another embodiment, the fake fingerprint recognition device may further include a color information extraction unit for the primary filtering. The color information extracting unit obtains an average brightness value of each of the first and second fingerprint images, and performs the first filtering to determine the fingerprint as a fake fingerprint when the difference between the average brightness values is smaller than the first reference value. Naturally, the color distribution analyzer does not obtain the cumulative histogram of the first filtered fingerprint.

According to an embodiment, the light source of the fingerprint sensor unit used for acquiring the first and second fingerprint images may use a visible light region of 500 to 700 nm band or an infrared region of 800 to 900 nm band.

The forgery fingerprint discrimination apparatus according to the present invention may generate a fingerprint image and determine whether the fingerprint is a bio fingerprint or a fake fingerprint by using a change in the average brightness value and the brightness value per pixel generated within the unit time of the images.

1 is a block diagram of a fake fingerprint discrimination apparatus of the present invention,

2 is a diagram showing an example of a color fingerprint image obtained from a bio fingerprint and a fake fingerprint; and

3 is a block diagram of a neural network learning system for discriminating forged fingerprints according to the present invention;

4 is a diagram illustrating an example of performing wavelet transformation on a fingerprint image;

5 is a diagram showing an example of a cumulative histogram for a bio fingerprint; and

6 is a diagram showing an example of a cumulative histogram for a fake fingerprint.

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

Referring to FIG. 1, the counterfeit fingerprint discrimination apparatus 100 of the present invention includes a fingerprint sensor unit 110, a biopsy unit 130, a feature point extractor 150, and a fingerprint authentication unit 170.

The fingerprint sensor unit 110 includes an optical refractor 111, a light source 113, a lens 115, and an image sensor 117, and the fingerprint contact surface 111a of the optical refractor 111 by an optical fingerprint authentication method. Obtain a fingerprint image of the fingerprint in contact with). Any method of generating an optical fingerprint image may be applied, including a scattering method or an absorption method known as a method of optically obtaining a fingerprint.

The photorefractor 111 is usually a triangular or trapezoidal prism in the shape of its cross-section, but can replace a prism of a broad concept of the optical refractor. The optical refractor 111 has a fingerprint contact surface 111a through which the fingerprint is in contact, an emission surface 111b through which light (fingerprint image) reflected or scattered from the fingerprint contact surface 111a is emitted, and the light source 113 therein. The incident surface 111c on which the emitted light is incident is provided.

According to an embodiment, the light source 113 preferably uses a visible light region of 500 to 700 nm band or an infrared region of 800 to 900 nm band.

First, the basic fingerprint image acquisition process of the fingerprint sensor 110 is as follows. When the user contacts the fingerprint with the fingerprint contact surface 111a, the light irradiated from the light source 113 passes through the incident surface 111c, the fingerprint contact surface 111a, and the exit surface 111b of the optical refractor 111. It is imaged at 115 and is input to the image sensor 117. In the scattering fingerprint sensor unit 110 as shown in FIG. 1, the light emitted from the light source 113 is incident on the fingerprint contact surface 111a at an angle smaller than a critical angle for right angle or total reflection. The light emitted from the light source 113 passes or scatters along the valleys and ridges of the fingerprint in contact with the fingerprint contact surface 111a to form a color fingerprint image. The image sensor 117 outputs a digital fingerprint image signal, which is an electrical signal corresponding to the incident fingerprint image, to obtain a color image of the fingerprint in contact with the fingerprint contact surface 111a. Therefore, the fingerprint image generated by the fingerprint sensor 110 becomes a color fingerprint image.

The biopsy unit 130 receives at least two first and second fingerprint images from the fingerprint sensor unit 110 and then prints the biometric fingerprint on the fingerprint contact surface 111a based on the color change of the fingerprint image. Determines whether fingerprints or fake fingerprints. At this time, the color change is to extract a small change in brightness according to the heartbeat of the living body. During the systolic and diastolic phases of the heartbeat, the blood flow in the finger or fingerprint changes, and this change in blood flow is represented by the change in brightness of the fingerprint image. In the systolic phase, the blood fills the finger finely, and in the diastolic phase, the blood escapes and becomes slightly brighter.

Forgery fingerprints of paper / rubber / film material have almost no such brightness change, and such brightness change is very small even in silicon having a color information similar to a living body. By using this feature the present invention is to determine whether or not the forgery fingerprint.

The biological determination unit 130 performs this process using an image processing technique such as a wavelet, and for this purpose, the object extractor 201, the color information extractor 203, the wavelet transform unit 205, and the color distribution are performed. The analysis unit 207 is included. The biodetermination unit 130 may perform a two-step forgery fingerprint determination process of primary filtering and secondary filtering. These operations will be described again below.

In terms of the determination of whether the forgery fingerprint, the feature point extraction unit 150 and the fingerprint authentication unit 170 is not an essential configuration of the present invention. However, most of the devices for acquiring the fingerprint image perform a fingerprint registration or fingerprint authentication process, and accordingly, the feature point extracting unit 150 and the fingerprint authentication unit 170 are provided.

If the current fingerprint is determined to be a bio fingerprint by the biopsy unit 130, the feature extraction unit 150 extracts the feature point from the original fingerprint image acquired by the fingerprint sensor unit 110, and the fingerprint authentication unit 170 extracts the feature point. The extraction unit 150 performs the registration process of the corresponding fingerprint by using the extracted feature points or determines whether or not it matches the pre-registered fingerprint. Feature point extraction or fingerprint registration / authentication can use a conventionally known method.

Hereinafter, with reference to Figure 3, the forgery fingerprint determination process of the present invention will be described.

The fingerprint sensor 110 obtains the first and second fingerprint images from the fingerprint in contact with the fingerprint contact surface 111a. Two first fingerprint images and a second fingerprint image are sufficient for the image processing of the present invention. The fingerprint sensor 110 may acquire a fingerprint image of 20 frames or more per second, but selects a fingerprint image from which all of the fingerprints are obtained, and uses the fingerprint image as the first and second fingerprint images. Accordingly, the first and second fingerprint images become fingerprint images acquired from the same fingerprint within a predetermined unit time.

<First-order filtering: S303 to S305>

First filtering is performed using the two fingerprint images acquired in step S301. In the primary filtering, a fake fingerprint is determined by determining that the difference between the average brightness values of the fingerprint regions (or the object regions) of the two fingerprint images is equal to or less than the first reference value. For example, in the case of a fake fingerprint made of a material such as paper, rubber or film, even if there is a parallax between the first and second fingerprint images, there is almost no difference in the color information, and similarly, the difference in the average brightness value of the fingerprint area. There is almost no. Therefore, the first reference value may be set to a very low value and may be determined by an experimental value. However, the primary filtering process is not an essential process of the present invention.

For primary filtering, the object extractor 201 first extracts an object region from each fingerprint image. The object area refers to an area in which the background is excluded, that is, an area corresponding to the fingerprint in the fingerprint image. As the image processing method for extracting the object region, a conventionally known method may be used as it is. The first object image is extracted from the first fingerprint image, and the second object image is extracted from the second fingerprint image.

Thereafter, the color information extracting unit 203 calculates average brightness values of the first and second object images. When the average brightness value is obtained for the entire first and second fingerprint images, the difference in the average brightness value obtained in the next step S305 is relatively small for the person having a small fingerprint size (in the case of non-inverted image, it becomes large). Significant error may occur in a simple comparison with the first reference value. Therefore, the color information extractor 203 extracts only average brightness values of the first and second object images, that is, the fingerprint region, obtained by the object extractor 201.

The color information extracting unit 203 obtains a difference between average brightness values of the first and second object images, and compares the difference with the first reference value. In the case of a biometric fingerprint, there is a slight change in the color (brightness) of the fingerprint due to blood flow in the fingerprint, which varies in the systolic and diastolic phases of the heartbeat, and the change appears as a difference in the average brightness value. However, in the case of forged fingerprints made of materials such as paper, rubber and film, the difference in average brightness is almost zero.

The color information extracting unit 203 determines that the difference in the average brightness value is smaller than the first reference value as a fake fingerprint. However, in a simple comparison such as first order filtering, a difference in average brightness occurs in a fake fingerprint of a silicon or gelatin material, and it is difficult to distinguish the difference from the average brightness value in a biometric fingerprint. Therefore, second order filtering is performed.

Second order filtering: S307 to S313

The wavelet transform unit 205 performs discrete wavelet transform on each of the first and second object images extracted by the object extractor 201, and as a result, the first LL including only the components of the low frequency band LL. The fingerprint image and the second LL fingerprint image are extracted.

In image processing, wavelet transform is performed by applying a pair of filters (high frequency filter and low frequency filter) to an image and separating it into a low frequency band and a high frequency band. Each band is subsampled with an element of 2, so it contains n / 2 samples. Applying a lowpass filter and a highpass filter to each row of the two-dimensional image and downsampling to two produces four sub-images LL, LH, HL, and HH. Recombining these four subband images again produces the original original image. For example, FIG. 4 is a diagram illustrating an example of performing one-step wavelet transformation on a first object image in horizontal and vertical directions.

Figure 4 (a) is the LL frequency band image is a subsampled to 2 by applying a low pass filter in the horizontal and vertical direction to the original image, (b) is a LH frequency band image is applied to the high pass filter in the horizontal direction It includes the error component of the horizontal frequency. (c) is the HL frequency band image, which is a high pass filter applied in the vertical direction, and includes an error component of the frequency in the vertical direction, and (d) is a HH frequency band image, which is a high pass filter applied in the horizontal and vertical directions.

Through wavelet transformation, high frequency components may be removed from a fingerprint image, and only low frequency (LL band) regions may be extracted. As in (d), most of the high frequency region is close to noise, and as in (a), most of the fingerprint information is included in the low frequency region, so the characteristics of the fingerprint can be read as it is in the low frequency region LL alone. Instead, in the process of removing the high frequency region, the size of the fingerprint image is down sampled (eg, 320 × 240 size → 160 × 120 size) to 1/4 to speed up subsequent image processing. Improve. Of course, when the feature point is extracted from the fingerprint image and the fingerprint authentication is performed, the first fingerprint image or the second fingerprint image as the original is used.

The wavelet transform unit 205 discards the remaining region for second order filtering and extracts the first LL fingerprint image and the second LL fingerprint image of the LL frequency band.

The color distribution analyzer 207 uses the following equation 1 to calculate the difference between the brightness values of each pixel for the two LL fingerprint images extracted by the wavelet converter 205 from the first and second object images, respectively. Difference) is obtained as an absolute value.

Here, x _p1 (n) is the brightness value of the n-th pixel of the first LL image, and x _p2 (n) is the brightness value of the n-th pixel of the second LL image. y (n) is the absolute value of the difference in brightness values between the n th two pixels of the first and second LL fingerprint images.

The color distribution analyzer 207 obtains a cumulative histogram as shown in FIG. 5 or 6 by obtaining y (n) of all pixels of the first and second LL fingerprint images. The cumulative histogram is a cumulative number of pixels having the same difference. For example, FIG. 5 is an example of a cumulative histogram obtained from a bio fingerprint, and FIG. 6 is an example of a cumulative histogram obtained from a fake fingerprint such as silicon. 5 and 6, the horizontal axis represents difference, that is, y, and the vertical axis represents the number of pixels.

When two or more peak points are identified in the cumulative histogram obtained in S311, the color distribution analyzer 207 determines the fingerprint as a bio fingerprint, and when only less than two peak points are identified, the fingerprint is a fake fingerprint. Judging by.

For example, referring to FIG. 5, three peak points are shown. The point at each peak point (y = 54, 95, 211) is a large number of pixels with that much brightness difference (y = 54, 95, 211), indicating that the brightness values of the two LL fingerprint images are changing. As a result, there is a slight difference in brightness value between the two fingerprint images.

Correspondingly, one peak point is shown in Fig. 6, and it can be seen that the number of pixels at the peak point is considerably small. As a result, there is almost no difference in the brightness values of the two LL fingerprint images, and thus, the brightness values of the two fingerprint images are almost the same.

As such, if the analysis is performed through steps S307 to S311, it may be determined whether there is a slight difference in brightness values between the two fingerprint images, and whether or not a fake fingerprint is determined using the point that the difference is confirmed only in the bio fingerprint.

Although the above has been illustrated and described with respect to preferred embodiments of the present invention, the present invention is not limited to the above-described specific embodiments, it is usually in the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

Claims

Acquiring first and second fingerprint images within a unit time from a fingerprint in contact with the optical fingerprint sensor unit;

Calculating a difference in brightness between pixels of the same coordinates of the first and second fingerprint images, and obtaining a cumulative histogram of the number of pixels having the same difference; And

And determining the fingerprint as a biometric fingerprint when at least two peak points are included in the cumulative histogram.
The method of claim 1,

The steps to find the cumulative histogram are

Extracting an object region having a fingerprint region from the first and second fingerprint images; And

Performing discrete wavelet transform on the object region to extract only LL frequency components to obtain a first LL fingerprint image and a second LL fingerprint image,

And calculating the difference in brightness between the pixels based on the first and second LL fingerprint images.
The method of claim 1,

Obtaining an average brightness value of each of the first and second fingerprint images; And

Performing primary filtering to determine the fingerprint as a fake fingerprint when the difference in the average brightness value is smaller than a first reference value;

The method of claim 1, wherein the step of obtaining the cumulative histogram is not performed on the first filtered fingerprint.
A fingerprint sensor unit having a photorefractor to acquire first and second fingerprint images within a unit time from a fingerprint optically contacting the fingerprint contact surface of the photorefractor; And

A color distribution analyzer configured to calculate a difference in brightness between pixels of the same coordinates of the first and second fingerprint images, and obtain a cumulative histogram of the number of pixels having the same difference;

And the color distribution analyzer determines the fingerprint as a bioprint when the cumulative histogram has at least two peak points.
The method of claim 4, wherein

An object extracting unit extracting an object region having a fingerprint region from the first and second fingerprint images; And

The apparatus may further include a wavelet transform unit configured to perform discrete wavelet transform on the object region to extract only LL frequency components to obtain a first LL fingerprint image and a second LL fingerprint image.

And the color distribution analyzer calculates a difference in brightness between the pixels based on the first and second LL fingerprint images.
The method of claim 4, wherein

The apparatus further includes a color information extracting unit configured to obtain an average brightness value of each of the first and second fingerprint images, and to perform primary filtering to determine the fingerprint as a fake fingerprint when a difference between the average brightness values is smaller than a first reference value. and,

And the color distribution analyzer does not obtain the first filtered fingerprint as the cumulative histogram.