JPH11339049A - Card authentication system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system hardly affected by the variation of light quantity, and capable of preventing illegality such as forgery of card and improving security of authentication by authenticating a card based on the value calculated by a correlation coefficient calculating means. SOLUTION: A face photograph 4 on an IC card 1 is fetched as original image data by a digital camera 21 included in an authenticating device 2 and on the other hand, an IC module 3 internal part on the card 1 preliminarily stores collation reference data in a user inaccessible area 35a of an EEPROM 35. The device 2 transmits the fetched original image data to the module 3 through an IC card interface 26 and the module 3 internal part performs authentication processing from the received original image data and the collation reference data of the area 35a based on the value calculated by a correlation coefficient operating part of an authentication program 36 of a ROM 33 and decides whether the card 1 is right or not.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a card authentication system and, more particularly, to an IC card having visible information for identifying an individual, such as a photograph, a signature character, and a real seal, for identifying the individual. To one screen n × m
(N, m are integers of 3 or more, respectively) In an IC card authentication system that reads as multi-gradation image information (image data) developed into pixels and authenticates an IC card based on this information, the amount of data processing is reduced. The present invention relates to a card authentication system capable of quickly authenticating the validity of an IC card.

[0002]

2. Description of the Related Art Under card authentication, commercial transactions using a credit card or the like, withdrawal of bank deposits using a cash card, and the like are performed. Alternatively, electronic commerce is performed on the Internet, and the cards
C cards are being used. The problem with this type of card is counterfeiting and unauthorized use. In order to prevent such a situation, conventionally, individual feature data is
An IC that records fingerprints and voiceprints in the memory of the C card for verification
Card authentication systems are known. As one of them,
A technology for collating a fingerprint image obtained by scanning a card user's fingerprint or the like with a scanner and fingerprint data sent from an IC card is disclosed in, for example, JP-A-61-15379.
No. 6, JP-A-61-276081 and the like, which are known. By performing such a process, card authentication can be performed with higher security than collation of the personal identification number.

[0003] Also, there have been proposed cards in which an image using a hologram is formed on the surface of the card, and a specially printed card is provided. For example, in the technology disclosed in Japanese Patent Application Laid-Open No. 4-338593, there is a technique in which a transparent protective layer having irregularities such that an image is formed by a thermal transfer method in an identification card or the like is provided on the image forming surface. Also, Japanese Patent Application Laid-Open No. Hei.
No. 3,883, No. 3883 obtains image data from a card, converts the image, and performs verification to perform card authentication, thereby ensuring security against forgery and unauthorized use of the card.
On the other hand, in Japanese Patent Application No. 9-219330, the present inventors divide the visual check area of the card into a plurality of blocks, perform outline emphasis processing, and perform position information and pixel count information on pixels within a preset range. Is disclosed as a collation code for each block to prevent forgery and falsification. Further, the inventors of the present invention disclosed in Japanese Patent Application No. 9-243.
No. 343 proposes a card authentication system which enables detection of a zero-cross point by a simple arithmetic processing, thereby reducing the amount of data processing and extracting features at high speed.

[0004]

Recently, various technologies have been proposed to prevent unauthorized use by using a card displaying a photograph of the owner's face, a signature, and the like on the card and a system therefor, which has better anti-counterfeiting properties than before. It is attracting attention. However, such a conventional technique has a problem that it is difficult to construct a card matching processing system that performs card matching using information corresponding to an image located in a visual confirmation area such as a face photograph or a signature. The aforementioned Japanese Patent Application No. 9-2
No. 19330 and Japanese Patent Application No. 9-243343 solve such a problem, and aim at authentication with high confidentiality by simple processing. These perform card authentication by calculating code data based on image data such as a face photograph and collating code data registered in advance.

However, when actually considering the application of the latter technique, it has been found that there is a problem that reading of image information such as a face photograph is affected by fluctuations in ambient light quantity.
Further, there are various environments in which the IC card is actually used, and the gradation at the time of imaging (the gradual simple increase of the face photograph in a specific direction according to the intensity of the irradiated light) is related to the optical reading. Or a simple decrease in concentration). If the risk of determining a forgery is avoided in consideration of such a situation, a contradictory condition that the possibility of determining a different face photograph as valid is increased. SUMMARY OF THE INVENTION An object of the present invention is to solve such a problem of the prior art, and is intended to be less susceptible to fluctuations in light quantity, to prevent unauthorized use such as forgery of a card, and to improve security in authentication. It is an object of the present invention to provide a card authentication system that can perform the above.

[0006]

SUMMARY OF THE INVENTION A feature of the card authentication system of the present invention for achieving the above object is that visible information for identifying the individual, such as a photograph for identifying the individual, a sign character, a real seal, or the like. , A reading device that expands visible information into n × m pixels (where n and m are integers of 3 or more) on one screen and reads multi-gradation image information from the card, and a reading device Divides one screen into a plurality of blocks based on image information obtained from a plurality of blocks, and corresponds to the number of pixels or pixels having specific characteristics for each block based on the multi-gradation detection value of each pixel in each of the divided blocks. Number calculation means for calculating the number of feature points obtained as a feature number, and a reference feature number stored in advance for each block and a block correspondence calculated by the feature number calculation means. Have a correlation coefficient calculating means for calculating a correlation coefficient between the feature number, and performs card authentication of the based on the value calculated by the correlation coefficient calculating means.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Alternatively, the visible information for identifying the individual, such as a photograph, a signature character, and a real seal, for identifying the individual engraved is read as image information, and the image information of each block divided into a plurality of blocks based on the read image information is read. The number of pixels having a specific feature corresponding to the block or the number of feature points obtained corresponding to the pixel is calculated by the feature number calculating means, and a correlation coefficient is calculated based on the calculated number and authentication is performed. Therefore, information for collation can be obtained by a simpler process than the calculation process of the entire image information. As a result, the load of arithmetic processing for obtaining collation information is reduced, and high-speed processing is enabled. In addition, since the number of features is obtained for each block, validity judgment is improved, and high security can be ensured. As a result, authentication with high confidentiality can be performed by simple arithmetic processing, and unauthorized use such as forgery of a card can be prevented, and security in authentication can be improved.

[0008] By the way, a specific configuration of the present invention comprises a card having a face photograph for identifying an individual on the surface thereof,
A reading device that reads the face photograph as image information that expresses a gradation corresponding to each pixel of one screen in which the screen is developed into n × m (n and m are each an integer of 3 or more) pixels; A means for calculating whether or not any pixel is zero-crossing from the image information of the screen; a means for determining to which block any pixel detected as zero-crossing belongs; and a zero for the determined block. Detection value updating means for sequentially updating the numerical value of the crossing detection, and correlation coefficient calculating means for calculating a correlation coefficient with a preset zero crossing number in each block from the calculated zero crossing number in each block. Means for comparing the calculated correlation coefficient with a preset threshold value. The correlation function indicates a value close to 1, for example, a value of 0.7 or more based on the correlation coefficient. It is intended to authenticate the validity of the card by.

An algorithm for encoding an original image according to the present invention for generating authentication code data used for such validity authentication includes extracting a feature amount from the original image for encoding and encoding the original image. Generate authentication data (authentication collation data). In this case, the time required for the arithmetic processing is short in the present invention. Further, when a program based on the encoding algorithm of the present invention is created, the program capacity can be reduced. Furthermore, since it is composed only of relatively simple arithmetic processing, it is easy to create a program. It is possible to construct a security system having such advantages. Therefore, according to the card authentication system of the present invention, when the same face photograph is read in an environment where the light amount fluctuates, the same face photograph is read based on the correlation coefficient. Unlike a simple comparison,
Misjudgment is unlikely to occur. On the other hand, when a different face photograph is read, a numerical value considerably smaller than 1 is calculated for the correlation coefficient.

[0010]

FIG. 1 is a block diagram showing one embodiment of an IC card authentication system to which the card authentication system of the present invention is applied. FIG. 2 calculates the number of pixels having specific features corresponding to blocks of the present invention. FIG. 3 is a flowchart of the process.
It is a flowchart of an authentication process. As described above, the card authentication system of the present invention shown in FIG. 1 includes a card having a face photograph for identifying an individual and a screen having nxm (where n and m are three or more, respectively). A reading device that reads a facial photograph as image information expressing gradation corresponding to each pixel of one screen developed into pixels of (integer), and image information for one screen read as an authentication program provided inside the IC card. A zero-crossing determining unit for calculating whether or not an arbitrary pixel is a zero-crossing, and a belonging-block determining unit for determining which block the arbitrary pixel detected as the zero crossing belongs to. And a counter variable provided corresponding to the number of each block in order to sequentially update the detected value of the number of zero crossings of the determined block. A counter variable updating unit for incrementing and updating, and a correlation coefficient calculation unit for calculating a correlation coefficient between the calculated number of zero crossings in each block from the calculated number of zero crossings (detected value) in each block. Department and
A true / false judgment unit for judging the validity of the IC card by comparing the calculated correlation coefficient with a preset threshold.

The contents will be described below with reference to FIG. In FIG. 1, a card 1 having a face photograph for identifying an individual on its surface is a card including an IC module 3, a face photograph 4 for identifying an individual, a magnetic stripe 1a, and an embossed area 1b. This I
The C card 1 has an I / O port unit 31, a CPU 32, and a ROM as a system control program logic, as shown as a block on the left side.
And a unit 33. The IC module 3 that contacts the IC card interface 26 of the authentication device 2
The I / O port unit 31 has a ground terminal (GND) 3
1a, I / O port (data input / output) 31b, clock signal terminal (CLK terminal) 3lc, reset signal terminal (R
(ST terminal) 31d, a power supply terminal Vcc 31e, and the like. The ROM section 33 stores system programs and the like executed by the CPU 32. In addition to the ROM section 33, a RAM 34 as a work area, various application programs and an EE as a data storage area thereof are stored.
And a PROM 35, which are mutually bus-connected.

The EEPROM 35 has a user inaccessible area 35a and a user area 35b for storing highly confidential data, IDs, and other management information. The user inaccessible area 35a has collation reference data ( Authentication data) 37 serving as a collation reference is stored. The ROM 33 inside the IC module 3 stores an authentication program 36 based on an authentication algorithm described later.
By executing this program, the processes performed by the zero-crossing determining unit, the belonging block determining unit, the counter variable updating unit, the correlation coefficient calculating unit, and the authenticity determining unit are realized collectively. In addition, used in the correlation coefficient calculation unit,
Information about the number of zero crossings in each block set in advance is stored in a two-dimensional array variable WRK [j] [k] described later.
Put in. 2 is an authentication device, which is a so-called IC
It consists of a card reader / writer. The digital camera 21 includes a digital camera 21 which is a reading device for reading the face photograph 4 attached to the IC card 1 as image information, a CPU 22, a memory 23, a printer 25, and an IC card interface 26 as an IC card communication interface. These are interconnected via a bus 24.

The photograph 4 of the face of the IC card 1 is read as original image data by the digital camera 21 of the authentication device 2, and is sent to the IC module 3 of the IC card 1 via the IC card interface 26. Is transmitted. Control of data and command transmission between the authentication device 2 and the IC module 3 includes transmission of original image data,
There are a start command transmission and a reset command transmission of the authentication program 36 in the C module 3 and a transmission of the authenticity judgment result.
The transmission control program 23b stored inside the CPU 3
22 executes the I / O of the IC module 3 that contacts the IC card interface 26 of the authentication device 2.
This is performed via the O port unit 31.

Next, in order to obtain collation target data or collation reference data 37 generated from the read image data,
The encoding algorithm will be described. Here, the face photograph 4 from which the image data is to be collected is divided into blocks as shown in FIG. In this case, as shown in (b), it is advantageous to make a division including a portion where the division length of each block is not equal. The division including the unequal length will be described later, but on the premise of such a block division, the image information and the block of the divided portion will be described below. FIG. 2 is a flowchart showing the algorithm of the encoding process. With reference to FIG. 2, an encoding algorithm for generating encoded data for authentication will be described first.
Here, a variable indicating an array corresponding to the C language will be described in the form of a variable name + []. That is, Cblk [i] is a one-dimensional array variable for storing numerical data of feature points of each divided block of the image data of the face photograph, and WRK [j] [k] is a two-dimensional array variable for zero-cross detection. It is assumed that each is set. Note that the array variable Cblk
Here, [i] is a variable that stores the number of zero crossings detected in each block, and corresponds to a data area in which the value is incremented and stored for each detection.
The count value stored in the array variable Cblk [i] is the number of feature points for generating code data for authentication. Also,
The array variable WRK [j] [k] is a data storage area for detecting a zero-cross pixel by obtaining a predetermined calculation value for two adjacent pixels around the pixel of interest and four pixels above, below, left and right around the pixel of interest. Has become. Further, a block Blk [i] is set as an array for storing the data of the detected values corresponding to the pixels. This is a block-divided image buffer area for temporarily storing data read as original image data by the digital camera 21.

First, in step 101, the height of the original image is set to a variable h (where h is the height converted into the number of pixels).
In FIG. 8), the width of the original image is substituted by a variable w (where w is the width converted into the number of pixels; see FIG. 8). Next, in step 102, an array variable Cblk [i] for setting a zero-cross count value of each block is initialized. FIG. 7 shows the data structure of the count value sequence variable Cblk [i] to be initialized. Each array element of the count value array variable Cblk 52 is assigned to each block of the original image 51 divided into blocks. If the number of block divisions is N for one screen of the original image, the i-th block obtained by dividing the original image into blocks Blk [i] for an integer i such that l ≦ i ≦ N Cb in order from the lower left of the image
Each area is assigned as an area for storing a zero-cross count value in each block divided into lk [i]. Thereby, the count value array variable Cblk [i]
Is a so-called soft counter corresponding to each divided block.

Next, in step 103, a variable i is initialized to 0 as a counter variable of the read row. Thereafter, in step 104, the variable i is compared with the variable h.
Proceed to. If i is larger than 0 and smaller than (h-1) in step 105, the process proceeds to step 106. In step 106, the data read start position of the i-th row is set. However, initially, i = 0. Then, in step 107, three rows are read from the original image and stored in WRK [j] [k], which is a working array. Here, the data for three rows is data for w × 3 pixels as shown in FIG. When three rows of data are stored in WRK [j] [k], a determination process is performed by the zero crossing determination unit in the processing in the ROM unit 33.
Here, the zero crossing determination unit determines that the previous array variable W
RK [j] [k], j is three from j = 0 to 2, and k
Are set as w from k = 0 to w−1.
Of these, the second row (j = 1) in the middle is the row of the target pixel to be subjected to the zero crossing determination. The detection value of each pixel is, for example, an 8-bit value of 256 gradations.

Therefore, the pixel data of one screen is sequentially read from the original image data in the memory for the first row in accordance with the processing, and the data of a plurality of blocks Blk [i] included in one row are sequentially processed by WRK. Data is stored from [0] [0] to WRK [2] [w-1]. As a result, the relationship between the data structure of the WRK and the acquired image information of one screen is as shown in FIG. Next, in step 108,
Substitute 1 for a counter variable k in the width direction within one row of the pixel of interest. Thereafter, if k is smaller than w−3 in step 109, the process proceeds to step 110. In step 110, a zero crossing determination function D (k) is calculated. The zero crossing determination function D (k) is a function for determining whether or not the pixels WRK [1] [k] (where k is 1 to w-2) of the original image are zero crossings. The first pixel of interest is k = 1, and WRK [1] [k] is the second pixel from the left in the second row in FIG. Hereinafter, the pixel of interest is updated by sequentially incrementing k, and the pixel of interest moves to the right. Around WRK [1] [k]
There are four pixels at the top, bottom, left and right. This attention pixel WRK
[1] [k], the detected values of four pixels above, below, left and right, and the next pixel of interest WRK [1] [k + 1] and the four pixels above, below, left and right
Here, zero crossing determination is made based on the relationship with the detected values of the pixels.

In the determination, the relationship between the calculated values of the two pixels, that is, the pixel of interest and the pixel adjacent thereto, is positive,
When it is different from negative or negative or positive, it is determined to be zero cross. This is detected when the product of the calculated values of both eye pixels is negative. The judgment formula is as follows. D (k) = {WRK [2] [k] + WRK [1] [k−1] + W
RK [1] [k + 1] + WRK [0] [k] -4WRK [1]
[k]｝ × {WRK [2] [k + 1] + WRK [1] [k] + WR
K [1] [k + 2] + WRK [0] [k + 1] -4WRK [1]
[k + 1]｝

If D (k) is a negative value, the pixel WRK [1] [k] of the original image is zero crossing. In the zero crossing, the pixel position in the original image corresponds to the inflection point of the change in luminance. FIG.
As shown in FIG. 2, the position of the inflection point itself becomes considerably stable when the light amount of the entire original image fluctuates. For example, when the luminance decreases in the entire original image as shown in FIG. 12, the luminance is offset in a direction in which the luminance disappears while the luminance change shape in a specific position direction is maintained. Therefore, the position of the inflection point itself becomes considerably stable. That is, the position of the zero crossing is considerably stable when the light amount of the entire original image fluctuates.
Here, attention is paid to this. In FIG. 11, the luminance is ± 20%.
7 shows a zero crossing image when changed. In this figure, a white portion represents a pixel determined to be zero crossing, and a pixel not determined to be zero crossing is represented by being binarized as black. From FIG. 11, it can be seen that the position of the zero crossing is quite stable as a whole with respect to the light amount fluctuation. So, here,
In order to realize authentication that does not make an erroneous determination even when there is a light amount fluctuation, the number of zero-crossing pixels is extracted from the original image for each block as a feature point that has a small numerical change even if there is a light amount fluctuation. According to the research of the inventor, it is determined whether or not zero crossing is performed using D (k) given by the above equation, and the conclusion that it is preferable to adopt the local density of zero crossing as a feature value. It has gained.

Returning to FIG. 2, in step 111, when D (k) is a negative value in the above equation, that is, when a certain pixel (WRK [1] [k]) of the original image is zero. If it is the crossing point, the process proceeds to step 112. In step 112, it is determined to which block the pixel subjected to the zero-cross detection belongs. Then, in the next step 113, the value of the array variable Cblk [i] corresponding to the block to which a certain pixel (WRK [1] [k]) of the original image belongs is incremented and its count value (code data as a feature value) is incremented. ) To update. For example, W
An integer 1 ≦ M ≦ N of the block Blk [i] in which the pixel of the target pixel row (second row) among the three rows read from RK [j] [k] is read in this one row When the WRK [1] [k] of the preceding zero-cross point is located in the M-th block and the block is located in that block, the numerical value of the corresponding code data Cblk [M] increases by one. Thereafter, in step 114, the numerical value of the width direction variable k is increased by one, and in order to proceed to the determination of the next target pixel, step 10 is executed.
Return to 9. As described above, when k increases and increases to k = w−3, it is determined that k <w−3 is not satisfied in step 109. In step 115, the value i on the row side is increased by 1 and the row is increased. Update and return to step 104. Similarly, i increases, and when i = h, step 1
If i <h in 04, the process ends. Cblk [i] of the finally calculated sequence block Blk [i] (however,
(From 1 to N) for each block,
It is stored as the collation reference data 37 in FIG.

In the above case, the data of WRK [j] [k] is pushed down by one line or read again according to the increase of the variable i on the line side, and the third line is The second row becomes the first row, and the first row is discarded.
Then, the next line is read to the third line. The one line set as the third line to be read next may be the first line composed of the next plurality of blocks. In this manner, the collation reference data 3 serving as the reference data serving as the collation reference
7 is generated and stored in the IC card 1, but the data to be verified is generated by the same processing from the image information read from the face photograph 4 at the time of authentication. By the way, in step 112, the block to which the zero-cross detection pixel belongs is determined from the pixel of interest in one row, but this is for reducing the processing load.
Of course, each pixel detection may be sequentially read into WRK [j] [k] for each divided block, a zero-cross point pixel in the block may be detected, and the detection value of the block may be incremented.

Next, the processing of an authentication algorithm for obtaining image information from the face photograph 4, generating data to be verified, and authenticating the IC card will be described with reference to FIG. In steps 201 to 215 after the start of the authentication processing, the processing of steps 101 to 115 of the encoding algorithm described above is performed to obtain data to be verified.
The only difference from the above is that the data to be verified is not stored in the IC card 1 as the verification reference data 37. Therefore, the explanation is omitted. Here, the collation reference data 37 stored in the IC card 1 is used.
In order to distinguish between the array data of the data to be collated and the array data of the data to be collated, the array variable of the data to be collated generated for the time of authentication is replaced with Dbl instead of the array variable Cblk in the previous encoding.
It is described as k. Therefore, the array variable Cblk in FIG. 7 becomes the array variable Dblk here, and the array data of the collation reference data 37 becomes the above-described array variable Cblk.

When i increases in step 215 and increases to i = h, it is determined in step 204 that i is not smaller than h, and the flow advances to step 216. Step 21
In step 6, the correlation coefficient Crrltn (Correlation) is calculated from the data of the array variable Dblk and the data of the array variable Cblk. The equation for calculating the correlation coefficient Crrltn is as shown in FIG.
It is one of the mathematical formulas. The correlation coefficient Crrltn can be calculated from each value of the array variable Dblk71 of the data to be collated and each value of the array variable Cblk52 of the collation reference data.
After that, in step 217, the correlation coefficient Crrltn is compared with the threshold value Th.
If not, the process proceeds to step 219 to make a true / false determination. Here, the closer the correlation coefficient Crrltn is to 1, the more the two coincide. The threshold Th is usually
A value of about 0.7 is adopted. That is, the correlation coefficient Crr
When ltn is 0.7 or more, the IC card 1 is determined to be valid.

The program based on the authentication algorithm according to FIG. 3 described above is the authentication program 36 in FIG. The zero crossing determining unit corresponds to step 210, the belonging block determining unit corresponds to step 212, and the counter variable updating unit corresponds to steps 213 and 214. The correlation coefficient calculation unit corresponds to step 215, and the authenticity determination unit performs steps 217 to 219.
With these, the processing is executed collectively. Next, an overall flow of the encoding process before distributing the card to the user will be described with reference to FIG. FIG. 1 described earlier
In the embodiment, for the sake of explanation, the explanation is made such that the IC card 1 receives the image data from the authentication device 2, generates the collation reference data 37, and stores it internally.
It is better to use a dedicated device that generates the collation reference data 37 from the image data and stores it in the IC card. Hereinafter, this will be described. The face photograph 4 on the IC card 1 is taken as original image data 43 into the encoding device 41 by a digital camera. In the encoding, the environment such as the amount of light when capturing the original image from the face photograph 4 is preferably as close as possible to the capturing of the original image during authentication. Therefore, it is desirable to take in the original image at the time of image encoding for obtaining the collation reference data using a digital camera 21 or a skinner equivalent to the authentication device 2, and then transfer the original image data 43 to the encoding device 41. .

As the encoding device 41, a personal computer or the like can be used simply. In this case, an encoding program based on the algorithm of FIG. 2 may be created, and the encoded data 45 may be calculated using the original image data 43 as input data. The value to be calculated is not limited to the zero-cross point calculated for each block as described above, and may be obtained from another characteristic value calculated for each block. The calculated code data 45 is the collation reference data 37 in FIG. 1 and is to be written into the IC module 3 of the IC card 1 using the IC card reader / writer 42. It is preferable to record the collation reference data 37 in the EEPROM 35 in the IC card, but it is not always necessary to write the collation reference data 37 in the IC card. RO using ROM writer 42
It is also possible to store it in M. In addition, EEPROM3
In the case of recording in No. 5, it is preferable to write in a user inaccessible area 35a which is a memory area in which the user cannot erase or rewrite data. This is because the user can check the collation reference data 37 stored in the IC module 3.
This is to prevent falsification of the file.

In the example shown in FIG. 4, the part for performing the authentication process from the collation reference data 37 and the original image data stored in the IC module 3 in advance is stored in the IC module 3. This is possible because the authentication algorithm shown in FIG. 3 has the following properties. That is, since the authentication algorithm includes only simple processing steps, the program capacity of the authentication program to be created can be small. Further, even if the data capacity of the code data is small, it is possible to perform the effect of improving the security and the authentication that is hardly affected by the fluctuation of the light amount. According to the present invention, since the authentication calculation itself can be performed in the IC chip of the IC card, data transmission / reception via the IC card can be performed.
Even if it can be intercepted, the authentication algorithm of the present invention is nearly unbreakable. In other words, it is difficult to break security behind the present invention.

Next, with reference to FIG. 5, the overall flow of the authentication process after the distribution to the user will be briefly described. The face photograph 4 on the IC card 1 is captured as original image data by a digital camera 21 built in the authentication device 2. On the other hand, inside the IC module 3 on the IC card 1, the user
The collation reference data 37 is stored in advance in a. The authentication device 2 transmits the captured original image data to the IC module 3 via the IC card interface 26. I
In the C module 3, the received original image data and the EEP
An authentication process based on the authentication algorithm shown in FIG. 3 is performed based on the collation reference data 37 in the user inaccessible area 35a of the ROM 35 to determine whether the IC card is a valid IC card or a forgery. In the above embodiment, the correlation coefficient Crr
ltn is compared with a threshold value Th (0.7), and Crrltn ≧ 0.7
Is determined to be valid. The legitimate / counterfeit information indicating whether or not it is legitimate is encrypted inside the IC module 3,
It is transmitted to the authentication device 2 side. The authentication device 2 decrypts the encrypted valid / forged information and displays whether the information is valid or forged.

Next, the flow of the authentication process after the IC card has been distributed to the user will be described with reference to FIG. 6 for an example different from FIG. FIG. 6 shows an example in which authentication is performed on the authentication device 2 side. In FIG. 1, the collation reference data 37 is stored inside the IC module 3 of the IC card 1,
The card 1 is regarded as a recording medium with a photograph 4 of the face. In this regard, the example in which all of the arithmetic processing is performed on the authentication device 2 side in FIG. 6 is inferior to the above in terms of security. However, the security of the example in FIG. 6 may be sufficient depending on the application, and the respective processes will be described below. The face photograph 4 on the IC card 1 is stored in a digital camera 21 built in the authentication device 2.
Is taken in as original image data. In parallel with the capture of the original image, the collation reference data 37 stored in the IC module 3 is also read from the IC card 1 into the authentication device 2. Using the original image data and the collation reference data 37 in the IC module 3, the authentication device 2 performs an authentication process based on the authentication algorithm of the present invention shown in FIG. As a result of the authentication processing, whether the IC card 1 is legitimate or counterfeit is displayed.

Next, a description will be given of the result of an experiment using an actual face photograph in the card authentication system of the present invention described above. FIG. 8A shows a zero crossing image calculated from the prepared original image. In the experiment, as shown in FIG. 8B, the number of divisions N = 5
The collation reference data was created with × 5 = 25, and the correlation coefficient with the code data calculated from the face photograph after the luminance change was obtained. If an unsigned short type used in the C language, that is, an array of an unsigned integer type is used, an encoding program and an authentication algorithm can be easily created. Division number N
Since = 25 is the number of elements of the unsigned short type array, the collation reference data eventually needs a very small capacity of 50 bytes.

In the program based on the authentication algorithm shown in FIG. 3, the belonging block judging unit is step 212. However, in actuality, as shown in FIG. The comparison was made in the case of block division including a long part. In the block division including the unequal length portion in FIG. 14B, the size of the block is set to be larger than the block inside the face in the background block. This helps to prevent deterioration due to gradation as described later. FIG. 15 shows the relationship between the luminance change and the correlation coefficient, which are the experimental results. In both the isometric block division and the block division including the non-equal length part, the correlation coefficient is a value of 0.7 or more even if the luminance change is ± 30%. In other words, if 0.7 is adopted as the threshold value in step 217 in FIG. Conversely, with respect to the risk of judging a different face photo as valid, there is no different face photo with a correlation coefficient exceeding 0.6 as far as the inventors have conducted experiments, and the risk of judging a different face photo as valid is very high. Small.

Next, the effect of the block division including the unequal-length portion in FIG. 14B will be described. As can be read from FIG. 15, the block division including the unequal-length portion in FIG. 14B has a correlation coefficient of 0.9 even for a luminance change of ± 30%.
And less degradation. This is because, by setting the size of the block in the background portion larger than the size of the block inside the face, the adverse effect due to the gradation existing in the background portion is averaged to reduce the influence. The mechanism of the adverse effect due to the gradation existing in the background will be described with reference to FIG. FIG. 13A schematically shows the relationship between the position of the gradation of the background portion and the luminance of the original image. Eight pits are assigned to each pixel, and luminance can be represented by an integer from 0 to 255. Because of the gradation, there is a portion where the luminance gradually increases as the position becomes farther away.
When it reaches 5, 255 is maintained. The inflection point of the luminance with respect to the position is at a position where the luminance reaches 255 as shown in FIG. FIG. 13B shows how such gradation changes due to light quantity fluctuation.

On the contrary, there is a case where the luminance gradually decreases. In this case, when it reaches 0, it is kept at 0 to avoid underflow. This also makes the zero crossing position unstable. Even if you do not intend to add a gradation intentionally, the gradation may naturally exist in the background, so it is desirable to take measures. Block division including non-equal length parts provides a measure for it. Block division including non-equal length portions is known in the field of image compression technology as a type of block division that improves the compression ratio by reflecting the properties of an original image. But,
The purpose of the block division including the unequal length part in the present invention is completely different. That is, an object is to reduce the adverse effect on the correlation coefficient of a block, which tends to have an adverse effect on the authentication, in performing the authentication without being affected by the light amount fluctuation. Blocks that include the gradation part described above are divided into blocks that include unequal lengths.
By increasing the block size of the background part, the contribution to the correlation coefficient is reduced. Accordingly, the number of zero crossing pixels in the block can be made closer between the reference image and the image after the luminance change. As a result, the decrease in the correlation coefficient is small, and such a result is obtained in FIG.

As described above, the background portion including the gradation portion is divided so as to include the non-equal-length blocks so that the size of the block of the background portion including the gradation portion is larger than that of the block inside the face. Can reduce the contribution of the block to the correlation coefficient. As a result, the number of zero-crossing pixels in the block approaches the reference image and the image after the luminance change, so that the correlation coefficient does not need to be reduced, and adverse effects due to gradation existing in the background part can be reduced. .

As described above, as described in the embodiment, a card having a face photograph for identifying an individual on the surface and one screen are developed into nxm (n and m are each an integer of 3 or more) pixels. A reading device for reading the face photograph as image information expressing a gradation corresponding to each pixel of one screen,
Means for calculating whether or not any pixel is a zero crossing from the image information for the image area, means for determining to which block any pixel for which zero-crossing is detected belongs, and Means for increasing a counter variable, means for calculating a correlation coefficient with a predetermined number of zero crossings in each block from the calculated number of zero crossings in each block, and means for calculating the calculated correlation coefficient and the number of zero crossings in each block. It has a comparison with a threshold value, and has a configuration in which the card is authenticated based on the correlation coefficient, so that when reading the same face photograph in an environment with light quantity fluctuation, the correlation coefficient is Since a numerical value close to 1 is calculated, it is possible to authenticate the same face photograph, and unlike a simple comparison / collation, it is not necessary to make an erroneous determination. On the other hand, when a different face photograph is read, a numerical value considerably smaller than 1 is calculated for the correlation coefficient, so that it is possible to authenticate a different face photograph. As a result, it is possible to provide a card authentication system that does not make an erroneous determination even when the light amount fluctuates when reading a facial photograph.

It should be noted that the present invention is not limited to the above-described optimum configuration, but uses the number of pixels having a specific feature corresponding to a block or the number of feature points obtained corresponding to a pixel as a feature number. A feature number calculating means for calculating, and a correlation coefficient calculating means for calculating a correlation coefficient between a reference feature number stored in advance for each block and a feature number for the block calculated by the feature number calculating means. Therefore, the capacity of the authentication program and the like can be reduced, the load of the authentication process can be reduced, the erroneous recognition can be reduced, and the accuracy of the authentication can be improved.

In particular, the encoding algorithm in the authentication program has the following advantages because the extraction of the feature amount from the original image for encoding and the creation of encoded data are integrated. . That is, the time required for the arithmetic processing is reduced. Further, when a program based on the encoding algorithm of the present invention is created, the program capacity can be reduced. Furthermore, since it is composed only of relatively simple arithmetic processing, it is easy to create a program.
A security system having such advantages can be constructed. Furthermore, since the authentication algorithm is devised, since the authentication algorithm is composed of only simple processing steps, the program capacity of the authentication program to be created can be small. In addition, even if the data customer amount of the code data is small, it is possible to improve security and perform authentication that is not easily affected by light quantity fluctuation.

[0037]

As can be understood from the above description, according to the present invention, a photograph, a sign character, a real seal image, etc. for identifying an individual attached or engraved on a card are provided.
Read the visible information that identifies the individual as image information,
Based on the read image information, the number of pixels having a specific feature corresponding to the block or the number of feature points obtained corresponding to the pixel is calculated by the feature number calculating means with respect to the image information of each block divided into a plurality of blocks based on the read image information. Then, based on this, the correlation coefficient is calculated and authentication is performed, so that information for collation can be obtained by a process that is simpler than a calculation process of the entire image information. As a result, the load of arithmetic processing for obtaining collation information is small and high-speed processing is possible. In addition, since characteristic values are obtained for each block, the validity of validity judgment is improved, and high security can be secured. As a result, authentication with high confidentiality can be performed by simple arithmetic processing, and unauthorized use such as forgery of a card can be prevented, and security in authentication can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of an IC card authentication system to which a card authentication system of the present invention is applied.

FIG. 2 is a flowchart of a process for calculating the number of pixels having a specific feature corresponding to a block according to the present invention;

FIG. 3 is a flowchart of an authentication process.

FIG. 4 is an explanatory diagram illustrating a flow of an encoding process before distributing a card to a user;

FIG. 5 is an explanatory diagram illustrating a flow of an authentication process after a card is distributed to a user;

FIG. 6 is an explanatory diagram illustrating an example different from FIG. 5 in the flow of authentication processing after a card is distributed to a user.

FIG. 7 is an explanatory diagram showing the structure of code data according to the present invention.

FIGS. 8A and 8B are explanatory diagrams illustrating the correspondence between a zero-crossing image and block division. FIG. 8A is an explanatory diagram of the original image, and FIG. 8B is a diagram illustrating detection data divided into blocks. FIG. 4 is an explanatory diagram of an array variable to be stored.

FIG. 9 is an explanatory diagram for explaining a relationship between a two-dimensional array variable and an original image.

FIG. 10 is an explanatory diagram for explaining the calculation of a correlation coefficient Crrltn;

FIG. 11 is a zero-crossing image when the luminance is changed during imaging.

FIG. 12 is an explanatory diagram of a relationship between a variation in light amount of the entire original image and an inflection point (detection pixel position of zero crossing).

FIGS. 13A and 13B are explanatory diagrams for explaining a shift of an inflection point position due to a gradation portion. FIG. 13A is a diagram for explaining the relationship between the position of the background portion in the gradation and the luminance of the original image. () Is an explanatory diagram when the luminance changes.

FIG. 14 is an explanatory diagram of block division;
(A) is an explanatory diagram of an isometric division, and (b) is an explanatory diagram of a division including non-equal length portions with a background portion as a large block.

FIG. 15 is a graph illustrating a relationship between a luminance change and a correlation coefficient using the same face photograph.

[Explanation of symbols]

1 IC card, 2 authentication device, 3 IC module,
4: photo of face, 21: digital camera, 22: CPU, 2
3 memory, 25 printer, 26 IC card interface, 31 I / O port unit, 32 CPU, 3
3 ROM section, 34 RAM, 35 EEPROM, 3
5a: User inaccessible area, 35: User area, 3
6: authentication program, 37: collation reference data.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code FIG06K 19/00 S

Claims (4)

[Claims] (1) a photograph, a sign character,
A card having a surface of visible information for identifying the individual such as a real seal image on the surface, and the visible information is represented by nxm (n,
(m is an integer of 3 or more, respectively)) A reading device that reads from the card as multi-gradation image information by developing into pixels, and divides the one screen into a plurality of blocks based on the image information obtained from the reading device. The number of pixels having a specific feature corresponding to the block or the number of feature points obtained corresponding to the pixel is calculated as the feature number based on the multi-tone detection value of each pixel in each of the divided blocks. Coefficient calculating means for calculating, and a correlation coefficient calculating means for calculating a correlation coefficient between a reference feature number stored in advance for each block and a feature number corresponding to the block calculated by the feature number calculating means A card authentication system for authenticating the card based on the value calculated by the correlation coefficient calculation means. 2. The card authentication system according to claim 1, wherein the number of features is the number of pixels representing an outline. 3. The visible information has a target portion and a background portion, and the feature number calculating means determines whether or not any of the pixels in the image information is a zero cross point in relation to an adjacent pixel. A determination unit, wherein the number of features is the number of pixels indicating a zero-cross point determined by the zero-cross determination unit, and the correlation coefficient calculation unit compares the calculated value with a predetermined reference value of 1 or less. 3. The card authentication according to claim 2, further comprising an authentication unit that authenticates the card when the value is equal to or more than a predetermined reference value, wherein the size of the block is set to be larger in the background portion than in the target portion. system. 4. The card is an IC card, and the reference number of features corresponding to the block is the IC card,
4. The card authentication system according to claim 3, wherein the information is recorded as one of a magnetic stripe provided on the IC card and a barcode provided on the IC card.