JPH0628459A - Fingerprint matching device - Google Patents

Abstract

PURPOSE:To provide a highly reliable fingerprint matching device capable of coping with the skill of a user and the change of a fingerprint in the lapse of time, etc., in the fingerprint matching device for executing pattern matching of a registered fingerprint dictionary and an input fingerprint and confirming the person himself. CONSTITUTION:The device is constituted of a main feature information storage part 11 for registering a prescribed number of feature patterns extracted by a feature extracting part 10, a sub-feature information storage part 12 for registering the prescribed number of feature patterns extracted by the feature extracting part 10 in the same way as auxiliary feature patterns, a matching part 13 for comparing an output of a binarizing circuit 2 and the feature patterns of the main feature information storage part 11 and the sub-feature information storage part 12, and a sequence updating part 14 for updating a use sequence at the time of matching by executing a replacing operation of a noncoincident pattern in the main feature information storage part 11 and a coincident pattern in the sub-feature information storage part 12.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fingerprint collation device for performing pattern matching between a registered fingerprint dictionary and an input fingerprint to confirm the person himself / herself.

In recent years, as computers have been introduced into a wide range of social systems, system security has attracted a lot of interested persons. Many questions have been raised in terms of ensuring security for the ID cards and passwords that have been used so far as means for confirming the identity when entering a computer room or using a terminal. On the other hand, fingerprints are considered to be the most effective means of personal identification because fingerprints have two major characteristics, that is, they are the same for all people and do not change for the rest of the life, and much research and development has been done on simple personal identification systems using fingerprints. There is.

[0003]

2. Description of the Related Art In a conventional fingerprint collation device, fingerprint features are registered in advance as a dictionary, and at the time of collation, the input image and the feature dictionary information are compared to verify the identity. FIG. 8 is a conceptual diagram of the configuration of a conventional fingerprint matching device. First, the operation at the time of registration will be described. A fingerprint pattern is detected by pressing a finger against the fingerprint sensor 1 and converted into digital data by an A / D converter (not shown) in the fingerprint sensor 1. Converted digital data (fingerprint data)
Is converted into binary data of “0” and “1” by the subsequent binarization circuit 2 and stored in the frame memory 3. The fingerprint data stored in the frame memory 3 is sequentially read out, and then enters the characteristic information extraction / collation circuit 4 to extract the characteristic information.

Here, the characteristic information is, for example, as shown in FIG.
It refers to a branch point as shown in FIG. The fingerprint can be specified by the number of such branch points and end points at which positions. The extracted feature information is
It is stored in the personal characteristic information data file 5. The above operation is repeated for a plurality of individuals, and the individual characteristic information is stored in the individual characteristic information data file 5.

When the registration of the amount of characteristic information in the personal characteristic information data file 5 is completed in this way, this time the fingerprint collation operation of the individual is started. In the case of collation, a finger (a finger used for registration in advance, for example, an index finger) is placed on the fingerprint sensor 1 and then an ID number of the user is input with a ten-key pad (not shown). As a result, the feature information extraction / collation circuit 4 can narrow the search range by determining the search range of the personal feature information data file 5 based on the ID number and reading it at the time of matching.

A fingerprint pattern is detected in the same manner as at the time of registration, and is converted into digital data by an A / D converter (not shown) in the fingerprint sensor 1. The converted digital data is converted into binary data of “0” and “1” by the subsequent binarization circuit 2 and stored in the frame memory 3.

The characteristic information extraction / collation circuit 4 reads out the fingerprint data for collation stored in the frame memory 3, reads out the characteristic information for each individual stored in the personal characteristic information data file 5, and collates both ( Pattern matching)
I do. When the number of matching feature information is more than a predetermined number, it is determined that the fingerprints match.

[0008]

However, in general, the fingerprint collation device is often not familiar with the fingerprint registration process, and it is common that the finger placement position and pressure are not constant.
In addition, it is often difficult to input a fingerprint image with high homogeneity at the time of matching at the time of verification due to changes in the work content of the user after registration, seasonal changes in the skin condition, and the like.

On the other hand, a method has been proposed in which a fingerprint image is input a plurality of times at the time of registration to create a dictionary for each image, and feature points commonly selected on each screen are preferentially adopted as the dictionary. There is. According to this method, it is possible to create a dictionary with a margin for the difference in the finger pressing position and the pressing force, but there is a problem that it is difficult to deal with changes over time in fingerprints. It was

The present invention has been made in view of the above problems, and an object of the present invention is to provide a highly reliable fingerprint collation device capable of coping with the user's proficiency level and changes in fingerprints over time. .

[0011]

FIG. 1 is a block diagram showing the principle of the present invention. The same parts as those in FIG. 8 are designated by the same reference numerals. In the figure, 1 is a fingerprint sensor that detects a fingerprint image, and 2 is an output that converts the output of the fingerprint sensor 1 into a binary image.
A digitizing circuit 10, a feature extracting unit that receives the output of the binarizing circuit 2 and extracts feature point information, and 11 a main feature information storage that registers a fixed number of feature patterns extracted by the feature extracting unit 10. 12 is a sub-feature information storage unit for registering a fixed number of feature patterns extracted by the feature extraction unit 10 as preliminary feature patterns, and 13 is an output of the binarization circuit 2 and the main feature information storage unit 11 and sub-features. A collating unit that compares the characteristic pattern of the information storage unit 12 with a matching pattern in the sub-feature information storage unit 12 based on the matching result of the matching unit 13. Is a rank update unit that replaces and updates the usage rank at the time of matching.

[0012]

In the dictionary registration, a preliminary feature pattern is registered in addition to the fixed number of feature patterns (main feature patterns) required for personal identification, and the collation unit 13 inputs the main feature pattern and the preliminary feature at the time of fingerprint input. Match any of the patterns with the input fingerprint image. Then, after the matching is successful, the matching feature pattern is stored in the rank updating unit 14 for each of the main feature pattern and the sub feature pattern as a result of the matching. Then, the rank updating unit 14 performs rank updating processing for replacing the non-matching pattern of the main feature patterns with the matching feature pattern of the sub-feature pattern based on the matching result.

In this way, the adequacy of the characteristic pattern of the dictionary is determined each time the collation succeeds, and the order of use of the characteristic points at the next collation is updated based on the result, whereby the user's proficiency level is improved. It is possible to provide a highly reliable fingerprint collation device that can cope with changes in fingerprints over time and the like.

[0014]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 2 is a flow chart showing the operation of the principle block diagram shown in FIG. The matching unit 13 performs matching (pattern matching) between the input fingerprint image and the characteristic pattern stored in the main characteristic information storage unit 11 (S1). The pattern matching is performed by sequentially scanning the input fingerprint image through a window having a predetermined size, and matching / mismatching the pattern read from the window with the dictionary pattern (feature pattern) stored in the main feature information storage unit 11. It judges by judging. The collation result is stored in the rank updating unit 14.

Next, the number of matching patterns (number of passing characteristic patterns) as a result of fingerprint collation is a predetermined reference value k1.
Identity verification is performed by determining whether or not it is larger than (S2). If the person is not confirmed, fingerprint verification is performed again. If the number of passing characteristic patterns ≧ k1, it is determined that the person is the person.

If the user is the person himself, the remaining feature patterns stored in the sub-feature information storage unit 12 are compared with the input fingerprint image (S3). The collation result is the rank update unit 1
Stored in 4. By the above operation, the order update unit 14 stores the match / mismatch information of the feature patterns stored in the main feature information storage unit 11 and the match / mismatch information of the feature patterns stored in the sub feature information storage unit 12. It will be stored.

The order updating unit 14 updates the use order using the pass / fail information based on these collation results (S4). Figure 3
FIG. 8 is an explanatory diagram of a usage order update operation. It is assumed that the number of characteristic patterns stored in the main characteristic information storage unit 11 is 12 and 12 and the number of characteristic patterns stored in the sub characteristic information storage unit 12 is 5 from 13 to 17. . As a result of the matching, it is assumed that the third and tenth main feature patterns do not match as shown in the figure. At this time, the order updating unit 14 moves the third characteristic pattern to the 17th position and replaces the 13th characteristic pattern (matching pattern), which is the sub-characteristic pattern, with the main characteristic information pattern side. As a result, the number of dictionaries that are likely to match the input fingerprint image increases on the main feature pattern side.

In this way, the main feature pattern for personal identification is constantly re-evaluated on the basis of the fingerprint matching result for a new feature pattern, so that the user's proficiency level is low or changes with time. Can also respond.

In the above description, as shown in FIG. 3, the case where the non-coincident feature pattern is sequentially turned to the end and the coincident feature pattern is incorporated into the main feature pattern has been described. However, the collation error of each feature pattern at the time of collation is used. It is also effective to update the order of use. That is, after the pass / fail judgment of all the feature patterns, sorting is performed in the order of the smallest matching error,
It is the order of use at the time of the next matching. That is, the main feature information storage unit 11 stores the feature patterns in order from the one with the smallest matching error, and the sub feature information storage unit 12 stores the feature patterns with the largest matching error.

FIG. 4 is a block diagram showing the configuration of an embodiment of the present invention. The same parts as those in FIG. 1 are designated by the same reference numerals. In the figure, reference numeral 15 is a use number counting means for counting the number of uses since registration. This utilization frequency counting means 1
5 is updated by 1 each time the matching operation is performed on the same person. The update information is given from the collation unit 13 to the usage number counting means 15.

Then, when the number of times of use reaches a predetermined constant value, the number-of-uses counting means 15 to the rank updating unit 14
Send a signal to. When the rank updating unit 14 receives this signal,
Unnecessary feature patterns stored in the sub-feature information storage unit 12 are deleted. As a result, only the highly reliable feature pattern necessary for identifying the person remains in the main feature information storage unit 11, and steps S3 and S4 of FIG. You can speed it up.

A frame memory (FIG. 8) for storing binarization information is provided between the binarization circuit 2 and the feature extraction unit 10.
3) is actually provided, and the characteristic information extraction unit 10 reads the binarized data from this frame memory to perform characteristic extraction, but it is omitted here (the same applies hereinafter).

FIG. 4 is a block diagram showing the configuration of an embodiment of the present invention. The same parts as those in FIG. 1 are designated by the same reference numerals. In the figure, 16 is a timer for counting the elapsed time from the registration. The timer 16 starts counting clocks when the fingerprint dictionary is registered. Then, when the count value of the timer reaches a predetermined constant value, the timer 16 sends a signal to the rank updating unit 14. Upon receiving this signal, the rank updating unit 14 deletes the unnecessary feature pattern stored in the sub feature information storage unit 12. As a result, only a highly reliable feature pattern necessary for identifying the person remains in the main feature information storage unit 11 after a certain period of time, and steps S3 and S4 of FIG. 2 are omitted in the subsequent matching operation. You can
The matching speed can be increased.

FIG. 6 is a block diagram showing the configuration of another embodiment of the present invention. The same parts as those in FIG. 1 are designated by the same reference numerals. In the figure, reference numeral 20 denotes a history information storage unit that stores history information of the matching result based on the matching result of the matching unit 13.

FIG. 7 is a flow chart showing the operation of the apparatus shown in FIG. As in the case of FIG. 2, the matching unit 13
Matching (pattern matching) between the input fingerprint image and the characteristic pattern stored in the main characteristic information storage unit 11 is performed (S1). The matching result is stored in the history information storage unit 20.

Next, the number of matching patterns (number of passing characteristic patterns) as a result of fingerprint collation is a predetermined reference value k1.
Identity verification is performed by determining whether or not it is larger than (S2). If the person is not confirmed, fingerprint verification is performed again. If the number of passing characteristic patterns ≧ k1, it is determined that the person is the person.

If the person is the person himself / herself, the remaining feature patterns stored in the sub-feature information storage unit 12 are compared with the input fingerprint image (S3). The matching result is stored in the history information storage unit 20.

The order updating unit 14 reads out history information based on the collation result stored in the history information storage unit 20,
In addition to the pass rate, history information is added to update the usage order (S4). As a result, more practical and highly reliable collation becomes possible as compared with the device shown in FIG.

Here, as the information to be stored in the history information storage section 20, the utilization order, the pass rate, the average matching error,
The number of trials (the number of times fingerprint collation is performed before the collation result is determined to be the person) may be considered. For example, the past pass rate and history information of the number of trials of each feature pattern, using the result of this pass, a new pass rate of each feature pattern is calculated, a method of determining the use order at the time of the next matching, A method is conceivable in which a new average matching error of each feature pattern is obtained from the past average matching error of each feature pattern, history information of the number of trials, and this matching error, and the new average matching error is used in ascending order. The rank updating unit 14 performs such processing.

According to the embodiment shown in FIG. 6, the rank updating unit 14
Replaces the feature pattern of the main feature information storage unit 11 with the feature pattern of the sub-feature information storage unit 12, so that the feature pattern stored in the main feature information storage unit 11 in which the feature pattern for matching is stored is The learning function further increases reliability and enables highly reliable fingerprint matching.

[0032]

As described above in detail, according to the present invention, a preliminary feature pattern is prepared in addition to the main feature pattern as the dictionary pattern, and the main feature pattern and the sub-feature feature are obtained based on the collation result. By updating the order of patterns, it is possible to use a pattern that is always easy to match with the input fingerprint image as a dictionary. Therefore, according to the present invention, it is possible to provide a highly reliable fingerprint collation device which can cope with the user's proficiency level and the change of fingerprints over time.

[Brief description of drawings]

FIG. 1 is a principle block diagram of the present invention.

FIG. 2 is a flowchart showing the operation of the principle block diagram.

FIG. 3 is an explanatory diagram of a usage order update operation.

FIG. 4 is a configuration block diagram showing an embodiment of the present invention.

FIG. 5 is a configuration block diagram showing another embodiment of the present invention.

FIG. 6 is a configuration block diagram showing another embodiment of the present invention.

7 is a flowchart showing the operation of the apparatus shown in FIG.

FIG. 8 is a conceptual diagram of a configuration of a conventional fingerprint matching device.

FIG. 9 is a diagram illustrating an example of fingerprint characteristic information.

[Explanation of symbols]

 1 Fingerprint Sensor 2 Binarization Circuit 10 Feature Extraction Unit 11 Main Feature Information Storage Unit 12 Sub-Feature Information Storage Unit 13 Collating Unit 14 Order Update Unit

Claims (7)

[Claims] 1. A fingerprint sensor (1) for detecting a fingerprint image.
A binarization circuit (2) for converting the output of the fingerprint sensor (1) into a binary image, and a feature extraction unit (10) for receiving the output of the binarization circuit (2) and extracting feature point information. ), A main feature information storage unit (11) for registering a fixed number of feature patterns extracted by the feature extraction unit (10), and a fixed number of feature patterns similarly extracted by the feature extraction unit (10). A sub-feature information storage unit (12) registered as a feature pattern, and collation for comparing the output of the binarization circuit (2) with the feature patterns of the main feature information storage unit (11) and the sub-feature information storage unit (12). And the matching pattern in the sub-feature information storage unit (12) and the non-matching pattern in the main feature information storage unit (11) are switched based on the matching result of the matching unit (13). With the order update unit (14) that updates the use order at the time of collation It made a fingerprint verification device. 2. A use number counting means (15) for counting the number of times of use from the time of registration is separately provided, and when the number of times of use reaches a predetermined constant value, an unnecessary characteristic pattern is updated by an order updating unit (14). The fingerprint collation device according to claim 1, wherein the fingerprint collation device is deleted. 3. A timer (16) for separately counting a period elapsed from the time of registration is separately provided, and when the elapsed period reaches a predetermined constant value, an unnecessary feature pattern is updated by the rank updating unit (1).
4. The fingerprint collation device according to claim 1, wherein 4) is deleted. 4. The fingerprint collation apparatus according to claim 1, wherein the rank updating unit (14) updates the use rank based on the pass / fail of each characteristic pattern at each matching. 5. The fingerprint matching apparatus according to claim 1, wherein the rank updating unit (14) updates the usage rank based on a matching error for each characteristic pattern at each matching. 6. A history information storage unit (20) for storing the matching result of the matching unit (13) as history information, storing the matching result of the matching unit (13), and updating the rank updating unit (14). ) Updates the order of use based on the pass rate newly calculated using the past pass rate and the history information of the number of trials read from the history information storage unit (20) and the pass / fail result of this time. The fingerprint collation device according to claim 1, wherein 7. The rank updating unit (14) uses the past average matching error of each characteristic pattern read from the history information storage unit (20), history information of the number of trials, and this matching error. The utilization order is updated based on the newly calculated average matching error.
The fingerprint matching device described.