JP3434768B2 - Sweat pore portion discriminating apparatus, sweat pore portion discriminating method, and recording medium - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sweat gland hole part discriminating apparatus, a sweat gland hole part discriminating method, and a recording medium, and more particularly to a sweat gland hole part discriminating part for discriminating a sweat gland hole from an island by utilizing a ridge direction. The present invention relates to an apparatus, a sweat gland hole part determination method, and a recording medium.

[0002]

2. Description of the Related Art Conventional fingerprint matching is a target of matching.
This is done by comparing the feature points of two fingerprints (ridge end points, bifurcation points). The feature points are extracted from the skeleton image obtained after binarizing the grayscale image of the fingerprint and further skeletonizing the image.

Further, regarding the ridge direction in fingerprint collation and how to obtain the reliability thereof, see "Japanese Patent Laid-Open No. 01-2718".
85 publication "and" JP-A-05-181956 publication ".

Further, the technique described in Japanese Patent Laid-Open No. 63-157286 discloses that "a distance between branch points is calculated for an image pattern obtained by subjecting an input fingerprint image to thinning processing, and the branch points are calculated. The field of view (survey area) is determined by finding the combination of objects that are in the vicinity of each other, and the pattern (fingerprint line) between the branch points is binarized to obtain a distribution map with or without the pattern. -When projected in both Y-axis directions, the distribution amount of no pattern (density 0) is obtained for the entire visual field,
It is determined that there is a sweat gland in the pattern when the distribution amounts on both the X and Y axes are both below the threshold value and there is no large difference. "

[0005]

In the prior art, when feature extraction is performed on a fingerprint image in which sweat pores are prominent, false feature points (false feature points) are detected in the sweat pore portions. Will end up. This will be described in detail with reference to FIG. FIG. 13 is an explanatory diagram showing an outline of fingerprint collation.

Referring to FIG. 13, the fingerprint has a groove (corresponding to a recess).
And ridges (convex parts), and sweat pores are scattered in the ridge like small holes. Therefore,
In the grayscale image of the fingerprint in which the sweat pores are significantly visible, the density of the sweat pores is lower than that of the ridges.

In the fingerprint feature extraction, a process of binarizing a grayscale image of the input fingerprint is required. Binarization refers to a grayscale image locally to determine a threshold value according to a density difference, and a pixel having a density equal to or higher than the threshold value is a black pixel (value 1),
This is a process of setting a pixel having a density lower than the threshold value as a white pixel (value 0).

In other words, the portion of the grayscale image corresponding to the ridge of the fingerprint is black pixel (value 1) and the portion of the fingerprint corresponding to the sweat pore is white pixel (value 0). Then,
The sweat hole portion (the portion corresponding to the sweat hole) of the obtained binary image has a state in which a white hole is embedded in a black ridge portion (a portion corresponding to the ridge). Further, if this binary image is converted into a core line, the core line, which is supposed to be one, would be duplicated in the sweat gland hole portion (FIG. 13).
(A)). The branch portion of this core line becomes a false feature point.

On the other hand, consider the case where there is a branch and a junction (hereinafter referred to as an island) of one ridge in a fingerprint. When binarization and skeletonization are performed on this fingerprint image, a skeleton very similar to that of the sweat gland hole is obtained in the island portion (FIG. 13B). However, this core line is a true core line unlike the case of a sweat pore, and therefore the feature point detected at the branch point of the core line is a true feature point.

As described above, it is difficult to distinguish between sweat gland holes and islands in the conventional fingerprint, so that it becomes difficult to reduce the collation rate by the sweat gland holes and the core wire (feature) correction work by the sweat gland holes in the automatic fingerprint identification system. There is. The sweat pores do not have stable characteristics depending on the fingerprint input device and the state of the finger (how to apply force, etc.), while the islands on the ridges have stable shapes because of the ridge shape. Therefore, it is necessary to distinguish between the sweat pores and the islands on the ridge to prevent only the feature detection by the sweat pores.

Further, in the technique disclosed in Japanese Patent Laid-Open No. 63-157286, since the "ridge direction" is not taken into consideration, there is a problem that the accuracy of sweat gland hole determination is low.

An object of the present invention is to find an algorithm for discriminating between sweat pores and islands by utilizing the ridge direction and prevent detection of false feature points due to sweat pores.

[0013]

A first sweat gland hole portion discriminating apparatus of the present invention creates a binary image from a storage unit and a fingerprint grayscale image, stores the binary image in the storage unit, and stores the binary image in the binarized image. extracting a plurality of white portions constituted by continuous white pixels, before
The black part surrounding each white part is regarded as a ridge, and each white part
Find the center pixel of the
Therefore, find both ends on a straight line in the ridge direction that passes through the center pixel.
Therefore, a ridge width difference between a center portion including a center pixel and both end portions of each white portion is calculated and stored in the storage unit, and based on the ridge width difference, each white portion is a sweat gland hole. And a sweat pore determination unit for determining whether or not

A second sweat gland hole part discriminating apparatus of the present invention comprises a storage unit and a binary image created from a fingerprint gray-scale image, stores the binary image in the storage unit, and comprises continuous white pixels in the binarized image. Extracted from a plurality of white portions, the pixels included in the corresponding portion of the fingerprint grayscale image corresponding to each white portion
Of the density values of each row in the X-axis direction and each column in the Y-axis direction
The minimum value is calculated, and each row in the X-axis direction and each row in the Y-axis direction
Calculates the variance of the minimum value corresponding to the column and stores it in the storage unit
However, the maximum corresponding to each column in the X-axis direction and each row in the Y-axis direction.
A sweat gland hole determination unit that determines whether or not each of the white portions is a sweat gland hole based on the variance of small values .

A third sweat gland hole part discriminating apparatus of the present invention comprises a storage unit and a binary image formed from a fingerprint grayscale image, stores the binary image in the storage unit, and comprises continuous white pixels in the binarized image. Multiple white parts to be extracted and surround each of the white parts
The black part is regarded as a ridge, and the center pixel of each white part is obtained.
Therefore, the ridge direction is calculated from the change in pixel density, and the
Both end portions are obtained on a straight line in the ridge direction, the ridge width difference between the center portion including the center pixel and both end portions is calculated for each white portion, and the difference is stored in the storage unit. Based on the above, it is determined whether or not each white portion is a sweat gland hole, and each row in the X-axis direction and Y-axis direction of the pixels included in the corresponding portion of the fingerprint grayscale image corresponding to each white portion. Each of
Find the minimum density value for each column, and use the X-axis
The variance of the minimum value corresponding to each row of columns and each column of the Y-axis direction
Stored in the storage unit, each column in the X-axis direction, Y-axis
A sweat gland hole determination unit that determines whether or not each of the white portions is a sweat gland hole based on the variance of the minimum value corresponding to each row in the direction .

According to a first method of discriminating sweat pores of the present invention, a binary image is created from a fingerprint gray-scale image, stored in a storage unit, and a plurality of white pixels composed of continuous white pixels in the binarized image. Extract the part, and the black part surrounding each white part as a ridge
Assuming that the center pixel of each white part is found,
The ridge line direction is obtained from the
Find both end parts above, and set the center pixel for each white part.
It includes a procedure of calculating a ridge width difference between the center portion and both end portions including the ridge width, storing the ridge width difference in the storage unit, and determining whether or not each of the white portions is a sweat pore based on the ridge width difference.

According to a second method for determining a sweat pore portion of the present invention, a binary image is created from a fingerprint gray-scale image, stored in a storage unit, and a plurality of white pixels composed of continuous white pixels in the binarized image. A portion is extracted, and each pixel in the X-axis direction of the pixel included in the corresponding portion of the fingerprint grayscale image corresponding to each white portion is extracted.
Obtain the minimum density value for each row and column in the Y-axis direction
Therefore, it corresponds to each row in the X-axis direction and each column in the Y-axis direction.
Then, the variance of the minimum value is calculated and stored in the storage unit.
Variance of the minimum value corresponding to each column in the axial direction and each row in the Y-axis direction
Based on the above, a procedure for determining whether or not each of the white portions is a sweat gland is included.

According to a third method of discriminating sweat pores of the present invention, a binary image is created from a fingerprint gray-scale image, stored in the storage unit, and a plurality of white pixels are formed in the binarized image. Extract the white part and remove the black part surrounding each white part.
Consider the ridge, find the center pixel of each white part,
The ridge direction is obtained from the change in density, and the ridge direction that passes through the center pixel
Find both ends on the straight line and center with respect to each white part
A first ridge width difference between a central portion including pixels and both end portions is calculated and stored in the storage unit, and whether or not each white portion is a sweat pore is determined based on the ridge width difference. Procedure and
The pixels included in the corresponding portion of the fingerprint grayscale image corresponding to each white portion are arranged in each row in the X-axis direction and each column in the Y-axis direction.
Obtain the minimum value of each density value,
Calculate the variance of the minimum value corresponding to each row and each column in the Y-axis direction
Stored in the storage unit, each column in the X-axis direction, Y-axis direction
And a second step of determining whether or not each of the white portions is a sweat pore based on the variance of the minimum value corresponding to each direction line .

The first computer readable form of the present invention
Such a recording medium extracts a plurality of white portions composed of continuous white pixels, and regards the black portions surrounding each white portion as ridges.
Then, find the center pixel of each white part,
Ridge direction is obtained from the
Obtaining both end portions, calculating a ridge width difference between the center portion including the central pixel and each end portion for each white portion, storing the ridge width difference in the storage unit, and based on the ridge width difference, A program for causing a computer to execute a procedure for determining whether or not each of the white portions is a sweat pore is recorded.

A second computer readable form of the invention
The recording medium creates a binary image from the grayscale image of the fingerprint, stores it in the storage unit, extracts a plurality of white portions composed of consecutive white pixels in the binarized image, and corresponds to each white portion. The density values of each row in the X-axis direction and each column in the Y-axis direction of the pixels included in the corresponding portion of the fingerprint grayscale image
The minimum value of is calculated for each row in the X-axis direction and for the Y-axis direction.
The variance of the minimum value corresponding to each column is calculated and stored in the storage unit.
It corresponds to each column in the X-axis direction and each row in the Y-axis direction
A program that causes a computer to execute a procedure for determining whether or not each of the white portions is a sweat pore based on the variance of the minimum value is recorded.

A third computer readable form of the invention
In this recording medium, a plurality of white portions composed of continuous white pixels are extracted, and the black portions surrounding the white portions are regarded as ridges.
None, find the center pixel of each white part, and change the density of the pixel
The ridge direction is obtained from the
Both end parts are obtained and the center pixel is included in each white part.
A first procedure of calculating a ridge width difference between the central portion and both end portions, storing the ridge width in the storage unit, and determining whether or not each white portion is a sweat pore based on the ridge width difference. , Each row in the X-axis direction and each column in the Y-axis direction of the pixels included in the corresponding portion of the fingerprint grayscale image corresponding to each of the white portions.
The minimum density value of X is calculated, and each line in the X-axis direction, Y
Calculate the variance of the minimum value corresponding to each column in the axial direction, and
It is stored in the memory and is stored in each column in the X-axis direction and each row in the Y-axis direction.
A program for causing a computer to execute a second procedure for determining whether or not each of the white portions is a sweat pore based on the variance of the corresponding minimum value is recorded.

A fourth sweat gland hole part discriminating device of the present invention creates a binary image from a storage part and a fingerprint gray-scale image, stores the binary image in the storage part, numbers white parts of the binary image, and counts the total number. Is stored in the storage unit as a count value, one white portion of the binary image is selected, and whether or not the area of the white portion is equal to or smaller than the threshold value A is determined. If the area of the white part is not less than the threshold value B, and if the area of the white part is not more than the threshold value B, the central pixel of the white part is determined, and whether the reliability of the central pixel is not less than the threshold value G is determined. When the reliability of the pixel is equal to or more than the threshold value G, the
The black area surrounding each white area is regarded as a ridge, and the pixel density changes
The ridge direction is obtained from the
Select endpoint candidate pixel of the white parts on the line, and stores in the storage unit, select the two endpoints pixel from the end point candidate pixel, and stores in the storage unit, that of the two endpoints pixels
Each of the end pixels separated by a certain distance from it is selected and stored in the storage unit, the ridge width difference is calculated from the ridge widths of the center pixel and the two end pixels, and stored in the storage unit, It is checked whether the ridge width difference is larger than the threshold value C and not more than the threshold value D, and if the ridge width difference is larger than the threshold value C and not more than the threshold value D, the fingerprint corresponding to the white portion. The density value of the pixel of the grayscale image is extracted, the median value is obtained and stored in the storage unit, and it is checked whether the median value is the threshold E or less. The minimum value in each row in the X-axis direction and each column in the Y-axis direction is selected from the density values of the pixels to be printed in the fingerprint gray-scale image, and the X-axis is selected from all the selected minimum values.
The variance in the Y-axis direction is calculated and stored in the storage unit,
Whether the variances in the X-axis and Y-axis directions are both equal to or less than a threshold value F, and the area of the white portion is equal to or less than the threshold value A, the median value is not equal to or less than the threshold value E, or the X-axis, Y
If the variance in the axial direction is not less than or equal to the threshold value F, the pixel value of the white portion is set to 1, the binary image is stored in the storage unit, and the area of the white portion is equal to or larger than the threshold value B. When the reliability of the central pixel is not more than a threshold value G, the ridge width difference is not greater than a threshold value C, is not a threshold value D or less, or the variances in the X-axis and Y-axis directions are both a threshold value F or less. Regards the white portion as an island, leaves the white portion as a white pixel, then subtracts 1 from the count value in the storage unit, stores the count value in the storage unit, and stores the reduced count. It is checked whether the value is 0, and if it is not 0, it is assumed that there is a white portion that has not been processed yet, and the processing from the selection is performed again on the remaining white portion. If the count value is 0, , Process all white parts It has a sweat gland hole determination unit to terminate the stuff, the.

According to a fourth sweat gland hole part discriminating method of the present invention, the first step of creating a binary image from a fingerprint gray-scale image and storing it in a storage section, and numbering the white part of the binary image,
A second step of storing the total number as a count value in the storage unit, a third step of selecting one white portion of the binary image, and a fourth step of checking whether the area of the white portion is less than or equal to a threshold value A. A step of checking if the area of the white portion is not less than the threshold value B if the area of the white portion is not less than the threshold value A, and a center of the white portion if the area of the white portion is not more than the threshold value B. A sixth step of determining a pixel, a seventh step of checking whether the reliability of the central pixel is equal to or higher than a threshold G, and a reliability of the central pixel being equal to or higher than the threshold G , a black surrounding each white portion. The part is regarded as a ridge
Then, the ridge direction is obtained from the change in pixel density, and the central pixel
Eighth step of selecting the end point candidate pixels of the white part on a straight line in the ridge direction passing through and storing them in the storage section; and selecting two end point pixels from the end point candidate pixels and storing them in the storage section. Step 9 and the two end pixel
Select each edge pixel a certain distance from each
A tenth step of storing in the storage unit, an eleventh step of calculating a ridge width difference from the ridge widths of the central pixel and the two end pixels, and storing the ridge width difference in the storage unit; A twelfth step of checking whether the ridge width difference is larger than the threshold value C and not more than the threshold value D, and the ridge width difference is larger than the threshold value C and not more than the threshold value D.
A thirteenth density value of a pixel of the fingerprint grayscale image corresponding to the white portion is extracted, a median value is obtained, and the result is stored in the storage unit.
And a fourteenth step of checking whether the median value is less than or equal to the threshold value E. If the median value is less than or equal to the threshold value E, the X-axis is calculated from the density value of the pixel corresponding to the white part in the fingerprint grayscale image. Direction in each row and each column in the Y-axis direction, the minimum value is selected, and the variances in the X-axis and Y-axis directions are calculated from all the selected minimum values, and stored in the storage unit. A 16th step of checking whether the variances in the X-axis and Y-axis directions are both equal to or less than a threshold value F, and whether the area of the white portion is equal to or less than a threshold value A,
The median value is not less than or equal to the threshold value E, or the X axis,
If the variance in the Y-axis direction is not less than or equal to a threshold F, a seventeenth step of setting the pixel value of the white portion to 1 and storing the binary image in the storage unit; Or more, the reliability of the central pixel is not greater than or equal to a threshold value G, the ridge width difference is not greater than a threshold value C, is not less than or equal to a threshold value D, or the variances in the X-axis and Y-axis directions are both If it is less than or equal to the threshold value F, the white portion is considered to be an island, and the white portion is left as a white pixel;
Alternatively, following the eighteenth step, the count value in the storage unit is decremented by 1 and stored in the storage unit, and it is checked whether or not the subtracted count value is 0. If there is a white part that has not ended, the remaining white part is executed again from the third step, and if the count value is 0, it is considered that the process has been executed for all the white parts, and the process ends. And a step.

A fourth computer readable form of the invention
The recording medium is a first step of creating a binary image from the grayscale image of the fingerprint and storing it in the storage unit, and the white parts of the binary image are numbered and the total number is stored in the storage unit as a count value. The second step and the white part of the binary image
And a fourth step of checking whether the area of the white portion is less than or equal to the threshold value A. If the area of the white portion is not less than or equal to the threshold value A, the area of the white portion is equal to or less than the threshold value B.
A fifth step of checking whether or not the above is true, and a sixth step of determining the central pixel of the white portion unless the area of the white portion is equal to or larger than the threshold value B, and the reliability of the central pixel is the threshold value G.
A seventh step of checking whether or not the above, and an eighth step of selecting the end point candidate pixel of the white portion when the reliability of the central pixel is equal to or more than a threshold value G and storing it in the storage section;
The black portion surrounding each white portion is regarded as a ridge, and the pixel density
The ridge direction is obtained from the change, and the ridge direction passing through the central pixel
Wherein selecting two end points pixel from the end point candidate pixels on a straight line, and the ninth step of storing in the storage unit, the two
The end points of each of the pixels at a certain distance from each
A tenth step of selecting a pixel and storing it in the storage unit; and an eleventh step of calculating a ridge width difference from the ridge widths of the center pixel and the two end pixels and storing the ridge width difference in the storage unit. A twelfth step of checking whether the ridge width difference is larger than a threshold value C and not more than a threshold value D, and the ridge width difference is larger than the threshold value C and not more than the threshold value D, the white A thirteenth step of extracting density values of pixels of the fingerprint grayscale image corresponding to a portion, obtaining a median value, and storing the median value in the storage unit; and a fourteenth step of checking whether the median value is equal to or less than a threshold value E; When the median value is equal to or less than the threshold value E, the minimum value in each row in the X-axis direction and each column in the Y-axis direction is selected and selected from the density values of the pixels corresponding to the white portion in the fingerprint grayscale image. From the minimum of all Calculating a variance in the axial direction, and a fifteenth step of storing in the storage unit, the X-axis, first the variance in Y-axis direction is examined whether both a threshold F or less
6 and the area of the white portion is less than or equal to the threshold value A, the median value is not less than or equal to the threshold value E, or the X
If the variance in the axial or Y-axis direction is not less than or equal to the threshold F, a seventeenth step of setting the pixel value of the white portion to 1 and storing the binary image in the storage unit; It is equal to or greater than a threshold value B, the reliability of the central pixel is not equal to or greater than a threshold value G, the ridge width difference is not greater than a threshold value C, is not less than or equal to a threshold value D, or the dispersion in the X-axis and Y-axis directions. Are both equal to or less than the threshold F, the white portion is regarded as an island, and the white portion is left as a white pixel, and the eighteenth step, the seventeenth step, or the eighteenth step is performed. Subsequently, the count value in the storage unit is decremented by 1 and stored in the storage unit,
Check whether the reduced count value is 0, and
If not, it is assumed that there is a white part that has not been processed, and the remaining white part is executed again from the third step. If the count value is 0, the process is executed for all white parts. And a program for causing a computer to execute the nineteenth step, which is terminated.

[0025]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the features of the embodiments of the present invention will be described with reference to the drawings. FIG. 3 is an explanatory diagram showing a difference in ridge width difference between a sweat pore and an island. Figure 4
FIG. 4 is an explanatory diagram showing a difference in concentration distribution between a sweat pore and an island.

In the present invention, the following two points should be noted when discriminating between sweat pores and islands. The first point is the ridge width of the sweat pore candidate portion in the binary image.

The ridge width is the length of a line segment perpendicular to the ridge direction (definition will be described later) in each pixel of the ridge (including a white portion). In the sweat pores, the ridge widths at both ends and the center of the hole are almost equal (Fig.
(A) Ridge width WS1, ridge width WM, ridge width WS2).
On the other hand, in the island, the ridge width at the center is slightly wider than at both ends (the ridge width WS1, the ridge width WM, and the ridge width WS2 in FIG. 3B). Therefore, by measuring the ridge width of the sweat gland hole candidate portion and calculating the difference, it is possible to determine whether or not the sweat gland hole is a sweat gland hole.

The second point is the density distribution of the sweat gland hole candidate portion in the grayscale image. On the island, low-density pixels are distributed almost uniformly along the ridge direction (FIG. 4B). On the other hand, in the sweat pores, the density distribution has less regularity than in the island, and the pixel values are more dispersed than in the island (FIG. 4).
(A)). Therefore, it is possible to determine whether the pore is a sweat gland hole based on the average value, the minimum value, the dispersion value, and the like of the density.

Next, a first embodiment of the present invention will be described in detail with reference to the drawings. FIG. 2 is a block diagram showing the first embodiment of the present invention. Referring to FIG. 2, the first embodiment of the present invention includes a sweat pore determination unit 10 that determines a sweat pore and a storage unit 20 such as a memory.

Next, the operation of the first embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows the first of the present invention.
3 is a flowchart showing the operation of the embodiment of FIG. Figure 5
FIG. 6 is an explanatory diagram showing a method of determining a center point of a white portion. FIG. 6 is an explanatory diagram showing a method of selecting an end point candidate of a white portion. FIG. 7 is an explanatory diagram of the ridge direction and the X and Y axes. Figure 8
FIG. 4 is an explanatory diagram showing a method of determining end point pixels. Figure 9
It is explanatory drawing which shows the determination method of an edge pixel. FIG. 10 is an explanatory diagram showing a method for obtaining the median density value of the white portion and the minimum density value in the X-axis and Y-axis directions. FIG. 11 shows
It is explanatory drawing which shows the calculation method of a density | concentration change.

Referring to FIG. 1, the sweat pore determination unit 10 according to the first embodiment of the present invention first stores in the storage unit 20 a fingerprint grayscale image 201 input from a fingerprint reading scanner (not shown). Then, binarization processing is performed on the stored fingerprint grayscale image 201 to create a binary image 202 and store it in the storage unit 20 (step A1 in FIG. 1). As a result of this binarization, a portion (mainly a ridge portion) of the fingerprint grayscale image 201 having a density equal to or higher than a certain threshold becomes a black pixel (value 1),
A portion having a density lower than the threshold value (mainly a groove portion) becomes a white pixel (value 0).

Next, from the obtained binary image 202, a white portion composed of continuous white pixels (a continuous portion of white pixels)
Find and number these one by one (labeling)
do. Further, the total number of numbered white portions is counted and stored as a count value 203 in the storage unit 20 (step A2 in FIG. 1).

Thereafter, one numbered white portion is selected from the binary image 202 (step A3 in FIG. 1). When the white portion is found, it is determined whether it is a sweat pore or an island by the following procedure (step A4 to step A18 in FIG. 1). In addition, threshold A to threshold F described below
Uses, for example, images of various sweat pore fingerprints to determine a statistically optimum value.

First, the area of the white portion (the number of pixels) is the threshold value A.
(For example, 20 [pixels]) or less is checked (step A4 in FIG. 1). If the area is equal to or smaller than the threshold value A (step A4 / Yes in FIG. 1), the white portion is regarded as a sweat pore or noise, and the white portion of the binary image 202 is replaced with a black pixel (value 1) and stored. It is stored in the unit 20 (step A17 in FIG. 1).

When the area of this white portion is larger than A (step A4 / No in FIG. 1), the area of this white portion is the threshold value B.
(For example, 150 [pixels]) or more is checked (step A5 in FIG. 1). When the area of this white portion is equal to or larger than the threshold value B (step A5 in FIG. 1 / Yes)
, The white part is regarded as an island, and the binary image 202
The white portion of is left as a white pixel (value 0) (step A18 in FIG. 1).

When the area of this white portion is smaller than the threshold value B (step A5 / No in FIG. 1), as shown in FIG. 5, a pixel M including the intersection of diagonal lines in the smallest rectangle surrounding this white portion. Is obtained, and this is set as the central pixel M of the white portion (step A6 in FIG. 1).

Next, the reliability of the central pixel M obtained above in the ridge direction is checked (step A7 in FIG. 1).

Here, the ridge direction means the fingerprint grayscale image 20.
1 is divided into, for example, blocks of 8 × 8 pixels, and in each block, a certain number of pixels in all 16 directions quantized in π / 16 angular steps as shown in FIG. The density change of (1) is determined, and the one with the smallest change is set as the ridge direction of the pixels in the block.

Referring to FIG. 11, a portion surrounded by a thick black frame is one block (8 × 8 pixels). In this block, for example, the density change in the direction 4 of FIG. 7A is calculated as follows.

(1) In a block, a strip-shaped mask 1 is provided along four directions of a certain length (for example, x) that passes through the central pixel (pixel shown in black) in the block.

(2) The pixel values of the grayscale image of the portion corresponding to the mask 1 are arrayed in order from the pixel at the end of the mask g [i].
(I = 0 to (x-1)).

(3) From the array g [i], for example, the difference between the elements is calculated every three, and the value obtained by adding them is si.
gma1. sigma1 = Σ {| g [k] -g [k + 3] | × δ}
However, k = 0- (x-4).

[0043] δ = 1 (| g [k] -g [k + 3] |> = R).

Δ = 0 (| g [k] -g [k + 3] | <
R) (R: a constant determined appropriately).

(4) In the block, the mask 1 given in (1) and, for example, a strip-shaped mask 2 and a mask 3 which are separated by three pixels and are parallel to each other are provided. ) And (3), then sigma
2. Calculate sigma3.

(5) sigma1, sigma2, si
From gma3, the change in concentration is obtained, for example, by the following formula.

<Density change> = (sigma1 + sigm
a2 + sigma3) / 4.

With this procedure, the density change is obtained in all 16 directions in the block, and the direction having the smallest density change is taken as the ridge direction in this block. Then, the difference between the density change in the determined ridge direction and the density change in the orthogonal direction (density accumulated variation value) is defined as the reliability of the ridge, and the larger this value is, the higher the reliability in the determined ridge direction is. I reckon.

When the reliability of the block including the central pixel M of the white portion is lower than the threshold value G (statistical optimum value is determined in advance) (step A7 / N in FIG. 1).
o), it is considered that it is impossible to determine whether the white portion is a sweat pore or an island, and the white portion of the binary image 202 is left as a white pixel (value 0) (step A18 in FIG. 1). If the reliability in the central pixel M of this white portion is equal to or higher than the threshold value G (step A7 / Yes in FIG. 1), each pixel in this white portion and the central pixel M are connected by a straight line, as shown in FIG. At that time, all the pixels in the white portion whose straight line can be regarded as parallel to the ridge direction are selected and stored as the end point candidate pixels (step A8 in FIG. 1).

The ridge direction is π / 16 as shown in FIG.
It is expressed by quantizing in 16 ways at each angle step. Here, what is considered to be parallel to the ridge direction is that when the central pixel M and a certain pixel in the white part are connected by a straight line, This is the case where the angle of the straight line is within the range of <angle in the ridge direction of the central pixel M> ± π / 16.

Next, among all the end point candidate pixels thus selected, the end point pixel S1 'and the end point pixel S2' as shown in FIG. 8 are selected by the following method and stored in the storage unit 20 ( 1 step A9). Referring to FIG. 7B, the X axis of the XY coordinates is an axis parallel to a straight line drawn from the origin O in FIG. 7A in the direction of 0, and the Y axis of the XY coordinates is shown in FIG. ) Is an axis parallel to a straight line drawn from the origin O in the direction of 8.

[0052] end point pixels S1 ', the end point pixel S2' method of determining the.

(A) The ridge direction of the central pixel M is shown in FIG.
In the direction of 0 in X, X is centered on the center pixel M.
Assuming the Y coordinate, the pixel having the largest X coordinate value and the pixel having the smallest X coordinate value are respectively the end point pixel S1 ′ and the end point pixel S2 ′.
And

(B) The ridge direction of the central pixel M is shown in FIG.
1 to 7 in the first quadrant (FIG. 7B) assuming the XY coordinate centering on the central pixel M.
In addition, the pixel that is most distant from the center pixel M is defined as the end point pixel S1 ′. Similarly, a pixel located in the third quadrant (FIG. 7B) and having the longest distance from the central pixel M is defined as an end point pixel S2 ′.

(C) The ridge direction of the central pixel M is shown in FIG.
In the case of 8 directions, assuming XY coordinates centering on the central pixel M, the pixel having the largest Y coordinate value and the pixel having the smallest Y coordinate are respectively defined as the end point pixel S1 ′ and the end point pixel S2 ′.

(D) The ridge direction of the central pixel M is shown in FIG.
9 to 15 in the direction of X, centering on the center pixel M
Assuming the Y coordinate, a pixel in the second quadrant (FIG. 7B) that is farthest from the central pixel M is defined as an end point pixel S1 ′. Similarly, a pixel located in the fourth quadrant (FIG. 7B) and having the longest distance from the central pixel M is defined as an end point pixel S2 ′.

Next, as shown in FIG. 9, the end point pixel S1 'and the end point pixel S2' are outside the white portion by several pixels (this value is preset as an optimum value. For example, 5 [pixel]). The shifted positions are set as the end pixel S1 and the end pixel S2, and are stored in the storage unit 20 (step A10 in FIG. 1). Here, the shift is performed in parallel with the ridge direction of the center pixel M from the end point pixel S1 ′ and the end point pixel S2 ′. These are the two end pixels S1 and S2 as the measurement base points. The reason why the positions obtained by shifting the end point pixel S1 ′ and the end point pixel S2 ′ by a few pixels are optimized is that by shifting the end point pixel S1 ′ and the end point pixel S2 ′, the measurement position moves away from the vicinity of a peculiar shape (island or sweat pore) This is because the ridge width becomes closer and the ridge width can be measured more stably.

In this way, the end pixel S of this white part
1. After the end pixel S2 is determined, the ridge widths of the central portion and both end portions in the white portion are calculated as shown in FIG. 3 using the center pixel M, the end pixel S1, and the end pixel S2 as base points (respectively). WM, WS1, and WS2), the ridge width difference Δwidth is calculated by the following formula, and the storage unit 20
(Step A11 in FIG. 1).

[0059] Δwidth = WM− (WS1 + WS2) / 2.

<Ridge width difference> = <ridge width at midpoint>-<average ridge width at two ends>.

Next, the ridge width difference Δwidth is checked (step A12 in FIG. 1). Ridge width difference Δwidth
Is less than or equal to the threshold C (statistically set to an optimum value in advance, for example, -0.5) (step A12 / N in FIG. 1).
o), considering that the distortion of the ridge near this white part is large,
This white portion is left as a white pixel (value 0) (step A18 in FIG. 1).

The ridge width difference Δwidth is the threshold value D
(The optimum value is set statistically in advance. For example, when it is larger than 3) (step A12 / No in FIG. 1), this white part is regarded as an island, and this white part of the binary image 202 is regarded as a white pixel. Leave it as it is (value 0) (step A1 in FIG. 1).
8).

When neither of the above two conditions is satisfied (threshold value C <Δwidth <= threshold value D) (step A12 / Yes in FIG. 1), the densities of all the pixels of the fingerprint grayscale image 201 corresponding to this white part A value (for example, in the case of 256 gradations, 0 to 255 is the density value) is extracted, the median value med is obtained as shown in FIG. 10, and the result is stored in the storage unit 20 (step A13 in FIG. 1). Here, the median value med is the density value that is exactly at the center position when all the density values are arranged in ascending order. In FIG.
1 is the median value med.

Next, the obtained median value med is checked (step A14 in FIG. 1). The median med is the threshold value E
(The optimum value is set statistically in advance. For example, 10
0) is larger (step A14 / No in FIG. 1),
This white part is regarded as a sweat pore, and this white part of the binary image 202 is replaced with a black pixel (value 1) and stored in the storage unit 20 (step A17 in FIG. 1).

When the median value med is less than or equal to the threshold value E (step A14 / Yes in FIG. 1), as shown in FIG. 10, from the density values in the fingerprint grayscale image 201 of all pixels corresponding to this white portion, the X-axis is calculated. Direction, the minimum value in each row and each column in the Y-axis direction is selected (step A1 in FIG. 1).
5). For example, in FIG. 10, the minimum value in each row in the X-axis direction is <0,0,0,0,0,0,0,0,
0,1,0,0,0,0,2>, and the minimum value in each column in the Y-axis direction is <0,0,0,0,0,0,8>.

Then, from all the selected minimum values, X
The variances Vx and Vy in the axis and Y-axis directions are calculated and stored in the storage unit 20 (step A15 in FIG. 1). For example, if the minimum value selected in the X-axis direction is Xi (i = 1
~ K), the variance Vx is calculated by the following equation.

Vx = {kΣX ² − (ΣX) ² } / k ² .

For example, in FIG. 10, k = 7, X
1 = 0, X2 = 0, X3 = 0, X4 = 0, X5 = 0, X
6 = 0 and X7 = 8.

Therefore, Vx = {7 × (0 ² +0 ² +0 ² +
0 ² +0 ² +0 ² +8 ² )-(0 + 0 + 0 + 0 + 0 + 0 +
8) ² } / 7 ² = (7 × 64−64) /49=7.8.

Next, the calculated variance Vx and variance Vy are checked (step A16 in FIG. 1). Both the variance Vx and the variance Vy are threshold values F (optimum values are set statistically in advance.
For example, if it is 2000 or less (step A16 / Yes in FIG. 1), the white part is regarded as an island, and the white part of the binary image 202 is left as a white pixel (value 0) (FIG. 1). 1 step A18). If this condition is not satisfied (Vx> threshold F or Vy> threshold F) (see FIG.
(Step A16 / No), the white part is regarded as a sweat pore, and the white part of the binary image 202 is replaced with a black pixel (value 1) (step A17 in FIG. 1).

In this manner, the determination as to whether one white portion is a sweat gland and the process of filling the sweat gland are completed. Next, the count value 203 in the storage unit 20 is decremented by 1,
The count value 203 stored in the storage unit 20 and reduced is 0.
(FIG. 1 step A19), if it is not 0 (FIG. 1 step A19 / No), it is determined that there is a white portion that has not been processed, and the white portion is searched again (FIG. 1 step A3). ) Is performed and the count value is 20
If 3 is 0 (step A19 / Yes in FIG. 1), it is assumed that the processing has been performed on all the white parts, and the process ends.

Next, a second embodiment of the present invention will be described with reference to the drawings. The second embodiment of the present invention creates a binary image 202 from a fingerprint grayscale image 201,
The first step of storing in the storage unit 20 (step A in FIG. 1)
1), the second step of numbering the white parts of the binary image 202 and storing the total number in the storage unit 20 as the count value 203 (step A2 in FIG. 1), and one white part of the binary image 202. Third step to select (step A3 in Figure 1)
And checking whether the area of this white portion is less than or equal to the threshold value A
Step (FIG. 1 step A4) and if the area of the white part is not less than the threshold A (FIG. 1 step A4 / No), the fifth step (step of FIG. 1 step) for checking whether the area of the white part is more than the threshold B or not. A5), and if the area of this white portion is not equal to or larger than the threshold value B (step A5 / No in FIG. 1), the sixth step (step A5 in FIG. 1) of determining the central pixel M of this white portion and the central pixel M A seventh step (step A7 in FIG. 1) of checking whether the reliability is equal to or higher than the threshold G, and that the reliability of the central pixel M is equal to or higher than the threshold G (step A7 in FIG. 1).
/ Yes), the eighth step of selecting the end point candidate pixel of this white portion and storing it in the storage unit 20 (step A8 in FIG. 1).
From the end point candidate pixel to the end point pixel S1 ′ and the end point pixel S2 ′
Is selected and stored in the storage unit 20 (FIG. 1).
Step A9), the end pixel S1 and the end pixel S2 are selected based on the end pixel S1 ′ and the end pixel S2 ′, and the storage unit 2
Tenth step of storing in 0 (step A10 in FIG. 1)
And the ridge width difference Δwid from the ridge widths WM, WS1, and WS2 at the center pixel M, the end pixel S1, and the end pixel S2.
Eleventh step of calculating th and storing it in the storage unit 20 (step A11 in FIG. 1) and twelfth step of checking whether the ridge width difference Δwidth is larger than the threshold value C and is equal to or less than the threshold value D (FIG. 1 step A12) and the ridge width difference Δwidth is larger than the threshold value C and equal to or less than the threshold value D (step A12 / Yes in FIG. 1), the densities of all the pixels of the fingerprint grayscale image 201 corresponding to the white part The thirteenth step (step A13 in FIG. 1) of extracting the values, obtaining the median value med, and storing the median value med, and the median value me
The fourteenth step of checking whether d is less than or equal to the threshold value E (see FIG.
Step A14) and if the median med is less than or equal to the threshold value E (Step A14 / Yes in FIG. 1), from the density values in the fingerprint grayscale image 201 of all the pixels corresponding to this white portion, each row in the X-axis direction, Y The minimum value in each column in the axial direction is selected, and from all the selected minimum values, the X-axis,
The fifteenth step of calculating the variance Vx and the variance Vy in the Y-axis direction and storing them in the storage unit 20 (step A1 in FIG. 1).
5) and the 16th step (step A16 in FIG. 1) of checking whether the variance Vx and the variance Vy are both equal to or less than the threshold F, and whether the area of the white part is equal to or less than the threshold A (step A in FIG. 1).
4 / Yes), whether the median med is not less than or equal to the threshold value E (step A14 / No in FIG. 1), or variance Vx, variance V
y is not less than either threshold F (FIG. 1 step A16 /
No), the pixel value of this white portion is set to 1 and the binary image 202 is stored in the storage unit 20 in the seventeenth step (FIG. 1).
Step A17), whether the area of this white portion is equal to or larger than the threshold value B (step A5 / Yes in FIG. 1), or whether the reliability of the central pixel M is equal to or larger than the threshold value G (step A7 / N in FIG. 1).
o), whether the ridge width difference Δwidth is not larger than the threshold value C or is less than or equal to the threshold value D (step A12 / No in FIG. 1),
Alternatively, when the variance Vx and the variance Vy are both equal to or less than the threshold F (step A16 / Yes in FIG. 1), the white portion is considered to be an island, and the white portion is left as a white pixel (value 0). 18 steps (step A18 in FIG. 1),
Subsequent to the 17th and 18th steps, the count value 203 in the storage unit 20 is decremented by 1 and stored in the storage unit 20, and it is checked whether or not the subtracted count value 203 is 0. If there is a white part that has not been processed, the process from the white part search (step A3 in FIG. 1) is performed again, and if the count value 203 is 0, it is assumed that the process has been performed on all white parts. The sweat gland hole part determining method includes a nineteenth step (step A19 in FIG. 1) to be ended.

Next, a third embodiment of the present invention will be described with reference to the drawings. FIG. 12 is a block diagram showing the third embodiment of the present invention. Referring to FIG. 12, the third embodiment of the present invention implements the steps of the second embodiment of the present invention by a computer 100 (eg,
The recording medium 120 stores a program to be executed by the sweat pore determination unit 10). This program is for recording medium 1
20 to a storage device of the computer 100 (for example,
It is loaded into the storage unit 20) and executed.

Specifically, the third embodiment of the present invention is
First step of creating binary image 202 from fingerprint grayscale image 201 and storing it in storage unit 20 (step A1 in FIG. 1)
And a second step of numbering the white parts of the binary image 202 and storing the total number in the storage unit 20 as the count value 203 (step A2 in FIG. 1), and selecting one white part of the binary image 202. A third step (step A3 in FIG. 1),
The fourth step (step A4 in FIG. 1) of checking whether the area of the white portion is equal to or smaller than the threshold value A. If the area of the white portion is not equal to or smaller than the threshold value A (step A4 / No in FIG. 1), the area of the white portion is A fifth step (step A5 in FIG. 1) of checking whether the threshold value is equal to or larger than the threshold value B. If the area of the white portion is not equal to or larger than the threshold value B (step A5 / No in FIG. 1), the sixth step of determining the central pixel M of the white portion. Step (Figure 1 Step A
5), and a seventh step (step A7 in FIG. 1) of checking whether the reliability of the central pixel M is equal to or higher than the threshold value G, and the central pixel M
Is higher than the threshold value G (step A7 / Y in FIG. 1).
es), the end point candidate pixel of this white portion is selected, and the storage unit 2
8th step of storing in 0 (step A8 in FIG. 1),
Based on the ninth step (step A9 in FIG. 1) of selecting the end point pixel S1 ′ and the end point pixel S2 ′ from the end point candidate pixels and storing the end point pixel S1 ′ and the end point pixel S2 ′, A tenth step (step A10 in FIG. 1) of selecting S1 and the end pixel S2 and storing them in the storage unit 20,
The ridge width difference Δwidth from the ridge widths WM, WS1, and WS2 at the center pixel M, the end pixel S1, and the end pixel S2.
Is calculated and stored in the storage unit 20 (step A11 in FIG. 1), and the ridge width difference Δwidth is the threshold C.
A twelfth step (step A12 in FIG. 1) of checking whether the ridge width difference Δ is larger than the threshold value D or less
When the width is larger than the threshold value C and equal to or smaller than the threshold value D (step A12 / Yes in FIG. 1), the density values of all the pixels of the fingerprint grayscale image 201 corresponding to this white portion are extracted to obtain the median value med. , Thirteenth stored in the storage unit 20
(Step A13 in FIG. 1), a fourteenth step (step A14 in FIG. 1) for checking whether the median med is less than or equal to the threshold value E, and that the median med is less than or equal to the threshold value E (step A14 / Yes in FIG. 1). , The minimum value in each row in the X-axis direction and each column in the Y-axis direction is selected from the density values in the fingerprint grayscale image 201 of all the pixels corresponding to the white portion, and X is selected from all the selected minimum values. axis,
The fifteenth step of calculating the variance Vx and the variance Vy in the Y-axis direction and storing them in the storage unit 20 (step A1 in FIG. 1).
5) and the 16th step (step A16 in FIG. 1) of checking whether the variance Vx and the variance Vy are both equal to or less than the threshold F, and whether the area of the white part is equal to or less than the threshold A (step A in FIG. 1).
4 / Yes), whether the median med is not less than or equal to the threshold value E (step A14 / No in FIG. 1), or variance Vx, variance V
y is not less than either threshold F (FIG. 1 step A16 /
No), the pixel value of this white portion is set to 1 and the binary image 202 is stored in the storage unit 20 in the seventeenth step (FIG. 1).
Step A17), whether the area of this white portion is equal to or larger than the threshold value B (step A5 / Yes in FIG. 1), or whether the reliability of the central pixel M is equal to or larger than the threshold value G (step A7 / N in FIG. 1).
o), whether the ridge width difference Δwidth is not larger than the threshold value C or is less than or equal to the threshold value D (step A12 / No in FIG. 1),
Alternatively, when the variance Vx and the variance Vy are both equal to or less than the threshold F (step A16 / Yes in FIG. 1), the white portion is considered to be an island, and the white portion is left as a white pixel (value 0). 18 steps (step A18 in FIG. 1),
Subsequent to the 17th and 18th steps, the count value 203 in the storage unit 20 is decremented by 1 and stored in the storage unit 20, and it is checked whether or not the subtracted count value 203 is 0. If there is a white part that has not been processed, the process from the white part search (step A3 in FIG. 1) is performed again, and if the count value 203 is 0, it is assumed that the process has been performed on all white parts. The computer system 100 includes a nineteenth step (step A19 in FIG. 1) to end.
The recording medium 120 stores a program to be executed by the recording medium 120.

[0075]

The first effect of the present invention is to reduce the number of false feature points due to the sweat pores.

The reason is that only the white part due to the sweat pores is corrected to the black part in the state of the binary image.

The second effect is that the fingerprint matching accuracy is improved.

The reason for this is that in the sweat pores, the ridge widths at both end portions and the center portion of the hole are almost equal, and in the island, the ridge width at the center portion is slightly wider than at both end portions. This is also because the distribution of the density values of the pixels in the portion of the fingerprint grayscale image corresponding to the white portion is used.

[Brief description of drawings]

FIG. 1 is a flowchart showing an operation of a first exemplary embodiment of the present invention.

FIG. 2 is a block diagram showing a first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a difference in ridge width difference between a sweat pore and an island.

FIG. 4 is an explanatory diagram showing a difference in concentration distribution between a sweat pore and an island.

FIG. 5 is an explanatory diagram showing a method of determining a center point of a white portion.

FIG. 6 is an explanatory diagram showing a method of selecting an end point candidate of a white portion.

FIG. 7 is an explanatory diagram of a ridge direction and an X axis and a Y axis.

FIG. 8 is an explanatory diagram showing a method of determining end point pixels.

FIG. 9 is an explanatory diagram showing a method of determining edge pixels.

FIG. 10 is the median of the density values of the white part, the X axis, and Y.
It is explanatory drawing which shows the method of calculating | requiring the minimum value of the density value of an axial direction.

FIG. 11 is an explanatory diagram showing a method of calculating a density change.

FIG. 12 is a block diagram showing a third embodiment of the present invention.

FIG. 13 is an explanatory diagram showing an outline of fingerprint collation.

[Explanation of symbols]

10 Sweat hole discrimination unit 20 memory 100 computers 120 recording media 201 Fingerprint gray image 202 Binary image 203 count value

Front page continuation (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06T 7/00 A61B 5/10

Claims (12)

(57) [Claims] And 1. A storage unit, to create a binary image from the fingerprint gray image, stored in the storage unit, extracts a plurality of white portions constituted by continuous white pixels in the binarized image, before
The black part surrounding each white part is regarded as a ridge, and each white part
Find the center pixel of the
Therefore, find both ends on a straight line in the ridge direction that passes through the center pixel.
Therefore, a ridge width difference between a center portion including a center pixel and both end portions of each white portion is calculated and stored in the storage unit, and based on the ridge width difference, each white portion is a sweat gland hole. And a sweat gland hole discriminating unit for determining whether or not the sweat gland hole portion discriminating device. 2. A storage unit and a binary image is created from a fingerprint grayscale image, stored in the storage unit, and a plurality of white portions formed of continuous white pixels in the binarized image are extracted. The pixels in the corresponding portion of the fingerprint grayscale image corresponding to each white portion are arranged in each row in the X-axis direction and each column in the Y-axis direction.
Obtain the minimum density value for each, and
Calculate the variance of the minimum value corresponding to each row and each column in the Y-axis direction,
Stored in the storage unit, each column in the X-axis direction, in the Y-axis direction
A sweat gland hole part determination device , comprising: a sweat gland hole determination part that determines whether or not each of the white portions is a sweat gland hole based on the variance of the minimum value corresponding to each row . 3. A storage unit, to create a binary image from the fingerprint gray image, stored in the storage unit, extracts a plurality of white portions constituted by continuous white pixels in the binarized image, before
The black part surrounding each white part is regarded as a ridge, and each white part
Find the center pixel of the
Therefore, find both ends on a straight line in the ridge direction that passes through the center pixel.
Therefore, a ridge width difference between a center portion including a center pixel and both end portions of each white portion is calculated and stored in the storage unit, and based on the ridge width difference, each white portion is a sweat gland hole. Each of the pixels in the X-axis direction included in the corresponding portion of the fingerprint grayscale image corresponding to each of the white portions.
Obtain the minimum density value for each row and column in the Y-axis direction
Therefore, it corresponds to each row in the X-axis direction and each column in the Y-axis direction.
Then, the variance of the minimum value is calculated and stored in the storage unit.
Variance of the minimum value corresponding to each column in the axial direction and each row in the Y-axis direction
A sweat gland hole part discriminating device which has a sweat gland hole discriminating portion which determines whether each white portion is a sweat gland hole based on the above. 4. A binary image is created from a fingerprint grayscale image, stored in a storage unit, a plurality of white portions formed by continuous white pixels in the binarized image is extracted, and each white portion is surrounded.
The black portion is regarded as a ridge, and the center pixel of each white portion is obtained.
Therefore, the ridge direction is calculated from the change in pixel density, and the
Both end portions are obtained on a straight line in the ridge direction, the ridge width difference between the center portion including the center pixel and both end portions is calculated for each white portion, and the difference is stored in the storage unit. Based on the above, a method for determining a sweat gland hole portion is characterized by including a procedure for judging whether or not each of the white portions is a sweat gland hole. 5. A binary image is created from a fingerprint grayscale image, stored in a storage unit, and a plurality of white portions composed of continuous white pixels in the binarized image are extracted and corresponding to each white portion. The density values of each row in the X-axis direction and each column in the Y-axis direction of the pixels included in the corresponding portion of the fingerprint grayscale image
The minimum value of is calculated for each row in the X-axis direction and for the Y-axis direction.
The variance of the minimum value corresponding to each column is calculated and stored in the storage unit.
It corresponds to each column in the X-axis direction and each row in the Y-axis direction
A sweat gland hole part discriminating method comprising a step of judging whether or not each of the white portions is a sweat gland hole based on the variance of the minimum value . 6. Create a binary image from the fingerprint gray image, stored in the storage unit, extracts a plurality of white portions constituted by continuous white pixels in the binarized image, the each white portion
The black part that surrounds is regarded as a ridge, and the center pixel of each white part
Then, the ridge direction is calculated from the change in pixel density, and the center pixel
Both end portions are obtained on a straight line in the ridge direction passing through, and the ridge width difference between the center portion including the central pixel and both end portions is calculated for each white portion and stored in the storage unit. A first procedure for determining whether each white portion is a sweat pore based on the difference, and each row in the X-axis direction of pixels included in a corresponding portion of the fingerprint grayscale image corresponding to each white portion. , Y
Find the minimum density value of each axial column, and
Minimum value corresponding to each row in the X-axis direction and each column in the Y-axis direction
Of the X-axis direction is calculated and stored in the storage unit.
Based on the variance of the minimum value corresponding to each column and each row in the Y-axis direction
And a second procedure for determining whether each of the white portions is a sweat gland hole or not. 7. Create a binary image from the fingerprint gray image, stored in the storage unit, the extracts a plurality of white portions constituted by continuous white pixels in the binarized image, the respective white portions min Fence
The black portion is regarded as a ridge, and the center pixel of each white portion is obtained.
Therefore, the ridge direction is calculated from the change in pixel density, and the
Both end portions are obtained on a straight line in the ridge direction, the ridge width difference between the center portion including the center pixel and both end portions is calculated for each white portion, and the difference is stored in the storage unit. A computer-readable recording medium for recording a program for causing a computer to execute a procedure for determining whether or not each of the white portions is a sweat pore based on the above. 8. A binary image is created from a fingerprint grayscale image, stored in a storage unit, and a plurality of white portions composed of continuous white pixels in the binarized image are extracted to correspond to each white portion. The density values of each row in the X-axis direction and each column in the Y-axis direction of the pixels included in the corresponding portion of the fingerprint grayscale image
The minimum value of is calculated for each row in the X-axis direction and for the Y-axis direction.
The variance of the minimum value corresponding to each column is calculated and stored in the storage unit.
It corresponds to each column in the X-axis direction and each row in the Y-axis direction
A computer-readable program for recording a program for causing a computer to execute a procedure for determining whether or not each of the white portions is a sweat pore based on the variance of the minimum value.
Removable recording medium. 9. A binary image is created from a fingerprint grayscale image, stored in the storage unit, and a plurality of white portions composed of continuous white pixels in the binarized image are extracted, and each of the white portions is extracted.
The black part that surrounds is regarded as a ridge, and the center pixel of each white part
Then, the ridge direction is calculated from the change in pixel density, and the center pixel
Both end portions are obtained on a straight line in the ridge direction passing through, and the ridge width difference between the center portion including the central pixel and both end portions is calculated for each white portion and stored in the storage unit. A first procedure for determining whether each white portion is a sweat pore based on the difference, and each row in the X-axis direction of pixels included in a corresponding portion of the fingerprint grayscale image corresponding to each white portion. , Y
Find the minimum density value of each axial column, and
Minimum value corresponding to each row in the X-axis direction and each column in the Y-axis direction
Of the X-axis direction is calculated and stored in the storage unit.
Based on the variance of the minimum value corresponding to each column and each row in the Y-axis direction
Te, wherein said computer readable to each white portion is characterized in that recording the program for executing a second routine to determine whether or not there in sweat pores on the computer
Do recording medium. 10. A storage unit and a binary image is created from a fingerprint grayscale image, stored in the storage unit, white portions of the binary image are numbered, and the total number is stored in the storage unit as a count value. Select one white part of the binary image and check whether the area of the white part is less than or equal to the threshold value A. If the area of the white part is not less than or equal to the threshold value A, check whether the area of the white part is greater than or equal to the threshold value B. If the area of the white portion is not greater than or equal to the threshold value B, the central pixel of the white portion is determined, it is checked whether the reliability of the central pixel is greater than or equal to the threshold value G, and the reliability of the central pixel is greater than or equal to the threshold value G. , See the black part surrounding each white part as a ridge
None, the ridge direction is calculated from the change in pixel density,
Select endpoint candidate pixel of the white parts in the ridge direction of the straight line passing through the element, is stored in the storage unit, select the two endpoints pixel from the end point candidate pixel, and stores in the storage unit, the 2
Each of the two end point pixels at a distance from each
Select an edge pixel , store it in the storage unit,
A ridge width difference is calculated from the ridge widths of the two end pixels, is stored in the storage unit, and it is checked whether the ridge width difference is larger than a threshold value C and equal to or less than a threshold value D. If the width difference is larger than the threshold value C and equal to or smaller than the threshold value D,
The density value of the pixel of the fingerprint grayscale image corresponding to the white portion is extracted, the median value is obtained and stored in the storage unit, and it is checked whether the median value is the threshold value E or less. Then, the minimum value in each row in the X-axis direction and each column in the Y-axis direction is selected from the density values of the pixels corresponding to the white part in the fingerprint grayscale image, and X is selected from all the selected minimum values. Calculate the variance in the axis and Y-axis directions,
It is stored in the storage unit, and it is checked whether the variances in the X-axis and Y-axis directions are both equal to or less than a threshold value F, and whether the area of the white portion is equal to or less than a threshold value A or the median value is less than or equal to a threshold value E.
Alternatively, when the variance in the X-axis or Y-axis direction is not less than or equal to a threshold F, the pixel value of the white portion is set to 1, the binary image is stored in the storage unit, and the area of the white portion is a threshold value. B or more, or the reliability of the central pixel is a threshold G
If the above is not the case, the ridge width difference is not greater than the threshold value C, is not less than the threshold value D, or the variances in the X-axis and Y-axis directions are both less than the threshold value F, the white portion is an island. It is assumed that the white portion is left as a white pixel, then the count value in the storage unit is decremented by 1 and stored in the storage unit, and it is checked whether the subtracted count value is 0. , 0, it is assumed that there is a white part that has not been processed yet, and the process from the selection is executed again for the remaining white part. If the count value is 0, the process is executed for all white parts. A sweat gland hole part discriminating apparatus comprising: a sweat gland hole discriminating unit that is executed and ends. 11. A first step of creating a binary image from a fingerprint grayscale image and storing it in a storage unit, and numbering white portions of the binary image, and storing the total number in the storage unit as a count value. The second step and the white part of the binary image
And a fourth step of checking whether the area of the white portion is less than or equal to the threshold value A. If the area of the white portion is not less than or equal to the threshold value A, the area of the white portion is equal to or less than the threshold value B.
A fifth step of checking whether or not the above is true, and a sixth step of determining the central pixel of the white portion unless the area of the white portion is equal to or larger than the threshold value B, and the reliability of the central pixel is the threshold value G.
A seventh step of checking whether or not the above, and if the reliability of the central pixel is equal to or more than a threshold value G, a black part surrounding each white part
Minutes are regarded as ridges, and the ridge direction is obtained from the change in pixel density.
Therefore, an eighth step of selecting an end point candidate pixel of the white portion on a straight line in the ridge direction passing through the central pixel and storing it in the storage section, and selecting two end point pixels from the end point candidate pixel, a ninth step of storing in the storage unit, the 2
Each of the two end point pixels at a distance from each
An eleventh step of selecting an edge pixel and storing it in the storage unit, and an eleventh step of calculating a ridge width difference from the ridge widths of the center pixel and the two edge pixels and storing the ridge width difference in the storage unit. And the ridge width difference is larger than the threshold value C, and
A twelfth step of checking whether it is less than or equal to a threshold value D,
When the ridge width difference is larger than the threshold value C and equal to or smaller than the threshold value D, the density value of the pixel of the fingerprint grayscale image corresponding to the white portion is extracted, the median value is obtained, and the result is stored in the storage unit. A thirteenth step, a fourteenth step of checking whether the median value is less than or equal to a threshold value E, and when the median value is less than or equal to the threshold value E, from the density value of the pixel corresponding to the white portion in the fingerprint grayscale image, Fifteenth minimum values are selected for each row in the X-axis direction and each column in the Y-axis direction, and the variances in the X-axis and Y-axis directions are calculated from all the selected minimum values and stored in the storage unit. Step and the X axis, Y
Sixteenth step of checking whether the variances in the axial directions are both equal to or less than a threshold value F, and whether the area of the white portion is equal to or less than the threshold value A, the median value is not equal to or less than the threshold value E, or the X-axis and Y-axis directions If either of the variances is less than or equal to the threshold F, the seventeenth step of setting the pixel value of the white portion to 1 and storing the binary image in the storage unit; and the area of the white portion is equal to or greater than the threshold B. Or the reliability of the central pixel is not greater than or equal to a threshold value G, the ridge width difference is not greater than a threshold value C, is not less than or equal to a threshold value D, or the X axis, Y
When the variances in the axial direction are both equal to or less than the threshold F,
Continuing from the eighteenth step of leaving the white portion as a white pixel by assuming that the white portion is an island and the seventeenth step or the eighteenth step, the count value in the storage unit is 1 is subtracted, the result is stored in the storage unit, and it is checked whether or not the subtracted count value is 0. If it is not 0, it is determined that there is a white part that has not been processed, and the remaining white part is again described above. The process is performed from the third step, and if the count value is 0, it is determined that the process has been performed on all the white parts, and the process ends.
And a step of discriminating a sweat pore part, which comprises: 12. A first step of creating a binary image from a fingerprint grayscale image and storing it in a storage unit, and numbering white portions of the binary image, and storing the total number in the storage unit as a count value. The second step and the white part of the binary image
And a fourth step of checking whether the area of the white portion is less than or equal to the threshold value A. If the area of the white portion is not less than or equal to the threshold value A, the area of the white portion is equal to or less than the threshold value B.
A fifth step of checking whether or not the above is true, and a sixth step of determining the central pixel of the white portion unless the area of the white portion is equal to or larger than the threshold value B, and the reliability of the central pixel is the threshold value G.
A seventh step of checking whether or not the above, and an eighth step of selecting the end point candidate pixel of the white portion when the reliability of the central pixel is equal to or more than a threshold value G and storing it in the storage section;
The black portion surrounding each white portion is regarded as a ridge, and the pixel density
The ridge direction is obtained from the change, and the ridge direction passing through the central pixel
Wherein selecting two end points pixel from the end point candidate pixels on a straight line, and the ninth step of storing in the storage unit, the two
The end points of each of the pixels at a certain distance from each
A tenth step of selecting a pixel and storing it in the storage unit; and an eleventh step of calculating a ridge width difference from the ridge widths of the center pixel and the two end pixels and storing the ridge width difference in the storage unit. A twelfth step of checking whether the ridge width difference is larger than a threshold value C and not more than a threshold value D, and the ridge width difference is larger than the threshold value C and not more than the threshold value D, the white A thirteenth step of extracting density values of pixels of the fingerprint grayscale image corresponding to a portion, obtaining a median value, and storing the median value in the storage unit; and a fourteenth step of checking whether the median value is equal to or less than a threshold value E; When the median value is equal to or less than the threshold value E, the minimum value in each row in the X-axis direction and each column in the Y-axis direction is selected and selected from the density values of the pixels corresponding to the white portion in the fingerprint grayscale image. From the minimum of all Calculating a variance in the axial direction, and a fifteenth step of storing in the storage unit, the X-axis, first the variance in Y-axis direction is examined whether both a threshold F or less
6 and the area of the white portion is less than or equal to the threshold value A, the median value is not less than or equal to the threshold value E, or the X
If the variance in the axial or Y-axis direction is not less than or equal to the threshold F, a seventeenth step of setting the pixel value of the white portion to 1 and storing the binary image in the storage unit; It is equal to or greater than a threshold value B, the reliability of the central pixel is not equal to or greater than a threshold value G, the ridge width difference is not greater than a threshold value C, is not less than or equal to a threshold value D, or the dispersion in the X-axis and Y-axis directions. Are both equal to or less than the threshold F, the white portion is regarded as an island, and the white portion is left as a white pixel, and the eighteenth step, the seventeenth step, or the eighteenth step is performed. Subsequently, the count value in the storage unit is decremented by 1 and stored in the storage unit,
Check whether the reduced count value is 0, and
If not, it is assumed that there is a white part that has not been processed, and the remaining white part is executed again from the third step. If the count value is 0, the process is executed for all white parts. A computer-readable recording medium having recorded thereon a program for causing a computer to execute the nineteenth step which is completed.