JP4735372B2 - Fingerprint verification device and fingerprint verification method - Google Patents

Description

  The present invention relates to a fingerprint collation apparatus and a fingerprint collation method that determine the identity by comparing a registered fingerprint image and a collation fingerprint image.

  In a conventional fingerprint verification device, after detecting a fingerprint ridge from the input fingerprint image, feature points such as branch points and end points of the fingerprint ridge are extracted, and a registered fingerprint image (registered) from the positional relationship of the feature points A method of determining the identity between a fingerprint image) and a fingerprint image (verification fingerprint image) input as a verification target is common.

  However, since the judgment based on only the feature points may cause the collation judgment to be wrong, a fingerprint collation device that judges the identity between the registered fingerprint image and the collation fingerprint image in consideration of elements other than the feature points is also considered ( For example, see Patent Documents 1 and 2).

  In the fingerprint collation device described in Patent Document 1, a wrinkle or a flaw of a finger is extracted from a fingerprint image, and feature point information related to the wrinkle or flaw is used as fingerprint data together with information of a feature point of a fingerprint ridge. To verify.

In addition, the fingerprint collation device described in Patent Document 2 extracts feature points (end points, branch points) from a fingerprint image, and detects the type and direction of singular points such as vortices in addition to the position and type of feature points. And verify.
JP 2005-352712 A Japanese Patent Laid-Open No. 11-39478

  As described above, in the conventional fingerprint collation apparatus, collation is performed using not only feature points but also other information.

  However, finger wrinkles and scratches that are features other than the feature points in Patent Document 1 do not necessarily exist in the fingerprint image, and therefore there are fingerprint images that cannot be collated using these features.

  Further, in Patent Document 2, singular points are detected and characterized by the type and direction of the singular points. However, since the singular points are not necessarily present in the fingerprint image, collation using these features cannot be performed. A fingerprint image exists.

  An object of the present invention is to provide a fingerprint collation apparatus and a fingerprint collation method capable of improving the accuracy of collation judgment by collating the relationship between a feature point extracted from a fingerprint image and a fingerprint ridge as a feature quantity. There is to do.

The invention according to claim 1 is a fingerprint collation device that compares a registered fingerprint image and a collation fingerprint image to determine identity, and the fingerprint ridges are obtained from the respective fingerprint images of the registration fingerprint image and the collation fingerprint image. a ridge extracting means for extracting the registered fingerprint image and the respective feature point extracting means for extracting a plurality of feature points from the fingerprint image of the collation fingerprint image, one of the features extracted by the front Symbol feature point extracting means eggplant and vector setting means for setting a vector directed to other feature points in a predetermined range from the point, and fingerprint ridge extracted by the ridge extracting means for crossing the vector and set vector by the vector setting means calculating the angles, set by the vector setting means in each of the ridge direction calculating means for this angle and ridge direction, and the registration fingerprint image the collation fingerprint image Vectors by comparing the direction of the calculated fingerprint ridge by the ridge direction calculating means with comparing, by comprising a determination means for determining identity of the collation fingerprint image and the registered fingerprint image Features.

According to a second aspect of the present invention, in the first aspect of the present invention, the horizontal component and the vertical component of the vector are used as the feature amount of the vector set by the vector setting unit from each of the registered fingerprint image and the verification fingerprint image. And a vector feature amount calculation means for calculating a vector length , wherein the determination means compares the feature amounts of the vectors calculated by the vector feature amount calculation means, and compares the registered fingerprint image and the verification fingerprint image. It is characterized by determining identity.

  According to a third aspect of the present invention, in the first aspect of the invention, the ridge direction calculation means selects a fingerprint ridge that is closest to the midpoint of the vector set by the vector setting means, and this selection is performed. A point is calculated by tracing the fingerprint ridge to the predetermined range from the intersection of the fingerprint ridge and the vector, and the angle between the vector from the intersection to the point and the vector set by the vector setting means is calculated. The direction of the fingerprint ridge is obtained.

According to a fourth aspect of the present invention, in the first aspect of the invention, the ridge direction calculation means may determine whether the one feature point among the plurality of fingerprint ridges crossing the vector set by the vector setting means. A fingerprint ridge that is closest to any one of the other feature points is selected, a point is calculated by tracing the fingerprint ridge from the intersection of each selected fingerprint ridge to the vector within a predetermined range, An angle formed by a vector from the intersection to the point and a vector set by the vector setting means is obtained and set as the direction of the fingerprint ridge.

According to a fifth aspect of the present invention, in the first aspect of the invention, the ridge direction calculation means selects a plurality of fingerprint ridges crossing the vector set by the vector setting means, and the plurality of selected fingerprints. For each ridge, a point is calculated by tracing the fingerprint ridge from the intersection of the fingerprint ridge and the vector to a predetermined range, and the vector from the intersection to the point is set by the vector setting means An average value of angles formed with the vector is obtained and set as the direction of the fingerprint ridge.

The invention according to claim 6 is a fingerprint collation method for comparing the registered fingerprint image and the collation fingerprint image to determine the identity, and the fingerprint image is obtained from each fingerprint image of the registration fingerprint image and the collation fingerprint image. a ridge extracting step of extracting a line, the feature point extraction step of extracting a plurality of feature points from each of the fingerprint images of said collation fingerprint image and the registered fingerprint image, which is extracted by the pre-Symbol feature point extraction process 1 A vector setting step for setting a vector from one feature point toward another feature point within a certain range, a vector set by the vector setting step, and a fingerprint ridge extracted by the ridge extraction step crossing the vector ; calculating an angle, and ridge direction calculation step for this angle and ridge direction, the vector setting means in each of the collation fingerprint image and the registered fingerprint image By comparing the direction of the fingerprint ridge calculated by the ridge direction calculating process as well as comparing the set vector that had a determining step the identity of the collation fingerprint image and the registered fingerprint image It is characterized by.

According to the present invention , the relationship between the feature points extracted from the fingerprint images of the registered fingerprint image and the verification fingerprint image and the fingerprint ridges, that is, the vector between the feature points and the direction of the fingerprint ridge crossing this vector are compared. By doing so, it is possible to improve the accuracy of collation determination.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram illustrating a configuration of a fingerprint collation apparatus according to the present embodiment.
As shown in FIG. 1, the fingerprint collation apparatus according to the present embodiment includes a control unit 10, a display unit 12, a key input unit 14, an image reading unit 16, a storage device 20, and a RAM 22.

  The control unit 10 controls the entire apparatus, and includes a processor, a memory for storing programs and data, and the like. The control unit 10 implements various functions by executing various programs by a processor. In the present embodiment, registration processing is executed by the registration program 20 a, whereby registration image data is generated based on the fingerprint image data detected by the image reading unit 16 and stored in the storage device 20. In addition, by performing collation processing by the collation program 20b, collation for determining whether the fingerprint image data for collation detected by the image reading unit 16 matches the registered image data registered by the registration processing. I do.

  The display unit 12 displays various data on, for example, a liquid crystal display under the control of the control unit 10.

  The key input unit 14 inputs various instructions according to user operations under the control of the control unit 10. The key input unit 14 inputs an instruction from a user according to an operation on another input device such as a pointing device in addition to an input operation on a plurality of keys (buttons), for example, and notifies the control unit 10 of the instruction.

  The image reading unit 16 detects the fingerprint image and outputs it to the control unit 10.

  The storage device 20 is for storing programs and various data, and includes a ROM, a hard disk device, a flash memory, and the like. The programs stored in the storage device 20 include a registration program 20a for registration processing for registering a fingerprint image, and a verification program 20b for verification processing for verifying a fingerprint image for verification and a registered fingerprint image. . The storage device 20 stores registered image data 20c registered by the registration process. The registered image data 20c includes registered image data generated based on fingerprint images collected from a plurality of persons. The registered image data 20c in the present embodiment includes data representing the feature amount extracted from the fingerprint image (details are shown in FIG. 6).

  The RAM 22 stores data to be processed by the control unit 10, for example, registered image data 22 a read from the storage device 20 for use in the verification process, or a verification target read from the image reading unit 16. In addition to the collation image data 22b, various data associated with the execution of the collation process are stored.

Next, the operation of the fingerprint collation device in this embodiment will be described.
First, registration processing for registering a fingerprint image will be described with reference to the flowcharts shown in FIGS. 2, 3, and 4.

  FIG. 2 is a flowchart for explaining the flow of the entire registration process.

  First, the control unit 10 reads a fingerprint image to be registered by the image reading unit 16 (step A1).

  The control unit 10 performs binarization and thinning image processing on the fingerprint image read by the image reading unit 16 to generate an image representing a fingerprint ridge pattern (step A2).

  Next, the control unit 10 extracts feature points from the fingerprint ridge image (step A3). As the feature points, for example, end points where the fingerprint ridges are interrupted and branch points where the fingerprint ridges branch are extracted as feature points.

  FIG. 5A is a diagram conceptually showing a plurality of feature points extracted from the fingerprint image. In FIG. 5A, a fingerprint ridge pattern is indicated by a broken line, and a plurality of feature points EM1, EM2,... EM8 are extracted. Usually, the feature point is when the quality of the fingerprint image read by the image reading unit 16 is poor, for example, when the fingertip has wrinkles or scratches, or when the fingerprint ridge at the singular point is not clear, Multiple extractions are possible. Also, even if the fingerprint ridge is partially unclear, it can be detected in a wide range of the fingerprint image. As will be described later, the fingerprint collation device according to the present embodiment uses a plurality of features and fingerprint ridges that are almost certainly extracted from a fingerprint image, and extracts the relationship between the plurality of features and the fingerprint ridges as feature amounts. The collation process is executed.

  Next, the control unit 10 executes a feature amount calculation process for extracting a feature amount for each feature point extracted from the fingerprint image (step A4). In the feature amount calculation processing in the present embodiment, a vector from one feature point to another feature point within a certain range is set, and the direction of the fingerprint ridge that crosses this vector is calculated as the feature amount. Details will be described later (FIG. 3).

  The control unit 10 causes the storage device 20 to store data relating to the feature amount calculated by the feature amount calculation process as the registered image data 20c (step A5).

  Next, the details of the feature amount calculation processing in the flowchart shown in FIG. 2 will be described with reference to the flowchart shown in FIG. FIG. 6 shows an example of feature amount data registered as registered image data. 7 and 8 are diagrams showing an example of a fingerprint ridge pattern for explaining the feature amount calculation process (ridge direction calculation process).

  Here, the case where the feature point data shown in FIG. 7 is used as a reference feature point and the feature amount data for the reference feature point M is calculated will be described as an example.

  First, the control unit 10 selects neighboring feature points whose distance from the reference feature point M in the fingerprint ridge pattern is equal to or smaller than a preset threshold value ThL, and sets the total number of neighboring feature points (total number of feature points) TNm. Calculate (step B1). In the example shown in FIG. 7, five feature points at a distance equal to or smaller than the threshold ThL from the reference feature point M are extracted, and the total feature point TNm is 5.

  The control unit 10 arranges the neighboring feature points in order from the shortest distance from the reference feature point M, and sets them as feature points m1 to m5 as shown in FIG. Here, the vectors V1 to V5 heading to the neighboring feature points m1 to m5 with the reference feature point M as a reference are set, and the horizontal component VX, the vertical component VY, and the vector length VL of each vector V1 to V5 are calculated ( Step B2).

  The control unit 10 calculates an angle θ between each vector V1 to V5 and a vector having the next longest vector length VL (step B3). Here, the clockwise direction is the positive direction, and the unit is [deg]. In the example shown in FIG. 7, the angle θ2 formed by the vector V1 and the vector V2, the angle θ3 formed by the vector V2 and the vector V3, the angle θ4 formed by the vector V3 and the vector V4, and the angle θ5 formed by the vector V4 and the vector V5. Are calculated respectively.

Next, the control unit 10 executes a ridge direction calculation process to calculate the direction of the fingerprint ridge that crosses the vector from the reference feature point M to each of the neighboring feature points m1 to m5 (step B4).

  FIG. 4 is a flowchart for explaining details of the ridge direction calculation processing. FIG. 8 shows an example of a fingerprint ridge pattern for explaining the ridge direction calculation processing. FIG. 8 shows a vector Vm of the reference feature point M and one neighboring feature point m.

  The control unit 10 selects the fingerprint ridge L that intersects the vector Vm from the reference feature point M to the neighboring feature point m. In FIG. 8, one fingerprint ridge L that is closest to the intermediate point x of the vector Vm is selected. The control unit 10 calculates the coordinates (Xc, Yc) of the intersection C, where C is the intersection of the fingerprint ridge L selected here and the vector Vm (step C1).

Next, the control unit 10 tracks the fingerprint ridge L from the intersection C in the same direction as the X direction of the vector Vm, and calculates a tracking point S that satisfies the following expression (1). Here, it is assumed that the coordinates of the tracking point S are (x, y).

  ThD is an upper limit value set in advance for the distance D from the intersection C to the tracking point S. The tracking point S at which the maximum value satisfying the equation (1) is obtained is calculated. Here, the coordinates of the tracking point S are (Xs, Ys).

  Next, the control unit 10 calculates a vector CS (step C3). That is, the vector CS = (Xs−Xc, Ys−Yc).

  The control unit 10 calculates an angle φ formed by the vector Vm and the vector CS (a clockwise direction is a positive direction and a unit is [deg]), and a vector from the reference feature point M to the neighboring feature point m is calculated. The direction LD of the fingerprint ridge that crosses is set (step C4). That is, LD = φ. However, if there is no fingerprint ridge that crosses the vector from the reference feature point M toward the neighboring feature point m, the fingerprint ridge direction LD = 180.

The control unit 10 generates feature amount data for the reference feature point M based on each data calculated by the feature amount calculation process (ridge direction calculation process).
The control unit 10 executes the above feature amount calculation processing for each feature point extracted from the fingerprint image. Then, as illustrated in FIG. 6, the control unit 10 collectively encodes the feature amount data and generates the feature amount data code. In the registration process, the feature data code is stored as registered image data.

  As shown in FIG. 6A, the feature amount data code includes the total number of feature points extracted from the fingerprint image (total feature point TNm) and feature amount data (F1,..., Fn (n) for each feature point. = TNm)).

  As shown in FIG. 6B, each feature amount data includes a horizontal coordinate (X), a vertical coordinate (Y), a feature point type (T) such as an end point or a branch point, a direction ( D), the number of feature points in the designated range (Nm), and feature quantities (VF1,..., VFm (m = Nm)) related to the vectors directed to the feature points in the designated range. As shown in FIG. 6C, the feature quantity related to the vector directed to the feature point within the designated range includes the type ET of the end point feature point (feature point within the designated range), the direction ED of the end point feature point, and the horizontal component VX of the vector. , Vertical component VY, vector length VL, vector direction VD, and ridge direction LD across the vector are included.

  Next, collation processing for collating fingerprint images will be described with reference to the flowcharts shown in FIGS.

  FIG. 9 is a flowchart for explaining the flow of the entire verification process.

  First, the control unit 10 reads a fingerprint image to be verified by the image reading unit 16 (step D1).

  The control unit 10 performs binarization and thinning image processing on the fingerprint image read by the image reading unit 16 to generate an image representing a fingerprint ridge pattern (step D2).

  Next, the control unit 10 extracts feature points from the fingerprint ridge image (step D3). As the feature points, for example, the end points where the fingerprint ridges are interrupted and the branch points where the fingerprint ridges branch are extracted as feature points in the same manner as in the registration process.

  FIG. 5B is a diagram conceptually showing a plurality of feature points extracted from the fingerprint image. In FIG. 5 (b), a fingerprint ridge pattern is indicated by a broken line, and a plurality of feature points VM1, VM2,... VM8 are extracted.

  Next, the control unit 10 executes a feature amount calculation process for extracting a feature amount for each feature point extracted from the fingerprint image (step D4). Note that the feature amount calculation process in the matching process is executed in the same manner as the registration process, and the description thereof is omitted (see FIGS. 2, 3, 4, 7, and 8). The control unit 10 causes the RAM 22 to store the feature amount code data generated based on the fingerprint image for collation by the feature amount calculation process as the collation image data 22b.

  Next, the control unit 10 compares the comparison image data 22b with the registered image data 20c read from the storage device 20 and stored in the RAM 22 (registered image data 22a) by executing a match determination process. The identity between the registered fingerprint image and the collation fingerprint image is determined, and the determination result is output (steps D5 and D6).

  Next, the match determination processing in the flowchart shown in FIG. 9 will be described with reference to the flowchart shown in FIG.

  In the following description, it is assumed that the feature quantity for the feature point EM included in the registered image data 20c (22a) is EMF, and the feature quantity for the vector toward the feature point in the specified range of the feature quantity EMF for the feature point EM is EVF. Further, the feature amount for the feature point VM included in the collation image data 22b is defined as VMF, and the feature amount for a vector directed to the feature point within the specified range of the feature amount VMF for the feature point VM is defined as VVF.

  First, the control unit 10 selects a feature point EM in the registered image to be compared with the collation image data 22b (Step E1).

  As condition 1, the control unit 10 compares the feature point type T, the direction D, and the number of feature points in the specified range (number of nearby feature points) TNm (step E2). That is, the control unit 10 selects the type T and direction D of the selected feature point EM, the number of feature points TNm in the designated range, the type T and direction D of the first feature point VM included in the collation image data 22b, and the features in the designated range. The score TNm is compared with each other.

  It is determined that the condition T is satisfied when the type T is the same, and the direction D and the number of feature points TNm within the specified range are within the threshold range set for each.

  When it is determined that the condition 1 is not satisfied (step E3, No), the control unit 10 sets the number of matching vectors MVN = 0 for the current feature point VM (step E6).

  On the other hand, when it is determined that the condition 1 is satisfied (step E3, Yes), the control unit 10 determines a vector from the feature point EM in the registered fingerprint image to another feature point in the specified range, and a feature point in the collation fingerprint image. A comparison is made with a vector from the VM toward another feature point within the specified range, that is, a comparison is made between the feature value EVF and the feature value VVF (step E4). In the comparison between the feature quantity EVF and the feature quantity VVF, the control unit 10 includes, as condition 2, the type ET of the end point feature point included in the feature quantity related to the vector toward the feature point in the specified range, the direction ED of the end point feature point, the vector Length VL, vector direction VD, and ridge direction LD crossing the vector, respectively.

  Same for type ET, end point feature point direction ED, vector length VL, vector direction VD, and ridge direction LD crossing the vector, the difference between them is within the threshold range set for each In some cases, it is determined that Condition 2 is satisfied.

  The control unit 10 determines the above-described condition 2 for all combinations of feature amounts relating to vectors directed to the feature points within each specified range, which are included in the feature amount EVF and the feature amount VVF, respectively. As a result, when there is a vector (match vector) determined to satisfy the condition 2, the control unit 10 stores the number of match vectors (step E5).

  Next, if there is a next feature point VM in the collation image data 22b (step E7, Yes), the control unit 10 executes the same processing as described above for the next feature point VM (steps E2 to E5). . Thereafter, the same processing is executed for all remaining feature points VM included in the collation image data 22b, and the number of matching vectors for each feature point VM is obtained.

  When the comparison with the feature point EM selected in step E1 is completed for all feature points VM (No in step E7), the control unit 10 selects the feature point VM having the maximum number of matching vectors. The number of matching vectors at this time is stored as the similarity S with the feature point EM (step E8).

  Next, if there is a next feature point EM in the registered image data 22a (step E9, Yes), the control unit 10 performs the same processing as described above for the next feature point EM (steps E2 to E8). . Thereafter, similarly, the same processing is executed for all the remaining feature points EM included in the collation image data 22b, and the similarity S for each feature point EM is obtained.

  Next, the control part 10 discriminate | determines as the condition 3 whether the similarity S calculated | required about each feature point EM is a preset threshold value. Then, the number of threshold values S that satisfy condition 3 is counted (step E10).

  Next, as the condition 4, the control unit 10 determines that the number of similarities S counted based on the determination result of the condition 3 (corresponding to the number of feature points EM) is the registered image data 22a and the collation image data 22b. Are determined to be equal to or greater than a threshold value set in advance as a reference value for determining that they are the same.

  If it is determined that the condition 4 is satisfied (step E11, Yes), the control unit 10 determines that the registered image data 22a and the matching image data 22b match (step E12).

  On the other hand, when it is determined that the condition 4 is not satisfied (No at Step E11), the control unit 10 determines that the registered image data 22a and the matching image data 22b do not match (Step E13).

  Thus, in the comparison process (match determination process), the ridge direction LD across the vector is included in the condition in the comparison between the feature quantity EVF and the feature quantity VVF. Even if it is not possible to make an accurate determination by comparison, it is possible to make a more accurate determination.

  In the above description, the direction of one fingerprint ridge that intersects the vector Vm from the reference feature point M to the neighboring feature point m is extracted as a feature quantity. However, a plurality of fingerprint ridges are included in the vector Vm. When intersecting (crossing), a direction (angle) may be calculated from each of the plurality of fingerprint ridges and used as a feature amount used for matching. In this case, all fingerprint ridges may be targeted, or some of the plurality of fingerprint ridges may be targeted. When calculating a direction by selecting a plurality of fingerprint ridges, each direction may be used as a feature amount, or an average value of directions (angles) may be calculated and used as a feature amount.

  In addition, an intermediate point that intersects the vector Vm and the nearest fingerprint ridge are selected, and the direction of the fingerprint ridge is extracted as a feature quantity. The direction may be extracted by selecting a nearby fingerprint ridge. In this case, an average value in the direction calculated from each fingerprint ridge may be calculated as the feature amount. Further, the direction may be extracted by selecting the nearest fingerprint ridge for either the reference feature point M or the neighboring feature point m.

  In addition, three fingerprint ridges, that is, an intermediate point intersecting the vector Vm and the nearest fingerprint ridge, each of the reference feature point M and the neighborhood feature point m, and the nearest fingerprint ridge are selected, and each fingerprint is selected. The direction calculated from the ridge may be used as the feature amount.

  In addition, a registration program and a verification program that can be executed by a computer that implements a fingerprint verification device are stored in a recording medium such as a magnetic disk (flexible disk, hard disk, etc.), an optical disk (CD-ROM, DVD, etc.), a semiconductor memory, etc. It can be written or provided through a communication medium. By controlling the operation of the computer by the registration program and the verification program, the functions in the above-described embodiments can be realized.

The block diagram which shows the structure of the fingerprint collation apparatus in embodiment of this invention. The flowchart for demonstrating the flow of the process of the whole registration process. The flowchart for demonstrating the detail of a feature-value calculation process. The flowchart for demonstrating the detail of a ridge direction calculation process. The figure which shows notionally several feature points extracted from a fingerprint image. The figure which shows an example of the feature-value data registered as registration image data. The figure which shows an example of the fingerprint ridge pattern for demonstrating a feature-value calculation process. The figure which shows an example of the fingerprint ridge pattern for demonstrating the feature-value calculation process (ridge direction calculation process). The flowchart for demonstrating the flow of the whole collation process. The flowchart for demonstrating a coincidence determination process.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Control part, 12 ... Display part, 14 ... Key input part, 16 ... Image reading part, 20 ... Memory | storage device, 20a ... Registration program, 20b ... Collation program, 22 ... RAM, 20c, 22a ... Registration image data, 22b ... collation image data.

Claims (6)

A fingerprint collation device that compares a registered fingerprint image and a collation fingerprint image to determine identity,
A ridge extraction means for extracting a fingerprint ridge from the fingerprint image of each of the registered fingerprint image and the verification fingerprint image ;
A feature point extracting means for extracting a plurality of feature points from each fingerprint image of the registered fingerprint image and the verification fingerprint image ;
And vector setting means for setting a vector from one feature point extracted by pre-Symbol feature point extracting means toward the other feature points in the predetermined range,
Calculating an angle between the vector set by the vector setting unit and the fingerprint ridge extracted by the ridge extraction unit crossing the vector, and setting the angle as a ridge direction ;
The registered fingerprint image is compared with the vector set by the vector setting unit in each of the registered fingerprint image and the direction of the fingerprint ridge calculated by the ridge direction calculating unit. A fingerprint collation apparatus comprising: a determination unit that determines the identity of the collation fingerprint image. A vector feature amount calculating unit that calculates a horizontal component, a vertical component, and a vector length of the vector as the feature amount of the vector set by the vector setting unit from each of the registered fingerprint image and the collation fingerprint image;
The determination means compares the feature quantity vector calculated by the vector feature quantity calculating means, a fingerprint according to claim 1, wherein the determining the identity of the collation fingerprint image and the registered fingerprint image Verification device. The ridge direction calculation means includes:
Select the fingerprint ridge that is closest to the midpoint of the vector set by the vector setting means,
Trace the point on the fingerprint ridge from the intersection of the selected fingerprint ridge and the vector to a predetermined range, and calculate the point.
2. The fingerprint collation apparatus according to claim 1, wherein an angle formed by a vector from the intersection to the point and a vector set by the vector setting unit is obtained and set as a direction of the fingerprint ridge. The ridge direction calculation means includes:
Among the plurality of fingerprint ridges crossing the vector set by the vector setting means , select the fingerprint ridge closest to any one of the one feature point or the other feature points ,
The point is calculated by tracing the fingerprint ridge to the predetermined range from the intersection of each selected fingerprint ridge and the vector,
Seeking an angle between the set vector by vector and the vector setting means to said point from the intersection of the fingerprint collation apparatus according to claim 1, characterized in that the direction of the fingerprint ridge. The ridge direction calculation means includes:
Selecting a plurality of fingerprint ridges across the vector set by the vector setting means;
For each of the selected plurality of fingerprint ridges, a point is calculated by tracing the fingerprint ridges from the intersection of the fingerprint ridges and the vector to a predetermined range,
2. The fingerprint collation apparatus according to claim 1, wherein an average value of an angle formed by a vector from the intersection to the point and a vector set by the vector setting unit is obtained and set as a direction of the fingerprint ridge. . A fingerprint verification method for comparing identity between a registered fingerprint image and a verification fingerprint image,
A ridge extraction step of extracting a fingerprint ridge from each fingerprint image of the registered fingerprint image and the verification fingerprint image ;
A feature point extraction step of extracting a plurality of feature points from each of the fingerprint images of the registered fingerprint image and the verification fingerprint image ;
And a vector setting step of setting a vector from one feature point extracted by pre-Symbol feature point extraction stroke towards the other feature points in the predetermined range,
Calculating an angle between the vector set in the vector setting step and the fingerprint ridge extracted in the ridge extraction step crossing the vector , and a ridge direction calculating step in which the angle is a ridge direction ;
Compare the vector set by the vector setting means in each of the registered fingerprint image and the verification fingerprint image and compare the direction of the fingerprint ridge calculated by the ridge direction calculation step, and the registered fingerprint image And a determination step of determining the identity of the verification fingerprint image.

Images