JP4030018B2 - Fingerprint personal authentication system - Google Patents

Description

  The present invention relates to a personal authentication system using fingerprints, and more particularly to an apparatus that enables accurate authentication in a short time.

  Personal authentication methods for confirming individuals include, for example, fingerprint management methods for building entrance / exit management, etc., personal authentication methods using passwords used for authorization in workflow systems using networks, the Internet and public In remote use via a line, there is a personal authentication method using a password when logging into the system (see, for example, Patent Document 1).

  Personal authentication using fingerprints is expected to be used for, for example, entrance and exit management in buildings using information processing devices, authentication of the right to use specific software, personal authentication such as purchasing goods in the Internet environment, etc. Used in limited departments such as building access control.

  In conventional fingerprint authentication, for example, fingerprint data is registered in a fingerprint authentication device as shown in FIG. The fingerprint data was collected and collated. In FIG. 5, reference numeral 102 denotes a keyboard for inputting the identification number ID of the person to be authenticated, and reference numeral 103 denotes a display unit for displaying guidance, displaying authentication results, and the like.

  Further, in the conventional workflow system, when performing an approval action, it is doubtful that only the password is really authorized by the person himself or herself because of personal authentication using a password.

  At the time of system login when using remotely via the conventional Internet or public line, personal authentication is performed using a password, but there are doubts about security.

Therefore, it is desired to provide a fingerprint authentication method that is more reliable than a password and easy to use.
JP-A-61-234478 (page 3, right column, lines 3 to 5)

  In the conventional fingerprint authentication system, it is necessary to update and register the fingerprint data every predetermined period because it does not consider the correspondence of aging, injury, and changes in the state at the time of fingerprint data collection. In addition, since there are cases where the acupressure at the time of registration and verification is not the same state, it is difficult to perform accurate verification, and, for example, there is a problem that it is impossible to enter the building.

  Accordingly, an object of the present invention is to provide a personal authentication system using fingerprints which has improved such problems.

The object of the present invention is achieved by a personal authentication system shown by the following configurations (1) and (2).
(1) Collectively managing registered fingerprint data, and comparing a fingerprint management server that compares a fingerprint for which verification has been requested with the registered fingerprint data, and reading a fingerprint to be authenticated and transferring the fingerprint to the fingerprint management server client and the humidity state of the person to be authenticated fingerprint, injuries, temperature, in order to reduce false matching accompanying changes by age or the like, the final verification initially and the original fingerprint data registered, Tsu Do the last collation success Holding means for holding data, and the holding means stores the data to be verified when the difference between the number of feature points of the final collation data and the number of feature points of the data to be collated is a predetermined value or less. It held as a final verification data, and the current verification fingerprint data transferred from said fingerprint reading client, the original fingerprint data, and finally becomes a verification success, feature points of the final verification data Personal authentication system fingerprint, wherein a number equipped with a learning matching unit for performing an overall rating of the fingerprint data to the fingerprint management server is less than the number of feature points of the matching data.
(2) In the personal authentication system of (1) above, the registered fingerprint data is collectively managed, the fingerprint management server that compares the fingerprint requested for verification with the registered fingerprint data, and the fingerprint of the person to be authenticated is read. A fingerprint reading client for transferring the fingerprint to the fingerprint management server, and a finger pressure measuring means for measuring the finger pressure at the time of fingerprint reading in the fingerprint reading means, and the finger pressure when the registered fingerprint is read is entered in the registered fingerprint data. The finger pressure when the person to be authenticated puts the finger on the fingerprint reading means is measured and compared with the finger pressure entered in the registered fingerprint data, and the difference state of the pressure is displayed on the fingerprint reading client. A personal authentication system using fingerprints, characterized by

  As a result, the following effects can be obtained.

  (1) Since the conventional fingerprint authentication system does not take into account the aging, injuries, and changes in the state at the time of fingerprint data collection, it is necessary to update the registration of fingerprint data many times. On the other hand, in the present invention, based on data at the time of initial registration, aging, injury, humidity change, temperature change and good fingerprint data of sampling state are taken into consideration and reflected in fingerprint verification, so the security level It is possible to save the trouble of updating fingerprint data without lowering the image quality.

  (2) By measuring finger pressure and comparing it with the one at the time of registration, it is possible to make the finger pressure at the time of registration and the finger pressure at the time of collation the same, and accurate fingerprint authentication under the same pressure is possible It becomes.

  The best mode for carrying out the present invention will be described with reference to FIGS. FIG. 1 is an embodiment of the present invention, FIG. 2 is a diagram for explaining the operation of the present invention, FIG. 3 is a diagram for explaining fingerprint reading means, and FIG. 4 is a diagram for explaining a registered fingerprint table.

  In FIG. 1, 1 is a fingerprint management server, 1-0 is a file server, 2 is a registered fingerprint data holding unit, 2-0 is a fingerprint management table, 3-0 to 3-N are fingerprint reading clients, 4-0 to 4- N is a keyboard, 5-0 to 5-N are fingerprint reading means, 6-0 to 6-N are registration threshold setting means, 10 is a collation means, 11 is a specific finger reading means, and 12 is a learning means.

  The fingerprint management server 1 performs processing related to fingerprint verification in an integrated manner, and the fingerprint transferred from the fingerprint reading clients 3-0 to 3-N matches the fingerprint registered in the registered fingerprint data holding unit 2. First, a fingerprint of a specific finger, for example, a forefinger is selected from ten registered fingerprints based on human habits and behavior in the environment. It performs processing and the like, and has collation means 10 and specific finger reading means 11. This outside fingerprint management server 1 is provided with a learning means 12 and, as will be described later, in order to reduce erroneous matching due to changes due to humidity, injury, temperature, age, etc., matching with the original registered fingerprint data was finally successful. The fingerprint data and the current collation fingerprint data are comprehensively evaluated, and when the collation of the current collation fingerprint data is successful, the current collation fingerprint data is registered as the fingerprint data that was the last collation success.

  The file server 1-0 accesses the registered fingerprint data holding unit 2 based on an instruction from the fingerprint management server 1. The file server 1-0 can be integrated with the fingerprint management server 1.

  The registered fingerprint data holding unit 2 stores fingerprints necessary for personal authentication. As shown in FIG. 4A, the registered fingerprint data holding unit 2 is data 1 that is an identification number ID for each individual and ten-fingerprint fingerprint data. -A registered fingerprint table 2-0 in which data 10, a collation level for each application, and the like are entered is held. The registration order of fingerprint data is determined in advance. For example, data 1 to data 10 are defined in the order of little finger of left hand, ring finger... Thumb, right thumb, index finger. When a user uses an application, it is necessary that the verification level of the fingerprint verification is high in order to increase security.

  For this reason, as shown in FIG. 4B, when the number of fingerprint feature points to be collated, which will be described later, is 21 or more, it is defined as state A, and it is defined that collation level 9 or 8 is possible. Similarly, when the number of feature points is 15 or more, state B is defined as collation level 7 or 6 being possible, and 10 or more is defined as collation level 5 or 4 being able to be state C.

  The collation level indicates the percentage of coincidence. The collation level 9 indicates that, for example, 90% or more of the number of feature points matches, and the collation level 8 indicates that 85% or more matches. Then, when the state of the fingerprint data required at this time is the state A having 21 or more feature points, it indicates that the collation level 9 or 8 can be collated.

  Then, as shown in FIG. 4 (A), when performing the necessary fingerprint collation for the application APL1 important for security measures, the collated fingerprint data needs to be in the state of the collation level 8. For the default ones that are not specified, it is obtained from this registered fingerprint table that the collation level should be 4 or more.

  It should be noted that the registered fingerprint data is not a single piece, but, as representatively shown for data 2 in FIG. 4C, the original data OR, which is the fingerprint data when it was first registered, and finally the collation succeeded. Data LD is entered. In FIG. 4C, CD is fingerprint data for which collation is requested.

  The fingerprint reading client 3-0 asks the user to verify his / her identity by checking his / her fingerprint against registered fingerprint data, and inputs his / her name, ID, etc., and displays guidance. A keyboard 4-0; fingerprint reading means 5-0 for reading a fingerprint and outputting fingerprint data; and a registration threshold setting means 6-0 for checking whether or not the quality of the read fingerprint data has reached a prescribed level. Yes.

  Other fingerprint reading clients 3-N are similarly configured and include a keyboard 4-N, fingerprint reading means 5-N, and registration threshold setting means 6-N.

  The fingerprint reading means 5-0 reads a fingerprint and creates fingerprint data, and is a commercially available product and is configured as shown in FIG. In FIG. 3A, 50 is a light emitting diode which illuminates the belly of the finger, 51 is a light guide plate such as a glass plate on which a finger is placed, and 52 is for reflection , 53 is a lens, 54 is a reflecting mirror, 55 is a CCD, 56 is a diaphragm, 57 is a finger, and when placed on the light guide plate 51, the fingerprint of the belly portion is irradiated by the LED 50 Then, a fingerprint image is formed on the CCD 55 by the lens 53. In FIG. 3A, 58 is a nail, 59 is a finger tip, and 60 is a fingerprint groove.

  In this manner, the spiral fingerprint shown in FIG. 3B is obtained from the CCD 53. From this, the center position indicated by a rectangle, the feature point indicated by a circle, the ridge direction indicated by an arrow, and the like are known, Extracted as feature fingerprint data. The quality of fingerprint data is determined based on the number of feature points. For example, as shown in FIG. 4B, if the number of feature points in one fingerprint is 21 or more, state A, C if it is 10 or more, and D if it is 5 or more.

  Since the read fingerprint state changes considerably depending on the finger pressure at the time of reading, it is desirable to read the fingerprint at the time of collation in the same state as the pressure at the time of registration.

  Therefore, a finger pressure measuring means is added to the fingerprint reading means 5-0, and the finger pressure at the time of registration is entered in the registered fingerprint data. By comparing the finger pressure at the time of matching with the finger pressure at the time of matching and displaying the difference state on the fingerprint reading client, the user adjusts the finger pressure accordingly and reads the fingerprint in a matching state. A highly accurate collation result can be obtained.

  The registration threshold setting means 6-0 is for setting the state of a fingerprint that can be registered when registering fingerprint data. If the quality of the registered fingerprint data is poor, the credibility with respect to the collation result is lost. Therefore, it is necessary to register a good condition without registering a bad condition.

  By the way, the fingerprint management server 1 is provided with collation means 10, specific finger reading means 11, learning means 12, and the like.

  The collation means 10 collates, for example, whether the fingerprint data requested to be collated from the fingerprint reading client 3-0 matches the registered fingerprint data held in the registered fingerprint data holding unit 2. This match is determined to match when the number of feature points matches, for example. The collation means 10 not only performs fingerprint data collation but also checks the fingerprint data collation level required according to the application. The collation level is defined as, for example, a collation level 9 when the feature points of the registered fingerprint data match 90% or more, and a collation level 8 when the feature points match 85% or more.

  For example, the application APL1 is an important application, and in order to obtain login or use permission for the application, fingerprint verification with high credibility is required. Therefore, in the registered fingerprint table 2-0 held in the registered fingerprint data holding unit 2, as shown in FIG. 4A, the collation level of the fingerprint to be matched corresponding to the application is entered.

  In this state, when the user wants to use the application APL1 from the fingerprint reading client 3-0, for example, the user inputs the name, ID, and APL1 from the keyboard 4-0, and puts a finger, for example, the right index finger on the fingerprint reading means 5-0. Place the fingerprint data.

  As a result, the fingerprint management server 1 reads the collation level 8 related to APL1 and the registered fingerprint data of the right-hand index finger from the registered fingerprint table 2-0 by ID. Then, the fingerprint data of the right-hand index finger obtained from the fingerprint reading means 5-0 is compared with the registered fingerprint data, and it is verified that the matching level 8 matches because the number of feature points matches 85% or more. The determination is made by the means 10 and a coincidence signal is output to the fingerprint reading client 3-0. As a result, the fingerprint reading client 3-0 outputs a log-in, and the user who has performed the previous fingerprint collation can access the application APL1.

  By providing a collation level corresponding to the fingerprint quality in this way, it is possible to set security according to the importance of the application to be used. Of course, the percentage of coincidence is not limited to this, and may be set not by% but by the number of feature point differences.

  The specific finger reading means 11 determines the finger to be collated first in the fingerprint collation, for example, as a right hand index finger. For collation, first, the registered fingerprint data of the right hand index finger is registered from the fingerprint management table 2-0 of the registered fingerprint data holding unit 2. Is read out.

  People have wrinkles, and when placing only one finger on the fingerprint reading means 5-0, which finger is placed first varies depending on the person, but generally they are the same. Accordingly, for such an example, statistics are taken by the actions of a large number of people, such as 100 people or 200 people, and if the fingerprint of the most frequent finger, for example, the registered fingerprint data of the right hand index finger, is read out first and collated, the first There is a high probability of matching in matching. Thus, the retrieval time can be shortened by determining the reading of registered fingerprint data based on human habits. When the first does not match, the second registered fingerprint data may be read in the same manner, or may be read in a predetermined order regardless of this.

  The learning means 12 reduces false collation due to fingerprint changes due to humidity, injury, temperature, age, etc., and the operation will be described with reference to FIG. In this case, in the registered fingerprint table, as shown in FIG. 4C, the original data OR which is the registered fingerprint data registered first and the fingerprint data LD which finally succeeded in collation are entered. In this example, a case will be described in which it is determined that the difference in the number of feature points is three or less.

  When the number of feature points in the original data OR is 21, for example, when a user who has a finger injured due to injury collates, if the number of feature points in the fingerprint data CD of the injured finger is 15, Since the comparison is performed, there are 6 differences in the number of feature points, which are inconsistent. Note that the crosses in the fingerprint data LD and CD in FIG. 4C indicate the scratch position due to injury.

  However, in the case of FIG. 4C, the fingerprint data LD that has been successfully collated last is also registered, so the learning means 12 collates the fingerprint data LD with the fingerprint data CD that has been requested to collate. In this case, the number of feature points of the fingerprint data LD that has been successfully collated last is 18, and the difference from the number of feature points 15 of the fingerprint data CD that has been requested to collate is 3. Therefore, the collation is now matched.

  Then, instead of the fingerprint data LD, the fingerprint data CD of this time is registered as the data LD which has been newly verified last.

  By configuring in this way, not only can the collation success rate be increased, but the last successful collation data can be rewritten at all times, and this is a countermeasure for cases in which good fingerprint data is not always obtained. In addition, it is not necessary to update the registered fingerprint data at regular intervals as in the prior art.

  Next, based on FIG. 2, the fingerprint collation operation in FIG. 1 will be briefly described.

  S1. For example, when a user who requests personal authentication uses a name, ID, and an application from the fingerprint reading client 3-0 in FIG. 1, the application name such as APL1 is input from the keyboard 4-0.

  S2. Then, the fingerprint reading means 5-0 is used to input the fingerprint of one finger.

  S3. Each data such as the user ID and fingerprint data is sent to the fingerprint management server 1.

  S4. The fingerprint management server 1 reads out the registered fingerprint data of the user from the registered fingerprint data holding unit 2 via the file server 1-0 based on the ID, and collates the two.

  S5. If they do not match as a result of collation, an ID and a fingerprint are input, and S1 and subsequent steps are repeated again. If they match, the fingerprint management server 1 outputs a match signal to the requesting fingerprint reading client 3-0.

  S6. The fingerprint reading client 3-0 outputs a system login signal in response to the coincidence signal, and enables the user to use the system, for example, the application APL1.

It is the best mode for carrying out the present invention. It is operation | movement explanatory drawing of this invention. It is fingerprint reading means explanatory drawing. It is a registered fingerprint table explanatory drawing. It is explanatory drawing of a prior art example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Fingerprint management server 2 Registered fingerprint data holding | maintenance part 3-0-3-N Fingerprint reading client 4-0-4-N Keyboard 5-0-5-N Fingerprint reading means 6-0-6-N Registration threshold value setting means

Claims (2)

A fingerprint management server that collectively manages registered fingerprint data and compares a fingerprint that has been requested for verification with the registered fingerprint data;
A fingerprint reading client that reads the fingerprint of the person to be authenticated and transfers the fingerprint to the fingerprint management server;
Humidity condition of the person to be authenticated fingerprint, injuries, temperature, in order to reduce false matching accompanying changes by age or the like, the original fingerprint data registered first, and a final verification data finally Tsu Do collation success And holding means for holding, when the difference between the number of feature points of the final collation data and the number of feature points of the data to be collated is equal to or less than a predetermined value, the data to be collated is final collation data. Hold as
The fingerprint management server includes learning collation means for comprehensively evaluating the current collation fingerprint data transferred from the fingerprint reading client, the original fingerprint data, and the fingerprint data that has been successfully collated last. Personal authentication system using fingerprints. A fingerprint management server that collectively manages registered fingerprint data and compares a fingerprint that has been requested for verification with the registered fingerprint data;
A fingerprint reading client that reads the fingerprint of the person to be authenticated and transfers the fingerprint to the fingerprint management server;
The fingerprint reading means comprises a finger pressure measuring means for measuring the finger pressure at the time of fingerprint reading,
The finger pressure when the registered fingerprint is read is entered in the registered fingerprint data, and the finger pressure when the person to be authenticated puts the finger on the fingerprint reading means is measured and entered in the registered fingerprint data. 2. The personal authentication system using fingerprints according to claim 1, wherein, compared with finger pressure, a difference state of the pressure is displayed on the fingerprint reading client.

Images