KR100443142B1 - Apparatus and method for collating image - Google Patents

Abstract

The first image 1a (registered image) registered in advance in the image matching device is stored. With respect to the second image 1b (contrast image) input when the collation is performed, the device determines the identity of the second image with the first image. At this time, the device 3 defines a plurality of regions each having a predetermined positional relationship with respect to the first image. The device 4 searches from the second image an image area which has a maximum correlation with the image of each defined area. The identity 5 between the first image and the second image is determined according to the difference between the positional relationship of each image region defined in the first image and the positional relationship of the maximum correlation region searched included in the second image.

Description

Image Control and Method {APPARATUS AND METHOD FOR COLLATING IMAGE}

In the past, identity verification was performed through input of a password, fingerprint, etc. in order to prevent illegal access of confidential information. Passwords are likely to be stolen by others, while fingerprints are less likely to be stolen by others because they are characteristic features of the body. From this point of view, the contrast achieved by the fingerprint is expected to be high in safety.

Conventional fingerprint matching was performed according to the following procedure. (1) A process of obtaining a fingerprint contrasted by an image reader with diversified image data. (2) The process of binarizing diversified image data. (3) A process of extracting features such as branching points or endpoints of wrinkles constituting a fingerprint from the binarized data. This feature is referred to as "feature appearance" or "maneusha". (4) A process of comparing the positional relationship between the extracted feature contour and the feature contour of the pre-registered fingerprint, thereby determining the identity between the two.

However, when the bones of the wrinkles of the skin of the subject's finger are shallow, or the skin is too soft, wrinkles occur at the moment of the input of the subject placing his finger on the glass surface of the image reader. As a result, an unclear image is obtained, which may result in the subject not being proved. The percentage of people who have such wrinkles on their skin is only a few percent, but given the fact that the fingerprint matching failure rate is 0.1%, such as rejecting me or misrecognizing myself as someone else, this value is not negligible. Value.

This also happens to people who have cut or cut their fingers. Even when the finger is slightly injured around the feature appearance, the result is that the person cannot be recognized as the person.

In the conventional fingerprint matching method described above, even if the input fingerprint image data is clearly output, these data cannot be accurately matched with the original fingerprint of the finger.

The present invention relates to a technique for collating image data, in particular to a technique for collating an image, which is suitable for identifying a person and the collation is carried out using fingerprint image data.

1 is a block diagram illustrating the principles of the present invention;

2 is a block diagram showing the structure of a device for matching a fingerprint according to the first embodiment of the present invention;

3 shows a recording medium which stores a fingerprint matching processing program and is read by a computer;

4 is a flowchart illustrating a fingerprint matching process according to the first embodiment of the present invention;

5A and 5B show fingerprint images used to illustrate the fingerprint matching process;

6A and 6B show a memory map of RAM associated with the fingerprint matching process;

7 is a flowchart showing a fingerprint matching procedure according to a second embodiment of the present invention;

8A and 8B show fingerprint images used to explain the fingerprint matching process;

9A and 9B show fingerprint images used to explain the fingerprint matching process;

10A and 10B show fingerprint images used to describe the fingerprint matching process;

11A and 11B show fingerprint images used to describe the fingerprint matching process;

12 is a diagram for explaining a process of setting an effective area;

FIG. 13 is a flowchart showing a resetting process of a sub-template.

Accordingly, it is an object of the present invention to provide an apparatus and method for collating image data, which enables the execution of fingerprint collation or the like without being affected by the features of the skin of the finger, the presence of a wound, or the like.

The present invention provides an image contrasting device for calculating the identity between a pre-registered first image and a second image input when contrasted, including:

A definition unit defining a plurality of regions in the first image;

A search unit searching for a region having a maximum correlation with each of the plurality of regions defined by the definition unit as a maximum correlation region from the second image; And

A calculation unit for calculating the identity between the first image and the second image according to the difference between the positional relationship of each of the plurality of areas defined by the definition unit and the positional relationship of the maximum correlation area.

According to the present invention, there is provided a computer readable recording medium storing a computer program for calculating identity between a second image input at a moment of contrast with a first registered first image, the computer program comprising:

A definition program for defining a plurality of regions in the first image;

A search program for searching an area having a maximum correlation with each of the plurality of areas defined by the definition program as a maximum correlation area from a second image; And

And a calculation program for calculating the identity between the first image and the second image according to the difference between the positional relationship of each of the plurality of regions defined by the definition program and the positional relationship of the maximum correlation region.

According to the present invention, there is provided an image contrast method for calculating the identity between a second image inputted at the moment of collation with a first registered first image, including:

A defining step of defining a plurality of regions in the first image;

A search step of searching for a region having a maximum correlation with each of the plurality of regions defined by the defining step as a maximum correlation region from a second image; And

Calculating the identity between the first image and the second image according to the difference between the positional relationship of each of the plurality of regions defined by the defining step and the positional relationship of the maximum correlation domain.

Preferred embodiments of the image contrast device and method according to the present invention will now be described with reference to the accompanying drawings.

First embodiment

1 is a block diagram illustrating the principles of the present invention.

The first image data 1a is image data registered in advance in the database. In the finger matching system, the registered fingerprint corresponds to the first image data.

The second image data 1b is image data obtained by using an image reader or the like. In a fingerprint matching system, the fingerprint to be matched corresponds to the second image data.

The defining unit 3 is defined as a plurality of regions through the first image data 1a. The definition of each area is performed by installing a template having a predetermined shape at a predetermined position in the image data. Each template may be the same shape and size or may be different sizes and shapes. Each template can also be placed with the other partially overlapping the image.

The extractor 4 sets an image area having the same size and shape as corresponding to each template arranged in the first image by the definer 3 with respect to the second image. This moves the set image area. This finds the region in the second image that has the greatest correlation with the first image data in the corresponding template. The found area is called the "maximum correlation area".

The determination unit 5 examines the positional relationship of the template defined by the definition unit 3 and the positional relationship of the maximum correlation region found by the extraction unit 4. In a range of two different positional relationships, the determining unit 5 determines the degree of similarity or the degree of similarity between the first image and the second image data.

Although it has been mentioned that binarized data can also be determined with respect to the image data, diversified image data is more preferred for accurate determination of the contrast. Moreover, it is also possible for the first image to be set as the contrasting fingerprint and the second image to be set as the registered fingerprint in which the first image and the second image data are interchanged with each other.

The functions of the definition section 3, the extraction section 4, and the determination section 5 described above can be executed by a computer. An embodiment given in the form of a fingerprint matching device using a computer will be described. In this case, the fingerprint matching processing described later is prepared by a program executed by a computer. This program is then stored in the storage device of the computer and executed by the central processing unit.

2 is a block diagram showing the structure of a fingerprint matching apparatus according to a first embodiment of the present invention. The device comprises a CPU 11, a memory 12, a RAM 13, an image reader 14, a display 15 and an input 16 that are connected to each other via a bus line 17. Consists of

The CPU 11 controls the entire operation of the fingerprint matching device (hereinafter referred to as " original device ") while using the RAM 13 as an operation area according to the control program stored in the storage device 12 as the central processing unit. . The CPU 11 executes the control program to execute the fingerprint matching process described later.

The storage device 12 is composed of a ROM, a hard disk, or the like. The control program described above, which is read from the CPU 11 immediately after the power of the apparatus is turned on in the storage device 12, is stored in advance. The storage device 12 also stores in advance the registered fingerprint of the object corresponding to the first image data of FIG.

The RAM 13 is an operating memory used when the CPU 11 executes the control program described above.

The image reader 14 is a device for obtaining a fingerprint of a contrasting object corresponding to the second image data 1b of FIG. 1. Examples of the image reader 14 include an image scanner or an image sensor such as a charge coupled device (CCD).

The display unit 15 is a display device such as a cathode ray tube (CRT) or a liquid crystal display for displaying the contrast determination result. The determination result corresponds to the determination result 2 shown in FIG.

The input unit 16 is an input device such as a keyboard device as a means for causing a user of the apparatus of this apparatus to acquire a fingerprint to be contrasted with the CPU 11 or to start a fingerprint matching process described later.

Incidentally, in the present invention, the fingerprint matching program is stored in the storage device 12 mentioned above. The fingerprint matching processing program can also be made and arranged to be read by a reading device installed in a computer corresponding to the storage medium from the storage medium having the fingerprint matching processing program stored therein. For example, the fingerprint matching processing program is temporarily stored in the main memory, and the program is executed by the CPU.

3 shows an example of a computer readable storage medium stored in a fingerprint matching processing program. As shown, a storage device 22 such as a ROM or a hard disk device is used as the storage medium, and a floppy disk, magneto-optical disk (MO) and CD-ROM (CD) are used in the computer 21. -ROM) or a portable storage medium 23 such as DVD-ROM, a computer connected to the computer 21 via a network 24, a storage device 26 which is a part of a program server, or the like. Is installed.

Next, the control operation performed by the CPU 11 will be described.

Fig. 4 is a flowchart of a control process of the apparatus which recognizes the whole by execution of the above-described control program by the CPU 11, and shows the contents of the fingerprint matching process related to the present invention. In the figure, the processing contents executed in steps 11 (S11) and 14 (S14) correspond to the processes executed in the defining section 3 of FIG. In addition, the contents of the processing executed in steps S12, S13, and S16 correspond to the processing executed in the extraction unit 4. Moreover, the processing contents executed in steps 18 (S18) and 19 (S19) correspond to the processes executed in the determination unit 5.

5A and 5B show examples of fingerprint images used to describe the fingerprint matching process. FIG. 5A shows a registered fingerprint of an object stored in the storage device 12 in advance. 5B shows the fingerprint of the contrasting object and was sampled by the image reader 14. Further, all of the images in this figure are assumed to be diversified images, respectively.

6A and 6B are memory maps relating to the fingerprint matching process showing the usage state of the RAM (RAM) 13, and are used when the CPU 11 executes the fingerprint matching process.

The fingerprint matching procedure executed by the CPU 11 will be described below with reference to FIGS. 4 to 5A, 5B, 6A, and 6B.

When the start command of the fingerprint matching process is made in the input unit 16, the main template mt is first positioned in the registered fingerprint image " A " as shown in Fig. 5A to define a blind area on the image of the registered fingerprint. (Step 11 (S11)). The size of the main template mt is arbitrarily long as the contrast precision required for the collation between the obtained fingerprints.

Next, the maximum correlation area MT corresponding to the area indicated by the main template mt is searched from the fingerprint image “B” of the fingerprint to be contrasted. This search operation is executed as follows. The setting step is executed with a rectangular area having the same shape and size as the main template mt with respect to the image "B" of the control fingerprint. This rectangular area is then scanned two-dimensionally in pixels. Then, the mutual coefficient between the rectangular area and the area indicated by the main template mt is calculated at each moment of movement. The rectangular area where the correlation coefficient is located at the maximum is set to the maximum correlation area MT (step 12 (S 12)). The correlation coefficient calculation will be described below.

In this case, the determination is performed with coordinates in which the main template mt and the rectangular area MT each refer to respective specific positions in both images, that is, coordinates of each vertex in the upper left corner of both rectangles. These coordinates are set in the main template (x ₀ , y ₀ ) and the maximum correlation region (X ₀ , Y ₀ ), respectively. These are stored in predetermined areas of the RAM 13 shown in Figs. 6A and 6B, respectively (step 13 (S 13)).

Next, as shown in FIG. 5A, each of the lower templates st1 to st4 is positioned in the image “A” of the registered fingerprint, and each vertex of the rectangular main template mt is positioned at the center of the main template, respectively. Thereby defining a corresponding rectangular area (step 14 (S 14)). It is said that the position of the lower template can be arbitrary here. In addition, the number of lower templates is not limited to four, and this number is freely discretionary as the contrasting accuracy of the contrast between fingerprints obtained as necessary. In addition, as mentioned above, the size of the rectangular region defined herein is similarly free at its discretion.

In the same way in the case of rectangular area MT extraction, with respect to each of the lower templates st1 to st4 (hereinafter typically referred to as st _i (i = 1, 2, 3 and 4)), the lower template ( The rectangular area having the same size as st _i ) is set in the image "B" of the control fingerprint as shown in FIG. 5B. This rectangular area is then scanned two-dimensionally in units of pixels. And the correlation coefficient between this rectangular area | region and the area | region pointed out by the lower template st _i is computed at each instant which a movement took place. As a result, the rectangular area located where the correlation coefficient is maximum is set to ST _i . The determination is then carried out with the coordinates referring to each particular position on the image of the lower template st _i and the rectangular area ST _i , ie the coordinates of each vertex at the upper left corner of both rectangles. These coordinates are set to st _i (x _i , y _i ) and ST _i (X _i , Y _i ). These are stored in predetermined areas of the RAM 13 shown in Figs. 6A and 6B, respectively (steps 15 (S 15) to 17 (S 17)). When searching for the maximum correlation region corresponding to the lower template st _i , using the position of the maximum correlation region MT specified in advance for reference, it is predicted at the position where ST _i is supposed to exist. This searches for the location and the surrounding area of the location. This fully accomplishes the purpose.

Subsequently, with respect to the relative distance between the main template mt and st _i and the relative distance between the maximum correlation region MT and ST _i , the difference Δ _i between the two relative distances is determined by i ( i = 1, 2, 3 and 4) (step 19 (S19)).

(One)

This determines whether all calculated values of Δ _i fall below a predetermined value. If all values fall below a predetermined value, it is determined that the registered fingerprint and the control fingerprint match each other. Otherwise, it is determined that the enrolled fingerprint and the control fingerprint do not match. And the determination result is displayed on the display part 15 by this (step 19 (S19) and step 20 (S20)). The predetermined value used herein is to be actually calculating the Δ _i values from the entries of the sampled fingerprint image data from a plurality of people, and obtains a value that will produce the desired control precision, depending on the result of this calculation breaks.

The process described above is a fingerprint matching process. Although the shape of the main template mt and the lower template st _i is rectangular in the fingerprint matching process described above, the shape of these templates is not limited to being square. That is, each template may be in various shapes. Also, it is preferable that the size and shape of the main template mt and the maximum correlation region MT are the same as the size and shape of st _i and ST _{i in} the control decision. However, even if there is a very small difference between the two, this difference can be tolerated by the required degree of contrast accuracy of the other fingerprints obtained.

Furthermore, in addition to the calculation method of Δ _i in the above-described processing, for example, it is executed based on the difference in shape or area between the contour formed using st _i as a vertex and the contour formed using ST _i as a vertex. It is also possible to adopt various methods such as making a decision.

Next, the correlation coefficient calculation used in step 12 (S 12) or step 16 (S 16) of the fingerprint matching procedure shown in FIG. 4 will be described. Here, the calculation of the correlation coefficient between the rectangular area RA and the rectangular area RB is described.

First, RA (i, j) and RB (m, n) represent pixels included in each of the provided rectangular area RA and rectangular area RB. It is also assumed that the total sum of the pixels included in the rectangular area RA and the rectangular area RB is the same. X _ij and Y _mn each represent a signal amount, which is a diversification value indicating the thickness of the corresponding pixel, respectively.

The equation for the generalized signal quantity Z _pq is

(2)

In the above equation, N represents the total number of pixels included in the corresponding rectangular area. In the above equation,? Represents the sum of signal intensities for all the pixels included in the corresponding rectangular areas. That is, the equation shows an average value of signal strengths for the pixels included in the corresponding region.

Next, the following equation is also defined.

(3)

The equation is the square root of the signal strength for the pixels contained in the corresponding rectangular regions.

Here, the correlation coefficient C _AB between the rectangular area RA and the rectangular area RB can be calculated using the following equation expressed using the equation described above.

(4)

here

Calculation of correlation coefficients between regions is performed using the above equation.

It calculates the correlation coefficient using the above equation instead of calculating the correlation coefficient using the signal strengths of all the pixels in the rectangular area. For example, the following calculation is performed. That is, a calculation performed using only pixels arranged in rows forming all rows in a rectangular area, a calculation performed using only pixels included in a part of a rectangular area, or a pixel selected by arbitrarily decreasing from within a rectangular area. Calculations to be executed are executed. Even when these calculations are performed, it is not a problem if the desired level of contrast accuracy of the fingerprint is obtained along with the others. By using this calculation process, the number of pixels used as objects related to the correlation coefficient calculation is reduced. As a result of this, the amount of calculation is reduced. So this calculation process is useful. In addition, other correlation coefficient calculation methods may be employed for the fingerprint matching procedure described above.

Another embodiment of an image contrast device and method according to the present invention will be described. The same parts as in the first embodiment are denoted by the same reference numerals, and their detailed description is omitted.

Second embodiment

In fingerprint matching, the object rests a finger on the fingerprint reader of the fingerprint matching device and has the contrasting fingerprint read with this fingerprint reader. When the object is out of the proper position of the finger, this results in an inappropriately wrong decision result.

In the above description, a matching method in which the matching can be made correctly when the target finger slightly deviates from the proper position at the time of sampling the control fingerprint will be described now as the second embodiment.

In this embodiment, it is assumed that both the main template and the sub-template are the same shape and size, and the sub-templates are respectively positioned to contact the main template and the corresponding contact.

Since the structure of this system is the same as that of the first embodiment, description thereof will be omitted. The processing operation will be described below with reference to FIGS. 7 to 13.

When the fingerprint matching process start command is made in the input unit 16, in step 101 (S 101), the main template mt representing the blind area is located in the registered fingerprint image " A " as shown in Fig. 8A. The size and shape of this main template mt is arbitrary as long as the required level of fingerprint matching precision is obtained along with the others. (X ₀ , Y ₀ ) is assumed to represent coordinates specifying the position of the registered image “A” of the main template mt defined herein as described above, where the weight of the main template mt The center coordinate.

Then, as shown in FIG. 8A, the sub-template st _i , _i ≧ ₁ defining a plurality of rectangular areas different from that defined by the main template mt in the image “A” registered in step 102 (S 102). ) Is located. In the example shown in FIG. 8A, i = 4 is set. The four rectangles, the lower templates st ₁ to st ₄ , are defined by the positional relationship shown in FIG. 8A in a manner surrounding the main template mt. In addition, the size and shape of these sub-templates st _i can be as arbitrary as the required level of fingerprint matching precision is obtained along with the others. (x _i , y _i ) represents the coordinates specifying the position in the registered image "A" of the sub-template st _i as defined above, where the weight of the sub-template st _i , which is square The center coordinate.

Next, in step 103 (S 103), a rectangular area having the same shape and size as the main template mt is set in the contrasting fingerprint image " B ". While the rectangular area is scanned in the image "B", the calculation is performed one by one with a correlation coefficient between the rectangular area and the area of the image "A" specified by the main template mt.

As a result of the successful correlation coefficient calculation, the decision in step 104 (S 104) is made up of the rectangular region described above, and the correlation coefficient between the rectangular region and the main template mt is considered to be maximum. This rectangular area is set to the area MT. Coordinates specifying the position of the image "B" of the area MT. Coordinates of the area MT corresponding to the coordinates (x ₀ , y ₀ ) of the main template mt are represented by (X ₀ , Y ₀ ). .

Then in step 105 (S 105), it is determined whether the area MT belongs to the effective area E set in the image " B ". Specifically, it is determined whether the coordinates X ₀ , Y ₀ specifying the position of the area MT are included in the range E _{x, y} represented by the effective area E. FIG. If the result of this determination is "yes", the flow proceeds to step 107 (S 107), while if "no", the flow proceeds to step 106 (S 106).

Here, the "effective area E" represents a range in the image "B", in which the area MT may exist, in which the maximum correlation area corresponding to the lower template st _i necessary for contrasting is obtained. can do. In the above embodiment, the range E _{x, y} represented by the following equation is set in the image "B".

(5)

here

FIG. 8B shows the range E _{x, y} of the effective area E in the case of having the main template mt and the lower template st _i defined in the registered image “A” as shown in FIG. 8A. The situation set in "B" is shown. In the equation, W _i and H _i are represented by the width of the lower template st _i (length in the x-axis direction, WIDTH in the equation) and height (length in the y-axis direction, HEIGHT in the equation). The width and the width represent the width (length in the x-axis direction) and the height (length in the y-axis direction) of each image "A" and the image "B" each having the same size. Also, x _i SEARCH and y _i SEARCH represent a range for searching for the maximum correlation region corresponding to the lower template st _i located in the image “A” in the image “B”. For each of these, for example, values corresponding to four to six pixels or values obtained in pixel units of image data are used. Also, the maximum value MAX [p, q] appears as a function for obtaining the larger of the p value and the q value (if the two values are the same, choose one of them).

A further explanation of the equation given above. 12 shows a process of setting the effective area E. As shown in FIG. FIG. 12 shows only the top surface of the image " A " for the purpose of explaining [alpha] _x and [beta] _x in the above formula (4). It is a brief description that assumes that the lower template (st ₃ , st ₄ ) is excluded from consideration.

Referring to equation (4) and FIG. 12 above, it is simply understood that α _x = L ₁ , β _x = width − L ₂ . The following can be explained by setting the effective area E shown on the x-axis using the values of α _x and β _x determined from the above equation. When each of the lower templates st ₁ and st ₂ are defined in such a manner as to be wider in the x-axis direction, the area MT is defined when the range of the effective area E, that is, the positional relationship described above is maintained. The range that can exist is narrow. On the other hand, when each of the lower templates st ₁ and st ₂ is defined in such a manner that they are located in the narrow direction in the x-axis direction, the range of the effective area E is widened.

Returning to the description of FIG. 7, reference will be made to step 105 (S 105) described above. In this step 105 (S 105), if the area MT is defined outside the effective area E, i.e., the image corresponding to the main template mt in the image "A" registered in the relationship between Figs. 9A and 9B. If the area MT in " B " is located outside the range E _{x, y} of the effective area E, the following process is executed. In other words, in step 106 (S 106), the lower template st _i is placed in the registered image " A " again after its position correction. Relocation of the lower template st _i is performed in accordance with the procedure shown in FIG.

Figure 13 shows the progress of the relocation process of relocating the lower template. This relocation process is a process of relocating the position of the lower template specified by the coordinates (x _i , y _i ) to the position specified by the coordinates (xr _i , yr _i ).

The process given is explained as follows.

First, it is determined whether or not the area MT is in the range of the effective area E with respect to the x-axis (step 201 (S 201) and step 203 (S 203)).

As a result of the determination, if the x coordinate (X ₀ ) specifying the location of the area MT is smaller than α _x representing the range shown on the x-axis of the effective area E, the contrasting fingerprint “B” is left Since they are arranged side by side, the x coordinate specifying the position of the lower template st _i is set to xr _i = x _i + (α _x -X ₀ ) and moved to the right (step 202 (S 202)). If the x coordinate (X ₀ ) is larger than β _x indicating the range shown in the x-axis direction of the effective area E, since the contrasting fingerprint “B” is located to the right, the lower template st _i The x coordinate specifying the position is set to xr _i = x _i- (X ₀ -β _x ) and moved to the left (step 204 (S 204)).

Here, if both of the determination results are outside, that is, if the value of X ₀ is larger than α _x and smaller than β _x , xr _i = x _i is set. That is, the position of the lower template st _i is maintained at the original value. The flow then proceeds to step 206 (S 206).

Then, along the y axis, it is similarly determined whether the area MT lies within the range of the effective area E (step 206 (S 206) and step 208 (S 208)).

The procedure executed according to the result of this determination is the same as the procedure described above executed along the x axis. If the y coordinate Y ₀ specifying the position of the area MT is smaller than α _y representing the range shown on the y axis of the effective area E, then the position of the lower template st _i after repositioning is determined. The y coordinate to be specified is set to yr _i = y _i + (? _Y- Y ₀ ) (step 207 (S 207)). And, if y coordinate Y ₀ is larger than β _y representing the range shown on the y axis of the effective area, yr _i = y _i − (Y ₀ − β _y ) is set (step 209 (S 209)). ). If the decision result is outside of both cases, that is, if the value of Y ₀ is greater than α _{y and} less than β _y , it is set to yr _i = y _i . That is, the position of the lower template st _i is maintained at the original value (step 210 (S 210)).

According to the above described process, the relocation is executed all the sub templates st _i predefined in the registered image "A". FIG. 10A shows the relocation result performed when the area MT of the image “B” lies outside the range E _{x, y} of the effective area E to the right as in FIG. 9B. In particular, FIG. 10A illustrates the result when the predefined lower templates st ₁ to st ₄ are moved to the left and rearranged as described above as shown in FIG. 8A. In addition, FIG. 10B illustrates an area MT corresponding to the main template mt in FIG. 10A. Since the position of the main template mt may not be moved, the position of the region MT is not moved from the position shown in FIG. 9B.

7 again, when the result of the decision process in step 105 (S 105) is "yes", that is, when repositioning of the lower template st _i is not necessary, or of the lower template st _i . After the relocation is executed through the process of step 106 (S 106), the next successful calculation is executed in step 107 (S 107). That is, in step 107 (S 107), a rectangular area having the same shape and size as the lower template st _i is set in the image "B". While this rectangular area is scanned on the above image "B", the correlation coefficient of this area and the lower template st _i is calculated one by one. Although the correlation coefficient calculation method uses the one used in the above-described step 103 (S 103), other calculation methods relating thereto may be applied.

As a result of the continuous calculation of the correlation coefficient in step 108 (S 108), the rectangular area is determined according to the correlation coefficient which is the maximum between the correlation coefficient and the lower template st _i . This rectangular area is set to the area ST _i .

Next, in step 109 (S 109), the area ST _i determines whether or not the detection relating to all the lower templates st _i is completed. If the determination result is "yes", proceed to step 110 (S 110). On the other hand, if the determination result is "no", the flow returns to step 107 (S 107) in which the process of determining the region ST _i described above is repeatedly executed. 11A and 11B show corresponding regions ST ₁ to ST ₄ and main templates mt and corresponding regions MT obtained in relation to all rearranged lower templates st ₁ to st ₄ . will be.

In step 110 (S 110), the position distribution of the main template mt and the lower template st _i of the image data “A” is determined by the position distribution of the region MT and the region ST _i in the image data “B”. Are compared. This comparison evaluates the identity between image "A" (registered fingerprint) and image "B" (reference fingerprint). The identity evaluation performed in this step is sufficiently the same as the process performed in step 19 (S 19) of FIG. 4 described as the first embodiment. When evaluating identity, the main template mt and the area MT may be excluded from the evaluated object. Also, by excluding this, the evaluation can be performed only by comparing the position distribution of the lower template st _{i with} the position distribution of the region ST _i .

The fingerprint matching procedure described so far is executed through the operation of the CPU 11, whereby fingerprint matching is executed in the apparatus.

Further, although in the above-described embodiment of the present invention, each main template mt, sub-template st _i , region MT, and region ST _{i are} formed in a square shape, the shape of these templates and regions Is not limited to squares. This shape can be anything you want. In addition, the size and shape of the main template mt and the region MT may be the same, and the size and shape of the lower template st _i and the region ST _i may be the same. However, even when there is a slight difference between the two, this difference is tolerated by the required level of contrast accuracy of one fingerprint with the other fingerprint obtained.

In addition, the positional relationship of the main template mt with respect to the two predefined templates, the lower template st _i , or vice versa, was determined in each embodiment of the invention described above as follows. That is, the two templates are located in a positional relationship where the main template mt and the lower template st _i contact each other, as shown in FIG. 8A. However, the two templates are also defined by and located in a positional relationship where the two partially overlap one on top of the other, or where the two are completely apart from each other.

Further, the area MT is placed outside the effective area E, and the progress is rearranged by moving the lower template st _i by the same distance in the lower template rearrangement process shown in FIG. However, considering that the area MT deviates from the effective area E in the upper, lower, left and right directions in the image “B” in the effective area E, the moving distance of one lower template st _i is adjusted. It may be arranged in a wrong way from the moving distance of another lower template coinciding with the above direction from its original position to its repositioning. As a specific example, after relocation, it will be the case to determine the coordinates (xr ₁ , yr _i ) that specify all positions of the lower templates (st ₁ to st ₄ ) defined in FIG. 8A.

First with respect to the x-coordinate, xr _i is derived from the coordinate X ₀ relationship of x specifying the position of the area MT with α _x and β _x described above representing the range of the x-axis of the effective area E as follows. Is determined.

When X ₀ <α _x

xr _i = x _i + (α _x -X ₀ ) (i = 2, 3)

xr _i = x _i + (α _x -X ₀ ) × u (i = 1, 4)

Where u represents a constant in the range 0 ≦ u <1.

When α _x ≤X ₀ ≤β _x

xr _i = x _i

when β _x <X ₀

xr _i = x _i- (X ₀ -β _x ) (i = 1, 4)

xr _i = x _i- (X ₀ -β _x ) × u (i = 2, 3)

Where u represents a constant in the range 0 ≦ u <1.

Next, with respect to the y coordinate, yr _i is next from the coordinate Y ₀ relationship of x specifying the position of the area MT with α _y and β _y described above representing the range of the y-axis of the effective area E. Is determined as follows.

When Y ₀ <α _y

yr _i = y _i + (α _y -Y ₀ ) (i = 1, 2)

yr _i = y _i + (α _y -Y ₀ ) × v (i = 3, 4)

Where v represents a constant in the range 0 ≦ v <1.

When α _y ≤Y ₀ ≤β _y

yr _i = y _i

when β _y <Y ₀

yr _i = y _i- (Y ₀ -β _y ) (i = 3, 4)

yr _i = y _i- (Y ₀ -β _y ) × v (i = 1, 2)

Where v represents a constant in the range 0 ≦ v <1.

In addition, in the case of executing the rearrangement of the lower template according to the above equation, especially if the rearrangement is performed under the condition of u = v = 0, the following advantages will be obtained.

That is, if so, even when the area MT is located outside the effective area E, the fingerprint matching can be performed without the relocation portion of the lower template according to the positional relationship of the two areas. That is, it is possible to have a sub-template that retains when it is first defined. In this case, the moment required for the rearrangement process of the lower template is shortened, and as a result, the process moment required for the fingerprint matching process is shortened.

Claims (18)

By calculating the identity with the second image input at the moment contrasted with the first registered first image: A definition unit formed to define a plurality of regions in the first image; A search unit configured to search for a maximum correlation region having a maximum correlation with each of the plurality of regions defined by the definition unit from the second image; And Calculating the identity between the first image and the second image according to the difference between the positional relationship of each of the plurality of regions defined by the definition unit and the positional relationship of each of the plurality of maximum correlation regions searched by the search unit Computational unit configured to; configured to. The method of claim 1, The definition unit defines a reference region and a plurality of subregions each having a predetermined positional relationship with the reference region; The searcher first searches for a first maximum correlation region having a maximum correlation with the reference region from a second image, and then searches for each of the subregions according to a predetermined positional relationship with respect to the first maximum correlation region. An image contrasting device for searching for a second associated maximum correlation area. The method of claim 2, A detector formed to detect whether the first maximum correlation region is in a predetermined relationship with the second image; And And a redefinition unit configured to redefine the plurality of regions defined by the definition unit when the detector detects that the first maximum correlation region is in a predetermined relationship to the second image. The method of claim 3, wherein The detector detects the first maximum correlation region disposed from the second image in an amount greater than a predetermined amount; And The redefinition portion moves the subarea based on the amount of placement of the first maximum correlation area, whereby the plurality of subareas is detected when the first maximum correlation area is disposed from the second image in an amount greater than a predetermined amount. Image contrast device to redefine. The image contrast device of claim 1, wherein the first image is a fingerprint image of an object and the second image is a contrasting fingerprint image. The apparatus of claim 1, wherein the searcher sets the plurality of areas corresponding to the first image defined by the definer to the second image area, and the data of the plurality of areas defined by the definer. And calculating a correlation coefficient using data of a plurality of corresponding regions while moving the corresponding region over the second image, thereby searching for the maximum correlation region, the maximum correlation coefficient. By storing a computer program for calculating the identity between the second image input at the moment contrasted with the first registered first image, said computer program: A definition program for defining a plurality of regions in the first image; A search program for searching from the second image a maximum correlation region each having a maximum correlation with each of the plurality of regions defined by the definition program; And And a calculation program that calculates an identity between the first image and the second image according to a difference between positional relationships of each of the plurality of maximum correlation regions searched by the search program. The method of claim 7, wherein The definition program defines one reference region and a plurality of subregions each having a predetermined positional relationship with the reference region; The search program is a program for searching for a first maximum correlation region having a maximum correlation with the reference region and a second maximum correlation region with respect to each subregion according to a predetermined positional relationship with respect to the first maximum correlation region. Recording medium for retrieving a code from the second image. The method of claim 8, A detection program for detecting whether the first maximum correlation region is in a predetermined relationship with the second image; And The detection program further comprises a redefinition program that redefines the plurality of areas defined by the definition program when detecting that the first maximum correlation area is in a predetermined relationship with the second image. The method of claim 9, The detection program detects the first maximum correlation region disposed from the second image in an amount greater than a predetermined amount; And The redefinition program moves the subregion based on the amount of placement of the first maximum correlation region, whereby the first maximum correlation region arranged in a larger amount than the predetermined one is detected from the second image. When redefining the plurality of sub-areas. The method of claim 7, wherein Wherein said first image is a fingerprint image of an object and said second image is a contrasting fingerprint image. The apparatus of claim 7, wherein the search program sets the plurality of areas corresponding to the first image to the second image by the definition program, and the data and data of the plurality of areas defined by the definition program. A recording medium for calculating a correlation coefficient using data of the plurality of corresponding regions while moving the corresponding region over a second image, thereby searching for the maximum correlation region, the maximum correlation coefficient. By calculating the identity between the second image input at the moment contrasted with the first registered first image: A defining step formed to define a plurality of regions in the first image; A search step formed to search from the second image a region having a maximum correlation with each of the plurality of regions defined by the defining step as a maximum correlation region; And The identity between the first image and the second image is dependent on the difference between the positional relationship of each of the plurality of regions defined by the defining step and the positional relationship of each of the plurality of maximum correlation regions searched by the searching step. An image contrast method that constitutes a calculation step formed to calculate. The method of claim 13, The defining step defines one reference region and a plurality of subregions each having a predetermined positional relationship with the reference region; The searching step may first search for a first maximum correlation region having a maximum correlation having the reference region from the second image, and then refer to the first maximum correlation region to determine each of the respective positions according to a predetermined position relationship. An image contrast method for searching for a second maximum correlation region associated with a subregion. The method of claim 14, A detection step formed to detect whether the first maximum correlation region is within a predetermined relationship with the second image; And The image contrast further comprising a redefinition step formed to redefine the plurality of areas defined by the defining step when the detecting step detects the first maximum correlation region having the predetermined relationship with the second image. Way. The method of claim 15, The detecting step detects the first maximum correlation region disposed from the second image in an amount greater than a predetermined amount; And The redefining step moves the subregion based on the placement amount of the first maximum correlation region and thereby detects the first largest correlation region disposed from the second image in a larger amount than the predetermined one. And image redefinition redefining the plurality of sub-areas. The method of claim 13, wherein the first image is a fingerprint image of an object and the second image is a contrasting fingerprint image. The method of claim 13, wherein the searching comprises setting the plurality of areas corresponding to the first image to the second image by the defining step, and the data of the plurality of areas defined by the defining step and the data. And a correlation coefficient is calculated using the data of the plurality of corresponding regions while moving the corresponding region over the second image, thereby searching for the maximum correlation region, the correlation coefficient which is the maximum.