JPH11253426A - Biometric feature authentication method and device, recording medium - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide a personal authentication device with improved authentication accuracy. A personal authentication apparatus 1 includes a plurality of different biometric feature matching units 11, a matching score integration processing unit 12, a matching score vector identification processing unit 13, and an authentication result processing unit 14. The matching score integration processing unit 12 integrates the individual matching scores and generates a matching score vector using the total number of matching scores for each biometric feature matching unit 11 as the number of dimensions. The matching score vector identification processing unit 13 identifies a matching score based on a predetermined identification rule in a matching score vector space. Each collation score is classified into either a “valid” or “illegal” category, and the authentication result processing unit 14 performs integrated authentication based on the category classification result in the collation score vector identification processing unit 13.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a biometric feature authentication technique, and more particularly to a technique for improving the authentication accuracy by integrating collation scores representing a plurality of different biometric features.

[0002]

2. Description of the Related Art In recent years, with the proposal of various forms such as electronic money in the electronic commerce field and Internet banking using a public network such as the Internet,
Research and development of personal authentication technology for authenticating users and the like are active. This personal identification technology is for identifying an individual based on biological characteristics such as a fingerprint and an iris of a human body. Among the biological characteristics, in particular, the face and voiceprint (voice)
Is known as an authentication method that can be constructed relatively easily, because it has little load and resistance on the user side and does not require special equipment. On the other hand, there is also known a method of improving authentication performance by integrating a plurality of different authentication methods. This method aims at improving the authentication performance by using the logical sum of the authentication results for each of the plurality of authenticators.

[0003]

When personal authentication is performed by a single authentication device (equipment), there is a problem that the authentication accuracy fluctuates due to, for example, the use environment described below. there were. (1) In the face authentication device, there is a lighting condition at the time of authentication and a characteristic change due to a change in the expression of the person to be authenticated. (2) In a voiceprint (voice) authenticator, there is a characteristic change due to a natural change with time. (3) In the fingerprint authentication device, there is a characteristic change due to the state of the authentication target person such as drying of the skin. The greater the feature variation, the lower the authentication accuracy,
In addition, satisfactory authentication accuracy cannot be obtained from the viewpoint of security.

On the other hand, when personal authentication is performed based on authentication results from a plurality of authentication devices, the exclusion rate increases due to the mere use of the logical sum, which is inconvenient for the user. There was a problem that would be.

An object of the present invention is to provide a method for authenticating a biological feature, which can improve the accuracy of authentication. Another object of the present invention is to provide an apparatus suitable for performing the above method. Another object of the present invention is to provide a recording medium for realizing the above method or apparatus on a computer device.

[0006]

According to the present invention, there is provided an authentication method comprising: a plurality of collation data each representing a biometric feature; and a plurality of collation scores representing a degree of similarity between the authentication data for each biometric feature. Is generated, and each matching score is integrated to perform a vector expression in which the number of scores is the number of dimensions. Based on a predetermined identification rule in a feature space of the matching score formed by the vector expression, the plurality of pieces of matching data are obtained. It is characterized by judging the integrated validity of.

[0007] In order to solve the above other problems, the present invention provides:
An authentication device for a biological feature and a personal authentication device having the authentication device are provided. The personal authentication apparatus of the present invention includes: means for generating a plurality of collation scores each representing a degree of similarity between a plurality of collation data each representing a biometric feature and a preset authentication data for each biometric feature; Means for generating a score vector with the number of scores as the number of dimensions, and matching for identifying the validity of each matching score based on a predetermined identification rule in a score vector space formed by the generated score vectors. And a score discriminating means, and a means for judging integrated validity of the plurality of collation data based on the validity of each identified collation score.

[0008] The collation score identification means is configured to identify the legitimacy of the collation score by classifying the category to which the collation score should belong, into one of "valid" and "illegal" categories. Is done.

In addition, the collation score identifying means may include, for example, a standard vector set in advance by a predetermined number for either “valid” or “illegal” category to which the collation score belongs based on a predetermined closest rule. , And a distance value from the distributed matching score is calculated, and is classified into a category corresponding to a standard vector when the calculated distance value is the minimum. The collation score collation means is a case where the collation score is linear,
In the case of a non-linear distribution, the distance between the standard vector and the distributed matching score is calculated based on a predetermined quadratic format identification function. In addition, based on the statistical probability calculated in accordance with the distribution of the matching score, a “valid” or “illegal” to which the matching score belongs
It is configured to be classified into any of the categories. Alternatively, for example, based on a predetermined artificial neural network, learning is performed by configuring an internal neuron so as to output a pattern corresponding to either the category of “valid” or “unfair” to the category to which the matching score belongs. , The matching score is classified into a category to which the matching score belongs from an output pattern obtained as an input pattern.

[0010] The personal authentication device of the present invention is configured such that the authentication device authenticates the legitimacy of the person to be authenticated based on collation data for each of the biological characteristics of the person to be authenticated.

A recording medium of the present invention for solving the above-mentioned other problems is a recording medium in which a program for causing a computer device to execute the following processing is recorded in a form readable by the computer device. (1) a process of generating a plurality of matching scores each representing a degree of similarity between a plurality of matching data each representing a biological feature and a preset authentication data for each biological feature; (2)
(3) processing of integrating each matching score to generate a score vector having the number of dimensions as the number of scores, and (3) validity of each matching score based on a predetermined identification rule in a score vector space formed by the generated score vectors. (4) a process of determining the integrated validity of the plurality of collation data based on the validity of each identified collation score.

[0012]

Embodiments of the present invention will be described below in detail with reference to the drawings. (First Embodiment) FIG. 1 is a functional block diagram showing an embodiment of a personal authentication apparatus to which the present invention is applied. In the figure, the solid line represents the flow of processing, and the broken line represents the flow of data. The personal authentication device 1 is realized by a computer device equipped with an operating system (OS). The computer device reads and executes a predetermined program to form N biometric feature matching units 1.
1A to 11N (hereinafter abbreviated as 11), a verification score integration processing unit 12, a verification score vector identification processing unit 13, and an authentication result processing unit 14.

The above program is usually stored in an internal storage device or an external storage device of a computer device, and is read and executed as needed. However, the present invention can be implemented as long as the functional blocks 11 to 14 can be formed on a computer device, and the recording form may be arbitrary. For example, it is recorded on a portable recording medium such as a CD-ROM or FD that can be separated from a computer device, or a program server connected to a private network, read at the time of use, and installed in the internal storage device or the external storage device. The program may be executed at any time.

The biometric feature matching section 11 performs matching of matching data such as fingerprints and voiceprints representing the biometric features of the person to be authenticated, and is provided corresponding to the type of the biometric feature. Although not shown, each authentication processing unit 11 includes:
A corresponding set of input interfaces is provided.
This input interface is known. The verification data is input to the biometric characteristic verification unit 11 via a predetermined input interface (not shown). From the biometric feature matching unit 11, the matching score is converted to the matching score integration processing unit 12.
Is input to The collation score indicates, for example, the degree of similarity of the collation data with the authentication data of the authentication subject set in advance for each biometric feature.

The collation score integration processing unit 12 integrates all acquired collation scores to generate a collation score vector described later, and inputs all collation scores and collation score vectors to the collation score vector identification processing unit 13. I do.

The collation score vector identification processing section 13 identifies each collation score based on a predetermined identification rule. This identification is performed by classifying each collation score into a category determined to belong to each, that is, any one of “valid” and “illegal”. The identification result is input to the authentication result processing unit 14.

The authentication result processing section 14 determines the legitimacy of the person to be authenticated based on the processing result in the collation score vector identification processing section 13. The authentication result is output via a predetermined output interface (not shown).

Next, an authentication method using the personal authentication device 1 will be described with reference to FIG. The person to be authenticated, that is, the person to be authenticated, is input via the corresponding input interface,
Input collation data for each biological characteristic (step S10)
1). Each input matching data is subjected to matching processing in the corresponding biometric feature matching unit 11 (Step S).
102). The matching score from each biometric feature matching unit 11 is input to the matching score integration processing unit 12 (step S103).

The collation score integration processor 12 integrates all acquired collation scores to generate a collation score vector (step S104). Specifically, for each collation score, a vector expression using the N outputs of each biometric feature collation unit 11 as the number of dimensions, that is, an N-dimensional collation score vector is generated. In addition, the matching score integration processing unit 12
Inputs the collation score vector and all the collation scores to the collation score vector identification processing unit 13.

In the collation score vector identification processing unit 13,
An identification process is performed for each collation score (step S10)
5). The identification rule in the identification process is a process of determining a category to which the input collation score belongs based on, for example, a nearest rule based on the Euclidean distance.
Specifically, using the average value of the matching score in the corresponding biometric feature and the like, a standard vector related to the category to which the matching score belongs is generated for each of two categories of “valid” or “illegal”, "Correct" for collation score
Alternatively, a distance value with respect to any of the “invalid” standard vectors is calculated, and a standard vector that is close is specified based on the distance value to determine a category. These standard vectors are indices having a template-like aspect when determining a category. The matching score for which the category has been determined and identified is input to the authentication result processing unit 15.
From this, the identification process is handled as a two-category identification problem of “valid” or “illegal” in the collation score vector space, that is, whether or not the person to be authenticated is the subject.

FIG. 3 is a conceptual diagram relating to identification of a collation score in the collation score vector space described above. This figure is a collation score vector space generated in the collation score integration processing unit 12, and “input data” indicated by black dots in the figure represents a collation score in the biometric feature collation unit 11. The matching score vector identification processing unit 13 determines whether the “input data” is close to one of a plurality of standard vectors, ie, a valid or invalid template, and determines a category to which the “input data” belongs.
In this case, the “input data” is close to the “valid template”, and the category to which the “input data” belongs is “valid”.

Next, the authentication result processing unit 15 performs integrated authentication based on each of the matching scores whose category has been determined in the matching score vector identification processing unit 13. If the authentication result is "illegal" (step S106: No), the person to be authenticated is rejected (step S107). On the other hand, if the authentication result is “valid” (step S106: Ye
s), a person to be authenticated is accepted (step S108). The authentication result is output to an output interface (not shown) as appropriate.

In the above description, the case where the closest rule based on the Euclidean distance is applied to the identification rule in the matching score vector identification processing unit 13 is described. Rules)
The law is equally applicable. In this case, the above-mentioned standard vector related to the category to which the matching score belongs is “k”
(K is a natural number), that is, a predetermined number is set in advance, and the configuration may be such that the category in which the distance value from these standard vectors is minimum is specified. If the collation score is other than the linear distribution as described above, that is, a non-linear distribution, for example, a distance value between the standard vector and the collation score is calculated based on a predetermined quadratic identification function. What is necessary is just to comprise. Furthermore, the present invention is not limited to the above example, and can be configured using a statistical method and an artificial neural network method. In the case of the statistical method, for example, the configuration may be such that the category of “valid” or “illegal” to which the matching score belongs is specified based on the probability calculated according to the distribution of the matching score.
In the case of the artificial neural network method, for example, the internal neural network unit is configured to output a pattern corresponding to the category to which the matching score should belong, either "valid" or "illegal". What is necessary is just to make it learn appropriately, and to make it possible to specify a category to which the matching score belongs from an output pattern corresponding to the matching score as an input pattern.

As described above, the personal authentication device 1 of the present embodiment
In the above, a plurality of different collation scores and authentication results are integrated, so that the authentication accuracy is improved as compared with the authentication processing by a single authentication device as in the conventional method.

Further, even when the characteristics of specific collation data greatly fluctuate due to the use environment of the apparatus and the authentication result is inconsistent, integrated authentication in accordance with other authentication results is performed. As a result, reliability is improved.

Also, by integrating the authentication results based on the characteristics of the collation score, high authentication accuracy can be achieved without increasing the personal exclusion rate due to the mere use of the logical sum as in the conventional method. And the convenience for the user is improved.

Further, by constructing a system in which the present invention is incorporated into an existing information system, for example, it can be expected that the reliability and confidentiality of the use of the system are greatly improved.

[0028]

As is apparent from the above description, according to the present invention, a personal authentication environment with higher recognition accuracy can be realized.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a personal authentication device according to an embodiment of the present invention.

FIG. 2 is a processing procedure diagram in the personal authentication device.

FIG. 3 is a conceptual diagram illustrating a collation score vector space.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 personal authentication device 11 biometric feature matching unit 12 matching score integration processing unit 13 matching score vector identification processing unit 14 authentication result processing unit

Claims (9)

[Claims] 1. A method for generating a plurality of collation scores each representing a degree of similarity between a plurality of collation data each representing a biometric feature and authentication data for each of the biometric features, and integrating the collation scores to dimension the number of scores. Performing a vector representation as a number and determining an integrated validity of the plurality of pieces of collation data based on a predetermined identification rule in a feature space of a collation score formed by the vector representation. Authentication method of the characteristic feature. 2. A means for generating a plurality of matching scores each representing a degree of similarity between a plurality of matching data each representing a biological feature and a preset authentication data for each biological feature, and integrating each matching score. Means for generating a score vector having the number of scores as the number of dimensions, and matching score identifying means for identifying the validity of each matching score based on a predetermined identification rule in a score vector space formed by the generated score vectors A biometric feature authentication device, comprising: means for determining integrated validity of the plurality of pieces of collation data based on the validity of each identified collation score. 3. The collation score identification means is configured to classify a category to which the collation score belongs to either a “valid” or “illegal” category to identify the validity of the collation score. The authentication device according to claim 2, wherein the authentication is performed. 4. A method according to claim 1, wherein said matching score identifying means includes a predetermined number of standard vectors set in advance for any one of “valid” and “invalid” categories to which said matching score belongs based on a predetermined closest rule. 3. The system according to claim 2, wherein distance values to the distributed matching scores are calculated, and the data is classified into a category corresponding to a standard vector when the calculated distance value is minimum. Or the authentication device according to 3. 5. The collation score identification means calculates a distance value between the standard vector and the distribution collation score based on a predetermined quadratic form identification function when the collation score has a non-linear distribution. The authentication device according to claim 2, wherein the authentication device is configured as follows. 6. The collation score identification means, based on a statistical probability calculated in accordance with the distribution of the collation score, determines whether the collation score should belong to “valid” or “illegal”.
The authentication device according to claim 2, wherein the authentication device is configured to be classified into any one of the following categories. 7. The collation score identification means outputs a pattern corresponding to a category to which the collation score belongs to either “valid” or “illegal” based on a predetermined artificial neural network. 4. The method according to claim 2, wherein learning is performed by forming an internal neuron, and the collation score is classified into a category to which the collation score belongs from an output pattern obtained as an input pattern. 5. Authentication device. 8. The authentication device according to claim 2, wherein the authentication device authenticates the legitimacy of the person to be authenticated based on collation data for each of the biological characteristics of the person to be authenticated. A personal authentication device characterized in that: 9. A process of generating a plurality of matching scores each representing a degree of similarity between a plurality of matching data each representing a biometric feature and a preset authentication data for each biometric feature; A process of generating a score vector with the number of scores as the number of dimensions, a process of identifying the validity of each of the collation scores based on a predetermined identification rule in a score vector space formed by the generated score vectors, an identified collation A recording medium in which a program for causing a computer device to execute a process of determining integrated validity of the plurality of pieces of collation data based on the validity of each score is recorded in a form readable by the computer device.