JPH10154231A - Person authentication device using biometrics information and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To maintain high security even if a collation system of low precision is used or a fingerprint pattern is copied by considering the order of fingers to be inputted as a password, combining finger inputs for plural times and collating them. SOLUTION: A fingerprint input device 100 is constituted of a television camera, an A/D converter and a display device. The fingerprint feature quantity of a user and the input order of the fingers are previously registered in a fingerprint feature quantity file 110. The fingerprints inputted by a person who desires entering are collated with the fingerprints registered by a fingerprint collation device 120. A finger input order management device 130 judges whether a collation result is matched with the input order of the fingers, which is previously registered or not. Only when the prescribed number of input times (three times, for example) are terminated and the fingerprints are inputted in the finger input order of an input order registration table, a locking management device 140 opens the locking of a door. Thus, collation precision can be improved by the combination of the input order and biometrics information, which only a person himself knows.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a personal authentication apparatus and method applied to a system for entering and exiting rooms and buildings using a computer, electronic commerce, and the like. Metric feature)
The present invention relates to a personal authentication apparatus and method applicable to a field of realizing high-precision security using a personal computer.

[0002]

2. Description of the Related Art Video information (1) published in September 1992, published by the Industrial Development Organization, as an authentication technique using biometric information (biometrics information) such as a fingerprint,
There is an entry management system described on pages 1 to 48.

In the room entry management system based on fingerprint collation described in this document, the fingerprint data of a user is registered in the system in advance, and the system extracts the fingerprint data of the user from a large number of registered fingerprints every time an input request is made. Search and collate with the input fingerprint. As a method of quickly searching for a user's fingerprint, a method in which a user inputs and specifies a registration number and a method in which an ID card is used are described. When a predetermined person presses a registered finger against the fingerprint input unit, the fingerprint is classified for each fingerprint pattern, subjected to fingerprint collation processing, and when the user is confirmed to be a correct user, the electric lock is unlocked. In particular, this system is characterized in that the comparison with the input fingerprint is performed only with respect to the registered fingerprint corresponding to the registration number or the ID card, so that quick comparison is possible.

[0004]

In the above-mentioned prior art, studies from the viewpoint of high-speed collation are progressing, but improvement in accuracy and security are not discussed. It can be considered that the accuracy of the collation can be increased if the algorithm is complicated. However, there is a problem of processing time and an increase of development cost. In the biometric personal identification technology, if the detection accuracy is increased, erroneous rejection occurs. On the contrary, if the detection accuracy is reduced, the erroneous detection increases, and it is difficult to adjust the threshold value in the matching process. For example, even if the detection accuracy is as high as 98%,
About two out of 100 people may enter the room by mistake, which is a security problem. It is necessary to increase the detection accuracy while suppressing the processing time.

In the case of using the fingerprint feature, since a fingerprint pattern remains in a fingerprint input device and a large number of remaining fingerprints exist in a living space, it is not impossible to imitate the fingerprint pattern. There is a problem with security.

An object of the present invention is to provide a personal authentication apparatus and method using ecological information that can achieve high security and high matching accuracy even if a low-accuracy matching method is used or a fingerprint pattern is duplicated. It is in.

[0007]

In order to solve the above-mentioned problems, according to the present invention, for example, when a fingerprint pattern is used as ecological information, not only is the fingerprint of one finger input, but also the order of the input fingers is changed. Considering a password, a method is used in which a plurality of finger inputs are combined and collated. This allows
Even if it is possible to imitate a fingerprint pattern, if you do not know the order of input of your fingers, you cannot follow the correct entry procedure.
In addition, the combination of a plurality of fingerprints can improve the accuracy of personal authentication from the accuracy of a single input.

[0008] That is, according to the first aspect of the present invention, there is provided a personal authentication apparatus using ecological information, comprising: input means for optically inputting ecological information; A storage unit that stores an input order, a matching unit that extracts a feature amount of the ecological information input by the user by the input unit, and matches the feature amount registered in the storage unit; Input order matching means for comparing the input order of the ecological information input by the user with the input order of each user's ecological information stored in the storage means, and matching by the matching means and the input order matching means As a result, only when the users registered in the storage means input the ecological information in the registered order,
Determining means for determining that the user is a correct user.

According to a second aspect of the present invention, there is provided a personal authentication apparatus using ecological information, wherein input means for optically inputting ecological information, feature amounts of ecological information of all users and each user input ecological information in advance. After extracting the characteristic amount of the ecological information input by the user by the storage unit storing the order and the input unit,
A matching unit that matches a feature amount registered in the storage unit and records a predetermined number of candidates obtained as a result of the matching in a matching candidate table; The content of the collation candidate table obtained as a result of the input is collated with the input order of the ecological information of each user stored in the storage means, and the collation candidate table is registered in the storage means in the collation candidate table. The input order collating means for collating whether there is a combination of candidates that can be recognized as having been input by the user in the registered order, and the result of collation by the collating means and the input order collating means, Only when there is a candidate combination in the matching candidate table that can be recognized as having entered ecological information in the registered order, the user is determined to be a correct user. Characterized by comprising a determining means.

According to a third aspect of the present invention, there is provided a personal authentication method using ecological information, wherein a feature amount of the ecological information of all users and an order in which each user inputs the ecological information are stored in a storage unit in advance. Optically inputting ecological information;
After extracting the characteristic amount of the ecological information input by the user, a collating step of collating with the characteristic amount registered in the storage means, an input order of the ecological information input by the user, and the order stored in the storage means. An input order matching step for checking the input order of the ecological information of each user, and a result of matching by the matching step and the input order matching step,
A determining step of determining that the user is a correct user only when the ecological information is input in the order in which the users registered in the storage means are registered.

According to a fourth aspect of the present invention, there is provided a personal authentication method using ecological information, wherein a feature amount of the ecological information of all users and an order in which each user inputs the ecological information are stored in a storage means in advance. Optically inputting ecological information;
A matching step of extracting a feature amount of the ecological information input by the user, comparing the feature amount with the feature amount registered in the storage unit, and recording a predetermined number of candidates obtained as a result of the matching in a matching candidate table; The contents of the collation candidate table obtained as a result of the user inputting the ecological information a plurality of times in a predetermined order are collated with the input order of the ecological information of each user stored in the storage means, and the collation candidate table is checked. An input order matching step for checking whether there is a combination of candidates that can be recognized as having been input in the order registered by the user registered in the storage means; and As a result of the collation, a combination of candidates that can be recognized as having input the ecological information in the order in which the users registered in the storage means are registered is stored in the collation candidate table. Only when it is in, it is characterized in that a determining step as a correct user.

[0012]

FIG. 1 is a block diagram showing an embodiment of the present invention.
7 will be described in detail with reference to FIG. This embodiment is an entry management system that performs door lock management by fingerprint collation.

FIG. 1 shows an apparatus configuration for realizing this embodiment. The room entry management system includes a fingerprint input device 100 including a TV camera, an A / D converter, a display device, and the like, a fingerprint feature file 110 in which a user's fingerprint feature and a finger input order are registered in advance. A fingerprint collation device 120 for collating the input fingerprint with the registered fingerprint, a finger input sequence management device 130 for determining the collation result and the input sequence of the pre-registered finger, and a collation result. The door lock management device 140 manages whether to leave the lock.

Hereinafter, referring to the flowchart of FIG.
The flow of the entry management process by the system of FIG. 1 will be specifically described.

First, before performing the entry management process of FIG. 2, fingerprint feature amounts of users A, B, and C who are users of the present system are registered in a feature amount registration table 150 shown in FIG. The order of fingers input by each user to the input device is registered in the input order registration table 160 shown in FIG. FIG.
4 and FIG. 4, information registered for one user is illustrated. Such information is registered in advance for the user. The feature amount registration table 150 and the input rank registration table 160 are stored in the fingerprint feature amount file 110 of FIG.

In this embodiment, the number of users is three in order to make the explanation easy to understand, but the number of registered persons is not limited. In addition, although only five fingers are registered with one hand, this point is also for the sake of simplicity of description, and both hands may be used. Also, the fingerprint input is performed three times, but this is also for the sake of easy understanding of the description, and is not limited thereto.

The finger input order is registered in the input order registration table 160 in a pattern as shown in FIG. 5, for example. A
Has registered the order of the middle finger → the index finger → the ring finger as shown in the pattern 170. Similarly, Mr. B
As shown in FIG. 0, registration was performed in the order of thumb → ring finger → index finger. Similarly, Mr. C registered in the order of the index finger → the middle finger → the ring finger as shown in the pattern 190. In the example shown in FIG. 5, one user does not use the same finger during three times of input, but the same finger may be used.
For example, the order may be the middle finger → the index finger → the middle finger.

The flow of the entry management process will be described with reference to the flowchart of FIG.

(1) Fingerprint Input Processing For example, a case will be described where Mr. A activates the entry management system shown in FIG. 1 in order to enter a room. When the entry management system is started, first, an input counter n is initialized (n = 1) (step 200).
The input counter n is a work variable for counting the number of finger inputs. Next, an instruction to input a first finger is displayed on a display device attached to the input device 100 (step 210). When Mr. A presses his / her finger against the fingerprint input device 100 according to the instruction, the system inputs the fingerprint of the finger pressed by Mr. A using the fingerprint input device 100 (step 220). Mr. A knows in which order he should enter his fingerprint to enter the room. That is, the user knows the input order pattern 170 registered in the input order registration table 160. Therefore, what A inputs here is the middle finger to be input first.

(2) Feature Extraction and Matching Process (Step 230) The feature is extracted from the first input fingerprint pattern, and is compared with the feature amount registered in the feature amount registration table 150 of the fingerprint feature amount file 110. As shown in FIG. 3, the feature amount registration table 150 stores information on the feature amount of the fingerprint pattern of each finger of each user. As described above, the user is registered in the feature amount registration table 150 in advance.

The method of feature extraction and collation processing may be performed, for example, by the method described in the section of the prior art. The outline of the conventional feature extraction and matching processing is as follows, for example. A fingerprint pattern of one registered finger is input, the input fingerprint pattern is classified based on the central core information, and thereafter, the fingerprint pattern is thinned, a feature amount is extracted, and the registered feature amount is collated. If the verification is successful, the electric lock is unlocked. Also in the present invention, the same method may be used for feature extraction and feature amount collation.

(3) Registration Table Matching Process As a result of the matching, for example, when it is recognized that the input finger is Mr. A's middle finger, the input order registration table 16 in FIG.
0 is compared with the contents of the first finger of Mr. A (that is, it is determined whether the contents of the first finger of Mr. A in the input order registration table 160 is the middle finger), and a criterion having a similarity is determined. It is determined whether or not the condition is satisfied (step 250). As the similarity, for example, a correlation value of a distance between feature points such as a minutia in a fingerprint may be used. Here, when the correlation value is 0.7 or more, it is determined that they are similar.

The collation result is the person's finger input order table 16
If the content is different from 0, it is determined that the user is not the user and that the key is not opened, a message is displayed on the display device (step 270), and the process ends. In the case of Mr. A's middle finger this time, it is determined that the collation with the finger input order table 160 matches, and the next determination processing is performed (step 260).

This determination process (step 260) is a process for determining whether or not the finger has been input a predetermined number of times since the number of finger inputs is set in advance. In this example, the number of inputs is three, while the input counter n is still n = 1.
= 2 (step 265), and processing for the next finger is performed.

(4) Fingerprint Input Processing An instruction to input a second fingerprint is displayed on the display device (step 210). When Mr. A presses the second finger against the fingerprint input device 100 according to the instruction, the system inputs the fingerprint of the finger pressed by Mr. A using the fingerprint input device 100 (step 220). This means that the fingerprint of the second finger has been input. The processing after the input is the same as the above (2) and (3). The same applies to the third fingerprint input. As a result, the processes (2) and (3) are repeated three times.

As described above, the predetermined number of times (three times) has been input.
The following processing is performed only when a fingerprint is input in the order of finger input in the input order registration table 160 shown in FIG.

(5) Processing to open the lock (step 28)
0) As a result of inputting three times, the fingerprint collation result is A's middle finger → index finger → ring finger, so the input person is recognized as correctly registered A, and the electric lock is opened and the door is opened ( Step 280).

According to the above-described embodiment of the present invention, by combining the input sequence of the finger which can be understood only by the user and the fingerprint collation, there is an effect of improving the fingerprint collation accuracy and an effect of improving security. For example, if the accuracy of one fingerprint matching process is 80% (that is, if 100 people are input, 20 people will enter the room by mistake), since the input is repeated three times, (1- (1-0) .8) (1-0.8) (1
−0.8)) = 99.2%, which can significantly improve the entry control accuracy. The accuracy of 99.2% is an accuracy in which only eight persons are mistaken for 1000 persons. Also, even if the fingerprint pattern is imitated by any method, it is not possible to enter the room unless the input order is clear, which is also effective in improving security.

In the embodiment of the present invention, on the assumption that the matching accuracy is high, the existence of the candidate is not assumed in the matching result. However, in an actual system, if a threshold value (threshold value used for collation) is set so as to increase collation accuracy, erroneous rejection increases, and there is a possibility that a correct user is rejected. For this reason, there arises a problem that even a correct user has to repeat the fingerprint input several times.

In the following, as a second embodiment, a method of ensuring high security and controlling entry to a room even if the matching accuracy is low will be described with reference to FIG.

A case will be described in which three persons A, B, and C are registered in advance and A enters the room. Also,
Only differences from the first embodiment will be described.
(1) Fingerprint input, (2) feature extraction and collation processing, and (5) lock management processing are the same as those in the above-described embodiment, and thus description thereof will be omitted.

(1) Fingerprint input processing (steps 200 to 2)
20) (2) Feature extraction and collation processing (Step 230) (3) Candidate table writing processing (Step 290) As a result of collating the input fingerprint with the characteristic quantity registered in the characteristic quantity registration table 150, the collation accuracy Three candidates from the highest order (the highest similarity) are recorded in the collation candidate table 300 as shown in FIG. The collation candidate table 300 is a table stored in the fingerprint feature file 110. The collation candidate table 300 includes first, second and third candidates obtained as a result of collation with respect to the first fingerprint input, and the second fingerprint. First and second obtained as a result of collating the input
9 is a table for recording first and second candidates and third candidates obtained as a result of performing collation with respect to a third fingerprint input. In the collation candidate table 300 of FIG. 7, the collation result for the first fingerprint input is that the first candidate is A's middle finger, the second candidate is A's ring finger,
It is assumed that there is no third candidate. In addition, the collation result for the second fingerprint input is that the first candidate is A's index finger, the second candidate is B's middle finger, and the third candidate is none.
The matching result for the third fingerprint input indicates that the first candidate is B
'S middle finger, 2nd candidate is A's index finger, 3rd candidate is A
Suppose that it became Mr. ring finger. In this example, the number of matching candidates is three, but the number of candidates is not limited to three. If the threshold for matching is not satisfied, the column of the candidate table 300 is left blank.

Candidate table writing process (step 29)
After 0), the next determination processing (step 260) is performed.
This determination process (step 260) is a process of determining whether or not the finger has been input a predetermined number of times since the number of times of finger input is set in advance. Here, only the number of times of finger input may be determined, and the order may be determined so that the same finger is input repeatedly. In this example, the number of inputs is three, while the input counter n is still n = 1, so that the counter n is incremented by 1 to n = 2 (step 26).
5) Perform processing for the next finger.

Next, an instruction to input a second fingerprint is displayed on the display device (step 210). When Mr. A presses the second finger against the fingerprint input device 100 according to the instruction, the system inputs the fingerprint of the finger pressed by Mr. A using the fingerprint input device 100 (step 220). This means that the fingerprint of the second finger has been input.
The processing after the input is the same as the above (2) and (3). The same applies to the third fingerprint input.
As a result, the processes (2) and (3) are repeated three times. After the predetermined number of times (three times) has been input, the following processing is performed.

(4) Candidate table collation processing (step 3)
20) It is assumed that a result of performing the above processing results in a collation candidate table 300 as shown in FIG. The comparison candidate table 300 is compared with the input order registration table 310, and the corresponding registrant is identified (step 320). Here, it is assumed that the fingerprint input order of each of Mr. A, Mr. B, and Mr. C has been registered as in the input order registration table 310 shown in FIG. The input order registration table 310 shown in FIG.
Here, the input order of each user is shown in the order of the first, second, and third times from the top. As a result of the candidate table matching process, if it is recognized that the registered users have input their fingerprints in the correct order, the process proceeds to step 280; otherwise, the process proceeds to step 270. In this example,
As a result of the candidate table collation, the shaded portion of Mr. A's input order registration table 310 and the collation candidate table 300 match, so the input person is regarded as Mr. A who is permitted to enter the room.

(5) Processing to open the lock (step 28)
0) Since it is recognized that the input person is correctly registered Mr. A, the electric lock is opened and the door is opened. If the registered user does not input the fingerprint in the correct order, it is determined that the user is not the user and that the key is not opened, and a message is displayed on the display device (step 270), and the process ends.

As described above, by deciding candidates, there is an effect that the present system can be operated efficiently even if the accuracy of feature extraction matching is low.

Although the fingerprint is used as the biometric information in the embodiment of the present invention, the present invention can be similarly realized using other biometric information. For example, the present invention can be applied by using a retinal pattern or an iris (iris) pattern and using the input order of the left and right eyes (for example, right eye → right eye → left eye) of these patterns. Although the present invention has been described with reference to the entry management system as an example of the embodiment, the invention is also applicable to personal authentication in the field of conducting business transactions via a network.

Further, a medium storing a program for realizing the flow of the personal authentication method using the ecological information according to the present invention is also one of the present invention.

[0040]

According to the present invention, there is an effect that the collation accuracy can be improved by combining the input sequence that can be known only by the user and the collation of the ecological information. In addition, for example, when a fingerprint pattern is used, the fingerprint pattern may be imitated by some method. However, even if the imitation is performed, it is not possible to enter the room unless the input order is clear, which is also effective in improving security. .

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an entrance management system using fingerprint collation according to the present invention.

FIG. 2 is a flowchart of an entry management process based on a plurality of fingerprint inputs.

FIG. 3 is a configuration diagram of a registered fingerprint feature amount table.

FIG. 4 is a configuration diagram of a registered finger input ranking table.

FIG. 5 is a diagram showing an example of three finger input rankings;

FIG. 6 is a flowchart of an entry management process in fingerprint matching when a plurality of candidates are allowed in the matching result.

FIG. 7 is a diagram showing a specific example of a collation candidate table and an input order registration table.

[Explanation of symbols]

130 finger input sequence management device 240 finger input sequence registration table collation process 250 sequence registration table collation result determination process 160 finger input sequence registration table 290 collation result candidate table writing process 320 collation process between collation candidate table and input sequence registration table

Continuation of the front page (72) Inventor Seiji Nasu 3-2-1 Sachicho, Hitachi City, Ibaraki Prefecture Inside Hitachi Engineering Co., Ltd.

Claims (4)

[Claims] An input means for optically inputting ecological information, a storage means for storing in advance a characteristic amount of ecological information of all users and an order in which each user inputs ecological information, A matching unit for extracting a feature amount of the ecological information input by the user, and then comparing the extracted feature amount with the feature amount registered in the storage unit; an input order of the ecological information input by the user by the input unit; Input order matching means for checking the input order of each user's ecological information stored in the storage means, and as a result of matching by the matching means and the input order matching means, the user registered in the storage means is registered. A personal identification device using ecological information, comprising: a determination unit that determines that the user is a correct user only when the ecological information is input in the specified order. 2. An input means for optically inputting ecological information, a storage means for storing in advance a characteristic amount of ecological information of all users and an order in which each user inputs ecological information, and Collating means for extracting a characteristic amount of the ecological information inputted by the user, collating the characteristic amount with the characteristic amount registered in the storage means, and recording a predetermined number of candidates obtained as a result of the collation in a collation candidate table; The user compares the contents of the matching candidate table obtained as a result of the user inputting the ecological information a plurality of times in a predetermined order with the input means and the input order of the ecological information of each user stored in the storage means. An input order matching unit that checks whether there is a combination of candidates in the matching candidate table that can be recognized as being input in the order in which the users registered in the storage unit are registered; As a result of the matching by the input order matching means, if the combination of candidates that can recognize that the user registered in the storage means has input the ecological information in the registered order is present in the matching candidate table, it is correct. A personal authentication device using ecological information, comprising: means for determining that the user is a user. 3. A step of storing in advance a feature amount of ecological information of all users and an order in which each user inputs ecological information in a storage means; a step of optically inputting ecological information; After extracting the characteristic amount of the input ecological information, a collating step of collating with the characteristic amount registered in the storage means, the input order of the ecological information input by the user, and the order stored in the storage means An input order matching step for checking the input order of each user's ecological information; and a result of the matching performed by the matching step and the input order matching step, in the order in which the users registered in the storage means are registered. A personal identification method using ecological information, comprising a determination step of determining that the user is a correct user only when the ecological information is input. 4. A step of storing in advance a feature amount of ecological information of all users and an order in which each user inputs ecological information in a storage unit; a step of optically inputting ecological information; After extracting the characteristic amount of the input ecological information, collating with the characteristic amount registered in the storage means, and recording a predetermined number of candidates obtained as a result of the collation in a collation candidate table; Compare the contents of the matching candidate table obtained as a result of inputting the ecological information a plurality of times in a predetermined order and the input order of the ecological information of each user stored in the storage means,
An input order matching step of checking whether there is a combination of candidates in the matching candidate table that can be recognized as having been input in the order in which the users registered in the storage means are registered; As a result of the collation in the order collation step, a correct user is determined only when there is a candidate combination in the collation candidate table that can recognize that the users registered in the storage means have input the ecological information in the registered order. A personal identification method using the ecological information, comprising: