JPS6349989A - Fingerprint collating device - Google Patents

Abstract

PURPOSE:To collate fingerprints with high positioning accuracy by determining a superposed position between a collating fingerprint and a registered fingerprint based on line number distribution and line width distribution obtained from a binary graphic pattern. CONSTITUTION:A collating pattern obtained from a reading part 11 is stored in a collating pattern memory 12. Registered patterns are previously stored in a registered pattern memory 13. The memories 12, 13 are scanned by a raster method in 1st and 2nd directions and the line number distribution and line width distribution of both the patterns are respectively extracted by a line number distribution extracting part 14 and a line width distribution extracting part 15. The line number distribution is compared with the line width distribution by a reference point extracting part 16 and a superposed position of both the patterns is extracted. A coincidence deciding part 17 superposes the collating pattern to the registered pattern based on the superposed position, finds out the degree of coincidence of both the patterns and decides the coincidence or discrepancy of both the patterns based on the degree of coincidence.

Description

[Detailed description of the invention] (Industrial application field) This invention relates to a fingerprint verification device.

(Prior Art) Conventionally, in automatic verification of seal stamps, fingerprints, etc., for example, a reference point such as the center coordinates of the object to be verified has been detected for the purpose of performing verification accurately and quickly.

For this purpose, various devices have been proposed for detecting the center of a figure such as a seal imprint. A figure recognition device suitable for use is disclosed.

The device described in the above-mentioned publication includes a reading section that reads a figure such as a seal imprint using a raster scanning method and provides a binary electric signal corresponding to a black point and a white point at each coordinate, and a reading section that reads a figure such as a seal imprint and provides a binary electric signal corresponding to a black point and a white point at each coordinate. It includes a fill pattern creation section that fills in the interior, and a center coordinate detection section that determines the barycenter coordinates of each weighted point of the filled pattern as the center coordinates of the figure. According to this device, the center point of the graphic pattern is detected by determining the center of gravity after filling the graphic pattern, so detection errors caused by fading or staining of the seal imprint pattern or the graphic pattern, unevenness of the internal pattern, etc. are extremely reduced. Therefore, it was possible to detect the center point coordinates with high accuracy.

(Problem to be Solved by the Invention) However, in the conventional device having the above-mentioned configuration, when trying to find the center point of a figure such as a fingerprint that does not have an outwardly convex closed curve, There was a problem that point detection accuracy decreased.

To be more specific, for example, stripes (line patterns) like a fingerprint.
When there is a break in one of the outermost lines of a geometric figure, or when one of the lines is concave inside the figure, it may have the same effect as when one of the lines is missing. In these cases, the center of gravity may shift due to these effects, which reduces the accuracy of detecting reference points such as the center point, making it difficult to accurately align the fingerprints to be compared. It turns out you can't do it.

An object of the present invention is to solve the above-mentioned problems and provide a fingerprint verification device that can perform fingerprint verification with high alignment accuracy.

(Means for Solving the Problems) In order to achieve this object, the fingerprint matching device of the present invention includes a reading section that reads a fingerprint to be matched as a binary matching pattern, and a matching pattern memory that stores this matching pattern. , The registered pattern memory that stores the fingerprints registered in advance as the aforementioned binary registered pattern, and the aforementioned verification pattern memory and registered pattern memory are scanned in a raster method in each of the first direction and the second direction. a line number distribution extraction unit that obtains the number of lines of both of these patterns for each scan and extracts the line number distribution of each of the above-mentioned first and second directions of both patterns; and the above-mentioned matching pattern memory and registered pattern memory. is scanned in a raster method in each of the first direction and the second direction, and the maximum line width of these two patterns is determined for each scan, and the line width distribution of these two patterns in each of the first and second directions is calculated. and the line width distributions of the matching pattern and the registered pattern in each of the first and second directions and comparing the line width distributions and the number of lines/ji same'' respectively. and a reference point extraction unit that extracts the overlapping position of both patterns, and a reference point extracting unit that overlays the above-mentioned matching and registration patterns based on the above-mentioned overlapping position to find the degree of coincidence between both patterns, and extracts the degree of coincidence between the two patterns based on this degree of coincidence. and a match determination unit that determines whether or not they match.

In carrying out the present invention, the reference point extraction section described above calculates the minimum dissimilarity obtained by comparing the line number distributions and the line width distributions in the first direction of the two patterns, respectively. Preferably, the two patterns are determined to be mismatched when at least one of the minimum dissimilarity values obtained by comparing the line number distributions and the line width distributions of the patterns in the second direction is larger than a predetermined value. be.

In carrying out the present invention, it is preferable that the above-mentioned line width is obtained by determining one of the above-mentioned binary values and the portion of the other value existing between the one value as one line width.

In carrying out the present invention, it is preferable that the number of lines described above is obtained by determining the portion of the other value between the two values 1 and 2 as one number of lines.

Furthermore, in carrying out the present invention, when the number of lines Q or the number of the other value existing between one of the two values is within a predetermined range. It can also be found as the number of lines in the tree.

(Function) According to such a configuration, the overlapping position of the verification fingerprint and the registered fingerprint is determined based on the line number distribution and the maximum line extracted from the respective binary graphic patterns stored in the verification and registration pattern memories, respectively. Determined based on width distribution. In this way, for example, even if one of the lines constituting a striped (line pattern) figure is recessed inside the figure, the number and width of this line can be confirmed. In addition, even if there is a break in one of the lines, the influence of that part on determining the overlapping position will be small, so both patterns can be overlapped quickly and with a high degree of M. After that, verification of stiffness will be carried out.

(Embodiment) Hereinafter, an embodiment of the fingerprint matching device of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the fingerprint matching device of the present invention.

In FIG. 1, reference numeral 11 indicates a reading unit, and this reading unit 11 optically reads the fingerprint to be compared, for example, as a binary graphic pattern (this will be referred to as a matching pattern). For example, a figure is converted into a binary electrical signal (also referred to as a black and white pattern for convenience of explanation) and output to the matching pattern memory 12.

Further, this matching pattern memory 12 is for storing this black and white pattern, and although not shown in the figure, in this case, it is used to specify addresses in the first direction (X direction) and second direction (Y direction) of the memory, respectively. The first and second nearest address counters are provided.

Reference numeral 13 indicates a registered pattern memory for storing previously registered fingerprints, and this has the same configuration as the verification pattern memory 12.

Reference numeral 14 indicates a line number distribution extraction unit, and in this embodiment, the first line number distribution extraction means +4a and the second line number distribution extraction means! 4b, and the first and second line number distribution extraction means +4a, 14b, although not shown, include a black dot line width counter, an X direction line number distribution counter Xtlist, and a Y
Each of them is provided with a direction line number distribution counter YHisj (details will be described later). In the case of this embodiment, the first line number distribution extraction means 14a scans the matching pattern memory 12 in each of the X direction and the Y direction using a raster scanning method. The pattern is read out line-by-line in each of the X and Y directions, and the number of lines for each scan is determined by counting the line segments by a method described later, thereby creating a line number distribution for each of the X and Y directions. On the other hand, the second line number distribution extraction means 14b uses the registered pattern memory 1
2 in each of the X and Y directions using a raster scanning method, and create the line number distribution of the registered pattern in each of the X and Y directions in the same way as the line number distribution of the matching pattern is created by the first extraction unit 14a. do.

Reference numeral 15 denotes a line width distribution extraction section, which in this embodiment is composed of a first line width distribution extraction means 15a and a second line width distribution extraction means 15b. Although not shown, the extraction means 15a and 15b each include a black dot line width counter, an X-direction line width distribution register X'1VDT, a Y-direction line width distribution register YWDT, and a line width maximum value register BMAX. (Details will be described later.)

In the case of this embodiment, the first line width distribution extraction means 15
a scans the verification pattern memory 12 in each of the X and Y directions using a raster scanning method, that is, reads out the verification patterns stored in the verification pattern memory 12 line-sequentially in each of the X and Y directions, as described below. The line widths of the line segments on the scanning line are counted to find the maximum line width for each scan, thereby creating line width distributions in each of the X and Y directions.

On the other hand, the second line width distribution extraction means 15b scans the 13 pattern memory 13 in each of the X and Y directions using a raster scanning method, and the first line width distribution extraction means 15a creates the line width distribution of the matching pattern. Similarly, the registered pattern
and create a line width distribution for each Y direction.

Reference numeral 16 indicates a reference point detection unit, and this reference point detection unit 16 detects each of the matching patterns and registered patterns created by the line number distribution extraction unit 14, in this case the first and second line number distribution extraction units 14a and 14b, respectively. Comparing the line number distribution in each of the X and Y directions, and comparing the matching pattern and registered pattern created by the line width distribution extraction unit 15, in this case, the first and second line width distribution extraction units 15a and +5b, respectively. The respective line width distributions are compared in each of the X and Y directions to determine the degree of difference between the two patterns, and based on this degree of difference, a reference point for matching and overlapping the registered patterns is extracted.

Reference numeral 17 denotes a match determination unit, which superimposes the matching pattern and registered pattern stored in the matching pattern memory 12 and the registered pattern memory 13, respectively, based on the reference point obtained by the reference point extraction unit 16, and determines whether the two patterns match. The degree of coincidence is extracted, for example, by a method described later, and it is determined whether both patterns are fingerprints obtained from the same finger based on the value of the degree of coincidence.

Voice Cloth 1tJ Tongue Day Next, the operation of the line number distribution extraction section will be explained.

In the case of this embodiment, the line number distribution extraction unit is explained as an example configured with first and second line number distribution extraction means 14a and 14b. Since the matching pattern memory has the same configuration except that the other one is a registered pattern memory, the first line number distribution extraction means 14a will be described with reference to the flowchart shown in FIG.
The operation will be explained, and the explanation of the second line number distribution extraction means 14b will be omitted.

When scanning the verification pattern memory 12, first, as shown in FIG. 3(A), the Y direction is the main scanning direction and the X direction is the sub-scanning direction. This will be explained using an example of counting numbers.

An initial value O is set in the X-direction line number distribution counter X+1ist provided in the first line number distribution extraction means 14a (step 21).

The X-direction starting point coordinate value x3 of the verification pattern memory 12 is set in the first address counter X of the verification pattern memory 12 (step 22).

A black dot line width counter (sometimes abbreviated as CNT) provided in the first line number distribution extraction means 14a is set to O (
Step 23).

The Y-direction starting point coordinate value Y3 of the verification pattern memory 12 is set in the second address counter Y of the verification pattern memory 12 (step 24).

After performing these initial settings, check pattern memory 1
The binary signal stored at point 2 (XS, YS) is read out, and if the signal is a black point, 1 is added to CNT (step 25.26). Also, if the point (xs, ys) is not a black point and the current CNT value is not 0, then
Hist (X) In this case, add 1 to list(Xs) (step 25.27), and finally CNT
is set to 0 (steps 25, 27, 28°29). Next, 1 is added to the coordinate Ys (step 30), and steps 25 to 31 are sequentially repeated until the value of the address counter in the Y direction exceeds the end point coordinate YE (step 30.31). In other words, in this example, the coordinates
When scanning the verification button memory 12 in the Y direction focusing on s, if a black point appears after a white point, 1 is added to CNT, and if black points continue, 1 is added to CNT, and this is repeated. Count the number of sunspots. This continuous portion of black dots is determined to be one line forming the matching pattern, and if a white dot appears again on the main scan, it is determined that the line segment has ended and CNT is initialized to 0. , and XHist(Xs
) is added by 1. By performing this process on the scanning line at the coordinate xS, the number of lines on the scanning line in the Y direction at the coordinate X3 can be extracted.

Next, when the Y coordinate value exceeds YE, the current CNT
If the value of is 0, add 1 to XHist (Xs), add 1 to the coordinate value
The processes of steps 23 to 35 are repeated until the value exceeds the value (steps 34 and 35).

By performing such processing, it is possible to obtain the number of lines after scanning in the Y direction for each X coordinate, and therefore it is possible to extract the number of lines corresponding to the number of X;A in the X direction.

Next, as shown in FIG. 3(B), the number of lines in the Y direction is determined by counting the number of lines for each Y coordinate of the matching pattern memory 12, with the X direction as the main scanning direction and the Y direction as the sub scanning direction. Extract the distribution (steps 41-55). Since this process can be performed basically in the same procedure as for obtaining the X-direction line number distribution described above, the explanation thereof will be omitted.

The second line number distribution extraction means 14b can be used to obtain the line number distribution of the registered pattern in the same manner as described above.

Shou-kun 1゜:0 Next, the operation of the line width distribution extraction section will be explained.

In the case of this embodiment, the line width distribution extracting section is explained as an example constituted by first and second line width distribution extracting means 15a and 15b, but these extracting means have only one target memory. has the same configuration except that the verification pattern memory and the other are registered pattern memories, so please refer to the flowchart shown in FIG.
5a will be explained, and the second line width distribution/extraction means 15 will be explained.
The explanation of b is omitted.

In addition, when scanning the verification pattern memory 12, first, as shown in FIG. 3(A), with the Y direction as the main scanning direction and the This will be explained using an example of counting line width.

The X-direction starting point coordinate value xs of the verification pattern memory 12 is set in the first address counter X of the verification pattern memory 12 (step 61).

The black dotted line width counter (
It is sometimes abbreviated as CNT. ) is set to 0 (step 62).

An initial value of 0 is set in the line width maximum value register BMAX, which stores the maximum value of line width that occurs on a scanning line when one scan is performed in the Y direction.
is set (step 63).

The Y-direction starting point coordinate value YS of the verification pattern memory 12 is set (step 64).

After performing these initial settings, check pattern memory 1
Read the binary signal stored at point 2 (XS, Ys), and if the signal is not a black point, turn the CNT
Then, if the signal is a black point, add 1 to CNT, and then add 1 to CN.
If the value of T is greater than the value stored in BMAX, the value of CNT is stored in BMAX (step 65.
67-69). Next, add 1 to the coordinate YS (step 7
0) Steps 65 to 71 are sequentially repeated until the value of the address counter in the Y direction exceeds the end point coordinate Yl: (Steps 70 and 71). In other words, in this example, when scanning the verification button memory 12 in the Y direction focusing on the coordinate xs, if a black point appears after a white point, 1 is added to CNT, and if there are consecutive black points, then Repeat the process of adding 1 to count the number of black dots. The continuous portion of the black dots is determined to be the line width of the line constituting the matching pattern, and if the white dot appears again on the main scan, it is determined that the line segment has ended. and,
Among the group of black dots that appear in the Y-direction scan of the coordinate xs, the one with the largest number of black dots is extracted as the maximum line width in this scan.

Next, when the Y coordinate exceeds YP, the address XWDT (
X ), :(7) If it is torn, then X WDT (X
s) kl: Stores the contents of BMAX, which is the maximum value of the line width in one scanning line obtained as described above (step 7).
1°72).

Next, 1 is added to the coordinate value Xs (step 73), and the processes of steps 62 to 74 are repeated until the end point coordinate XE in the X direction of the matching pattern memory is exceeded (step 73.
74).

By performing such processing, the maximum line width of line segments appearing in the Y direction for each X coordinate can be determined, and therefore the line width distribution in the X direction can be extracted.

Next, as shown in FIG. 3(B), the line width in the Y direction is calculated by counting the line width for each Y coordinate in the matching pattern memory 12, with the X direction as the main scanning direction and the Y direction as the sub scanning direction. Extract the distribution (steps 81 to 94). This process can be performed using basically the same procedure as for determining the X-direction line width distribution described above, so the explanation thereof will be omitted.

Further, the line width distribution of the registered pattern can be obtained in the same manner as described above using the second line width distribution extraction means +5b.

The terrorist reference point extraction unit 16 in the standard fortune-telling mountain section first detects the X direction of the registered pattern!
Number distribution XHistt and the number of lines in the X direction of the matching pattern @
XHist, and the registered pattern's X-direction line width distribution X■DTt and the matching pattern's X-direction line width distribution Find the position that best overlaps the two.

In other words, the line number distribution when the matching pattern is moved by ΔX in the X direction is XHistt, and the line width distribution is XWDT.
When t is the degree of difference in the number of lines and line width distribution between the registered pattern and the matching pattern, βAx is calculated from the following formula, for example, and -LT≦ΔX≦LT. The value ΔX for which the value of βAX is the minimum within the range is set as the movement iDx in the X direction of the verification pattern necessary for superimposing it on the registered pattern. Furthermore, this L
T may be set to a value larger than the maximum value of the amount of deviation between the registered pattern and the matching pattern.

Next, the Y-direction line number distribution of the registered pattern is Yll+sLt
When the line width distribution is YWDTt, the Y-direction line number distribution of the matching pattern is Yllist, the line width distribution is YVDT, and the degree of difference in the number of lines and line width distribution between the registered pattern and the matching pattern is βllY, Let this dissimilarity β4Y be βAY”Σ(l Y I(istt(Y IY His
t, (Y+ΔY)IhYs+IY WDTt(Y)−Y WDT−(Y+ΔY
)l)/((YE -YS +1)-1ΔYl), and calculate ΔY that minimizes the value of βAY within the range of -LT≦ΔY≦LT as the Y required for superimposing the matching pattern with the registered pattern. Let the amount of movement in the direction be DY.

In this way, it is possible to extract a reference point at which the matching pattern and the registered pattern are superimposed.

In the "binary" title J parking reflection matching determination section 1F, the X and Y obtained in the reference point extraction section 1B are
The degree of coincidence between the matching pattern and the registered pattern when the matching pattern is moved by the overlapping movements iDX and DY in each direction is investigated for all points (addresses) in the pattern memory, and the degree of uniformity α is determined. This degree of coincidence α can be determined, for example, using the following formula.

Here, TP (X, Y) is the value of the point (X, Y) (7) of the registered pattern memory 13, and St' (X, Y) is the value of the point (X, Y) of the matching pattern memory 12, respectively. Representation,
For example, if the point is a black point, it has a value of 1, and if it is a white point, it has a value of 0. Further, the symbol Δ represents a summation and the symbol V represents a logical sum.

The coincidence determination unit 17 further determines that the degree of coincidence α is a predetermined threshold value 1.
When it is 2 or more, it is determined that the registered pattern and the matching pattern match, and α is T.

When the value is less than 1, it is determined that the two patterns do not match.

In addition, in the operation of the matching determination unit 17 described above, there is no difference in slope between the recorded pattern and the matching pattern, and if the matching pattern is moved by the movement iDX and DY obtained by the reference point extraction unit 15, The explanation has been given using an example in which the matching pattern completely overlaps the registered button. However, when actual verification is performed, it is not possible to superimpose the images simply by such a simple movement, but it is necessary to take into consideration the inclination and amount of movement and make fine adjustments. If the degree of coincidence in such a case is γ, this γ can be determined, for example, by the following formula.

DY + y in Y0)) In addition, RX = [(X-Xo) cos θ ten (Y-Yo
) sin θ coRY = [(Y-Y,) cos θ-(X
-X. ) sin θ]. Here, x and y are
Representing the amount of fine adjustment movement in the direction and Y direction, Xo,
Yo represents the X and Y coordinates of the rotation center of the registered pattern, and θ represents the rotation angle. Also, [ ko represents integerization.

Then, -03≦θ≦θ, -LG≦X≦+LG, -L
Find the maximum value γ8 of γ within the range of G≦y≦+LG, and find this γ. If the value is greater than or equal to the above-mentioned Ta, it is determined that the registered pattern and the matching pattern match, and if it is less than γ8 or Ta, it is determined that the two patterns do not match.

By the way, the value of θ may be set as an allowable value for the difference in inclination angle between the W31 pattern and the matching pattern, and if all angles are allowed, it will be θ52180 (degrees).

Further, the value of LG may be, for example, the maximum error in extraction of the movement amounts DX and DY obtained by the reference point extraction section 15.

Further, the value of the registered pattern rotation center (Xo, Yo) may be the center of gravity of the registered pattern, or the midpoint of a circumscribed quadrilateral surrounding the registered pattern.

As described above, according to the fingerprint matching device of the present invention, the fingerprint patterns to be matched can be quickly superimposed on each other at a position where both patterns substantially overlap, so that the fingerprint matching time can be shortened.

By the way, in the above embodiment, when comparing the line number distribution and line width distribution of the registered pattern with the line number distribution and line width distribution of the matching pattern, regardless of the magnitude of the minimum value of the degree of difference βAX, β4Y, the difference between the two is This is explained using an example of superimposing patterns to find the degree of uniformity α, but when these minimum values exceed a certain value Tb, it is immediately determined that the two patterns do not match.'1
- You may also do so. By doing so, the time required for verification can be further reduced.

In addition, in the embodiment described above, the line number distribution and line width distribution are
6! This is explained using an example of extracting from both the pattern and the matching pattern, but regarding the line number distribution and line width distribution of the registered pattern, when registering the registered pattern, the line number distribution and line width distribution of this pattern are extracted accordingly. The same effect as in the embodiment can be obtained by extracting the lines and registering the number of lines and the line width distribution together with the registered pattern data.

Furthermore, in the operation of the line number distribution extraction unit 14 of the embodiment described above,
During scanning performed in the X or Y direction, it is determined that a line segment exists regardless of the line width, that is, regardless of the number of black dots between white points, and the line is Therefore, 1 is added. However, if the reading unit 11 reads dust and stores it in the pattern memory as, for example, black dots, or conversely, if the line of the figure is scanned in the longitudinal direction of the line, a very large number of black dots may appear in a row. There may also be cases where it is stored away. Taking this situation into consideration, it may be possible to count the line as one line only if the line width is within a predetermined range. That is, in the case of the embodiment, step 2B of the flowchart shown in FIG.

In 33.48.53, regardless of the size of the number of black dots (value of the black dot line width counter CNT) existing between white points, when a black dot disappears and a white dot appears, Xtlist
Or, 1 was added to YHist, but the CNT value was W.
By adding 1 to XHist or YHisj only when I≦CNT≦W2 is satisfied, it is possible to deal with situations that occur when scanning dust or a line.

Note that Wl and W2 are predetermined numbers, and since the width of the protrusions (ridges) of a fingerprint is generally 0.2 to 0.5 mm, the reading resolution of the reading unit 11 is set to 20 pixels/m.
If the width of the ridge is 0.5 mm, it corresponds to 10 pixels, and if it is 0.2 mm, it corresponds to 4 pixels, so if W1 = 4 and w2 = io, the ridge width is This makes it suitable for making judgments.

In addition, in the operations of the line number distribution extracting section 14 and the line width distribution extracting section 15 of the embodiment described above, the pattern memory 12.1
3 along the X, Y orthogonal coordinate system to obtain the line number and line width distribution. Even if the scanning is performed along the YD axis tilted by 0.2 degrees and the other processing is performed in the same manner as in the example to obtain the line number and line width distribution for each X and Yo, the same effect as in the example can be obtained. You can get it. still,
In this case, it is preferable that the XD and YD axes are not parallel.

Furthermore, in the operations of the line number distribution extraction unit 14 and the line width distribution extraction unit 15 of the above-described embodiment, the number of lines and the line width distribution are extracted by focusing on the black points existing between white points. However, the present invention is not limited to this embodiment. For example, by focusing on a white point that exists between two black points, and using this white point as one line segment, the invention is not limited to this embodiment. By performing the same processing as in the embodiment, the number of lines and line width distribution can also be determined.

Further, in the above-described embodiment, the line number distribution extraction section is configured by the first and second line number distribution extraction means, and the line width distribution extraction section is configured by the first and second line width distribution extraction means. Although the configuration examples are described, the configuration of the line number and line width distribution extraction section is not limited to this embodiment and can be changed according to the design. For example, the line number and line width distributions of the verification and registration patterns may be created by time-division and one line number and line width extraction means, respectively.

Furthermore, it is clear that the fingerprint verification device of the present invention can be applied to verification devices for figures other than fingerprints.

(Effects of the Invention) As is clear from the above description, the fingerprint verification device of the present invention calculates the line number distribution and line width distribution in each of the first and second directions from the patterns of two fingerprints to be verified. and compares these line number distributions and line width distributions in each of the first and second directions, respectively, and detects the optimal position for superimposing the two fingerprint patterns. Therefore, since the overlapping position of two fingerprints can be determined accurately and quickly, the time required for verification can be shortened.

Therefore, it is possible to provide a fingerprint verification device that can perform fingerprint verification with high alignment accuracy.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the fingerprint matching device of the present invention, FIG. 2 is a flowchart showing the operation of the line number distribution extraction unit according to the invention, and FIGS. 3 (A) and (B) are the invention. FIG. 4 is an explanatory diagram showing a scanning process for a pattern memory to explain the line number distribution extraction section and the line width distribution extraction section according to the present invention. FIG. 4 is a flowchart showing the operation of the line width distribution extraction section according to the present invention. 11...Reading unit, 12...Verification pattern memory 13...Registered pattern memory 14...Line number distribution extraction unit 14a...First line number distribution extraction means +4b-...Second line number distribution Extracting means 15...Line width distribution extraction section 15a...First line width distribution extraction means 15b...Second line width distribution extraction means 16...Reference point extraction section, 17.. Matching determination section.

Claims (5)

[Claims] (1) a reading unit that reads a fingerprint to be matched as a binary matching pattern; a matching pattern memory that stores the matching pattern; and a registered pattern memory that stores a pre-registered fingerprint as the binary registered pattern; The verification pattern memory and the registered pattern memory are scanned in a raster method in each of the first direction and the second direction, the number of lines of both patterns is determined for each scan, and the number of lines of the first and second patterns of the two patterns is A line number distribution extraction unit that extracts the line number distribution for each direction, and scans the matching pattern memory and registered pattern memory in a raster method in each of the first direction and the second direction, and calculates the maximum line width of both patterns. a line width distribution extraction unit that obtains each scan and extracts a line width distribution in each of the first and second directions of both patterns; and the matching pattern and the registered pattern in each of the first and second directions. a reference point extraction unit that compares the line width distributions and line number distributions of the two patterns and extracts the overlapping position of both patterns; and a match determination unit that determines whether the two patterns match or do not match based on the degree of match. (2) The reference point extraction unit calculates the minimum difference obtained by comparing the line number distributions in the first direction and the line width distributions of both the patterns, and the line number distributions in the second direction of both the patterns. Claim 1, wherein the two patterns are determined to be inconsistent when at least one of the minimum dissimilarity values obtained by comparing the number distributions and the line width distributions is larger than a predetermined value. The fingerprint verification device described. (3) The line width is obtained by determining one value of the binary values and a portion of the other value existing between the one value as one line width. Fingerprint verification device according to item 2. (4) The number of lines is determined by calculating the part of the other value existing between one of the two values as one line number, or 2. Fingerprint verification device according to item 2. (5) The number of lines is determined as one line number when the number of values between one of the binary values is within a predetermined range. A fingerprint verification device according to claim 1 or 2.