WO2011062096A1

WO2011062096A1 - Multimodal authentication device

Info

Publication number: WO2011062096A1
Application number: PCT/JP2010/069990
Authority: WO
Inventors: 中村　陽一; 亀井　俊男
Original assignee: 日本電気株式会社
Priority date: 2009-11-17
Filing date: 2010-11-10
Publication date: 2011-05-26

Abstract

Disclosed is a multimodal authentication device wherein both safety and convenience are improved in personal authentication in which at least two types of physical features are used. The multimodal authentication device is provided with: a first matching means which applies a plurality of types of matching algorithms to determine whether or not a first physical feature of an authentication subject that is input is that of an authorized user; a second matching means which applies at least one type of matching algorithm to determine whether or not a second physical feature of the authentication subject that is input is that of an authorized user; and an authentication means which determines whether the authentication subject is an authorized user if a match is established in the first and second matching means. The first matching means determines that the first physical feature of the user that is input is that of an authorized user when a match is established in matching in which at least any one of the matching algorithms from among the plurality of types of matching algorithms is applied.

Description

Multimodal authentication device

The present invention relates to a multimodal authentication technique using a plurality of physical features.

In recent years, authentication technology using various physical features such as fingerprints and finger vein patterns has attracted attention. In such authentication technology, improvement of authentication accuracy is a research subject.

As one method for improving the authentication accuracy, a multimodal biometric authentication system that combines a plurality of physical features has been proposed (see, for example, Patent Document 1).

In the authentication system proposed by Patent Document 1, two or more types of authentication methods (for example, fingerprint authentication method, face authentication method, etc.) are combined, and a theoretical value representing the probability of whether or not the user is a registrant is used for each authentication method. It is determined whether or not the user is the registrant based on the result of the logical operation.

JP 2000-242786 A

However, when it is determined that the user is a registered person when the authentication is successful with any of the authentication methods, the attacker only has to forge a physical feature (for example, fingerprint) corresponding to any one of the authentication methods, There is a problem that safety is lowered compared to the case where it is determined that the user is the registrant only when the authentication is successful in all the adopted authentication methods.

On the other hand, if it is determined that the registrant is the only person when authentication is successful in all authentication methods, the user is rejected if authentication fails in any of the authentication methods, which reduces convenience. There's a problem.

The present invention has been made in view of the above-described circumstances, and provides a multimodal authentication apparatus capable of improving both safety and convenience in personal authentication using two or more types of physical features. The purpose is to provide.

The multimodal authentication apparatus according to the present invention applies a plurality of types of matching algorithms to the first physical characteristics of the input authentication target person to determine whether or not the authentication target person belongs to a legitimate user. A second verification unit that applies one or more verification algorithms to determine whether or not the second physical feature of the input person to be authenticated is a regular user; Authentication means for determining that the person to be authenticated is a regular user when the first and second matching means are matched, and the first matching means includes the plurality of types of matching algorithms. Of these, when collation is established in the collation using at least one of the collation algorithms, it is determined that the first physical feature of the input user is that of the regular user. The features.

In addition, the multimodal authentication apparatus according to the present invention applies a plurality of types of verification algorithms to determine whether or not the first physical feature of the input authentication target person belongs to the authorized user. And a second verification unit that determines whether or not the input second physical feature of the person to be authenticated is an authorized user by applying one or more verification algorithms. And an authentication means for determining that the person to be authenticated is a regular user when the first and second matching means are matched, wherein the first matching means includes the plurality of types of matching. Based on the result of combining each of the weighted theoretical values by weighting each theoretical value that represents whether or not the user is a regular user, obtained by combining algorithms and applying each matching algorithm If the verification is determined to be established, the first physical characteristic of the object's input to and determines that is legitimate user.

Further, the multimodal authentication method according to the present invention is a collation for determining whether or not the first physical feature of the input authentication target person belongs to a regular user by applying a plurality of types of collation algorithms. A first physical feature of the user that is input when the collation is established in the collation using at least one of the plural collation algorithms among the plurality of collation algorithms A first collation step for judging that the second physical feature of the person to be authenticated is applied and one or more collation algorithms are applied to determine whether or not the legitimate user belongs. When verification is established in the second verification step to be performed and the first and second verification steps, it is determined that the person to be authenticated is a regular user. Characterized in that it comprises a testimony step.

According to the present invention, it is possible to improve both safety and convenience in identity authentication using two or more types of physical features.

It is a figure which shows the structure of the multimodal authentication apparatus which concerns on this embodiment. It is a figure which shows the functional block of the computer which concerns on the same embodiment. It is a figure which shows the principle of the reflected light sensing which concerns on the same embodiment. It is a figure which shows the principle of the transmitted light sensing which concerns on the same embodiment. It is a figure which shows the fingerprint image which concerns on the same embodiment. It is a figure which shows the finger vein image which concerns on the same embodiment. It is a figure which shows the outline of the personal authentication operation | movement which concerns on the same embodiment. It is a flowchart which shows the fingerprint collation process which concerns on the embodiment. It is a flowchart which shows the finger vein collation process which concerns on the same embodiment.

Hereinafter, preferred embodiments of a multimodal authentication device according to the present invention will be described with reference to the accompanying drawings.

A. Embodiment (1) Configuration of Embodiment FIG. 1 is a diagram illustrating a configuration of a multimodal authentication device 100 according to the present embodiment.

The multimodal authentication device 100 is a device that performs authentication by combining a plurality of physical features (modalities). In the present embodiment, it is assumed that fingerprints and finger veins are used as a plurality of modalities. For example, a feature (shape feature) representing a human body shape such as a fingerprint, a face image, and an iris may be used. Moreover, you may utilize the characteristic (behavioral characteristic) based on a person's action, such as a voice, a handwritten signature, and a key touch. Note that the combination of physical features of fingerprints and finger veins and the combination of physical features of faces and iris are obtained from the same physical part (finger or face). There are merits such that many parts can be used in common as a sensor for acquiring the information, and the apparatus itself can be miniaturized. Furthermore, since the user who is the subject of authentication needs only one input operation for authentication, there is an advantage that the authentication can be performed easily and speedily.

As shown in FIG. 1, the multimodal authentication device 100 includes an upper housing 10a, a lower housing 10b, a first light source 30a provided in the upper housing 10a, and a first light source 30a provided in the lower housing 10b. 2 light sources 30 b, an imaging device (input means) 40, and a computer 50. In the multimodal authentication apparatus 100, two types of collation methods are prepared for each physical feature of the finger and the finger vein. When the collation is established by any one of the collation methods, it is determined that this physical feature is the physical feature of the person (details will be described later). Hereinafter, functions of the computer 50 serving as the center of the multimodal authentication device 100 will be described.

FIG. 2 is a functional block diagram of the computer 50.

The computer 50 includes a fingerprint collation unit 510a, a fingerprint collation database 520a, a finger vein collation unit 510b, a finger vein collation database 520b, and a personal authentication unit 530.

The fingerprint collation unit 510a determines whether or not the user's fingerprint corresponding to the fingerprint image output from the imaging device 40 is that of the user. The fingerprint collation unit 510a includes a plurality of types of fingerprint collation algorithms. In the present embodiment, the fingerprint verification unit 510a includes a first fingerprint verification module MF1 for verifying a user's fingerprint by an image pattern matching method and a second fingerprint verification module for verifying a user's fingerprint by a minutia method. MF2.

In the fingerprint verification database 520a, a first fingerprint verification template TF1 for the first fingerprint verification module and a second fingerprint verification template TF2 for the second fingerprint verification module are registered. In the first fingerprint matching template TF1 and the second fingerprint matching template TF2, a line image corresponding to the fingerprint of the authorized user (here, the principal) and minutia information of the principal's fingerprint are registered. Yes. In the present embodiment, it is assumed that there is one authorized user, but there may be a plurality of authorized users. When there are a plurality of regular users, a unique user ID may be assigned to each regular user, and a line image may be registered in association with the user ID (this is the same for other templates described later).

The finger vein collating means 510b determines whether or not the user's blood vessel pattern (vein pattern) corresponding to the finger vein image output from the imaging device 40 is that of the user. The finger vein matching unit 510b includes a plurality of types of finger vein matching algorithms. In the present embodiment, the finger vein matching unit 510b includes a first finger vein matching module for matching the user's finger vein by the pattern matching method and a second finger vein for matching the user's finger vein by the feature amount method. A finger vein verification module.

In the finger vein matching database 520b, a first finger vein matching template TV1 for the first finger vein matching module and a second finger vein matching template TV2 for the second finger vein matching module are registered. Yes. In the first finger vein matching template TV1 and the second finger vein matching template TV2, an image representing the blood vessel pattern of the person and the feature amount of the blood vessel pattern of the person are registered.

The personal authentication unit 530 receives the verification result of the fingerprint verification unit 510a and the verification result of the finger vein verification unit 510b, and performs personal authentication. Specifically, the personal authentication unit 530 performs AND type integration of the fingerprint matching result and the blood vessel pattern matching result (see γ in FIG. 5), and when both fingerprint matching and vein matching are established. Only when the person to be authenticated is determined to be the person. Hereinafter, features of fingerprint matching and finger vein matching according to the present embodiment will be described.

<Fingerprint verification>
FIG. 3A is a diagram illustrating the principle of reflected light sensing realized by the multimodal authentication device 100.

The multimodal authentication device 100 performs reflected light sensing with the second light source 30b and the imaging device 40 provided in the lower housing 10b, and obtains a fingerprint image as shown in FIG. 4A. Specifically, a fingerprint image in which the ridge RL where the skin is raised is bright and the valley line VL in between is dark (see FIG. 4A).

The matching algorithm of the image pattern matching method extracts a ridge pattern as a line image by performing image processing such as binarization on the fingerprint image obtained by the imaging device 40, and performs matching using the extracted line image. It is an algorithm to do.

As is well known in the art, the image pattern matching method uses an image corresponding to the fingerprint of the person registered in the first fingerprint matching template TF1 and a line obtained by the first fingerprint matching module MF1 at the time of matching. By directly comparing the image and calculating the similarity, it is determined whether or not it is the fingerprint of the user.

On the other hand, the matching algorithm of the minutia method is an algorithm that extracts minutiae (feature points) from the fingerprint image obtained by the imaging device 40 and collates using minutiae information such as the position, type, and direction of the minutiae.

As well known in the art, the minutia method is also well known in the art, the minutia information corresponding to the fingerprint of the person registered in the second fingerprint collation template TF2, and the minutiae obtained by the second fingerprint collation module MF2 at the time of collation. By comparing the information, it is determined whether or not it is the fingerprint of the person.

In this embodiment, two types of fingerprint matching methods (specifically, the image pattern matching method and the minutiae method) are OR-integrated (see α in FIG. 5), and matching is established by any one of the fingerprint matching methods. Then, it is determined that the fingerprint is the person's fingerprint (first physical characteristic).

<Venous verification>
FIG. 3B is a diagram illustrating the principle of transmitted light sensing realized by the multimodal authentication device 100.

The multimodal authentication device 100 performs transmitted light sensing using the first light source 30a and the imaging device 40 provided in the upper housing 10a, and obtains a finger vein image as shown in FIG. 4B. Specifically, a finger vein image in which the blood vessel portion BA is dark and the other portions are bright is obtained by absorbing near-infrared light by hemoglobin in the blood (see FIG. 4B).

The authentication algorithm of the image pattern matching method is an algorithm for extracting a blood vessel pattern by performing image processing on the finger vein image obtained by the imaging device 40 and collating an image representing the extracted blood vessel pattern.

As is well known in the art, the image pattern matching method includes an image representing the blood vessel pattern of the person (user) registered in the first finger vein matching template TV1 and the first finger vein matching module at the time of matching. By directly comparing the line image obtained by the MV1, it is determined whether or not the blood vessel pattern (vein pattern) of the user.

On the other hand, the collation algorithm of the feature amount method performs image processing on the finger vein image obtained by the imaging device 40, thereby thinning the blood vessel portion and then sequentially connecting the lines between the branch points from the branch points to the branch points. Are extracted, feature quantities such as branch point coordinates, length, branch angle at the branch point, etc. are extracted, and the extracted feature quantities are collated.

As is well known in the art, the feature amount method is also obtained by the second finger vein matching module MV2 at the time of matching with the feature amount of the person's blood vessel pattern registered in the second finger vein matching template TV2. It is determined whether or not the blood vessel pattern (vein pattern) of the person is by comparing the characteristic amount of the blood vessel pattern to be obtained.

The two types of finger vein matching methods described above (specifically, the image pattern matching method and the feature amount method) are also integrated by OR type integration (β in FIG. If the collation is established by any one of the finger vein collation methods, it is determined that the blood vessel pattern (vein pattern; second physical feature) of the person.

Finally, the fingerprint matching result and the blood vessel pattern matching result are AND-type integrated (see γ in FIG. 5), and only when it is determined that the person is the person in both fingerprint matching and vein matching. It is determined that the user (authenticated person) who acquired the image and the finger vein image is the person himself / herself.

As described above, when a plurality of collation methods are prepared for each physical feature (modality) and collation is established by any of the methods, it is determined that the physical feature (for example, fingerprint) belongs to the person ( (Refer to OR type integration indicated by α and β in FIG. 5). This makes it possible to reduce the probability of rejecting the person as compared with the conventional example in which the physical features are collated using one collation method.

In addition, only when it is determined that all physical characteristics (for example, fingerprints and finger veins) belong to the person in question, the attacker determines that the person to be authenticated is the person himself. Therefore, impersonation becomes difficult, and safety can be improved.

(2) Operation of Embodiment A user who is an authentication target puts a finger between the upper housing 10a and the lower housing 10b of the multimodal authentication device 100, and the belly of the finger is placed on the template TE of the lower housing 10b. Place (see FIG. 1). When the multimodal authentication device 100 detects that the user's finger belly is placed on the template TE, the multimodal authentication device 100 performs reflected light sensing using the second light source 30b and the imaging device 40 provided in the lower housing 10b. A fingerprint image as shown in 3A is acquired. On the other hand, transmitted light sensing is performed by the first light source 30a and the imaging device 40 provided in the upper housing 10a, and a finger vein image as shown in FIG. 3B is acquired. The imaging device 40 transmits the acquired fingerprint image and finger vein image to the computer 50. In this embodiment, it is assumed that an image is acquired by bringing a finger into contact with the sensor surface (template TE) (when a contact sensor is used), but the sensor is a non-contact sensor (that is, a non-contact sensor). ). When using a non-contact sensor, the template TE is not necessary. As a method for determining whether or not the finger is present at a predetermined position when there is no template TE, for example, a method in which images obtained successively are arranged in time series and compared can be employed.

<Fingerprint verification process>
FIG. 6 is a flowchart showing fingerprint collation processing using the multimodal authentication device 100.

When the fingerprint verification unit 510a of the computer 50 receives the fingerprint image of the user from the imaging device 40 (step S1a), the fingerprint output from the imaging device 40 using the first fingerprint verification module (image pattern matching method) MF1. It is determined (collated) whether or not the user's fingerprint corresponding to the image is that of the user (step S2a). When the fingerprint collation unit 510a determines that the fingerprint of the user who is the person to be authenticated is the fingerprint of the user (ie, collation is established) by fingerprint collation using the first fingerprint collation module MF1 (step S3a; YES) The process proceeds to step S6a without performing collation using the second fingerprint collation module (maneuver method) MF2.

As described above, when collation using the first fingerprint collation module MF1 is established, the collation using the second fingerprint collation module MF2 is not performed as follows. In the present embodiment, when the collation is established by any one of the fingerprint collation methods, the collation using the first fingerprint collation module MF1 is performed in order to determine that the user's fingerprint is the original person (see FIG. 5α). This is because there is no need for further verification if the above holds. By adopting such a configuration, the verification speed can be improved and the processing can be simplified.

On the other hand, when collation using the first fingerprint collation module MF1 is not established (step S3a; NO), the fingerprint image output from the imaging device 40 using the second fingerprint collation module (maneuver method) MF2 is used. It is judged (verified) whether or not the user's fingerprint corresponding to is the person's own (step S4a). When the fingerprint collation unit 510a determines that the fingerprint of the user who is the person to be authenticated is the fingerprint of the user (that is, collation is established) by fingerprint collation using the second fingerprint collation module MF2 (step S5a; YES) Then, the personal authentication means 530 is notified that the verification has been established (step S6a), and the process is terminated. On the other hand, when the collation using the second fingerprint collation module is not established (step S5a; NO), the fingerprint collation means 510a notifies the personal authentication means 530 that the collation is not established (step S7a). The process is terminated.

<Finger vein verification process>
FIG. 7 is a flowchart showing finger vein collation processing using the multimodal authentication device 100.

When the finger vein collating unit 510b of the computer 50 receives the finger vein image of the user from the imaging device 40 (step S1b), the finger vein matching unit 510b outputs the image from the imaging device 40 using the first finger vein matching module (image pattern matching method) MV1. It is determined (collated) whether or not the user's blood vessel pattern (vein pattern) corresponding to the finger vein image to be processed is the user's own (step S2b). When the vein verification using the first finger vein verification module MV1 determines that the blood vessel pattern of the user who is the person to be authenticated is the blood vessel pattern of the user (that is, the verification is established) S3b; YES), the process proceeds to step S6b without performing collation using the second finger vein collation module (feature quantity method) MV2.

As described above, when collation using the first finger vein collation module MV1 is established, the reason why collation using the second finger vein collation module MV2 is not performed is the same as in the case of the fingerprint collation described above. When the collation is established by the finger vein collation method, the user's blood vessel pattern is determined to be the person's own (see FIG. 5β), so that the collation using the first finger vein collation module MV1 is established. This is because there is no need for further verification.

On the other hand, when collation using the first finger vein collation module MV1 is not established (step S3b; NO), the second finger vein collation module (feature amount method) MV2 is used and output from the imaging device 40. It is determined (collated) whether or not the user's blood vessel pattern (vein pattern) corresponding to the finger vein image is that of the user (step S4b). When the finger vein matching unit 510b determines that the blood vessel pattern of the user who is the person to be authenticated is the blood vessel pattern of the user by the vein matching using the second finger vein matching module MV2 (that is, the matching is established) (step S1). S5b; YES), the personal authentication means 530 is notified that the verification has been established (step S6b), and the process is terminated. On the other hand, when the verification using the second finger vein verification module MV2 is not established (step S5b; NO), the finger vein verification unit 510b notifies the personal authentication unit 530 that the verification has not been established ( Step S7b), the process is terminated.

The personal authentication unit 530 AND-integrates the fingerprint verification result received from the fingerprint verification unit 510a and the finger vein verification result received from the finger vein verification unit 510b (see [gamma] in FIG. 5). Only when both of the above verification and finger vein verification are established, the user who is the person to be authenticated is determined to be the person himself / herself.

When the personal authentication is successful, the computer 50 executes processing when the personal authentication is successful, such as unlocking the door. On the other hand, when the personal authentication fails, the computer 50 executes a process when the personal authentication fails, for example, displaying a message indicating that the personal authentication failed on a display (not shown).

In addition, regarding the fingerprint collation process (see FIG. 6) and the vein collation process (see FIG. 7) described above, both processes may always be executed regardless of each collation result. When collation is not established in the collation process), the other collation process (in this case, vein collation process) may be omitted.

B. Modifications (1) In the above-described embodiment, two types of collation methods are prepared for each physical feature (modality), and collation is established for each physical feature using one of the collation methods (that is, the person to be authenticated) (If it is determined that the physical characteristics of the user are the physical characteristics of the user), collation with other collation methods is omitted (see FIGS. 6 and 7). It is good also as performing two types (all) collation systems about a physical feature.

Here, when all the collation methods (here, two types of collation methods are assumed), the theoretical values obtained by the respective collation methods are weighted, and based on the values obtained by integrating the weighted theoretical values. It may be determined whether or not the verification is established. Specifically, as shown in the following formula (A), theoretical values T1 and T2 representing the probability that the physical feature of the user who is the person to be authenticated is the physical feature of the user at the time of matching in each matching method are as follows. Ask for each. Then, the obtained theoretical values T1 and T2 are weighted, and when the value obtained by integrating the weighted theoretical values αT1 and (1-α) T2 is larger than the threshold value S, it is determined that the collation is established. The weight and threshold value can be set arbitrarily.

αT1 + (1-α) T2> S (A)
T1, T2; theoretical value α; weight (0 <α <1)
S: threshold

(2) In this embodiment, for convenience, the pattern matching method is used as the first fingerprint matching method, the minutia method is used as the second fingerprint matching method, the direct comparison method is used as the first finger vein matching method, and the second finger vein is used. Although the feature amount method is exemplified as the matching method, it is needless to say that other matching methods (for example, a frequency analysis method in the case of the fingerprint matching method) may be adopted. In this embodiment, the first fingerprint verification method (or the first finger vein verification method) and the second fingerprint verification method (or the second finger vein verification method) are executed in this order. It is possible to arbitrarily set whether or not to execute. For example, a collation method with a high collation speed may be executed first, or a collation method with high collation accuracy may be executed first.

(3) Furthermore, in this embodiment, two types of collation methods are prepared for each of the fingerprint and the finger vein, but three or more types may be used. Furthermore, any one of physical features (for example, fingerprints) may be adopted by using a plurality of types (ie, two or more types) of collation methods, and any other physical features (for example, finger veins) may be selected. The verification method may be adopted. Of course, the physical features used for collation are not limited to two types, and may be three or more types. In this embodiment, reflected light sensing is used for fingerprint matching and transmitted light sensing is used for finger vein matching. Conversely, transmitted light sensing is used for fingerprint matching and finger vein matching is performed. Reflected light sensing may be used, and how it is combined can be arbitrarily set.

In addition, the steps of each process shown in the present embodiment can be executed in any order or in parallel within a range where no contradiction occurs in the processing contents. Furthermore, in this specification and the like, the term “means” does not simply mean a physical means, but includes a case where the functions of the means are realized by software. Furthermore, the function of one means may be realized by two or more physical means, or the functions of two or more means may be realized by one physical means. In addition, the software according to the present invention can be installed in a computer by downloading it through various recording media such as an optical disk such as a CD-ROM and a DVD-ROM, a magnetic disk, and a semiconductor memory, or via a communication network. Can be loaded.

This application claims priority based on Japanese Patent Application No. 2009-262167 filed on November 17, 2009, the entire disclosure of which is incorporated herein.

The present invention has been described above with reference to the embodiments, but the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

The multimodal authentication device according to the present invention is suitable for improving both safety and convenience in personal authentication using two or more types of physical features.

DESCRIPTION OF SYMBOLS 100 ... Multimodal authentication apparatus, 40 ... Imaging device, 50 ... Computer, 510a ... Fingerprint collation means, 520a ... Fingerprint collation database, 510b ... Finger vein collation means, 520b ... Finger vein collation database, 530 ... Personal authentication means, MF1 ... First fingerprint verification module, MF2 ... second fingerprint verification module, TF1 ... first fingerprint verification template, TF2 ... second fingerprint verification template, MV1 ... first finger vein verification module, MV2 ... second finger Vein verification module, TV1... First finger vein verification template, TV2.

Claims

A first verification unit that determines whether or not the first physical feature of the input person to be authenticated belongs to an authorized user by applying a plurality of types of verification algorithms;
A second verification unit that determines whether or not the second physical feature of the authentication target person is an authorized user by applying one or more verification algorithms;
Authentication means for determining that the person to be authenticated is a regular user when the first and second matching means are matched;
The first collating means is configured such that, when collation is established in collation to which at least one of the plural collation algorithms is applied, the first physical feature of the user input is normal Determine that it belongs to the user,
Multimodal authentication device.
The first matching means is applied in order from the fastest matching speed among the plurality of types of matching algorithms, stops the matching operation at the time when one matching is established, The multimodal authentication device according to claim 2, wherein the one physical feature is determined to be that of an authorized user.
The multimodal authentication device according to claim 1 or 2, further comprising input means for inputting the first and second physical characteristics of the person to be authenticated from the same physical part.
The multimodal authentication device according to claim 3, wherein the input means inputs from the same physical part using the same sensor.
The second collating unit determines whether the user is a legitimate user by applying a plurality of collating algorithms,
The second verification unit is configured to input the second physical feature of the person to be authenticated that is input when verification is established in verification using at least one of the plurality of verification algorithms. The multimodal authentication device according to any one of claims 1 to 4, wherein the multi-modal authentication device determines that is an authorized user.
A first verification unit that determines whether or not the first physical feature of the input person to be authenticated belongs to an authorized user by applying a plurality of types of verification algorithms;
A second verification unit that determines whether or not the second physical feature of the authentication target person is an authorized user by applying one or more verification algorithms;
Authentication means for determining that the person to be authenticated is a regular user when the first and second matching means are matched;
The first matching means weights each theoretical value representing whether or not a regular user is obtained by combining the plurality of types of matching algorithms and applying each matching algorithm. When it is determined that collation is established based on the result of integrating theoretical values, it is determined that the first physical feature of the authentication target person to be input is that of a regular user.
Multimodal authentication device.
A verification step for determining whether or not the first physical feature of the authentication target person is an authorized user by applying a plurality of types of verification algorithms, wherein the plurality of types of verification algorithms A first collation step of determining that the first physical feature of the input user is that of a regular user when collation is established in collation using at least one of the collation algorithms;
A second collation step for determining whether or not the second physical feature of the input person to be authenticated is that of an authorized user by applying one or more collation algorithms;
An authentication step of determining that the person to be authenticated is a regular user when the verification is established in the first and second verification steps;
Including multi-modal authentication methods.