KR102139548B1 - System and method for decentralized identifier based on face recognition - Google Patents

Abstract

The present invention discloses a decentralized identification system. More specifically, the decentralized identity verification system and method based on facial recognition technology, not controlled by a central system unlike the existing identification method, and allows individuals to have complete control over information of their own. According to an embodiment of the present invention, by using an artificial intelligence algorithm based on deep learning, the false recognition rate of the existing facial recognition technology is reduced to within 1% to 3%. Therefore, the facial recognition can be used as a value as a main means, not as an auxiliary means of biometrics.

Description

Decentralized identification system and method based on facial recognition technology {SYSTEM AND METHOD FOR DECENTRALIZED IDENTIFIER BASED ON FACE RECOGNITION}

The present invention relates to a decentralized identification system, and in particular, to a decentralized identification system and method based on facial recognition technology that is not controlled by a central system unlike an existing identification method, and allows individuals to have full control over their information. It is about.

With the development of electronic information technology, people's lifestyle is changing. As portable high-performance mobile terminals such as smartphones become the center of modern people's lifestyles, the use of financial applications through smartphones is becoming frequent, and users who easily authenticate themselves through biometric authentication are gradually increasing.

Accordingly, when using a mobile service such as shopping and finance using a smartphone, the perception that people want to be safe from the threat of information leakage due to hacking of personal information related to the user and other personal information is increasing. Security is becoming more and more important.

Among data security technologies, in particular, interest in authentication solutions based on various biometric technologies that the user himself has control over is increasing, and known biometric technologies include fingerprint, iris, and facial recognition.

Among them, the face recognition technology refers to a technology that extracts the feature points of a face and compares it with data in a stored database to identify the user. Because the user's skin does not directly contact the authentication device, It is more hygienic and convenient, so it is expected to be used in various fields.

However, at the current level of technology, the face recognition technology tends to have a higher false acceptance rate than other biometric technologies such as fingerprint recognition, and has been used as an auxiliary means rather than an identity verification method. In order to solve the misrecognition rate of facial recognition technology, it can be said that continuous research is needed to increase accuracy using artificial intelligence (AI) technologies such as deep learning.

On the other hand, decentralized identity (DID) technology is an electronic identity verification technology that stores personal information on a user's terminal and selects and submits only necessary information when authenticating personal information. This is a feature that enables individuals to self-sovereign identity that can verify their identity without going through a centralized institution by directly managing their data by using a blockchain network.

Therefore, it is expected to increase the accuracy of the above-described face recognition technology and, at the same time, apply the face recognition technology to the decentralized identity authentication platform, to implement a more personalized and safe identity authentication solution.

Published Patent Publication No. 10-2019-0127676 (published on: November 13, 2019)

The present invention relates to an identity authentication system based on facial recognition technology, and there is a problem in solving a low false recognition rate of facial recognition technology compared to conventional biometrics using artificial intelligence technology.

In addition, the present invention has another challenge to provide a platform with improved security through more convenient and reliable identification of users through self-sovereign identity authentication based on blockchain.

In addition, the present invention is to spread the popularization of biometric authentication technology by expanding the scope of the fingerprint authentication method, which is frequently used in the current online authentication technology, FIDO (Fast Identity Online), so that users can utilize a variety of biometric authentication data. There is another challenge.

In order to solve the above-described problem, a distributed identification system based on facial recognition technology according to an embodiment of the present invention receives a face image of one or more users' faces, and a learning model derived through a plurality of learning data Apply to the face image to generate the user's authentication information, a learning server to generate a user's face in real-time when verifying the identity, one or more user terminals performing local authentication using the authentication information and user-generated FIDO information By registering on the blockchain network, it may include an identity authentication server that performs identity authentication for a locally authenticated user terminal.

The learning server is a collection unit that collects a facial image for a user photographed by a camera device from a facial registration terminal, and a facial, emotional, and gender-age recognition model for an analysis target using a plurality of learning data for the facial image Among them, a learning unit generating a case specific value for a face by performing a neural network-based machine learning procedure for one or more may be included, and an authentication information generating unit generating authentication information according to the generated case measurement value for the user.

The learning unit comprises: a similar distance calculating unit calculating a similar distance to the learning data prepared in advance in units of frames of the facial image, a facial data extracting unit extracting facial data by applying a HOG algorithm to the facial image, from the facial data The user detects the direction of the face by tracking the major landmarks of the face and uses the feature point extractor and the converted face image to convert the face image so that the face's eyes and mouth correspond to the center of the screen. It may include an analysis unit for generating a plurality of case measurements for the face.

The plurality of learning data may include data analyzing the degree of deformation of the pattern according to the shape of the subject by emitting infrared dots of a plurality of specific patterns on the subject by the ToF sensor.

The learning unit may apply facial data according to a plurality of facial angles of a subject whose direction is changed on the X, Y, and Z axes in the learning data to the learning model.

The identification server includes a network participant accessing the blockchain network as a peer, a FIDO processing unit for requesting to record the generated FIDO information to the blockchain network, and an identification result for a user registered with the FIDO information from the outside City, it may include a server storage for storing the information required for the identification information excluding the ID and the registered user ID and the FIDO information, and a query unit for returning the identification result using the FOD information stored in the blockchain network.

In addition, in order to solve the above-described problem, a decentralized identification method based on face recognition technology according to an embodiment of another aspect of the present invention comprises: acquiring a face image of a user who wants to register FIDO information by a face recognition terminal; A learning server collects the facial image from the facial recognition terminal, and analyzes the collected facial image using a prepared learning model to generate authentication information, and the user terminal receives and stores authentication information from the learning server. , The user terminal recognizes the user's face and performs user local authentication through authentication information, as the user terminal normally completes user local authentication, generating a public key pair, and the user terminal verifying the identity Requesting FIDO registration by providing a public key to a server, recording the FIDO information including the public key for the user in the blockchain network by the identity authentication server, and registering the identity document to the user terminal by the identity authentication server And returning completion.

The learning server collecting the facial image from the facial recognition terminal, and analyzing the collected facial image using the prepared learning model to generate authentication information, the facial image in frame unit with the pre-prepared learning data Calculating a similar distance, extracting face data by applying a HOG algorithm to the facial image, tracking a major landmark of the face from the face data, determining a face direction, and determining the face direction according to the determined face direction And converting the facial image so that the mouth corresponds to the center of the screen and generating a plurality of case measurements for the user's face using the converted facial image.

In the step of calculating a similar distance to the learning data prepared in advance in units of frames of the facial image, in the learning data, facial data according to a plurality of facial angles of a subject whose direction is changed on the X, Y, and Z axes is fed to a learning model. And applying.

After the ID server returns the FIDO registration completion to the user terminal, the user terminal recognizes the user's face and performs user local authentication through authentication information, and the user terminal is authenticated to the ID server Requesting proof, the identification server checks the FIDO list stored in the blockchain network to check whether the user is registered, the identification server checks the FIDO information of the user, and tokens to the user terminal Requesting FIDO authentication by sending, the step of the user terminal signing the token through the private key stored in the storage, the step of the user terminal sending the signed token to the identity authentication server to return the FIDO authentication , The step of verifying the signature by using the public key included in the FIDO information for the user, and when the normal authentication is completed, the step of reporting the FIDO authentication success to the user terminal by the identity server. can do.

According to an embodiment of the present invention, by using a deep learning-based artificial intelligence algorithm, the recognition rate of the existing facial recognition technology is reduced to within 1% to 3%, so that facial recognition is a value as a primary means rather than a secondary means of biometric recognition. It has an effect to be used as.

In addition, the present invention has the effect of securing the reliability based on the integrity and transparency of the user's authentication data by using a blockchain network based on a private blockchain.

In addition, by using Hyperledger Indy, it is possible to establish and utilize a DID authentication system that conforms to the W3C standard, which is an international certification authority, thereby providing an identification solution with a wide range of applications.

1 is a diagram schematically showing the overall structure of a distributed identification system based on facial recognition technology according to an embodiment of the present invention.
FIG. 2 is a diagram showing the components participating in the authentication information registration procedure in the distributed identification system based on the facial recognition technology according to an embodiment of the present invention.
3 is a view for explaining a facial recognition technology of a distributed identification system based on the facial recognition technology according to an embodiment of the present invention.
4 is a diagram illustrating the configuration of a learning server of a distributed identification system based on facial recognition technology according to an embodiment of the present invention.
5 is a diagram illustrating the configuration of an identity authentication server of a distributed identity authentication system based on facial recognition technology according to an embodiment of the present invention.
6 is a view showing a procedure for registering authentication information in a distributed identification method based on facial recognition technology according to an embodiment of the present invention.
7 is a view showing an identification method in a distributed identification method based on facial recognition technology according to an embodiment of the present invention.

When a certain part of the specification refers to a "equipment" or "includes" a certain component, it means that the component may further include other components, not to exclude other components, unless otherwise specified. In addition, terms such as "... unit", "... terminal" and "... system" described in the specification refer to a unit that processes one or more functions or operations. Means, it may be implemented in hardware, software, or a combination of hardware and software.

Also, in this specification, the term "an embodiment" means to serve as an example, example, or illustration, but the subject matter of the invention is not limited by such examples. In addition, “comprising”, “having”, “having” and other similar terms are used, but when used in the claims “includes” as an open transition word that does not exclude any additional or other components. "Comprising" can be used generically in a similar way.

The various techniques described herein may be implemented with hardware or software, or combinations of all of them where appropriate. Terms such as "...Unit", "...Terminal" and "...System" as used herein are likewise computer related entities, i.e. hardware , It can be treated as a combination of hardware and software, software, or software at runtime. In addition, each function executed in the program of the present invention may be configured in a module unit, or may be recorded in one physical memory or may be distributed and recorded between two or more memories and recording media.

In the following description, the term “distributed identity system based on facial recognition technology” of the present invention may be abbreviated as “identity system” or “system” for convenience of description.

Hereinafter, a distributed identification system and method based on facial recognition technology according to an embodiment of the present invention will be described in detail with reference to the drawings.

1 is a diagram schematically showing the overall structure of a distributed identification system based on facial recognition technology according to an embodiment of the present invention.

Referring to FIG. 1, a distributed identification system based on facial recognition technology according to an embodiment of the present invention receives a face image of one or more users' faces and faces a learning model derived through a plurality of learning data A learning server (100, one or more user terminals 200 performing local authentication using authentication information by capturing a user's face in real time during identification) by applying the image to the learning server (100) and FIDO information for the user It may include an identity authentication server 300 to generate and register on the blockchain network 400 and perform identity authentication for the locally authenticated user terminal 200.

The learning server 100 serves to improve the accuracy of the biometric information about the user through machine learning. In the system according to an embodiment of the present invention, in using biometric information to implement a decentralized identification service, it is characterized by using Face Recognition rather than body recognition such as fingerprint and iris. It is characterized by using a deep learning technique to increase recognition accuracy and lower the recognition rate.

In detail, the learning server 100 acquires a facial image of the user's face at various angles when registering the user's first FIDO information, calculates a similar distance by comparing with simple face images prepared in advance for each frame, and calculates the facial distance. Extracts face data corresponding to the face part and derives case measurement values that reflect the characteristics of the face using a learning model, and generates authentication information that recognizes the user's facial recognition, emotional recognition, and gender-age based on this. It can be provided to the user terminal 200.

The user terminal 200 is a terminal possessed by a user who wants to distribute and manage authentication information through a blockchain network through the system of the present invention and perform identity authentication when necessary. When the first system is registered, the user terminal 200 is faced through the learning server 100. It generates and stores authentication information for recognition, and then, when a user's identity is required, it is possible to perform identity authentication using FIDO information stored in the blockchain network through a local authentication procedure.

The above-described user terminal 200 includes PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication) )-2000, Code Division Multiple Access (CDMA)-2000, W-Code Division Multiple Access (W-CDMA), Wireless Broadband Internet (Wibro) terminal, smartphone, smart pad, tablet PC (Tablet) PC), such as all kinds of handheld (Handheld)-based wireless communication device and a stationary PC, a computing device such as a laptop can be used.

The identity authentication server 300 serves to register FIDO information by a plurality of user terminals 200 through an information communication network and to process the identity authentication if necessary, and in particular, the blockchain network (FIDO information registered by the user terminal 200) 400) is characterized by storage and management.

In particular, the identity authentication server 300 may participate as a single peer in the blockchain network 400, sign using a private key in the user terminal 200, and FIDO stored in the blockchain network 400 The signed information can be verified by using the public key included in the information to process the user's identification.

In addition, the identity authentication server 300 may provide an inquiry function stored in the blockchain network 400 at the request of the user terminal 200.

Blockchain network 400 consists of a group by a trusted organization, one or more consortiums with a small number of nodes that want to participate in the channel, with administrator privileges to configure the fabric network and change network settings, to the group. Another group creates a peer by sending a join request to obtain access authority.

In addition, a group to which the above-described identity verification server 300 belongs to establishes a guarantee policy, which is an agreement algorithm of the block chain network 400 hyperledger fabric according to an embodiment of the present invention. The policy can be set to'total peer consent'.

In particular, the blockchain network 400 according to an embodiment of the present invention registers, authenticates, and cancels FIDO information in a block by executing a smart contract at the request of peers, including the peer 410 of the identity verification server 300 In addition, it is possible to provide a search function for registered personal information and other information.

In detail, the peer 410 of the blockchain network 400 may include a blockchain storage in which a plurality of blocks in which FIDO information is recorded as a general node is stored and linked, and the same blocks are stored in all blockchain storages. As they are linked and stored together, they form a blockchain.

The blockchain network 400 may include nodes such as CA 420 and orderer 430 in addition to peers. The CA (Certificate Authority) 420 is a public key of a participant in the blockchain network 400, and when the participant wants to authenticate himself/herself to another person, the other person confirms the public key of the participant through the CA. It plays a role of blocking the attack method (MITM; Man in the Middle attack) disguised as a hacker, etc. as an authorized participant.

In addition, the orderer 430 collects and sorts the transactions requested by the blockchain network 400 and generates a latest block for information recording.

In addition, the blockchain network 400 enables identity verification of users belonging to each other through channel connection between consortiums according to the operator. For example, when a user purchases a product in a shopping mall belonging to a shopping mall consortium, the user does not perform a separate identification procedure requested by the shopping mall, but rather performs FIDO information of the blockchain network 400 of the registered financial institution consortium associated with the shopping mall. It can be used to perform the identity verification process of the shopping mall.

According to the above-described structure, the distributed identity authentication system based on the facial recognition technology according to the embodiment of the present invention, the user who wants to use the identity authentication service registers FIDO information in the blockchain network operated by the system, and thereafter When identification is required, local authentication based on facial recognition through one's own terminal may be performed based on facial recognition technology, and identification may be performed using FIDO information.

Hereinafter, a procedure for registering authentication information of a distributed identity authentication system based on facial recognition technology according to an embodiment of the present invention will be described with reference to the drawings.

FIG. 2 is a diagram showing the components participating in the authentication information registration procedure in the distributed identification system based on the facial recognition technology according to an embodiment of the present invention.

Referring to FIG. 2, a user who intends to use a distributed identification system based on facial recognition technology according to an embodiment of the present invention must perform initial authentication information generation and FIDO information registration procedures upon registration in the system, and the present invention It may be performed by the learning server 100 and the terminal associated with it.

First, a user who wishes to use the identification service of the present invention, photographs his or her face through the face registration terminal 70 to which the camera device 30 capable of photographing the face is connected, or is mounted on the user terminal 200 It is necessary to take a picture of one's face through the camera and provide it to the learning server 100 through the information communication network.

Preferably, according to an embodiment of the present invention, a camera device 30 and a camera device 30 capable of emitting 30,000 or more specific patterns of infrared dots to a subject are mounted with a Time of Flight (ToF) sensor. It is preferable to use a dedicated face registration terminal 70 to control. In particular, in the case of using the above-described camera device 30, more accurate facial analysis is possible as the modified information of the pattern can be analyzed according to the shape of the subject surface.

According to this structure, the face registration terminal 70 may control the camera device 30 to capture the user's face and generate a face image.

In particular, referring to FIG. 3, the camera device 30 may obtain facial data for facial recognition from various angles by dividing the user's facial angle into a plurality of faces on the X, Y, and Z axes with respect to the user's face as a subject. To make.

As an example, a face image for each angle may be obtained when the user switches the direction of looking at the left ↔ right direction, the upper ↔ lower direction, the left upper ↔ lower right direction and the upper right ↔ lower direction based on the front.

Referring back to FIG. 2, when using the user terminal 200 owned by a user other than a dedicated device in the registration procedure, the functions of the above-described face registration terminal are implemented in the same way as a mobile application, and the user terminal 200 Can be installed and run on.

The face registration terminal 70 may transmit one or more face images taken from the camera device 30 to the learning server 100 through the information communication network.

The learning server 100 may generate authentication information for facial recognition by applying the received one or more facial images to a previously prepared learning model, and transmit the authentication information for the facial recognition to the user terminal 200 of the corresponding user in the user terminal 200 Make it available for local authentication.

That is, in the present invention, the biometric information is not stored in the learning server or the identity authentication server, but is stored only in the user terminal 200 so that the user can directly manage his/her biometric authentication information.

Hereinafter, a learning server of a distributed identification system based on facial recognition technology according to an embodiment of the present invention will be described in detail with reference to the drawings.

4 is a diagram illustrating the configuration of a learning server of a distributed identification system based on facial recognition technology according to an embodiment of the present invention.

Referring to Figure 4, the learning server 100 according to an embodiment of the present invention, a collection unit 110 for collecting a facial image for a user photographed by the camera device from the facial registration terminal, a plurality of facial images, Generated by the learning unit 120 and the user generating case-specific values for the face by performing a neural network-based machine learning procedure for one or more of the facial, emotional, and gender-age recognition models for the analysis target using the learning data It may include an authentication information generating unit 130 for generating authentication information according to the measured case value.

The collection unit 110 may collect one or more facial images of the user photographed by the camera device from the facial registration terminal, or may collect facial images photographed by the user terminal.

The image storage 115 may store a plurality of training data required for machine learning. The learning data may be photographs of faces of people photographed at various angles.

The learning unit 120 performs a neural network-based machine learning procedure for one or more of the facial, emotional, and gender-age recognition models for an analysis target using a plurality of learning data stored in the image storage 115 for the facial image. By creating a case-specific value for the face.

In order to implement this function, the learning unit 120 is based on a deep learning (deep learning) technology to analyze the image similar distance calculation unit 121, face data extraction unit 123, feature point extraction unit 125 And an analysis unit 127.

In detail, the similar distance calculating unit 121 may calculate the similar distance between the training data registered in the image storage 115 in units of video frames for the user's face image. The similar distance calculation is a procedure for quantifying the similarity between images, and a commercialized program, for example, a TensorFlow library may be used.

The face data extractor 123 may extract face data, which is a set of face parts, by applying a Histograms of Oriented Gradients (HOG) algorithm to generalize, or simplify, the faces appearing in the face image.

The feature point extraction unit 125 may extract a plurality of feature points capable of representing features of the face in each region included in the extracted face data.

Here, the plurality of feature points described above may be referred to as a landmark that can most clearly represent the features of the face in the face data, and the feature point extracting unit 125 tracks the main landmark in the face data to determine the face direction. I can judge.

Then, the feature point extracting unit 125 may convert the facial image so that the eyes and mouth of the face correspond to the center of the screen according to the determined face direction.

The analysis unit 127 may calculate and store a plurality of case measurements of the user's face using the converted face image. The calculated case measurement value can be used to generate authentication information that enables the user terminal to authenticate that the user is a registered user, that is, a user whose face is currently detected.

The authentication information generation unit 130 may generate authentication information according to the generated case measurement value and provide it to the user terminal. Here, as the authentication information is not stored in the learning server 100 and the identity authentication server, the risk of leakage of personal information such as a face is reduced.

According to the above-described structure, the learning server according to the embodiment of the present invention may analyze the facial image using the machine learning technique for the user, and generate authentication information corresponding to the user's face.

Hereinafter, an identification server according to an embodiment of the present invention will be described in detail with reference to the drawings.

5 is a diagram illustrating the configuration of an identity authentication server of a distributed identity authentication system based on facial recognition technology according to an embodiment of the present invention.

Referring to FIG. 5, the identification server 300 according to an embodiment of the present invention includes a network participant 310 accessing as a peer of the blockchain network 400, and the generated FIDO information in the blockchain network ( 400) FIDO processing unit 320 for requesting a record, and when an ID verification result for a user registered with FIDO information from the outside, an inquiry unit 330 that returns the identification result using FOD information stored in the blockchain network 400 ) And server storage 340.

The network participant 310 allows the identity authentication server 300 of the present invention to participate as a peer in the blockchain network 400 to which the present invention belongs, and requests and performs a smart contract for the identity authentication service to the blockchain network 400. Make it possible.

The FIDO processing unit 320 may register FIDO information to the blockchain network 400 at the request of the user terminal, or provide an identity verification service of a registered user using the FIDO information.

When the registration of FIDO information is requested from the user terminal, the FIDO processing unit 320 requests the smart contract for recording FIDO information to the blockchain network 400 by receiving the public key and generating the FIDO information, and the blockchain network ( In accordance with the agreement of the participants of 400), the verification of FIDO information and the distributed recording procedure into blocks may be performed.

In addition, when the authentication using the FIDO information is requested from the user terminal, the FIDO processing unit 320 identifies the user ID, retrieves a list of FIDO information stored in the blockchain network 400, extracts the FIDO information, and generates a temporary token. And a user terminal. Then, when the decryption is normally performed by decrypting the token encrypted with the private key from the user terminal using the public key included in the FIDO information, it is possible to perform identification authentication as a registered user and return the result to the user terminal.

The inquiry unit 330 may provide a inquiry function such as whether to register a user ID and FIDO information according to a request of the user terminal. The inquiry unit 330 requests a smart contract for this to the blockchain network 400 when FIDO information inquiry is requested from the user terminal, receives the FIDO information distributed in the block from the blockchain network 400, and receives the user terminal Can be delivered to.

The server storage 340 is a participant of the blockchain network 400 and may store information required for identification of the user except for the block-registered user ID and the FIDO information.

In an embodiment of the present invention, the identification server 300 does not store the FIDO information of the registered user in a storage, but is the minimum identification means required to perform the identification service, such as a basic user ID. It is characterized in that only the information is stored and the FIDO information for performing the actual identification procedure is distributed and stored in the blockchain network 400.

Accordingly, the server storage 340 may store basic information such as a registered user ID and setting values for operating the identity authentication server.

Hereinafter, the authentication information registration procedure will be described in detail in the distributed identification method based on the facial recognition technology according to an embodiment of the present invention with reference to the drawings.

6 is a view showing a procedure for registering authentication information in a distributed identification method based on facial recognition technology according to an embodiment of the present invention.

To use the identity verification service of the present invention, the user must firstly store authentication information for facial recognition in the user terminal 200 such as his or her smartphone, and for this purpose, the face registration terminal 70 equipped with a camera device Alternatively, it is necessary to acquire a face image photographed using the user's user terminal 200. In the following description, a method of registering authentication information using the face registration terminal 70 will be described.

Referring to FIG. 6, the face registration terminal 70 obtains one or more face images of the user's face through the connected camera device (S100), and transmits it to the learning server 100 through the information communication network to request analysis It will be (S110).

In addition, the learning server 100 may collect facial images to be analyzed from one or more facial registration terminals 70 and apply the collected facial images to a learning model derived through previously prepared learning data to perform image analysis. (S120).

As a result of the analysis, the learning server 100 can derive 128 case measurement values, generate and provide authentication information including the same to the user terminal 200 (S130), and accordingly the user terminal 200 receives The authentication information is stored in the storage in the terminal (S140).

Next, the user terminal 200 may perform a procedure for first registering FIDO information according to the face recognition method to the identity authentication server 300, and for this purpose, the user may recognize his/her face to the user terminal 200 User local authentication is performed through a comparison with the stored authentication information (S150).

Next, when the user local authentication is completed, the user terminal 200 generates a public key pair including a private key and a public key (S160), stores the double private key in a storage, and the public key is an identity authentication server 300 ) To request registration of FIDO information (S170). At this time, information for user registration, such as a user ID, may also be transmitted.

Accordingly, the identity authentication server 300 may register a user ID, generate FIDO information including a public key, and record it on the blockchain network for each user (S180). Particularly, step S180 described above in the present invention is performed by execution of a smart contract defined in the blockchain network, and the identity verification server 300 participates as a peer of the blockchain network to perform procedures such as transaction request and verification. After that, FIDO information is distributed and recorded in blocks.

Next, if the FIDO information for the user is normally stored, the identity authentication server 300 may return the registration of the FIDO information to the user terminal 200 (S190), and then the user faces the user terminal 200 Identification service can be used through recognition.

Hereinafter, an identification procedure will be described in detail in a distributed identification method based on facial recognition technology according to an embodiment of the present invention with reference to the drawings.

7 is a view showing an identification method in a distributed identification method based on facial recognition technology according to an embodiment of the present invention.

Referring to FIG. 7, in the distributed identification method based on the facial recognition technology of the present invention, when the user needs to authenticate himself/herself when using various online services such as finance and shopping, FIDO information registered in the system of the present invention may be used. .

First, the user photographs his or her face using the user terminal 200, and accordingly, the user terminal 200 performs user local authentication according to whether or not it matches through authentication information stored in the storage (S200).

Next, when the facial recognition is normally completed, the user terminal 200 requests the identity authentication server 300 to authenticate the user (S210), and thus the identity authentication server 300 identifies the user through the user ID. Then, a list is searched to see if FIDO information registered in the blockchain network exists (S220).

Next, in the case of a registered user, the identity authentication server 300 transmits a temporary token to the user terminal 200 to request FIDO authentication (S230), whereby the user terminal 200 tokens through a private key stored in the storage. FIDO authentication is returned by signing and replying to (S250).

Next, the identity authentication server 300 decrypts the signed token using the public key of the user included in the FIDO information to verify the signature (S260).

Subsequently, the identity authentication server 300 reports to the user terminal 200 that the FIDO authentication is successful (S270), so that the user whose identity authentication is completed can use a service such as financial transaction or product purchase.

Although many matters are specifically described in the above description, this should be interpreted as an example of a preferred embodiment rather than limiting the scope of the invention. Therefore, the invention should not be determined by the described embodiments, but should be determined by equivalents to the claims and claims.

30: camera device 70: face registration terminal
100: learning server 110: collection unit
115: image storage 120: learning department
121: similar distance calculation unit 123: face data extraction unit
125: feature point extraction unit 127: analysis unit
130: authentication information generation unit 200: user terminal
300: Identity authentication server 310: Network participant
320: FIDO processing unit 330: inquiry unit
340: server storage 400: blockchain network
410: peer 420: CA (Certificate Authority)
430: orderer

Claims (10)

A learning server that receives face images of one or more users' faces and applies a learning model derived from a plurality of learning data to the face images to generate user authentication information;
One or more user terminals that perform a local authentication using the authentication information by photographing a user's face in real time during identification; And
Includes an identity verification server that generates FIDO information for the user and registers it on the blockchain network, and performs identity verification for the locally authenticated user terminal,
The learning server,
A collection unit for collecting a facial image of the user photographed by the camera device from the facial registration terminal;
A learning unit that generates a case-specific value for a face by performing a neural network-based machine learning procedure for one or more of the facial, emotional, and gender-age recognition models for an analysis object using a plurality of learning data on the facial image ; And
It includes an authentication information generating unit for generating authentication information according to the case measurement value generated for the user,
The learning unit,
A similar distance calculating unit for calculating the similar distance to the learning data prepared in advance in units of frames of the face image;
A face data extraction unit extracting face data by applying a HOG algorithm to the face image;
A feature point extracting unit that tracks a main landmark of the face from the face data to determine a face direction, and converts the facial image such that the face's eyes and mouth correspond to the center of the screen according to the determined face direction; And
An analysis unit that generates a plurality of case measurements for the user's face using the converted face image
Decentralized identification system based on facial recognition technology comprising a. delete delete According to claim 1,
The plurality of learning data,
Data that analyzes the degree of deformation of a pattern according to the surface shape of a subject by emitting infrared dots of a plurality of specific patterns on the subject by a ToF sensor
Decentralized identification system based on facial recognition technology comprising a. According to claim 1,
The learning unit,
A distributed identification system based on facial recognition technology that applies facial data according to a plurality of facial angles of a subject whose direction is changed on the X, Y, and Z axes in the training data to a learning model. According to claim 1,
The identity server,
A network participant that accesses the blockchain network as a peer;
A FIDO processing unit for requesting to record the generated FIDO information to the blockchain network;
An inquiry unit that returns an identification result using FOD information stored in the blockchain network when an identification result of a user registered with the FIDO information is external; And
Server storage to store the information required for the identity authentication except the registered user ID and the FIDO information
Decentralized identification system based on facial recognition technology comprising a. In the decentralized identification method based on the facial recognition technology by the system according to claim 1,
Acquiring a facial image of the user who wants to register the FIDO information by the facial recognition terminal;
A learning server collecting the facial images from the facial recognition terminal, and analyzing the collected facial images using a prepared learning model to generate authentication information;
A user terminal receiving and storing authentication information from the learning server;
The user terminal recognizing the user's face and performing user local authentication through authentication information;
Generating a public key pair as the user terminal normally completes user local authentication;
The user terminal requesting FIDO registration by providing a public key to the identity server;
Recording, by the identity verification server, FIDO information including a public key for the user in a blockchain network; And
And the step of returning the completion of FIDO registration to the user terminal by the identification server,
The learning server collecting the facial image from the facial recognition terminal, and analyzing the collected facial image using the prepared learning model, generating authentication information,
Calculating a similar distance from the face image to the learning data prepared in advance in units of frames;
Extracting facial data by applying a HOG algorithm to the facial image;
Determining a face direction by tracking a main landmark of the face in the face data, and converting the facial image so that the face's eyes and mouth correspond to the center of the screen according to the determined face direction; And
Generating a plurality of case measurements for the user's face using the converted face image
Distributed identification method based on facial recognition technology comprising a. delete The method of claim 7,
In the step of calculating the similar distance to the learning data prepared in advance by the frame unit in the face image,
Applying facial data according to a plurality of facial angles of a subject whose direction is changed on the X, Y, and Z axes in the learning data to the learning model
Distributed identification method based on facial recognition technology comprising a. The method of claim 7,
After the ID server returns the FIDO registration completion to the user terminal,
The user terminal recognizing the user's face and performing user local authentication through authentication information;
Requesting an identification certificate from the user terminal by the user terminal;
Confirming whether the user is registered by querying the FIDO list stored in the blockchain network by the identification server;
Checking, by the identity verification server, FIDO information of the corresponding user and sending a token to the user terminal to request FIDO authentication;
The user terminal signing a token through a private key stored in the storage;
The user terminal sending a signed token to the identification server to return a FIDO authentication;
Verifying a signature using the public key included in the FIDO information for the user by the identity verification server; And
When normal authentication is completed, the identity authentication server reports the success of FIDO authentication to the user terminal.
Distributed identification method based on facial recognition technology comprising a.

Images