JP2006133979A - Fingerprint collation device and fingerprint collation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fingerprint collation device capable of easy fingerprint collation without imposing any troublesome operation on a user. <P>SOLUTION: A fingerprint collation device records the registered fingerprint image of a person to be authenticated and quality information in an IC card 6. Then, an MPU 5 calculates the security quality level, concretely, decision level of collation matched with the fingerprint based on quality information recorded in the IC card 6. A fingerprint data inputting/collating part 2 calculates matching level information showing the matching level of the input fingerprint image and registered fingerprint image of the person to be authenticated, and decides whether or not the input fingerprint image is the same as the registered fingerprint image based on the matching level information and the calculated decision level. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a fingerprint collation device and a fingerprint collation method.

  In a conventional fingerprint collation apparatus, a collation security quality level (a parameter that determines the strictness of collation, for example, a determination level) used when performing fingerprint collation is matched with a normal quality fingerprint. The normal quality fingerprint is a fingerprint taken from an average finger (for example, an average number of fingerprint feature points and an average finger wetness).

  Therefore, a fingerprint image read from a wet finger of a person who sweats easily, a fingerprint image read from a dry finger of a person with dry skin, or a fingerprint image read from a finger having a fingerprint with fewer feature points than the average However, when used for fingerprint collation, collation errors increase and accurate fingerprint collation cannot be performed.

  Also, in order to reduce collation errors for all fingerprints, if the collation security quality level (judgment level) is lowered and fingerprints are collated, the fingerprint collation device acceptability rate increases, and the fingerprint collation device is practically used. It becomes unusable.

  In Patent Document 1 (Japanese Patent Laid-Open No. 2002-298126), it is determined whether or not a fingerprint image (hereinafter referred to as “input fingerprint image”) input by a user is suitable for fingerprint verification, and the input is performed. If the fingerprint image is not suitable for fingerprint verification, fingerprint verification is not performed, and the fingerprint that informs the user why the input fingerprint image is not suitable for fingerprint verification (eg dry finger or wet finger) and countermeasures A verification device is described.

  According to this fingerprint collation device, even if the collation security quality level is set to a normal quality fingerprint, the user can perform fingerprint collation by taking measures known from the fingerprint collation device.

Further, in Patent Document 2 (Japanese Patent Laid-Open No. 11-306352), a plurality of types of authentication are performed using a plurality of types of collation data each representing a biometric feature, the plurality of types of authentication results, and an authentication threshold value. Based on the above, a personal authentication device that performs integrated authentication is described. The personal authentication device further changes the authentication threshold based on the integrated authentication result.
JP 2002-298126 A JP-A-11-306352

  In the fingerprint collation device described in Patent Document 1, the user who has a dry finger or a wet finger recognizes the reason why the input fingerprint image is not suitable for fingerprint collation and the countermeasure, and then reprints his / her fingerprint again. You have to do the troublesome work of making it read.

  Moreover, in the personal authentication device described in Patent Document 2, a plurality of types of collation data representing biometric features are used. For this reason, a plurality of collation processes are required, and the collation process becomes complicated.

  An object of the present invention is to provide a fingerprint collation apparatus and a fingerprint collation method capable of easily performing fingerprint collation without forcing a user to perform troublesome work.

  To achieve the above object, a fingerprint collation device of the present invention reads a user's fingerprint and outputs an input fingerprint image indicating the read fingerprint, and an image for storing a registered fingerprint image registered in advance. The input fingerprint image output from the storage unit and the reading unit is collated with the registered fingerprint image stored in the image storage unit to generate the matching degree information indicating the matching degree, and the matching degree information and the registration A verification determination unit that determines whether or not the input fingerprint image is the same as the registered fingerprint image based on a determination level corresponding to the quality of the fingerprint image.

  The fingerprint collation method of the present invention is a fingerprint collation method performed by a fingerprint collation apparatus including an image storage unit that stores a registered fingerprint image registered in advance, and reads a user's fingerprint and indicates the read fingerprint. A reading step of outputting an input fingerprint image, a matching degree output step of collating the input fingerprint image with a registered fingerprint image stored in the image storage unit and outputting matching degree information indicating the degree of matching thereof, And a collation determination step of determining whether or not the input fingerprint image is the same as the registered fingerprint image based on the matching degree information and a determination level corresponding to the quality of the registered fingerprint image.

  According to the above invention, the determination level depends on the quality of the registered fingerprint image. The determination level is used to determine whether the input fingerprint image is the same as the registered fingerprint image. For this reason, even if fingerprint images of various qualities are used for collation, it is possible to determine whether or not the input fingerprint image is the same as the registered fingerprint image at a determination level corresponding to the quality. Therefore, the number of verification retries and verification errors due to the difference between the actual fingerprint quality and the quality of the fingerprint used to set the judgment level is reduced. Fingerprint verification can be performed.

  Note that quality information indicating the quality of the registered fingerprint image is received, a determination level is set according to the quality information, and the input fingerprint image is the registered fingerprint when the matching level information exceeds the determination level. It is desirable to determine that the image is the same.

  According to the above invention, it is not necessary for the fingerprint collation apparatus to always store and manage quality information.

  Further, it is desirable that the determination level decreases as the quality of the user's fingerprint indicated by the quality information deteriorates.

  According to the above invention, it is possible to accurately determine whether or not the input fingerprint image is the same as the registered fingerprint image according to the quality.

  Further, the quality information and the registered fingerprint image are read from the portable recording medium on which the quality information and the registered fingerprint image are recorded, and the read registered fingerprint image is stored in the image storage unit. desirable.

  According to the above invention, the quality information and the registered fingerprint image can be read from a common portable recording medium. Further, the user can manage quality information together with the registered fingerprint image by managing the portable recording medium.

  The portable recording medium is preferably an IC card.

  According to the present invention, even when fingerprint images of various qualities are used for collation, it is possible to determine whether or not the input fingerprint image is the same as the registered fingerprint image at a determination level corresponding to the quality. Become. Therefore, the number of verification retries and verification errors due to the difference between the actual fingerprint quality and the fingerprint quality used to set the judgment level is reduced, and the user is not forced to perform any troublesome work. Fingerprint verification can be performed.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing a fingerprint collation apparatus according to an embodiment of the present invention.

  1, the fingerprint collation apparatus includes a communication control unit 1, a fingerprint data input / collation unit 2, an IC card input / output control unit 3, a storage unit 4, and an MPU 5.

  The communication control unit 1 controls communication with an external device (for example, an input device or a host device) and monitors connection with the external device. Further, the communication control unit 1 analyzes a command command provided from an external device.

  The fingerprint data input / collation unit 2 controls the input of fingerprint data and extracts feature points from the fingerprint data. The fingerprint data input / collation unit 2 collates the input fingerprint data (hereinafter referred to as “input fingerprint image”) with fingerprint data stored in the storage unit 4 (hereinafter referred to as “registered fingerprint image”). .

  The IC card input / output control unit 3 records a registered fingerprint image and quality information indicating the quality of the registered fingerprint image on the IC card 6. Further, the IC card input / output control unit 3 reads a registered fingerprint image and quality information from the IC card 6.

  The IC card 6 is an example of a portable recording medium, and includes a registered fingerprint image storage unit 61 and a quality information storage unit 62. A registered fingerprint image is recorded in the registered fingerprint image storage unit 61, and quality information is recorded in the quality information storage unit 62.

  The storage unit 4 stores the registered fingerprint image read by the IC card input / output control unit 3.

  The MPU 5 is responsible for all processing of the fingerprint collation device.

  The MPU 5 sets a determination level according to the quality information read by the IC card input / output control unit 3. This determination level is a security quality level of collation (a parameter that determines the strictness of collation). If this determination level is low, the rejection rate of the principal decreases, but the acceptance rate of others increases. The MPU 5 lowers the determination level as the quality of the registered fingerprint image indicated by the quality information deteriorates. Therefore, the MPU 5 changes the determination level according to the quality information.

  In this embodiment, the MPU 5 sets the determination level using values from 0 to 100.

  The MPU 5 registers the determination level in the fingerprint data input / collation unit 2.

  FIG. 2 is a block diagram showing an example of the fingerprint data input / collation unit 2. In FIG. 2, the same components as those shown in FIG.

  In FIG. 2, the fingerprint data input / collation unit 2 includes a reading unit 21 and a collation determination unit 25. The collation determination unit 25 includes a collation unit 22, a level storage unit 23, and a determination unit 24.

  The reading unit 21 is a fingerprint sensor, reads a user's fingerprint, and outputs an input fingerprint image indicating the read fingerprint to the collation unit 22.

  The collation determination unit 25 collates the input fingerprint image output from the reading unit 21 with the registered fingerprint image stored in the storage unit 4 and generates coincidence degree information indicating the degree of coincidence thereof. In addition, the collation determination unit 25 determines whether or not the input fingerprint image is the same as the registered fingerprint image based on the matching degree information and the determination level corresponding to the quality of the registered fingerprint image.

  The collation unit 22 receives the input fingerprint image output from the reading unit 21, collates the input fingerprint image with the registered fingerprint image stored in the storage unit 4, and outputs coincidence information indicating the degree of coincidence. .

  Specifically, the collation unit 22 extracts feature points (hereinafter referred to as “input feature points”) from the input fingerprint image received from the reading unit 21.

  The matching unit 22 reads the registered fingerprint image stored in the storage unit 4 and extracts feature points (hereinafter referred to as “registered feature points”) from the read registered fingerprint image.

  The collation unit 22 collates the input feature points with the registered feature points, and outputs coincidence information indicating the coincidence. For example, the collation unit 22 outputs the ratio of the number of registered feature points corresponding to the input feature points to the total number of registered feature points as the matching degree information.

  In the present embodiment, the matching unit 22 divides the number of registered feature points corresponding to the input feature points by the number of all registered feature points, and outputs a value obtained by multiplying the value by 100 as matching degree information.

  The level storage unit 23 stores the determination level set by the MPU 5.

  The determination unit 24 determines that the input fingerprint image is the same as the registered fingerprint image when the coincidence degree information output from the collation unit 22 exceeds the determination level stored in the level storage unit 23.

  Next, an outline of the operation will be described.

  In the present embodiment, when the person to be authenticated operates an external device to register his / her fingerprint on the IC card 6, the fingerprint collation device will print the fingerprint image (registered fingerprint image) of the person to be authenticated and the quality information of the fingerprint. Is recorded on the IC card 6.

  When the IC card input / output control unit reads the registered fingerprint image and quality information recorded on the IC card 6 for personal authentication, the MPU 5 uses the quality information to verify the security quality of the matching for the registered fingerprint image. A level, specifically, a determination level is calculated.

  In addition, when the reading unit 21 reads the fingerprint of the person to be authenticated and outputs an input fingerprint image for personal authentication, the collation unit 22 collates the input fingerprint image with the registered fingerprint image read from the IC card 6. Then, the degree of coincidence information indicating the degree of coincidence is output.

  The determination unit 24 determines whether or not the input fingerprint image is the same as the registered fingerprint image based on the matching degree information output from the collation unit 22 and the calculated determination level.

  Next, the operation will be described.

  First, an operation for registering registered fingerprint information and quality information in the IC card 6 will be described.

  When the communication control unit 1 receives a fingerprint registration request from an external device, the fingerprint data input / verification unit 2 reads the fingerprint of the person to be authenticated, and based on the fingerprint image indicating the read fingerprint, quality information of the fingerprint Is calculated.

  For example, when the collation unit 22 accepts a fingerprint image composed of a plurality of pixels from the reading unit 21, the collation unit 22 takes the difference between the maximum pixel output and the minimum pixel output in the fingerprint image, and uses the value as the pixel. The value obtained by dividing the output gradation (predetermined maximum pixel output difference) and multiplying the value by 100 is rounded down to the quality information. In this case, the quality information is a value from 0 to 100, and the quality information value is closer to 100 as the quality of the fingerprint is higher.

  When the fingerprint data input / collation unit 2 calculates the quality information, the MPU 5 writes the quality information and the fingerprint image received by the reading unit 21 to the IC card 6 via the IC card input / output control unit 3.

  Specifically, the MPU 5 causes the IC card input / output control unit 3 to record the registered fingerprint image storage unit 61 of the IC card 6 using the fingerprint image received by the reading unit 21 as a registered fingerprint image. Further, the MPU 5 causes the IC card input / output control unit 3 to record the quality information calculated by the verification unit 22 in the quality information storage unit 62 of the IC card 6.

  Next, the fingerprint collation operation will be described.

  FIG. 3 is a flowchart for explaining the fingerprint collation operation. Hereinafter, the fingerprint collation operation will be described with reference to FIG.

  In step 31, when the communication control unit 1 receives a fingerprint collation request from the external device, the fingerprint data input / collation unit 2 reads the fingerprint of the person to be authenticated, and based on the input fingerprint image indicating the read fingerprint, Extract fingerprint feature points (input feature points).

  Specifically, the reading unit 21 reads the fingerprint of the person to be authenticated, and outputs an input fingerprint image indicating the read fingerprint to the verification unit 22. The collation unit 22 extracts input feature points from the input fingerprint image received from the reading unit 21.

  When step 31 ends, the IC card input / output control unit 3 executes step 32.

  In step 32, the IC card input / output control unit 3 reads a registered fingerprint image and quality information from the IC card 6. The IC card input / output control unit 3 stores the read registered fingerprint image in the storage unit 4 and outputs the read quality information to the MPU 5. When step 32 ends, the MPU 5 executes step 33.

  In step 33, the MPU 5 calculates a determination level according to the quality information read by the IC card input / output control unit 3.

  For example, if the quality information is included in 0 to 29, the MPU 5 sets the determination level to 15, if the quality information is included in 30 to 70, the determination level is 50, and the quality information is included in 71 to 100. The determination level is 85. In this case, a fingerprint with a higher quality information value is subjected to a strict collation. If the determination level is set according to the quality information, the determination level calculation method is not limited to the above, and can be changed as appropriate.

  The MPU 5 registers the determination level in the fingerprint data input / collation unit 2 (specifically, the level storage unit 23).

  When step 33 ends, the fingerprint data input / collation unit 2 executes step 34.

  In step 34, the fingerprint data input / collation unit 2 reads the registered fingerprint image stored in the storage unit 4, and extracts feature points (registered feature points) from the read registered fingerprint image. The fingerprint data input / collation unit 2 calculates the degree of coincidence between the input feature point and the registered feature point, and the input fingerprint image matches the registered fingerprint image based on the calculation result and the determination level calculated by the MPU 5. Judgment is made.

  Specifically, the collation unit 22 reads a registered fingerprint image stored in the storage unit 4 and extracts registered feature points from the read registered fingerprint image. The collation unit 22 collates the input feature points extracted in step 31 with the registered feature points, and outputs coincidence information indicating their coincidence.

  The determination unit 24 determines that the input fingerprint image is the same as the registered fingerprint image (fingerprint authentication OK) when the degree of coincidence information output by the collation unit 22 exceeds the determination level stored in the level storage unit 23. When the matching degree information output by the collation unit 22 is equal to or lower than the determination level stored in the level storage unit 23, it is determined that the input fingerprint image is different from the registered fingerprint image (fingerprint authentication NG).

  According to the present embodiment, the following effects can be obtained.

  Just as fingerprints vary from person to person, the state of the finger varies from person to person. For example, there are a finger having a fingerprint having fewer feature points than the average, a wet finger of a person who sweats easily, and a dry finger of a person having dry skin.

  For this reason, in a situation where fingerprint authentication is performed at an authentication level (judgment level) that matches the average finger state, a fingerprint read from a finger, wet finger, or dry finger with fewer feature points than the average If is used for verification, there is a high possibility of an authentication error.

  In this case, it cannot be used practically apart from a scene that requires very high security.

  In this embodiment, when a fingerprint is registered in advance, the fingerprint image and the quality level of the fingerprint corresponding to the state of the finger are stored in the IC card 6, and the fingerprint stored in the IC card 6 at the time of fingerprint authentication. Verification is performed at the authentication level (judgment level) that matches the quality level. For this reason, fingerprint verification of normal quality fingerprints, fingerprints with few feature points, wet fingerprints, dry fingerprints, etc. can be performed as efficiently as possible without causing an authentication retry or an authentication error.

  In other words, in this embodiment, the determination level depends on the quality of the registered fingerprint image. For this reason, even if fingerprint images of various qualities are used for collation, it is possible to determine whether or not the input fingerprint image is the same as the registered fingerprint image at a determination level corresponding to the quality.

  Therefore, there are fewer verification retries and verification errors due to the difference between the actual fingerprint quality and the quality of the fingerprint used to set the judgment level, and it is easy to perform fingerprint verification without compromising the user. It becomes possible to do.

  In the present embodiment, quality information indicating the quality of the registered fingerprint image is received, a determination level is set according to the quality information, and the degree of coincidence information calculated by the collation unit 22 exceeds the determination level. The input fingerprint image is determined to be the same as the registered fingerprint image. In this case, it is not necessary for the fingerprint collation device to always store and manage quality information.

  In the present embodiment, the determination level decreases as the quality of the user's fingerprint indicated by the quality information deteriorates. Therefore, it is possible to accurately determine whether the input fingerprint image is the same as the registered fingerprint image according to the quality.

  In the present embodiment, the quality information and the registered fingerprint image are read from the IC card 6 (portable recording medium) on which the quality information and the registered fingerprint image are recorded, and the read registered fingerprint image is stored in the storage unit. 4 is stored. For this reason, the quality information and the registered fingerprint image can be read from the common IC card 6. The user can manage the IC card 6 to manage the quality information together with the registered fingerprint image.

  In the embodiment described above, the illustrated configuration is merely an example, and the present invention is not limited to the configuration.

  For example, the method of calculating quality information is not limited to the above and can be changed as appropriate. For example, the collation unit 22 may extract fingerprint feature points (for example, branch points of fingerprint ridges) from the input fingerprint image received from the reading unit 21, and the number of the extracted feature points may be used as quality information. In this case, the quality information value (number of feature points) increases as the quality of the fingerprint is improved.

It is the block diagram which showed the fingerprint collation apparatus of one Example of this invention. It is the block diagram which showed an example of the fingerprint data input and collation part. It is a flowchart for demonstrating operation | movement of a fingerprint collation apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Communication control part 2 Fingerprint data input and collation part 21 Reading part 22 Collation part 23 Level storage part 24 Judgment part 25 Collation judgment part 3 IC card input / output control part 4 Memory | storage part 5 MPU
6 IC Card 61 Registered Fingerprint Information Storage Unit 62 Quality Information Storage Unit

Claims (10)

A reading unit that reads a user's fingerprint and outputs an input fingerprint image indicating the read fingerprint;
An image storage unit for storing a registered fingerprint image registered in advance;
The input fingerprint image output from the reading unit is collated with a registered fingerprint image stored in the image storage unit to generate coincidence information indicating the degree of coincidence thereof, and the coincidence information and the quality of the registered fingerprint image And a verification determination unit that determines whether or not the input fingerprint image is the same as the registered fingerprint image based on a determination level according to the fingerprint verification device. The fingerprint verification apparatus according to claim 1, wherein
An input unit for receiving quality information indicating the quality of the registered fingerprint image;
A setting unit that sets the determination level according to the quality information received by the input unit,
The collation determination unit
A collation unit that collates the input fingerprint image output by the reading unit with a registered fingerprint image stored in the image storage unit and outputs matching degree information indicating the degree of matching;
And a determination unit that determines that the input fingerprint image is the same as the registered fingerprint image when the matching degree information output by the verification unit exceeds a determination level set by the setting unit. . The fingerprint collation device according to claim 2,
The setting unit lowers the determination level as the quality of a registered fingerprint image indicated by the quality information deteriorates. In the fingerprint collation device according to claim 2 or 3,
The input unit reads the quality information and the registered fingerprint image from a portable recording medium in which the quality information and the registered fingerprint image are recorded;
The image storage unit is a fingerprint collation device that stores a registered fingerprint image read by the input unit. The fingerprint collation device according to claim 4,
The portable recording medium is a fingerprint collation device, which is an IC card. A fingerprint collation method performed by a fingerprint collation apparatus including an image storage unit for storing a registered fingerprint image registered in advance,
A reading step of reading a user's fingerprint and outputting an input fingerprint image indicating the read fingerprint;
A matching degree output step of collating the input fingerprint image with a registered fingerprint image stored in the image storage unit and outputting matching degree information indicating the matching degree;
A collation determining step for determining whether or not the input fingerprint image is the same as the registered fingerprint image based on the matching degree information and a determination level corresponding to the quality of the registered fingerprint image. . The fingerprint collation method according to claim 6,
Accepting quality information indicating the quality of the registered fingerprint image;
A setting step for setting a determination level according to the quality information,
The collation determination step includes determining that the input fingerprint image is the same as the registered fingerprint image when the degree-of-match information exceeds the determination level. In the fingerprint collation method according to claim 7,
The fingerprint collating method, wherein the setting step includes lowering the determination level as the quality of a registered fingerprint image indicated by the quality information deteriorates. The fingerprint collation method according to claim 7 or 8,
The accepting step includes reading the quality information and the registered fingerprint image from a portable recording medium recording the quality information and the registered fingerprint image;
A fingerprint collation method further comprising a storage step of storing the registered fingerprint image read in the input step in the image storage unit. The fingerprint collation method according to claim 9, wherein
The fingerprint collation method, wherein the portable recording medium is an IC card.

Images