Identity verification system, identity verification method, and identity verification program

Abstract

We provide an identity verification system that determines whether the person using the avatar on a designated service is the same person as the actual user. [Solution] The personal authentication system comprises: a personal feature point information generation unit 2 that extracts multiple feature points from a personal image and generates personal feature point information regarding the positional relationships between the feature points of the personal image; a target image input unit 12 that inputs a target image, which is an image of the person to be authenticated during the authentication process; a target feature point information generation unit 13 that extracts multiple feature points from a target image and generates target feature point information regarding the positional relationships between the feature points of the target image; an image information comparison unit 15 that compares the target image information generated by the target image information generation unit 14 with the personal image information generated by the personal image information generation unit 3; an authentication processing unit 17 that performs personal authentication processing using the personal feature point information / target feature point information; and an authentication display instruction unit 18 that instructs the authentication processing unit 17 to display an authentication display consisting of content corresponding to the authentication result.

Description

Technical Field

[0001] The present invention relates to a technique for realizing highly accurate personal authentication processing with a small amount of information.

Background Art

[0002] In recent years, with the improvement of processing capabilities in electronic computers such as computers, two-dimensional and three-dimensional computer graphics have been widely used in IT services such as games and SNS (Social Networking Service). In these services, not only icons and avatars that imitate fictional characters are used, but the use of avatars that abstractly represent real people such as users while taking into account their characteristics is increasing. As a result, for example, in a game, when a protagonist consisting of an avatar representing the characteristics of the user battles an enemy character, the immersion of the user in the game world is improved. By using an avatar that abstractly represents the user himself / herself as an icon indicating the user in SNS, it is expected that the communication between users will be promoted in a virtual space in the same sense as the real world.

[0003] Both Patent Documents 1 and 2 disclose technologies for using avatars that imitate the actual appearance of the player himself / herself or co-players in a computer game that represents a virtual space using a head-mounted display.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] These avatars are primarily used online, and it cannot be definitively concluded that the person using a particular avatar is the same person as the original image from which the avatar was designed. Therefore, identity verification procedures are necessary when using online services. However, authentication using passwords and the like cannot completely prevent impersonation by third parties due to information leaks, and biometric authentication presents challenges in that obtaining biometric information in advance is not easy.

[0006] On the other hand, even if identity verification is to be done using image recognition, if a reference image of the person is to be stored in advance, there is a problem in terms of the burden on data storage, especially when high-resolution images are used to improve authentication accuracy. Furthermore, there is a risk that the reference image of the person may be tampered with due to information leakage, etc., and from that perspective, identity verification using image recognition is not foolproof. In addition, when using image recognition for identity verification, there is a problem in that the authentication accuracy decreases if, for example, a long period of time passes between obtaining the reference image of the person and the actual authentication process, or if the person is authenticated with a completely different facial expression from the reference image.

[0007] Furthermore, while performing identity verification constantly (for example, every minute) would place a significant burden on the system and is not practical, if identity verification is performed periodically (i.e., at regular time intervals), other users of online services that utilize avatars would have difficulty determining whether the person using the avatar has not undergone any identity verification, whether they were verified recently, or how much time has passed since the verification process.

[0008] This invention has been made in view of the above-mentioned problems, and aims to provide a technology that enables highly accurate personal authentication processing with a small amount of information, effectively prevents forgery of materials used in authentication processing, and allows a third party to easily understand the authentication results. [Means for solving the problem]

[0009] To achieve the above objective, the identity authentication system according to claim 1 is an identity authentication system that determines whether the person using an avatar in a predetermined service is the same person as the person for which a correspondence relationship with the person has been set, and comprises: an identity feature point information generation means that extracts a plurality of feature points from an identity image which is an image of the person and generates identity feature point information which is information relating to the positional relationships between the plurality of feature points; a target feature point information generation means that extracts a plurality of feature points from a target image which is an image of a target person which is a person to be authenticated and generates target feature point information which is information relating to the positional relationships between the plurality of feature points; an authentication processing means that compares the positional relationships between the plurality of feature points in the identity image included in the identity feature point information with the positional relationships between the plurality of feature points in the target image included in the target feature point information and performs authentication processing regarding the identity of the target person and the person based on the comparison result; and an authentication display instruction means that instructs to display an authentication display consisting of content corresponding to the authentication result by the authentication processing means in conjunction with the display of the avatar in the predetermined service.

[0010] Furthermore, in order to achieve the above objective, the personal authentication system according to claim 2 is characterized in that, in the above invention, the authentication display instruction means instructs to display an authentication indication corresponding to the fact that the person has been authenticated when the result of the authentication process by the authentication processing means recognizes the identity between the target person and the person concerned, and also instructs to change the content of the said authentication indication according to the time elapsed since the completion of the authentication process.

[0011] Furthermore, in order to achieve the above objective, the personal authentication system according to claim 3 further comprises: personal image information generation means for generating personal image information which includes at least one of the following: personal shooting condition information which includes information on at least one of the personal facial expression and health condition at the time of shooting and information on the date and time the personal image was taken; and personal shooting device information which is information on the shooting device used to take the personal image; target image information generation means for generating target image information which includes at least one of the following: information on at least one of the facial expression and health condition of a target person at the time of shooting and information on the date and time the target image was taken; and target shooting device information which is information on the shooting device used to take the target image; image information comparison means for comparing the personal image information and the target image information; and feature point information modification means for modifying the positional relationship between feature points in at least one of the personal feature point information and the target feature point information according to the difference between the personal image information and the target image information based on the comparison result by the image information comparison means, wherein the authentication processing means performs authentication processing based on the personal feature point information and / or the target feature point information modified by the feature point information modification means.

[0012] Furthermore, in order to achieve the above objective, the personal authentication method according to claim 4 is a personal authentication method that determines whether the person using an avatar in a predetermined service is the same person as the person for which a correspondence relationship with the person has been set, and is characterized by including: a personal feature point information generation step of extracting a plurality of feature points from a personal image which is an image of the person and generating personal feature point information which is information relating to the positional relationships between the plurality of feature points; a target feature point information generation step of extracting a plurality of feature points from a target image which is an image of a target person which is a person to be authenticated and generating target feature point information which is information relating to the positional relationships between the plurality of feature points; an authentication processing step of comparing the positional relationships between the plurality of feature points in the personal image included in the personal feature point information with the positional relationships between the plurality of feature points in the target image included in the target feature point information, and performing authentication processing regarding the identity of the target person and the person based on the comparison result; and an authentication display instruction step of instructing to display an authentication display consisting of content corresponding to the authentication result in the authentication processing step in conjunction with the display of the avatar in the predetermined service.

[0013] Furthermore, in order to achieve the above objective, the personal authentication method according to claim 5 further includes, in the above invention, an encryption processing step of encrypting the personal characteristic point information; an output information generation step of generating output information including encrypted personal characteristic point information relating to the person corresponding to the avatar, which is recorded in a distributed ledger on a distributed network corresponding to a non-fungible token that corresponds to the avatar; an output instruction step of instructing the distributed network to output the output information generated in the information generation step; an output information acquisition step of acquiring the output information from information recorded in a distributed ledger on the distributed network; and a decryption processing step of decrypting the personal characteristic point information included in the output information acquired in the output information acquisition step, wherein the personal characteristic point information decrypted in the decryption processing step is used in the authentication process in the authentication process.

[0014] Furthermore, in order to achieve the above objective, the personal authentication program according to claim 6 is a personal authentication program that causes a computer to determine whether the person using an avatar that has been set to correspond with a person in a predetermined service is the same person as the person in question, and is characterized by causing the computer to execute: a personal feature point information generation function that extracts a plurality of feature points from a personal image which is an image of the person in question and generates personal feature point information which is information relating to the positional relationships between the plurality of feature points; a target feature point information generation function that extracts a plurality of feature points from a target image which is an image of a person to be authenticated and generates target feature point information which is information relating to the positional relationships between the plurality of feature points; an authentication processing function that compares the positional relationships between the plurality of feature points in the personal image included in the personal feature point information with the positional relationships between the plurality of feature points in the target image included in the target feature point information and performs authentication processing regarding the identity of the target person and the person in question based on the comparison result; and an authentication display instruction function that instructs the computer to display an authentication display consisting of content corresponding to the authentication result by the authentication processing function in conjunction with the display of the avatar in the predetermined service. [Effects of the Invention]

[0015] According to the present invention, highly accurate personal authentication processing can be achieved with a small amount of information, and the authentication results can be easily understood by a third party. [Brief explanation of the drawing]

[0016] [Figure 1] This is a schematic diagram showing the configuration of the user authentication system according to Embodiment 1. [Figure 2] This is a schematic diagram showing the configuration of the user authentication system according to Embodiment 2. [Modes for carrying out the invention]

[0017] The embodiments of the present invention will be described in detail below with reference to the drawings. The following embodiments describe the most appropriate examples of the present invention, and naturally, the content of the present invention should not be limited to the specific examples shown in these embodiments. It goes without saying that any configuration other than the specific configurations shown in the embodiments that produces similar functions and effects is also included in the technical scope of the present invention.

[0018] (Embodiment 1) First, the identity authentication system according to Embodiment 1 will be described. As shown in Figure 1, the identity authentication system according to Embodiment 1 includes: an identity image input unit 1 for inputting an image of the person to be used for generating identity feature point information; an identity feature point information generation unit 2 for extracting multiple feature points on the image from the identity image and generating identity feature point information, which is information about the positional relationship between the feature points of the identity image; an identity image information generation unit 3 for generating information about the identity image; an avatar generation unit 5 for generating an avatar to be used by the person on services such as games and social networking services; an avatar token generation unit 6 for generating an avatar token, which is a non-fungible token that corresponds to the avatar used by the person; a transaction generation unit 7 for generating a transaction to be stored in a block in a distributed ledger corresponding to the avatar token as output information including identity feature point information (preferably encrypted information); an electronic signature generation unit 8 for generating an electronic signature, which is data that proves that the generation and output of the information included in the transaction is in accordance with the intentions of the holder of the avatar token; and an output unit 9 for outputting the transaction and electronic signature to a distributed network associated with the avatar token. A transaction acquisition unit 10 acquires transactions recorded in a distributed ledger corresponding to a token; a personal feature point information extraction unit 11 extracts personal feature point information from the transactions acquired by the transaction acquisition unit 10; a target image input unit 12 inputs a target image, which is an image of the person to be authenticated during the authentication process; a target feature point information generation unit 13 extracts multiple feature points from the target image and generates target feature point information, which is information about the positional relationship between the feature points of the target image; a target image information generation unit 14 generates target image information, which is information about the target image; an image information comparison unit 15 compares the target image information generated by the target image information generation unit 14 with the personal image information generation unit 3; a feature point information modification unit 16 modifies the personal feature point information and / or target feature point information based on the comparison result by the personal image information comparison unit 15; an authentication processing unit 17 performs personal authentication processing using the personal feature point information / target feature point information modified by the feature point information modification unit 16; and an authentication display instruction unit 18 instructs the authentication processing unit 17 to display an authentication display consisting of content corresponding to the authentication result.An authentication display generation unit 19 that generates a specific authentication display based on an instruction, and an authentication display output unit 20 that outputs the authentication display generated by the authentication display generation unit 19 to the service where the avatar is used.

[0019] The personal image input unit 1 is for inputting a personal image, which is an image of the person to be authenticated and is used for generating personal feature point information. The form of the personal image may be either a still image or a moving image, and it may be composed not only of a single image data but also of a plurality of image data taken from different directions. Regardless of color or black and white, it may be either a 2D image or a 3D image. As a specific configuration of the personal image input unit 1, it may be configured to directly input a personal image by having an imaging function such as a camera, or for example, it may be configured to input a personal image taken by a portable terminal owned by the person.

[0020] The personal feature point information generation unit 2 extracts a plurality of predetermined feature points corresponding to each part of the face such as eyes, nose, mouth, ears, etc. and characteristic parts such as the connection points of the jaw skeleton in the human body from the personal image, and generates personal feature point information, which is information on the positional relationship between the plurality of extracted feature points. The personal feature point information generated by the personal feature point information generation unit 2 is used as reference information for the personal authentication process by the authentication processing unit 17 described later. Although the generated personal feature point information may be used directly as reference information, preferably, the personal feature point information subjected to necessary correction processing by the feature point information correction unit 16 described later is used.

[0021] The specific configuration of the personal feature point information generation unit 2 may utilize image recognition technology realized by deep learning, machine learning, etc., or it may combine multiple image recognition technologies. Furthermore, the "feature points" in this embodiment 1 indicate the location of parts on the body surface or inside the body, and the parts to be set as feature points can be arbitrarily determined. For example, feature points may be set only for parts on the face surface that are visually recognizable (eyes, nose, mouth, ears, eyebrows, etc.) or points corresponding to the contours, or feature points may be set corresponding to the facial muscles inside the face or the locations of joints, which are internal structures. Also, for example, when setting feature points for the eyes, only the central position of the eye may be set as a feature point, or feature points may be set in detail for the inner corner of the eye, the outer corner of the eye, the pupil, and the white of the eye. Furthermore, each feature point extracted by the personal feature point information generation unit 2 is recorded as feature point information that includes information about the location of the feature point (which may be the location in the original image itself, but is preferably the location in a normalized coordinate system for convenience during personal authentication processing, etc.) and the significance of the feature point (e.g., it is a point corresponding to the corner of the eye, it is a point corresponding to the jaw joint, etc.).

[0022] The personal image information generation unit 3 is for generating personal image information about the person who is the subject of the personal image input by the personal image input unit 1, as well as information about the person such as the shooting time and shooting conditions of the personal image. The specific content of the personal image information includes at least one of personal shooting condition information including information about at least one of the expression and health condition of the person at the time of shooting and information about the shooting date and time of the personal image, and personal shooting device information which is information about the shooting device used for shooting the personal image. More preferably, it includes personal attribute information such as the age (or date of birth) and gender of the person. The specific content of the personal image information includes, in addition to personal attribute information about the target person such as the age (or date of birth) and gender of the target person, information about the shooting device such as the focal length, f - value of the lens used for shooting, and the sensitivity of the light - receiving part (film, image sensor) as personal shooting device information, and includes at least one piece of information such as the shooting date and time, the health condition and expression of the person at the time of shooting as personal shooting condition information. As a specific mode of the information generation process by the personal image information generation unit 3, it may be configured to input character information describing these information from the outside, or for information that can be estimated by image analysis processing on the personal image, such as the health condition, expression of the person at the time of shooting, and the age and gender of the person, it may be configured to obtain it by image analysis.

[0023] The avatar generation unit 5 is for generating an avatar used by a user in services such as games and SNS. Specifically, the avatar generation unit 5 extracts feature points (such as the center - of - gravity position of the eyes, the end of the eyes, the start of the eyes, etc.) set according to the surface features (eyes, eyebrows, nose, mouth, ears, hairstyle, etc.), internal features (joints, etc.), and the position and shape of external elements such as clothes, accessories, and items held, based on, for example, a full - body photo of the user, and has a function of generating a realistic avatar that reflects the physical characteristics of the user based on the position information of the feature points. As the configuration of the avatar, it is preferably composed of skeleton information (bones), surface information (skin), and weight information that defines the relationship between the two, but a simpler configuration such as a configuration consisting only of surface information may also be used. Furthermore, instead of using a full-body photograph of the user as reference material when generating the avatar, only a face photograph may be used, and generic data may be used for the part from the neck down, or the design may be unrelated to the user's characteristics. The avatar design is not limited to realistic designs, but may also be pictorial or animated, and only the hairstyle, clothing, accessories such as glasses, rings and bracelets, and possessions such as bags and umbrellas may be made to differ from the user's actual appearance. The avatar generation process by the avatar generation unit 5 may be performed by taking a full-body photograph of the user in addition to the user's image and generating the avatar based on that full-body photograph, but it may also be performed using the user's image input via the user image input unit 1 and / or information on feature points extracted from the user's image by the user feature point information generation unit 2.

[0024] The avatar token generation unit 6 is for generating avatar tokens, which are non-fungible tokens that correspond to the avatars generated by the avatar generation unit 5. A "non-fungible token" is a so-called NFT (Non-Fungible Token), that is, a token that has the property of being infungible with other tokens by possessing unique data, and is issued based on the Ethereum® standard ERC721, for example. The avatar token in this embodiment 1 is issued based on ERC721 or other predetermined standards, and is configured to store information such as information about the user of the corresponding avatar and personal characteristic point information (preferably encrypted personal characteristic point information, as described later) on the blockchain.

[0025] Blockchain is a technology that uses cryptographic techniques to synchronize data among multiple computers that make up a decentralized network. Specifically, each block is composed of a collection of token information, such as agreed-upon transaction records, and information to connect to other blocks (information from the previous block). A blockchain is formed by linking multiple such blocks together. Even if data is tampered with on some of the computers, the correct data is selected by majority vote among the other computers, making it extremely difficult to destroy or tamper with the data.

[0026] The correspondence between avatar tokens and avatars should be in the form of a one-to-one link between the avatar token identifier and the avatar's identification information (identification information may be assigned to each individual avatar, or identification information related to the person the avatar targets (such as a name) may be used). More directly, the avatar token identifier and the avatar's identification information may be identical, or a correspondence may be established for items composed of different strings or other elements. Alternatively, the correspondence may be linked to the URL information on the internet where the data constituting the avatar is stored.

[0027] Furthermore, in this embodiment 1, the blocks constituting the distributed ledger corresponding to the generated avatar token will record transaction records relating to the transfer of ownership of the avatar token, as well as personal characteristic point information (more preferably encrypted personal characteristic point information) relating to the person corresponding to the avatar. This information is converted by the transaction generation unit 7 into a transaction, which is information output to the distributed network corresponding to the avatar token.

[0028] Furthermore, the generation of avatar tokens may be performed by the avatar token generation unit 6 itself, or by an external system that is directly or indirectly connected to the avatar token generation unit 6, by issuing predetermined commands to the external system, thereby enabling the external system to generate the tokens. In addition, the specific format of the avatar token is not limited to that conforming to the Ethereum® standard ERC721, but may be any format as long as it has non-fungibility and its transaction history is stored on the blockchain.

[0029] The transaction generation unit 7 has the function of generating transactions to be stored in blocks in a distributed ledger, and corresponds to an example of output information generation means in the claims. Specifically, the transaction generation unit 7 generates information as the content of a transaction, including information regarding the change of the name of the holder of an avatar token, as well as personal characteristic point information about the person corresponding to the avatar. The transaction generation unit 7 may generate a transaction that includes the personal characteristic point information in its original state, but in this embodiment 1, an encryption processing unit 21 is newly provided to encrypt the personal characteristic point information, and the transaction generation unit 7 generates a transaction that includes the personal characteristic point information encrypted by the encryption processing unit 21. Any method of encryption by the encryption processing unit 21 can be used, for example, the personal characteristic point information may be encrypted using a secret key as described later.

[0030] The electronic signature generation unit 8 has the function of generating an electronic signature, which is data that proves that the creation and output of the information included in the transaction is in accordance with the intentions of the avatar token holder. Specifically, the electronic signature generation unit 8 has the function of generating a hash value of the transaction generated by the transaction generation unit 7 and encrypting the hash value with a private key held by the bromide token holder to generate an electronic signature. The "hash value" is a fixed-length value obtained by applying a certain calculation procedure to the original data. Since this calculation procedure is irreversible, it is considered impossible to restore the original data from the hash value. The "private key" is a sequence of numbers used in the encryption process and is configured to be decryptable by the corresponding "public key". In this embodiment 1, the transaction generation unit 7 and the electronic signature generation unit 8 are configured to create the transaction and electronic signature themselves, which are one form of ownership name change information, but it is not necessary to limit it to this, and the transaction generation unit 7 and the electronic signature generation unit 8 may have the function of only instructing an external predetermined device to create the transaction and electronic signature.

[0031] The output unit 9 is for instructing the distributed network associated with the avatar token to output the transaction generated by the transaction generation unit 7 and the digital signature generated by the digital signature generation unit 8. The output unit 9 functions as one aspect of the output instruction means (output instruction step in the method invention), and in this embodiment, it is configured to directly output the transaction and digital signature, but it may also be configured to only instruct other components (including those provided outside the identity authentication system according to Embodiment 1) to output. In this Embodiment 1, the output unit 9 is directly or indirectly connected to the distributed network and is configured to output predetermined data to the distributed network. The distributed network generates a hash value of the outputted transaction and compares it with the data obtained by decrypting the digital signature with the public key (= hash value of the transaction). If the two values ​​are different, it determines that the transaction is not by the owner of the tangible object in question and rejects registration. If the two values ​​match, it determines that the transaction is by the owner and stores the information composed of the transaction in a block that constitutes the distributed ledger. The transaction output by output unit 9 contains information about the individual's characteristic points, and when the transaction is output to the distributed network, this information is stored in blocks that constitute a distributed ledger established on the distributed network.

[0032] The transaction acquisition unit 10 is for acquiring transactions recorded in the distributed ledger and corresponds to the output information generation means (or output generation step in the invention of the method). Specifically, the transaction acquisition unit 10 has the function of accessing the distributed ledger corresponding to the avatar that is the subject of user authentication processing and acquiring user characteristic point information (or information obtained by encrypting the user characteristic point information by the encryption processing unit 20). As a mechanism for accessing corresponding transactions, for example, identification information common to the corresponding avatar may be included in the transaction, or information regarding the correspondence with the avatar may be added to the distributed ledger that stores the transactions.

[0033] The personal feature point information extraction unit 11 is for extracting personal feature point information corresponding to the avatar to be authenticated from the transaction acquired by the transaction acquisition unit 10. If the personal feature point information is included in the transaction as is, the personal feature point information is extracted by selecting the relevant part from the transaction. However, in this embodiment 1, since the personal feature point information is included in the transaction after being encrypted by the encryption processing unit 21, the personal feature point information extraction unit 11 has the function of performing decryption processing on the encrypted part by the newly provided decryption processing unit 22 before extracting the personal feature point information.

[0034] The target image input unit 12 is for inputting a target image, which is an image of the person to be authenticated during the authentication process. The target image can be a still image or a video, and may consist of a single image or multiple image data taken from different angles, and may be in color or black and white, and may be a two-dimensional or three-dimensional image. However, from the viewpoint of avoiding complexity in the authentication process, it is preferable that the format of the target image be the same as the format of the personal image input via the personal image input unit 1. Furthermore, the specific configuration of the target image input unit 12 may be such that it inputs a target image taken with a portable terminal carried by the target person, but from the viewpoint of preventing forgery of the target image, it is also preferable that it inputs only video footage taken with a shooting device (including both devices located inside and outside the system) that incorporates an anti-fraud mechanism, as provided in the personal authentication system according to this embodiment 1.

[0035] The target feature point information generation unit 13, like the personal feature point information generation unit 2, has the function of extracting a predetermined number of feature points from a target image input via the target image input unit 12, corresponding to various parts of the face such as the eyes, nose, mouth, and ears, as well as characteristic parts such as the connection points of the jawbone, and generating target feature point information, which is information about the positional relationships between these multiple feature points. The specific configuration of the target feature point information generation unit 13 may utilize image recognition technology realized by deep learning, machine learning, etc., or it may be a combination of multiple image recognition technologies. Furthermore, while the parts to be set as feature points for the feature points extracted by the target feature point information generation unit 13 may be arbitrarily determined, it is preferable to extract feature points that are the same as or common to the feature points extracted by the personal feature point information generation unit 2.

[0036] The target image information generation unit 14 generates target image information for a target image input by the target image input unit 12, which includes at least one of the following: target shooting condition information, which includes information on at least one of the subject person's facial expression and health condition at the time of shooting and information on the date and time the target image was taken; and target shooting equipment information, which is information on the shooting equipment used to take the target image. Preferably, it also includes target attribute information such as the subject person's age (date of birth) and gender. Specifically, the target image information includes target attribute information such as the subject person's age (or date of birth) and gender, as well as information on the shooting equipment used for shooting, such as the focal length, f-number, and sensitivity of the light-receiving part (film, image sensor) of the lens used for shooting, and at least one of the following information as target shooting condition information, such as the date and time of shooting, the subject person's health condition at the time of shooting, and facial expression. Preferably, it includes the same items as the personal image information. As for specific forms of information generation processing by the target image information generation unit 14, it may be configured to input text information describing this information from an external source, or it may be configured to acquire information that can be estimated by image analysis processing of the target image, such as the person's age, gender, and facial expression at the time of shooting, through image analysis.

[0037] The image information comparison unit 15 compares the subject image information generated by the subject image information generation unit 3 with the target image information generated by the target image information generation unit 14, and outputs the comparison result to the feature point information correction unit 16. Specifically, the image information comparison unit 15 compares common items (for example, the shooting date) for both the subject image information and the target image information, and has the function of extracting the differences if there are differences in content (or if the differences exceed a predetermined threshold) and outputting them as a comparison result. As for the specific method of extracting the differences, for example, it may be to simply extract the difference value for numerical information, but more preferably, it has the function of extracting the content of the information in both the subject image information and the target image information, and outputting a list of the values ​​of both, rather than the difference value, for example, the focal length of the lens as information related to the shooting equipment, as a comparison result. If there are no differences in each item or if the differences are below a predetermined threshold, the comparison result may be output as "no differences", or no comparison result may be output at all.

[0038] The feature point information correction unit 16 corrects the positional relationships between feature points in at least one of the subject's feature point information and the target's feature point information, based on the comparison results from the image information comparison unit 15, according to the differences between the subject's image information and the target image information. Specifically, the feature point information correction unit 16 has the function of correcting the feature point information related to an image after inferring the positional relationships between feature points if one of the images were taken under the conditions of the other image information instead of the conditions of its own image information, based on the comparison results for one or more of the information from the shooting equipment information, shooting condition information, and attribute information in the image information.

[0039] For example, the feature point information modification unit 16 compares the shooting date and time of the person's image included in the person shooting condition information with the shooting date and time of the target image included in the target shooting condition information. For example, if the shooting date and time of the target image is 10 years after the shooting date and time of the person's image, the unit estimates how the positional relationship between feature points would change due to aging and other age-related changes if the person's image were taken 10 years after the actual shooting date and time, and modifies the content of the person's feature point information based on the estimation result (in this case, it is also preferable to estimate the age-related changes after calculating the specific age based on the attribute information). Furthermore, for example, if the person's facial expression is recorded as smiling in the person shooting condition information and the target person's facial expression is recorded as angry in the target shooting condition information, the feature point information modification unit 16 estimates how the positional relationship between feature points would change if the target image were taken with a smiling expression instead of an angry expression, and modifies the content of the target feature point information based on the estimation result. Similarly, if the health status differs to such an extent that the facial shape and other appearance changes, the feature point information correction unit 16 estimates how the positional relationships between feature points on the person's image would change if the health status of the subject in the target image were that of the person in question, rather than the person's health status at the time the image was taken. Based on the estimation results, the unit then changes the positional relationships between feature points in the person's feature point information.

[0040] Furthermore, the feature point information correction unit 16 has the function of correcting the positional relationship information of feature points included in the personal feature point information, based on a comparison result between personal image equipment information, which is information about the characteristics of the camera used to take the personal image, and target image equipment information, which is information about the characteristics of the camera used to take the target image, so that the positional relationship is the same as if the image was taken with the camera used to take the target image, rather than the camera used to take the personal image. For example, if the personal image is taken with a short focal length, a so-called wide-angle lens, the captured image of the face will appear more bulging compared to the case of a long focal length lens, and accordingly, the positions of feature points in the personal image and target image will also change from the positions when taken with a long focal length lens. Similarly, if the focal length and f-number of the lens are different, the depth of field will be different, and accordingly, the positions of feature points in the personal image and target image will change. In light of the fact that the position of feature points derived from a captured image changes depending on the characteristics of the shooting device, even for the same subject, the feature point information modification unit 16 estimates, based on the differences in the characteristics of each shooting device, how the positional relationship between feature points on the subject's image would change if, for example, the subject's image were not taken under the conditions recorded in the subject's shooting device information but under the conditions recorded in the target shooting device information, and then changes the positional relationship between feature points in the subject's feature point information based on the estimation result.

[0041] Furthermore, the specific algorithm for predicting how the positional relationship between feature points changes based on the differences in each piece of information may be generated using, for example, machine learning, deep learning, or reinforcement learning. Alternatively, a large number of specific examples, such as changes in captured images due to differences in imaging equipment, may be collected, and when correcting the personal feature point information, the example of the image that is closest to the comparison result to be referenced is extracted from the collected examples, and the correction content accordingly is determined. In addition, if there are no differences between the image information, or if the differences are only slight, the feature point information correction unit 16 will output each feature point information to the authentication processing unit 17 as is without making any special corrections.

[0042] The authentication processing unit 17 performs authentication processing, which involves comparing the feature point information of the subject in the target image with the subject in the subject image by using the subject feature point information and target feature point information corrected by the feature point information correction unit 16 as needed. Specifically, the authentication processing unit 17 compares the positional relationship information of the feature points contained in the subject feature point information and target feature point information corrected by the feature point information correction unit 16 (or the original subject feature point information and target feature point information if no correction was made), and performs authentication processing by detecting the similarities and differences in the positional relationships between multiple corresponding feature points. In other words, the authentication processing unit 17 has the function of determining that the subject in the target image is the same person as the subject in the subject image if the positional relationships between corresponding feature points are the same, or if they are not the same but the degree of difference is less than a predetermined standard value and is considered to be within the range of identity. Furthermore, the authentication processing unit 17 has the function of making a determination to deny identity if the difference in the positional relationship between corresponding feature points is greater than or equal to a predetermined standard value. The authentication processing unit 17 may perform a determination of similarity or difference in positional relationship between all feature points included in the personal feature point information and the feature points of the target image, or it may perform a determination of similarity or difference only for the positional relationship of some predetermined important feature points. Furthermore, if the determination results differ for each feature point, such as when the difference in positional relationship between some feature points is above a threshold, and the positional relationship between other feature points is the same, the authentication process may be performed to determine the presence or absence of identity based on whether the sum of the difference amounts exceeds a predetermined threshold, or the proportion of pairs of feature points that differ by more than a threshold. The determination result by the authentication processing unit 17 shall be output to the authentication display instruction unit 18 along with information regarding the date and time the determination process was performed. The timing of the authentication process by the authentication processing unit 17 may be at any time, but it is preferable to perform it during interactions with other users in a service where the avatar is used (message exchange, delivery of items or items with monetary value, or money, etc.).

[0043] The authentication display instruction unit 18 is for instructing the authentication processing unit 17 to display an authentication display attached to the avatar when an avatar corresponding to the user is displayed in a service such as an SNS or game, with content corresponding to the authentication result of the authentication processing unit 17. In a more preferred embodiment, the authentication display instruction unit 18 has the function of instructing the authentication processing unit 17 to display a message to that effect and to change the display content according to the time elapsed since the completion of the authentication process when the target person is authenticated as being the same as the user. Specifically, the authentication display instruction unit 18 has the function of instructing the authentication display unit 18 to generate an authentication display that is displayed attached to the avatar when it is displayed in a service such as a game or SNS, and is a visual display that can be seen by other users as indicating that the user has been authenticated. Note that "attached to the display of the avatar" means that it is displayed in the vicinity of the avatar displayed in a service in which the avatar is used, but it may be displayed attached integrally with the avatar, or it may be a display that modifies part or all of the avatar itself (for example, putting a special hat on the head, giving the whole thing a glossy finish, etc.). Furthermore, the authentication display instruction unit 18 may instruct the system to simply display a message indicating that the authentication process has been completed as the authentication display. More preferably, the instruction unit will notify the system of information regarding the date and time the authentication process was completed, and will also instruct the system to display a visually changing display according to the time elapsed since the date and time the authentication process was completed. Possible examples of such changes include changes in the hue, saturation, and brightness of the authentication display according to the passage of time (for example, the color of the authentication display may fade and eventually disappear as time passes), or changes in shape, but other forms of change are also possible.

[0044] The authentication display generation unit 19 is for generating a specific authentication display based on the instructions of the authentication display instruction unit 18. The authentication display generation unit 19 determines the specific content of the authentication display that will be visually displayed along with the avatar and generates image data. The generated image data of the authentication display is output to the service on which the avatar is used via the authentication video output unit 20. If the authentication display is simply an indication that authentication processing has been completed, the authentication display generation unit 19 only generates a single image data. However, if the authentication display changes over time from the completion of the authentication processing, it prepares multiple image data and also generates information regarding the display timing of these multiple image data (e.g., until one month has passed since the completion of authentication processing, from one month after the completion of authentication processing until six months after the completion of authentication processing, etc.).

[0045] The authentication display output unit 20 is for outputting information related to the authentication display generated by the authentication display generation unit 19 to external services such as games and social networking services where the avatar is used. Specifically, the authentication display output unit 20 has the function of outputting image data generated as information related to the authentication display (if multiple image data is generated, the multiple image data and data indicating the display time of each image data). The output image data, etc., is displayed in conjunction with the avatar on the external service, allowing other users to visually understand information such as whether the avatar has been authenticated, whether it has been authenticated but a certain period of time has passed since the authentication process (= there is a certain possibility that the user has changed since then), or whether the authentication process has not been performed. As a modification, the authentication display generation unit 19 may be omitted, and the authentication display output unit 20 may be configured to directly output the display instructions related to the authentication display generated by the authentication display instruction unit 18 to the external service, and the external service may be configured to generate specific authentication display images, etc., in the same way as the authentication display generation unit 19.

[0046] Next, the advantages of the personal authentication system according to this embodiment 1 will be explained. First, the personal authentication system according to this embodiment 1 employs a configuration in which personal feature point information, which is information about the positional relationships between multiple feature points extracted from the personal image, is compared with target feature point information extracted from the target image as reference information used during authentication. While such feature point information consists of an extremely small amount of information compared to the image data itself, it provides a detailed description of the facial features of a person. Therefore, the personal authentication system according to this embodiment 1 has the advantage of being able to achieve highly accurate personal authentication processing with a small amount of information.

[0047] Furthermore, the personal authentication system according to this embodiment 1 employs a configuration in which, during the authentication process, the personal image information and the target image information are compared, and the personal feature point information and / or target feature point information are appropriately modified based on the differences before the personal authentication process is performed. By adopting such a configuration, the personal authentication system according to this embodiment 1 can correct the information regarding the positional relationship between feature points to the appropriate value if it differs from the appropriate content during the authentication process due to differences in the shooting equipment, facial expressions, age, etc. at the time of shooting. By adopting such a configuration, the personal authentication system has the advantage of achieving more accurate personal authentication processing.

[0048] Furthermore, the personal authentication system according to this embodiment 1 has a configuration in which the personal feature point information used as reference information during the authentication process is stored in a distributed ledger. Because the use of blockchain makes it extremely difficult to destroy or tamper with the data constituting the personal feature point information, the personal authentication system according to this embodiment 1 has the advantage of effectively preventing the forgery of documents used in the authentication process. In addition, since this embodiment 1 is configured to encrypt the personal feature point information and store it in a distributed ledger, even in a system accessible to an unspecified number of people, such as a public blockchain, a third party cannot grasp the contents of the personal feature point information in advance. This prevents fraud, such as manipulating a target image so that target feature point information that matches the personal feature point information is generated in advance by the target person, and from this perspective as well, it has the advantage of effectively preventing the forgery of documents used in the authentication process.

[0049] Furthermore, the user authentication system according to this embodiment 1 displays an authentication message corresponding to the result of the authentication process in conjunction with the display of the avatar being used. By adopting this configuration, a third party other than the avatar user (for example, another avatar user using the same service) can determine whether the avatar has been authenticated (i.e., whether there is proof that the avatar is being used by the person in question) and how much time has passed since the authentication was performed (while there is proof that the person used it in the past, there is a certain degree of risk of impersonation by someone else depending on the passage of time), and this has the advantage that a third party can easily grasp the authentication result.

[0050] (Embodiment 2) Next, the user authentication system according to Embodiment 2 will be described. In Embodiment 2, components that have the same name and reference numeral as those in Embodiment 1 will perform the same functions as those in Embodiment 1 unless otherwise specified.

[0051] Figure 2 shows that the personal authentication system according to Embodiment 2 includes a target image storage unit 24 that stores target images used in past authentication processes that have been determined to be the image of the person through the authentication process, and a variation pattern detection unit 25 that detects the variation patterns of the positions of common points in the target images and the person's image stored in the target image storage unit 24. The feature point information correction unit 26 has the function of estimating the changes in the positional relationship of feature points due to differences in shooting time based on the variation patterns derived by the variation pattern detection unit 25 and shooting condition information.

[0052] The target image storage unit 24 is for storing target images that have been determined to be the person's image through the authentication process (more precisely, target images in which the person who is the subject of the person's image and the person who is the subject of the target image are determined to be the same person). Specifically, the target image storage unit 24 has the function of storing target images that have been determined to be the person's image by the authentication processing unit 17, along with target image information including information about the time the target image was taken and the camera used to take it.

[0053] The variation pattern detection unit 25 uses past target images stored in the target image storage unit 24, which have been identified as being of the same person, to derive variation pattern information, which is information about how the positional relationships of feature points extracted from the person's image change over time. Specifically, the variation pattern detection unit 25 first derives the positional relationships of feature points on the target image when the target image is taken with the same camera as the person's image and when the same facial expression as the person's image is captured, based on the target image information associated with each target image. Next, the variation pattern detection unit 25 has the function of detecting how the positional relationships of feature points under the same conditions and with the same facial expression change according to the time the person's image was taken and the time each target image was taken, and generating the detection result as variation pattern information.

[0054] The feature point information correction unit 26 has the same functions as the feature point information correction unit 16 in Embodiment 1, and in relation to a new target image input via the target image input unit 12, it uses the variation pattern information derived by the variation pattern detection unit 25 when changing the position of each feature point based on the person's image to a position that takes into account changes due to aging and other age-related changes, based on the difference between when the person's image was taken and when the target image was taken. By utilizing the variation pattern information, which is information about the variation in the positional relationship between actual feature points related to the person, the feature point information correction unit 26 can more accurately derive the positional relationship of the feature points of the person's image at the time the newly input target image was taken.

[0055] Although the content of the present invention has been described above in terms of Embodiments 1 and 2, the technical scope of the present invention should not be interpreted as being limited to the specific configurations described in the embodiments. Various modifications and applications of the above embodiments are also, of course, within the technical scope of the present invention, as long as they can realize the functions of the present invention. [Industrial applicability]

[0056] This invention can be used as a technology that enables highly accurate personal authentication processing with a small amount of information, effectively prevents forgery of documents used in the authentication process, and allows third parties to easily understand the authentication results. [Explanation of Symbols]

[0057] 1. Image input section for the user 2. Personal Characteristic Point Information Generation Unit 3. Personal Image Information Generation Unit 5. Avatar Generation Section 6. Avatar Token Generation Unit 7. Transaction Generation Unit 8 Electronic signature generation section 9 Output section 10 Transaction Acquisition Section 11. Personal Characteristic Point Information Extraction Unit 12 Target image input section 13 Target Feature Point Information Generation Unit 14 Target Image Information Generation Unit 15 Image Information Comparison Section 16, 26 Feature Point Information Correction Section 17 Authentication Processing Unit 18 Authentication display instruction section 19 Authentication display generation unit 20 Authentication display output section 21 Cryptographic Processing Unit 22 Decoding Processing Unit 24 Target Image Storage Unit 25 Variation Pattern Detection Unit

Claims

1. A personal authentication system that determines whether the person using an avatar, for which a correspondence relationship has been established with the actual person, is the same person as the person using the avatar in a designated service. A means for generating personal feature point information that extracts multiple feature points from a personal image, which is a picture of the person, and generates personal feature point information, which is information about the positional relationships between those multiple feature points. A target feature point information generation means extracts multiple feature points from a target image, which is an image of a person to be authenticated, and generates target feature point information, which is information regarding the positional relationships between those multiple feature points. Authentication processing means that compares the positional relationships between multiple feature points in the image of the person included in the person feature point information with the positional relationships between multiple feature points in the target image included in the target feature point information, and performs authentication processing regarding the identity of the target person and the person based on the comparison result, Authentication display instruction means that instructs the display of an authentication display consisting of content corresponding to the authentication result by the authentication processing means to be displayed in conjunction with the display of the avatar in the predetermined service, A personal authentication system characterized by having the following features.

2. The authentication display instruction means is characterized in that, when the result of the authentication process by the authentication processing means confirms the identity between the target person and the person concerned, it instructs to display an authentication indication corresponding to the fact that the person has been authenticated, and also instructs to change the content of the authentication indication according to the time elapsed since the completion of the authentication process.

3. A means for generating personal image information that generates personal image information including personal shooting condition information, which includes information on at least one of the person's facial expression and health condition at the time of shooting and information on the date and time the personal image was taken, and personal shooting equipment information, which is information on the shooting equipment used to take the personal image. A target image information generation means generates target image information which includes at least one of the following: information regarding the facial expression and health condition of the subject person at the time of shooting, and information regarding the date and time the target image was taken; and information regarding the shooting equipment used to take the target image. Image information comparison means for comparing the aforementioned personal image information with the aforementioned target image information, A feature point information modification means that modifies the positional relationship between feature points in at least one of the feature point information and the feature point information, based on the comparison results by the image information comparison means, in accordance with the differences between the person's image information and the target image information. Furthermore, The authentication system according to claim 1, characterized in that the authentication processing means performs authentication processing based on the personal feature point information and / or target feature point information modified by the feature point information modification means.

4. A method for authenticating an avatar that has been associated with a person, and determining whether the person using the avatar in a designated service is the same person as the person in question. A step of generating personal feature point information, which involves extracting multiple feature points from a personal image, which is a picture of the person, and generating personal feature point information, which is information about the positional relationships between those multiple feature points. A target feature point information generation step involves extracting multiple feature points from a target image, which is an image of a person to be authenticated, and generating target feature point information, which is information regarding the positional relationships between those multiple feature points. Authentication processing step: Compares the positional relationships between multiple feature points in the image of the person included in the person feature point information with the positional relationships between multiple feature points in the target image included in the target feature point information, and performs authentication processing regarding the identity of the target person and the person based on the comparison result. An authentication display instruction step that instructs the display of an authentication display consisting of content corresponding to the authentication result in the authentication processing step to be displayed in conjunction with the display of the avatar in the predetermined service, A method of authenticating an individual, characterized by including the following:

5. The encryption processing step involves encrypting the aforementioned personal characteristic point information, Output information generation step: Generates output information including encrypted personal characteristic point information relating to the person corresponding to the avatar, which is recorded in a distributed ledger on a distributed network corresponding to a non-fungible token that corresponds to the avatar. An output instruction step that instructs the distributed network to output the output information generated in the information generation step, An output information acquisition step, which involves acquiring the output information from information recorded in a distributed ledger on the distributed network, A decoding process step which decodes the personal characteristic point information contained in the output information acquired in the output information acquisition step, The method for authenticating an individual according to claim 4, further comprising using the individual characteristic point information decrypted in the decryption step for authentication in the authentication step.

6. A personal authentication program that causes a computer to determine whether the person using an avatar, which has a correspondence relationship set with the actual person, is the same person as the actual person, in a designated service. To the aforementioned computer, A feature point generation function that extracts multiple feature points from an image of the person and generates feature point information that is information about the positional relationships between those multiple feature points, A feature point information generation function that extracts multiple feature points from a target image, which is an image of a person to be authenticated, and generates target feature point information, which is information about the positional relationships between those multiple feature points. An authentication processing function that compares the positional relationships between multiple feature points in the image of the person included in the person feature point information with the positional relationships between multiple feature points in the target image included in the target feature point information, and performs authentication processing regarding the identity of the target person based on the comparison result, An authentication display instruction function that instructs the display of an authentication display consisting of content corresponding to the authentication result by the authentication processing function in conjunction with the display of the avatar in the predetermined service, A user authentication program characterized by prompting the execution of an action.

Images