KR20040057210A - Apparatus and Method for Recognizing Hologram Fingerprint - Google Patents

Abstract

PURPOSE: A device and a method for recognizing a holographic fingerprint are provided to discriminate the person by comparing the holographic fingerprint information played from a hologram card with the actual fingerprint information of the user through an optical mode. CONSTITUTION: The hologram card(140) records a fingerprint of a user as a hologram(142). A holographic fingerprint obtaining tool(114) obtains an image of the holographic fingerprint from a hologram on the hologram card. An actual fingerprint obtaining tool(126) obtains the image of the actual fingerprint from the user. A fingerprint recognizer(170) certifies the identity between the holographic fingerprint and the actual fingerprint when a calculated initial correlation value is larger than a reference initial correlation value, and the similarity between the first and the second spatial frequency value is large.

Description

Apparatus and Method for Recognizing Hologram Fingerprint}

The present invention relates to a fingerprint recognition device and method, and more particularly, a fingerprint recognition device for determining mutual identity by comparing a holographic fingerprint image obtained from a hologram card with an actual fingerprint image of a user of the hologram card by using an optical method. And to a method.

As the informatization of society develops, rapid informatization is taking place in the financial related business area, and most of the financial settlement is made by personal credit card. In addition, as the Internet becomes more popular, an electronic financial settlement system that orders and pays products on the Internet that meets consumer needs is emerging in all fields.

Currently, personal credit cards and corporate credit cards are used for payment through a card payment system using a card reader. In addition, the government has continuously promoted policies to promote the use of credit cards. With the introduction of electronic payment systems such as electronic money and electronic cards, transactions using credit cards are rapidly increasing not only offline but also online.

As such, the rapid increase in the use of credit cards has severely affected both personal and related financial companies due to the loss of cards, and the economic damage to the country from counterfeiting and piracy of the cards is already significant. Is reaching. As a card company, there are many risk factors associated with credit card services, so many efforts have been made to prepare a countermeasure realistically, but faced with technical difficulties, and therefore, studies are being made to solve such problems at home and abroad. In order to solve this problem, a method of reducing the risk of loss, forgery, forgery, and illegal copying when an individual or a corporation uses a credit card should be considered as a top priority, and an institutional solution should be combined. The solution must be preempted.

An important consideration in this technical approach is to make it easy for individuals to use their credit cards, to ensure that they can be used with peace of mind, and to protect themselves from theft or loss of credit cards. There is an urgent need for a technical approach to implementing the security features of credit cards so that they cannot be used. In particular, in order for individuals to use the electronic payment system securely online without the risk of exposing credit information, a technical solution that can solve the problem using information technology is needed.

Currently, credit cards generally used recognize a user by reading personal information recorded on the magnetic tape by a card reader and comparing it with a password entered by the user.

Identifying the owner of a card on a credit card includes fingerprints, handwritten signatures, seals, and passwords. Therefore, if you print and use fingerprints, handwritten signatures, stamps, and passwords on your card, the card reader recognizes the fingerprints, handwritten signatures, stamps, and passwords printed on the card, and the user's actual fingerprints, handwritten signatures, stamps, and secrets. How to identify yourself compared to numbers is a commonly used technique. In particular, recently, many studies have been conducted on a fingerprint recognition card system that accurately identifies a user using a fingerprint and thus increases security, and various kinds of products have been introduced.

However, the problem with the conventional fingerprint recognition method is that in order to find the feature points representing the features of the fingerprint and to reduce the error rate of the fingerprint recognition, it is necessary to database and store as many data representing the features of the fingerprint as possible.

Therefore, in order to store such a large amount of data, an additional process such as data compression is required, and there is a problem in applying to other physical storage media having capacity limitations.

Therefore, in order to solve the above problem, a fingerprint recognition system capable of identifying a fingerprint image representing a feature of a fingerprint by using optical correlation is needed.

Therefore, it is an object of the present invention to provide a fingerprint recognition device and method for identifying that the user's fingerprint information is the same person by optically comparing the hologram fingerprint information reproduced from the hologram card recorded with the hologram with the user's actual fingerprint information in an optical manner. The purpose.

1 is a block diagram of a holographic fingerprint recognition device according to the present invention,

2 is a system configuration diagram of an authentication algorithm by comparing a playback image with an actual fingerprint acquisition image;

3 is a flowchart of an algorithm according to an embodiment of the present invention;

4 is a conceptual diagram for applying a grid method to a cropped image.

<Explanation of symbols for the main parts of the drawings>

100: frame 102: card insertion slot

110: first light source 112: first imaging lens

114: first image sensor 120: second light source

122: prism 124: second imaging lens

126: second image sensor 150: first image processor

160: second image processing unit 170: fingerprint recognition unit

310: branch point 320: disadvantage

330 center point

Fingerprint recognition apparatus of the present invention for achieving the above object is a hologram card in which the fingerprint of the user is recorded as a hologram; Hologram fingerprint acquisition means for acquiring an image of the hologram fingerprint from the hologram on the hologram card; Actual fingerprint acquisition means for acquiring an image of the actual fingerprint of the user; A center point of the holographic fingerprint image acquired by the holographic fingerprint acquisition means and the actual fingerprint acquisition means and the image of the image of the actual fingerprint, and the first center of the holographic fingerprint including the main information of the fingerprint based on the center point A nonlinear Joint Transform Correlation (NJTC) is applied to an image and a second center image of the real fingerprint, and a correlation peak value is calculated, and the fingerprint grid on the closed curve surrounding the first center image is calculated. The first spatial frequency distribution is generated by applying diffraction of the Fraunhofer at the intervals and positions, and the second spatial frequency distribution is applied by diffraction of the Fraunhofer at the intervals and positions of the fingerprint grating on the closed curve surrounding the outside of the second central image. Generating the generated correlation peak value greater than a predetermined reference correlation peak value; It characterized in that it comprises a first group and the second spatial frequency value when a similarity degree of the spatial frequency is greater the hologram fingerprint and fingerprint to authenticate the identity of the real parts of the fingerprint.

According to the present invention for achieving the above object, a fingerprint authentication method for determining the mutual identity of the hologram fingerprint recorded as a hologram on the hologram card and the actual fingerprint of the user of the hologram card, (a) the image of the hologram fingerprint Obtaining an image of the actual fingerprint; (b) finding a center point of the image of the holographic fingerprint and the image of the actual fingerprint by using the gradient of the image of the holographic fingerprint and the image of the actual fingerprint; (c) selecting first and second center block images of predetermined sizes based on the center points of the holographic fingerprint image and the actual fingerprint image; (d) calculating a correlation peak value by applying a nonlinear joint transform transformation (NJTC) to the first center block image and the second center block image; (e) extracting a first frequency component according to the distance and position of the fingerprint grid on the closed curve surrounding the outer edge of the first center block image, and the distance of the fingerprint grid on the closed curve surrounding the outer edge of the second central block image; Extracting a second frequency component according to a position; And (f) a case where the correlation peak value calculated in step (d) is larger than a predetermined reference correlation peak value, and the similarity between the first spatial frequency value and the second spatial frequency value obtained in step (e). Authenticating identity of the holographic fingerprint and the actual fingerprint when large.

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

1 is a block diagram of a fingerprint recognition system according to the present invention. The fingerprint recognition system of the present invention includes a frame 100, a card insertion slot 102, a first light source 110, a first imaging lens 112, a first image sensor 114, a second light source 120, and a prism. And 122, a second imaging lens 124, a second image sensor 126, a first image processor 150, a second image processor 160, and a fingerprint recognition unit 170.

The card insertion slot 102 is composed of a long groove into which the hologram card 140 is inserted, as in a conventional card reader. The hologram card 140 includes a hologram 142 in which fingerprint information for identifying a user is recorded. Attached. In addition, as will be described in detail below, a reference value to be compared with fingerprint information is printed on the hologram 142 or in the margin portion of the hologram 142.

In general, a narrow band laser such as a semiconductor laser or a light emitting diode is used as the first light source 110. The first light source 110 irradiates a portion of the hologram 142 on the hologram card 140 inserted into the card insertion slot 102 of the frame 100 to reproduce the hologram fingerprint 118. The hologram fingerprint 118 thus reproduced is projected onto the space as shown, and is imaged on the first imaging lens 112.

Typically, holograms appear as stereoscopic images when viewed with the naked eye, which corresponds to the virtual image of the hologram. Generally, when recording a hologram and reproducing the recorded hologram, it can be recognized as a virtual image and a real image of the originally recorded image. For example, in the hologram card 140, the virtual image refers to an image seen when looking at the hologram 142 attached to the hologram card 140, and in fact, the virtual image is attached to the hologram card 140 as in the present invention. The hologram 142 refers to an image irradiated with light and reflected therefrom to form an image in space.

The first image sensor 114 picks up the actual holographic fingerprint 118 formed on the first imaging lens 112. The first image sensor 114 is a CCD or CMOS sensor for outputting an electrical signal corresponding to the amount of light input. Therefore, the amount of light incident on the first image sensor 114 is different depending on the ridges and valleys of the hologram fingerprint 118, and thus, the first image sensor 114 is different from each other according to the pattern of the hologram fingerprint. It will output different levels of electrical signals. The hologram fingerprint 118 obtained by the first image sensor 114 is provided to the first image processor 150 in the form of image information.

Meanwhile, like the first light source 110, the second light source 120 uses a narrow band laser such as a semiconductor laser or a light emitting diode, and serves to irradiate light to the prism 122.

Prism 122 is a right triangle and includes a triangular prism that causes total reflection inside the prism when there is no input of a fingerprint. The inclined surface of the prism 122 is used as a fingerprint input window 132 that receives a user's actual fingerprint to confirm that the hologram card 140 is the same person as the user who inserted the hologram card 140 into the card insertion slot 102. The inner surface of 132 is a total reflection surface causing total reflection. For example, the user inputs the user's fingerprint information by placing his finger on the fingerprint input window 132.

In the state where the fingerprint is not applied to the fingerprint input window 132, the light from the second light source 120 is totally reflected inside the prism 122 to the second image sensor 126 through the second imaging lens 124. Incident. When the finger is placed on the fingerprint input window 132, the light is totally reflected from the inner surface of the fingerprint input window 132 to the second image sensor 126 because the bone of the fingerprint is finely spaced apart from the surface of the fingerprint input window 132. Since the ridge of the fingerprint is in close contact with the fingerprint input window 132, the light is partially absorbed without total reflection on the inner surface of the fingerprint input window 132 so that the remaining part of the light reaches the second image sensor 126. .

Like the first image sensor 114, the second image sensor 126 may be a CCD or a CMOS sensor. Therefore, the amount of light incident on the second image sensor 126 outputs electrical signals having different levels according to the valleys and ridges of the fingerprint. The actual fingerprint 128 acquired by the second image sensor 126 is provided to the second image processor 160 in the form of image information.

Each of the first and second image processing units 150 and 160 may format a hologram fingerprint image and an actual fingerprint image output from the first and second image sensors 114 and 126 into digital signals.

As described above, in the case of the optical fingerprint recognition module including the prism 122, the second imaging lens 124, and the second image sensor 126, which receives the user's actual fingerprint, the fingerprint of the finger is displayed on the fingerprint input window 132. As the contact is made, the fingerprint input window 132 leaves residual fingerprints due to sebum or contaminants. The optical fingerprint recognition module should be able to obtain a clear raw fingerprint image without distortion, without being affected by the potential fingerprint of the previous user, but when using the prism 122 between the fingerprint and the second imaging lens 124 Trapezoidal Distortion is generated by the optical path difference, and quantization noise is generated in the process of converting an analog image into a digital image by a potential fingerprint of a previous user. Similarly, dust or contaminants may be deposited on the hologram 142 on the hologram card 140 to generate noise light, or quantization noise may be generated during the conversion into a digital signal.

Therefore, the first image processor 150 and the second image processor 160 respectively generate the fingerprint image using the spatial domain filtering, the frequency domain filtering, or the wavelet transform on the holographic fingerprint image and the actual fingerprint image. Eliminate noise components.

In addition, the first image processor 150 and the second image processor 160 perform a binarization process for extracting ridges and valleys of the image of the hologram fingerprint from which the noise is removed and the image of the actual fingerprint. This binarization process divides the image of the hologram fingerprint and the image of the actual fingerprint into blocks of N × N size, calculates an average of pixels in the divided N × N block image, and applies a hologram by applying a predetermined threshold value. The binarization process is performed by extracting the features of the ridges and valleys of the fingerprint and the actual fingerprint. As a result of the binarization process, the ridges and the valleys representing the characteristics of the fingerprint can be clearly separated, and the reliability of the calculation result generated in the next image comparison process can be improved.

The fingerprint recognition unit 170 may be implemented as a conventional microcomputer having a microprocessor, a basic memory, a basic peripheral device, and the like, and executes a fingerprint recognition program according to the present invention. The fingerprint recognition program executed by the fingerprint recognition unit 170 may include the hologram fingerprint 118 obtained from the hologram 142 attached to the hologram card 140 and the actual fingerprint of the user acquired from the fingerprint information input window 132. Comparing and judging (128) by applying a nonlinear joint transform transformation correlation (NJTC) method and a lattice method to identify the same person.

The operation of the fingerprint recognition unit 170 according to the present invention will be described below with reference to the flowchart of FIG. 2 and FIGS. 3A to 3D illustrating the configuration of a fingerprint of a user.

As illustrated in FIG. 3A, the structure of the ridges and valleys of the fingerprints provided from the first and second image processing units 150 and 160 is a major source of information obtained from the fingerprint and is the most important element in fingerprint recognition. Can be. The information that can be obtained from the ridges and valleys of the fingerprint includes a directional field representing the basic shape of the fingerprint and a minutiae that provides more detailed information of the fingerprint. The directional field is defined as the Local Orientation of the ridge-bone structure and is usually used for classification of fingerprints. In addition, the feature point provides detailed information of the ridge-bone structure, and includes a branch point 310, a disadvantage 320, and a center point 330, and is mainly used in a process of comparing fingerprints.

The basic concept for determining the center point 330 of the fingerprint is to calculate the directional field of the ridge-bone structure to find the most curved point. Several methods of calculating the directional field in a fingerprint are known. In the present invention, a gradient method of calculating the directional field of the fingerprint to find the center point 330 of the fingerprint is used.

The slope can be best described in the figure representing the contour line, where the direction of the gradient vector is the direction in which the slope changes most drastically, and the magnitude is measured by the angle of inclination. The inclination in the fingerprint image may be regarded as elementary orientation in each pixel of the image. However, since the directional field represents the directionality of the ridge-valve structure, which is a much larger scale compared to the gradient, which is the directional gradient of the pixel scale, the directional field should be obtained by averaging the gradient including the adjacent adjacent pixels. When the average of these gradient vectors is calculated, the angle of the gradient vectors is doubled, the magnitudes are squared, and the average is obtained by averaging squared gradients so that the opposite vectors are not canceled. This is determined.

The process performed to determine the center point of the fingerprint is described in detail as follows.

First, as illustrated in FIG. 3B, a block having an N × N pixel size (N is a positive integer) for the holographic fingerprint image from which noise components are removed and the actual fingerprint image in the first and second image processing units 150 and 160. For example, in order to obtain a directional field for finding a center point for a binarized image divided into 10 x 10 pixel blocks, the gradient of the fingerprint in each block, that is, The slope of the direction Wow The slope of the direction Is calculated (step S202). The slope is defined as in Equation 1.

Thus obtained from the rectangular coordinate system , Converts to polar coordinates and back to Cartesian coordinate system. The value obtained through this process may be expressed by Equation 2 if the angle of the gradient vector is doubled and the magnitude is squared to prevent the effect of canceling the vectors in opposite directions.

In Equation 2, , Are the slopes for each axis, each squared and doubled in angle.

Then, the average of 10 × 10 pixel blocks , Is calculated as in Equation 3 below (step S204).

In Equation 3, i and j each represent a pixel in a block. Wow Are respectively gradient vectors Wow Indicates the variance of Is Wow Means covariance. The variance and covariance of the gradient vector are shown in Equation 4 below.

Next, the average Gradient Direction ( ) Is calculated as in the following equation (5) (step S206).

here, ego, Average ridge and valley direction perpendicular to ) Is expressed by the following equation (step S208).

In FIG. 3C, the thick solid line indicates the direction of the ridges and valleys.

The next step is to determine the center point 330 of the fingerprint, within each block. Value from 0 Going down the -y direction only along the slope in the range of / 2, When a block is encountered that is not in the / 2 range, it stops searching and marks the block (see Fig. 3c). Similarly, within each block The value of from 0 to- Climb in the + y direction along the slope, which is in the range of / 2, from 0 to- If a block is encountered that is not in the / 2 range, it stops searching and marks the block. By accumulating in this way, the block marked with the most marks becomes a block corresponding to the center of the fingerprint, and the x and y coordinates of the block are determined as the center point 330 of the fingerprint (step S210).

Next, the hologram image and the actual fingerprint image are cut into M × M blocks, for example, 100 × 100 based on the center point 330 of the fingerprint determined as described above (step S212). The criteria for performing this process is that most of the information of the fingerprint is included in the range of a certain pixel of 100 × 100 size based on the center of the fingerprint when the criteria of the external classification of the fingerprint are considered. 3D illustrates a truncated fingerprint image block.

Each cut hologram image extracted as described above is compared with an actual fingerprint image. In order to compare these two images, nonlinear combined transform correlation method and lattice method are applied to make a comparison decision.

First, the nonlinear joint transform correlation method will be described.

First, each cut holographic fingerprint image is referred to as a reference image, and the actual fingerprint image is defined as an input image. , It is expressed as a function to compose the image provided through different input channels into one new input image. In this case, the center interval of each of the two images is 2 Spaced apart. When the Fourier transform is performed on the newly configured input video signal, it is expressed as a Joint Power Spectrum (JPS) as shown in Equation 7 below.

Where E is the combined power spectrum, I is the intensity distribution of light, S is the Fourier transformed signal of the input signal, R is the Fourier transformed signal of the reference signal, Is the spatial frequency for each axis when Fourier transformed to the corresponding values for the x and y axes, respectively. Denotes the phase of the input signal and the reference signal, respectively.

Since the combined power spectrum includes a strong DC term other than the component necessary to measure the correlation between two images, unnecessary similar characteristics occur. In other words, when the combined power spectrum is used, the effects of phase and amplitude components on similarity determination cannot be effectively processed. Considering only the amplitude component in the degree of similarity, the ratio of peak to side-lobe is lowered, so that the components for noise may be included in the similarity. On the other hand, if only the phase signal is emphasized, it is possible to generate correlation error and loss of correlation in the multiple target image, so proper selection becomes important.

Therefore, it is necessary to remove DC components and unnecessary correlation signals, adjust phase and amplitude characteristics, and generate correlations. Therefore, the similarity is determined using the modified power spectrum. The power spectrum of each of the two images represented by the input image, R and S, is obtained, and the cross-correlation signal of the autocorrelation and the same plane is obtained from the equation represented by Equation (7). Algebraic removal can yield a modified combined power spectrum (E _NEW ).

Equation (8) consists purely of cross-correlation signals between the reference image and the input image, but it is difficult to determine the similarity as it has a low correlation peak value ratio depending on the amplitude signal in determining the similarity between each other. Therefore, in order to adjust the phase and amplitude signals in Equation 8 and to show a correlation result, nonlinear functions such as Equation 9 and Equation 10 are used.

here, Denotes a coupling power spectrum expressed using a nonlinear function, and p is a nonlinear parameter that determines the characteristics of the nonlinear coupling transformation, and has a range of 0 ≦ p ≦ 1.

Therefore, the forms represented by Equations 9 and 10 can obtain the original linear output result and the phased output result depending on the pure phase according to the change of the value of p. However, although phased output results that rely on pure phase can yield better correlation peak values than the original linear output results, they can cause correlation errors and loss of correlation depending on the location of the image for multiple targets. Since a high correlation may be generated for an input image having a form very similar to, the similarity is determined by having a complementary relationship in consideration of an appropriate amplitude relationship (step S214).

The process of determining the similarity is compared with the correlation peak value obtained in equations (9) and (10) and the reference correlation peak value (TH) read from the hologram card 140 (step S216).

The reference correlation peak value TH is an average of correlation peak values obtained by using the NJTC method after receiving the user's fingerprint several times when the user's fingerprint is hologramd. The average of these correlation peaks is printed on the hologram 142 of the hologram card 140 as a reference correlation peak or recorded in the blank margin portion of the hologram 142. The reference correlation peak value printed on the hologram 142 or recorded in the blank margin portion of the hologram 142 is the first light source 110, the first imaging lens 112 and the first image sensor described with reference to FIG. 1. The hologram fingerprint acquisition module, which is configured as 114, is provided to the fingerprint recognition unit 170 through the first image processing unit 150 together with the hologram fingerprint.

The fingerprint recognition unit 170 determines that the correlation peak value obtained in the above-described equations (9) and (10) is the same when the correlation peak value is larger than the reference correlation peak value T _H (step S218). However, if the correlation peak value obtained in equations (9) and (10) is smaller than the reference correlation peak value (T _H ), it is determined that they are not the same person (step S220).

Next, the grid comparison method will be described.

First, when the grid comparison method takes a closed curve surrounding the image of the cut fingerprint, it can be regarded as an object having an arbitrary pattern, such as a lattice whose spacing is not constant, and the Fraunhofer diffraction (Fraunhpfer) with respect to the lattice. It is a concept that the similarity is determined by extracting the values of the direction and angle of the ridge of the fingerprint in a certain sample block through the process of frequency analysis and resampling of the grid image.

FIG. 4 illustrates a configuration in which fingerprint lines are expressed in the form of a grid by arranging pixels of the edge portion surrounding the cut fingerprint image of FIG. 3D in one dimension. In FIG. 4, each grating 410 represents fingerprint lines on a closed curve formed by an edge, and the empty space between the two gratings 410, that is, the fingerprint lines, is defined as the opening 420. Reference numeral 430 denotes the origin of the closed curve, reference numeral 440 denotes the width of the opening 420, and reference numeral 450 denotes the position of the grid spaced apart from the origin 430.

The amplitude transmitted function of the rectangular opening 420 can be defined by Equation 11 as follows.

here, Represents the permeability function, Wow Denotes the width of the opening on the x- and y-axes.

Since the diffraction equation of the Fraunhofer of the opening 420 having this transmittance function is in turn proportional to the Fourier transform of the function representing the distribution of the opening 420, the amplitude distribution of light at the observation position away from the opening 420 by a certain distance z (42) is a product of the Fourier transform of the amplitude transmittance function and the second phase term, and the term preceding the second phase term corresponds to the impulse response in free space as shown in Equation 12 below.

Where U (x, y) is the amplitude distribution of light, Is the wavelength of light, z is the viewing distance, sinc is the sinc function, Is the wavenumber, A is the area between the grids ( ).

Therefore, in order to determine the similarity using the optical method using the diffraction of the grating with such a cut image, the width 440 of the opening 420 and the position information of the opening 420 are found and the width of the corresponding opening 420 is determined. 440 and the values of the spatial frequencies from the position are extracted (step S222).

That is, the spatial frequency distributions generated according to the width 440 of the opening 420 and the position of the opening 420 with respect to the input image and the reference image are extracted, and how the same distribution of each spatial frequency distribution is extracted. The similarity is measured by judging whether there is a (step 224). A criterion for measuring similarity may be determined as a threshold through an experiment using an experimental fingerprint image.

If the similarity satisfies the threshold, it is identified as the same person (step S218). Otherwise, it is determined that they are not the same person (step S226).

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the equivalent scope should be interpreted as being included in the scope of the present invention.

As described above, according to the present invention, by identifying the feature point of the fingerprint and accurately identifying the user, it is possible to realize the fingerprint recognition with improved security.

Since the data that characterizes the fingerprint is made into a database, the fingerprint can be easily identified based on the data.

Claims (22)

In the fingerprint recognition device, A hologram card in which a fingerprint of a user is recorded as a hologram; Hologram fingerprint acquisition means for acquiring an image of the hologram fingerprint from the hologram on the hologram card; Actual fingerprint acquisition means for acquiring an image of the actual fingerprint of the user; A center point of the holographic fingerprint image acquired by the holographic fingerprint acquisition means and the actual fingerprint acquisition means and the image of the image of the actual fingerprint, and the first center of the holographic fingerprint including the main information of the fingerprint based on the center point A nonlinear Joint Transform Correlation (NJTC) is applied to an image and a second center image of the real fingerprint, and a correlation peak value is calculated, and the fingerprint grid on the closed curve surrounding the first center image is calculated. The first spatial frequency distribution is generated by applying diffraction of the Freunhofer at the interval and the position, and the diffraction of the Freunhofer is applied to the interval and the position of the fingerprint grid on the closed curve surrounding the outside of the second central image. Generating a spatial frequency distribution, wherein the calculated correlation peak value is greater than a predetermined reference correlation peak value; Group a first spatial frequency value and the second time the degree of similarity of the second spatial frequency is greater fingerprint authenticating the hologram fingerprint and identity of the actual fingerprint recognizer Fingerprint recognition device comprising a. The method of claim 1, The hologram fingerprint acquisition means, A card insertion slot into which the hologram card is inserted; A first light source for reproducing the hologram fingerprint by irradiating a hologram portion on the hologram card inserted into the card insertion slot; A first imaging lens in which the hologram fingerprint is formed; And a first image sensor configured to photograph the holographic fingerprint formed on the first imaging lens and to output electrical signals having different levels according to patterns of valleys and ridges of the holographic fingerprint. The actual fingerprint acquisition means, A prism having a fingerprint input surface on which the finger of the user is located; A second imaging lens for irradiating light onto the prism to form light reflected from the fingerprint input surface of the prism; A second image sensor which picks up the actual fingerprint formed on the second imaging lens and outputs electrical signals having different levels according to the patterns of the valleys and ridges of the actual fingerprint; Fingerprint recognition device comprising a. The method of claim 1, The reference correlation peak value is printed on the hologram or in the blank portion of the hologram, and the reference correlation peak value is acquired by the hologram fingerprint acquisition means and provided to the fingerprint recognition unit. The method of claim 2, And the hologram fingerprint is an image of a real image reflected from the hologram by light emitted by the first light source to form an image in space. The method of claim 2, And the first image sensor and the second image sensor are Charge Coupled Device (CCD) or Complementary Metal Oxcide Semicondiuctor (CMOS) sensors. The method of claim 2, The fingerprint recognition device, Performing a binarization process to remove noise included in the image of each of the holographic fingerprints and the image of the actual fingerprint, and extract the features of the ridges and valleys of the image of the hologram fingerprint and the image of the actual fingerprint from which the noise is removed Fingerprint recognition apparatus further comprises an image processing unit. The method of claim 6, And the image processor removes the noise by applying spatial domain filtering, frequency domain filtering, or wavelet transform to the image of the holographic fingerprint and the image of the actual fingerprint. The method of claim 6, The image processor divides the image of the holographic fingerprint and the image of the actual fingerprint into an N × N block, calculates an average of pixels in the divided N × N block image, and applies a predetermined threshold value. And performing a binarization process to extract features of the ridges and valleys of the actual fingerprint. The method of claim 8, wherein the fingerprint recognition unit, The direction of the ridges and valleys of the hologram fingerprint and the real fingerprint by averaging the inclination of the pixels in each block of the image of the holographic fingerprint and the image of the real fingerprint ), A point where the direction perpendicular to the direction of the ridge and the valley converge as a center point, and a predetermined size block is set as the center image based on the center point, The slope is represented by the following equation 1, [Equation 1] The slope average of the pixel blocks is calculated by Equation 3 below. [Equation 3] i and j each represent a pixel in the block, Wow Are respectively gradient vectors Wow Indicates the variance of Is Wow Represents the covariance of, and the variance and covariance of the gradient vector are as shown in Equation 4 below. [Equation 4] Average Gradient Direction ) Is calculated as in Equation 5 below. [Equation 5] here, And the slope direction ( Average ridge and valley direction perpendicular to ) Is calculated as in Equation 6 below. [Equation 6] Fingerprint recognition device, characterized in that. The method of claim 9, The tilt direction ( ) Doubles the angle of the inclination and squares the magnitude so that the inclination vectors in opposite directions do not cancel each other, and takes the average, here, Is a squared and double angle representing each of the inclinations of the x and y axes. The method of claim 1, The correlation peak value is calculated from a joint power spectrum for the first center image and the second center image, and the combined power spectrum is calculated as in Equation 7 below. [Equation 7] Where E is the combined power spectrum, I is the intensity distribution of light, S is the Fourier transformed signal of the input signal, R is the Fourier transformed signal of the reference signal, Is the spatial frequency for each axis when Fourier transformed to the corresponding values for the x and y axes, respectively. Is a phase of the first center image and the second center image. The method of claim 1, The correlation peak value is calculated from a joint power spectrum for the first center image and the second center image, and the combined power spectrum is calculated as in Equation 8 below. [Equation 8] Where E is the combined power spectrum, R is the Fourier transformed signal of the first center image, S is the Fourier transformed signal of the second center image, Is a phase of the first center image and the second center image. The method of claim 12, The correlation peak value is calculated from a modified combined power spectrum for the first center image and the second central image, and the modified combined power spectrum is calculated as in Equations 9 and 10 below. , [Equation 9] [Equation 10] here, Denotes a coupling power spectrum expressed using a nonlinear function, and p is a nonlinear parameter for determining a characteristic of the nonlinear coupling transformation, wherein the fingerprint recognition device has a range of 0 ≦ p ≦ 1. The method of claim 1, The spatial frequency distribution is calculated by the following equation (11), [Equation 11] Where U (x, y) is the amplitude distribution of light, Is the wavelength of light, z is the viewing distance, sinc is the sinc function, Is the wavenumber, A is the area between the grids ( Fingerprint recognition device characterized in that. A fingerprint authentication method for determining a mutual identity between a hologram fingerprint recorded as a hologram on a hologram card and an actual fingerprint of a user of the hologram card, (a) obtaining an image of the holographic fingerprint and an image of the actual fingerprint; (b) finding a center point of the image of the holographic fingerprint and the image of the actual fingerprint by using the gradient of the image of the holographic fingerprint and the image of the actual fingerprint; (c) selecting first and second center block images of predetermined sizes based on the center points of the holographic fingerprint image and the actual fingerprint image; (d) calculating a correlation peak value by applying a nonlinear joint transform transformation (NJTC) to the first center block image and the second center block image; (e) extracting a first frequency component according to the distance and location of the fingerprint grid on the closed curve surrounding the outer edge of the first center block image, and the distance between the fingerprint grid on the closed curve surrounding the outer edge of the second central block image; Extracting a second frequency component according to a position; And (f) when the correlation peak value calculated in step (d) is larger than a predetermined reference correlation peak value, and the similarity between the first spatial frequency value and the second spatial frequency value obtained in step (e) is large. When authenticating the identity of the holographic fingerprint and the actual fingerprint Fingerprint recognition method comprising a. The method of claim 15, The method after step (a), An image processing step of removing noise by applying spatial domain filtering, frequency domain filtering, or wavelet transform to each of the holographic fingerprint image and the actual fingerprint image; And Binarization processing step for extracting the features of the ridges and valleys of the noise-free hologram fingerprint and the actual fingerprint Fingerprint recognition method further comprising. The method of claim 16, The binarization processing may divide an image of the holographic fingerprint and the image of the actual fingerprint into an N × N block image, calculate an average of pixels in the divided N × N block image, and set a predetermined threshold value. And extracting features of the ridges and valleys of the holographic fingerprint and the actual fingerprint. The method of claim 17, Step (b) is, (b-1) obtaining an average of inclinations of pixels in the image block of the holographic fingerprint and the image of the actual fingerprint to obtain the direction of the ridges and valleys of the hologram fingerprint and the actual fingerprint; (b-2) from 0 along the direction perpendicular to the direction of the ridge and the valley of the hologram fingerprint and the real fingerprint in the image block of the holographic fingerprint and the real fingerprint; Down in the -y direction within the range of / 2, If a block is encountered that is not in the range of / 2, stopping the visit and marking the block; And (b-3) accumulating the displayed values of the image block and determining the x and y coordinates of the most marked block as the center points of the hologram fingerprint and the actual fingerprint; Fingerprint recognition method comprising a. The method of claim 18, In the step (b-1), the angles of the inclination are doubled, squared in magnitude, and averaged so that the inclination vectors in opposite directions do not cancel each other. The method of claim 15, In the step (c), characterized in that the first center block image and the second center block image is a block in which the information of the holographic fingerprint and the information of the actual fingerprint is concentrated. The method of claim 15, Step (d) is, Obtaining a joint power spectrum of two inputs by allocating the first center block image and the second center block image as inputs of a nonlinear combined transform correlation method; And Inverse Fourier transforming the combined power spectrum to yield the correlated peaks Fingerprint recognition method comprising a. The method of claim 15, Step (e), (e-1) the fingerprints included in the first center block image and the second center block image are arranged in one dimension by arranging pixels of an edge portion surrounding the first center block image and the second center block image in one dimension; Expressing in the form of; And Extracting the first and second spatial frequency values generated by diffraction according to Fraunhpfer Diffraction at the intervals and positions of the gratings of the image of the holographic fingerprint and the image of the actual fingerprint. Fingerprint recognition method characterized in that it comprises a.