Avatar output device, avatar output method, and avatar output program

Abstract

To present an appropriate avatar candidate in an avatar output technique that outputs an avatar using an avatar candidate selected from a plurality of avatar candidates based on the original image.SOLUTION: An avatar output device comprises: an original image input unit 1 for inputting the original image; a feature point extraction unit 3 that extracts feature points of the original image input from the original image input unit 1; an avatar candidate generation unit 4 that generates a plurality of avatar candidates on the basis of the original image and the feature point extracted from the original image; an avatar candidate selection unit 5 for selecting one avatar candidate from the plurality of avatar candidates; an avatar output unit 6 that outputs an avatar using the selected avatar candidate; a storage unit 7 that stores avatar candidate information of the selected avatar candidate, and feature-point information in association with each other; and a related avatar candidate information provision unit 8 that outputs, to the avatar candidate generation unit 4, related avatar candidate information that is information on avatar candidates having the feature-point information in common when the original image is input.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a technology for generating an avatar of an avatar user after presenting multiple design candidates based on an original image related to the user. [Background technology]

[0002] In recent years, with the improvement in the processing power of computers and other electronic devices, IT services such as games and social networking services (SNS) are increasingly using two-dimensional and three-dimensional computer graphics. These services not only use icons and avatars that resemble fictional characters, but also increasingly use avatars that are somewhat abstract and based on the characteristics of real people, such as users. This has the effect of increasing the user's sense of immersion in the game world, for example, by having the protagonist, whose avatar represents the user's characteristics, compete against enemy characters. Similarly, by using an avatar that represents the user in an abstract way as an icon representing the user on SNS, it is expected that interactions between users in a virtual space will be promoted in a way that feels similar to the real world.

[0003] Patent documents 1 and 2 both disclose technology for using avatars that resemble the actual appearance of the player or co-player in computer games that use head-mounted displays to represent virtual spaces. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2019-012509 [Patent Document 2] Japanese Patent Application Publication No. 2019-139673 Summary of the Invention [Problem to be solved by the invention]

[0005] When generating an avatar that resembles the actual appearance of a player, it is generated based on an original image related to the player's appearance, such as a facial image or a full-body image. Rather than using the original image as is, it is common to modify the shape and other features of the original image to a certain extent, while retaining the original image's characteristics to reflect the worldview of the social networking site where the avatar is used and the player's preferences. However, it is not realistic for a designer to manually create an avatar design while listening to the player's preferences. Therefore, changes to the shape and other features of the original image must be automatically or mechanically generated using software processing, etc. However, when using automatic or mechanical processing, there is a problem in that the avatar design, which changes the original image to a certain extent, does not necessarily reflect the player's preferences.

[0006] Furthermore, since these avatars are primarily used online, it is not possible to be certain that the person using a particular avatar is the same person as shown in the original image that the avatar is designed from, and identity verification procedures are required when using online services. However, authentication using passwords or other methods cannot completely prevent third parties from impersonating someone due to information leaks, and biometric authentication has the problem that it is not easy to obtain biometric information in advance.

[0007] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a technology that presents appropriate candidates in a technology that presents multiple avatar candidates based on an original image and then determines the avatar candidate to be used for an avatar, and also provides a technology that easily verifies the identity of the avatar user. [Means for solving the problem]

[0008] In order to achieve the above object, the avatar output device according to claim 1 is an avatar output device that outputs an avatar using one avatar candidate selected from a plurality of avatar candidates that are semi-finished or finished avatars, and includes: feature point extraction means that extracts feature points of an original image; avatar candidate generation means that generates a plurality of avatar candidates that are semi-finished or finished avatars from the original image; avatar output means that outputs an avatar using one avatar candidate selected from the plurality of avatar candidates; storage means that stores avatar candidate information that is information about the selected avatar candidate in association with feature point information that is information about the feature points extracted by the feature point extraction means and attribute information that is information about attributes of a person that is associated with the generated avatar through the original image; and storage means that stores, from the avatar candidate information stored in the storage means, feature point information that has commonality with the feature points extracted by the feature point extraction means, and attribute information that is information about the attributes of a person that is associated with the generated avatar. The system comprises: a related avatar candidate information providing means for extracting related avatar candidate information, which is avatar candidate information, and providing the avatar candidate generation means with avatar candidates based on the extracted related avatar candidate information; and an attribute avatar candidate information providing means for providing the avatar candidate generation means with attribute avatar candidate information, which is avatar candidate information associated with attribute information that has commonality with attributes of a person associated with a newly generated avatar through an original image, from the avatar candidate information stored in the storage means, wherein the avatar candidate generation means generates avatar candidates including designs that are identical and / or similar to all or a part of the designs of avatar candidates corresponding to the related avatar candidate information provided by the avatar candidate information providing means, and avatar candidates including designs that are identical and / or similar to all or a part of the designs of avatar candidates corresponding to the attribute avatar candidate information provided by the attribute avatar candidate information providing means."Semi-finished avatar" refers to an intermediate product in the process of generating an avatar using feature points extracted from an original image, and includes all intermediate products, such as electronic data resulting from image processing of the original image, electronic data that converts a two-dimensional original image into three-dimensional data, and electronic data related to surface information generated based on the original image when the completed avatar is constructed using computer graphics consisting of skeletal information (bones), surface information (skin), and weight information that defines the relationship between the two. Furthermore, "semi-finished avatar" includes not only intermediate products related to the entire avatar, but also intermediate products related to a partial region of the entire avatar, and the avatar itself related to a partial region. Furthermore, "avatar using avatar candidates" refers to an avatar completed using the avatar candidates if the avatar candidates are semi-finished, and refers to an avatar made from the avatar candidates themselves if the avatar candidates are finished. Furthermore, in the context of feature point information, "having commonality" most simply means that the positions of all corresponding feature points (preferably normalized to the same coordinate system) are the same, but it is not necessary to interpret it as being limited to this, and it may also be that "having commonality" occurs when the distance between corresponding feature points is equal to or less than a certain value. Furthermore, "having commonality" can be interpreted not only when all feature points are compared, but also when the above relationship is achieved when only some feature points are compared.

[0009] In order to achieve the above object, the avatar output method according to claim 2 is an avatar output method for outputting an avatar using one avatar candidate selected from a plurality of avatar candidates that are semi-finished or finished avatars, the method including: a feature point extraction step for extracting feature points of an original image; an avatar candidate generation step for generating a plurality of avatar candidates that are semi-finished or finished avatars from the original image; an avatar output step for outputting an avatar using one avatar candidate selected from the plurality of avatar candidates; a storage step for storing avatar candidate information that is information about the selected avatar candidate in association with feature point information that is information about the feature points extracted in the feature point extraction step and attribute information that is information about attributes of a person associated with the generated avatar through the original image; and a storage step for selecting an avatar from the avatar candidate information stored in the storage step that is associated with feature point information that has commonality with the feature points extracted in the feature point extraction step. The method includes a related avatar candidate information providing step of extracting related avatar candidate information, which is candidate information, and providing avatar candidates based on the extracted related avatar candidate information to the avatar candidate generation step; and an attribute avatar candidate information providing step of providing attribute avatar candidate information to the avatar candidate generation step, which is avatar candidate information associated with attribute information that has commonality with attributes of a person associated with a newly generated avatar through an original image from the avatar candidate information stored in the storing step, wherein the avatar candidate generation step generates avatar candidates including designs that are identical and / or similar to all or a part of the designs of avatar candidates corresponding to the related avatar candidate information provided in the avatar candidate information providing step, and avatar candidates including designs that are identical and / or similar to all or a part of the designs of avatar candidates corresponding to the attribute avatar candidate information provided in the attribute avatar candidate information providing step.

[0010] In order to achieve the above object, the avatar output program according to claim 3 is an avatar output program that causes a computer to output an avatar using one avatar candidate selected from a plurality of avatar candidates that are semi-finished or finished avatars, and includes the following functions: a feature point extraction function that extracts feature points from an original image; an avatar candidate generation function that generates a plurality of avatar candidates that are semi-finished or finished avatars from the original image; an avatar output function that outputs an avatar using one avatar candidate selected from the plurality of avatar candidates; a storage function that stores avatar candidate information that is information about the selected avatar candidate in association with feature point information that is information about the feature points extracted by the feature point extraction function and attribute information that is information about attributes of a person that is associated with the generated avatar through the original image; and a storage function that stores, from the avatar candidate information stored in the storage function, feature points that have commonality with the feature points extracted by the feature point extraction function. The system executes a related avatar candidate information providing function that extracts related avatar candidate information, which is avatar candidate information associated with feature point information, and provides the avatar candidate generation function with avatar candidates based on the extracted related avatar candidate information, and an attribute avatar candidate information providing function that provides the avatar candidate generation function with attribute avatar candidate information, which is avatar candidate information associated with attribute information that has commonality with attributes of a person associated with a newly generated avatar through an original image, from the avatar candidate information stored in the storage function, and the avatar candidate generation function generates avatar candidates including designs that are identical and / or similar to all or a part of the designs of avatar candidates corresponding to the related avatar candidate information provided from the avatar candidate information providing function, and avatar candidates including designs that are identical and / or similar to all or a part of the designs of avatar candidates corresponding to the attribute avatar candidate information provided from the attribute avatar candidate information providing function. [Effects of the Invention]

[0011] According to the present invention, an avatar output technology that presents multiple avatar candidates based on an original image and outputs an avatar using an avatar candidate selected from these can present appropriate candidates, and also has the effect of providing a technology that can easily verify the identity of the user of the avatar. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a schematic diagram illustrating a configuration of an avatar output device according to a first embodiment. [Figure 2] 1 is a flowchart showing an avatar output process in the avatar output device according to the first embodiment. [Figure 3] FIG. 10 is a schematic diagram showing a configuration of an avatar output device according to a second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following embodiments, examples that are considered to be the most appropriate embodiments of the present invention will be described, and it goes without saying that the contents of the present invention should not be interpreted as being limited to the specific examples shown in the present embodiments. Of course, any configuration that achieves similar functions and effects, even if it is a configuration other than the specific configuration shown in the embodiments, is included in the technical scope of the present invention.

[0014] (Embodiment 1) First, an avatar output device according to a first embodiment will be described. As shown in Fig. 1, the avatar output device according to the first embodiment includes an original image input unit 1 for inputting an original image, an attribute information input unit 2 for inputting attribute information that is information about the attributes of the person who will use the avatar, a feature point extraction unit 3 for extracting feature points of the original image input from the original image input unit 1, an avatar candidate generation unit 4 for generating a plurality of avatar candidates based on the original image and the feature points extracted from the original image, an avatar candidate selection unit 5 for selecting one avatar candidate from the plurality of avatar candidates, an avatar output unit 6 for outputting an avatar using the selected avatar candidate, a storage unit 7 for storing avatar candidate information that is information about the selected avatar candidate, feature point information that is information about feature points of the original image used to generate the avatar candidate, and attribute information of the person who will use the avatar in association with each other, a related avatar candidate information providing unit 8 for, when a new original image is input, outputting related avatar candidate information that is avatar candidate information of avatar candidates having common feature point information to the avatar candidate generation unit 4, and a new avatar candidate information providing unit 9 for providing a new avatar candidate information providing unit 10 for providing a new avatar candidate information providing unit 11 for providing a new avatar candidate information providing unit 12 for providing a new avatar candidate information providing unit 13 for providing a new avatar candidate information providing unit 14 for providing a new avatar candidate information providing unit 15 for providing a new avatar candidate information providing unit 16 for providing a new avatar candidate information providing unit 17 for providing a new avatar candidate information providing unit 18 for providing a new avatar candidate information providing unit 19 for providing a new avatar candidate information providing unit 20 for providing a new avatar candidate information providing unit 21 for providing a new avatar candidate information providing unit 22 for providing a new avatar candidate information the attribute avatar candidate information providing unit 9 that outputs attribute avatar candidate information related to avatar candidates that share attribute information when an original image is input to the avatar candidate generation unit 4; a token generation unit 10 that generates an avatar token, which is a non-fungible token that corresponds to the avatar output by the avatar output unit 6; a transaction generation unit 11 that generates a transaction to be stored in a block in a distributed ledger corresponding to the avatar token, as information including information about the holder of the avatar token and feature point information; an electronic signature generation unit 12 that generates an electronic signature, which is data that proves that the generation and output of information included in the transaction is in accordance with the intention of the holder of the avatar token; an output unit 13 that outputs the transaction and the electronic signature to a distributed network associated with the avatar token; a feature point information acquisition unit 14 that accesses the distributed network to acquire feature point information included in the transaction; and an identity authentication unit 15 that performs this authentication process based on the acquired feature point information.

[0015] The original image input unit 1 is used to input an original image to be used as a reference image for the avatar to be generated. The original image is an image of a person to be associated with the avatar. Specifically, the original image is preferably a two-dimensional or three-dimensional image including the face image of the person, but may also be an image including only a portion of the face image, or an image of parts other than the face that have some kind of relationship with the person to be associated with the avatar. The original image input unit 1 may be specifically configured to input image data as electronic information, or may be a scanning device that reads the image displayed on a medium such as paper, or may be configured with a photographing device or three-dimensional photographing device that directly photographs the person to be associated with the avatar.

[0016] The attribute information input unit 2 is used to input attribute information, which is information about the attributes of a person associated with an avatar. The attribute information includes information about one or more of the target person's gender, age, nationality, ethnicity, educational background, and work history, and may also include information about the usage history of IT services such as SNS in which the avatar is used, and behavioral history in the IT services. As a specific configuration of the attribute information input unit 2, the person associated with the avatar may input the attribute information themselves, or the attribute information may be generated based on information about the person published on the Internet or the like.

[0017] The feature point extraction unit 3 extracts, from the original image, predetermined feature points corresponding to distinctive features on the human face, such as the eyes, nose, mouth, and ears, and skeletal connections in the limbs. The feature point extraction unit 3 may be configured using image recognition technology implemented by deep learning, machine learning, or a combination of multiple image recognition technologies. Furthermore, the "feature points" in the first embodiment indicate the positions of body surface or internal body parts, and the specific parts to be set as feature points may be arbitrarily determined. For example, feature points may be set only to visually recognizable parts on the face (eyes, nose, mouth, ears, eyebrows, etc.) or points corresponding to the contours. Alternatively, feature points may be set corresponding to facial muscles within the face or the positions of joints throughout the face. Furthermore, when feature points are set for the eyes, for example, only the center of the eye may be set as a feature point, or feature points may be set in detail for the inner corner and outer corner of the eye, the pupil, and the white of the eye. Each feature point extracted by the feature point extraction unit 3 is recorded as feature point information including information on the position of the feature point (which may be the position in the original image itself, but for reasons such as convenience in generating avatar candidates, it is preferable to use the position in a normalized coordinate system) and the significance of the feature point (for example, the point corresponds to the corner of the eye, the point corresponds to the knee joint, etc.).

[0018] The avatar candidate generation unit 4 generates avatar candidates, which are semi-finished or finished avatars, based on an original image and feature point information, which is information about feature points extracted from the original image. Specifically, the avatar candidate generation unit 4 in the first embodiment includes a part design storage unit 18 that stores part design proposals for each region corresponding to each part of the entire human body, a part design extraction unit 19 that extracts each part design from the part design storage unit 18, a part design processing unit 20 that modifies and processes the part designs extracted by the part design extraction unit 19, an overall design construction unit 21 that constructs an overall design by combining the part designs that have been modified and processed as necessary, and an avatar generation unit 22 that generates a finished avatar using the overall design. Note that, although the first embodiment employs a configuration that generates finished avatars as "avatar candidates," the avatar candidate generation unit may be configured to generate only semi-finished products rather than finished products. The term "semi-finished product" refers to an intermediate product in the avatar generation process that uses feature points extracted from an original image, and includes all intermediate products, such as electronic data resulting from image processing of the original image, electronic data obtained by converting a two-dimensional original image into three-dimensional data, and electronic data related to surface information generated based on the original image when the finished avatar is constructed using computer graphics consisting of skeletal information (bone), surface information (skin), and weight information that defines the relationship between the two. For example, the term "semi-finished product" includes electronic data generated solely by the partial design storage unit 18 and partial design extraction unit 19, electronic data that has undergone design processing by the partial design processing unit 20, and electronic data representing the overall design constructed by the overall design construction unit 21. Therefore, the avatar candidate generation unit 4 may be configured without the avatar generation unit 22 or even without the overall design construction unit 21. When the avatar candidate generation unit 4 generates avatar candidates as semi-finished products, the avatar candidates selected by the avatar candidate selection unit 5 are subjected to a predetermined process to generate avatars as finished products, which are then output by the avatar output unit 6.

[0019] The part design storage unit 18 is for storing part designs, which are designs for each part obtained by dividing the entire human body into multiple parts. In the first embodiment, a "part" may be set for each feature point, for example, or a certain area that groups multiple feature points together may be set as a "part." It is also possible to set a "part" that does not include any feature points within the area, but even in this case, it is preferable to define the relationship with feature points outside the area.

[0020] As an example of a partial design, the simplest configuration may be a design that combines a fairly wide range of elements, such as a "face." Alternatively, partial designs may be stored as separate "parts" for the eyes, nose, and even for smaller parts of the eyes, such as the pupils, eyelashes, and wrinkles at the corners of the eyes. Furthermore, since each part is a division of the entire human body into multiple parts, the partial design need not be limited to the design of the human body itself (i.e., a naked human body without clothes). It may also be a design of the human body wearing clothes, shoes, etc., or of the human body wearing accessories, or of the human body holding a foreign object in its hand. Similarly, the design of (part of) the face may be, for example, a design of the human body wearing glasses, or of the human body wearing makeup such as lipstick or blush, or of the human body wearing a mask. Furthermore, partial designs are not limited to being realistic, but may be pictorial, cartoon-like, or animated. For example, partial designs may be separated into separate parts, such as the skin and clothing, that can be detached.

[0021] In the first embodiment, the partial design storage unit 18 stores partial designs in association with similarity information, which is information indicating whether a specific partial design is similar to other partial designs. By adopting this configuration, in the partial design extraction process by the partial design extraction unit 19 based on each piece of design information provided by the related avatar candidate information providing unit 8 and the attribute avatar candidate information providing unit 9, it is possible to quickly extract not only the partial design indicated in each piece of design information, but also partial designs similar to it.

[0022] The partial design extraction unit 19 extracts partial designs necessary for generating an avatar design for each part that makes up the overall image from the partial designs stored in the partial design storage unit 18. The partial design extraction by the partial design extraction unit 19 extracts partial designs that have feature points that are somehow related to the feature points of the original image, at least for parts that correspond to feature points extracted from the original image.

[0023] For example, when extracting a partial design for a part of the face that includes the eyes, it is preferable to extract a partial design that includes the eyes that matches the size and shape of the eyes indicated by the feature points extracted from the original image, but even if there is not such a strong correlation, the partial design may be extracted based on similarity, such as similarity in the position of the center of gravity or similarity in the shape of the area formed by multiple feature points.In addition, the partial design extraction unit 19 extracts partial designs based on each avatar candidate information provided by the related avatar candidate information providing unit 8 and the attribute avatar candidate information providing unit 9 (described later).

[0024] Since the avatar candidate generation unit 4 according to the first embodiment generates multiple avatar candidates, the partial design extraction unit 19 will, in principle, extract multiple patterns for each part. However, it is possible to reuse some partial designs; for example, if the avatar candidate generation unit 4 generates 10 avatar designs, it may select 9 or fewer partial designs for some parts.

[0025] The partial design processing unit 20 enlarges or reduces the partial designs extracted by the partial design extraction unit 19 to match the size of the overall avatar design, and processes the connections between adjacent parts so that they are continuous. The partial design processing unit 20 may perform the size adjustment and boundary adjustment automatically, or the processing may be performed manually by a designer or the like. Furthermore, the partial design processing unit 20 may be configured to not only adjust the size and boundary, but also to change or modify the partial design itself under certain conditions (for example, to have a function to change the shape to match the design trends of the SNS used).

[0026] The overall design construction unit 21 constructs an overall avatar design consisting of multiple parts by combining the partial designs extracted by the partial design extraction unit 19 and processed as necessary by the partial design processing unit 20. The overall design construction unit 21 determines which of the extracted partial designs to combine and then constructs an avatar design, which is an overall design that integrates the partial designs. The partial designs may be combined randomly or according to a certain rule (e.g., preferentially combining partial designs that have commonalities). However, the avatar output device according to the first embodiment generates one or more avatar designs based on the design information provided by the related avatar candidate information providing unit 8 and the attribute avatar candidate information providing unit 9, which will be described later.

[0027] Avatar generation unit 22 is used to generate an avatar as a finished product using the overall design generated by overall design construction unit 21. Specifically, the avatar generated by avatar generation unit 22 is configured using image data such as a portrait of a target person, and in the first embodiment, it is preferable that the image includes at least a facial image of the target person. There are no limitations on the data format of the avatar, and it may be, for example, simple image data or computer graphics consisting of skeletal information (bones), surface information (skin), and weight information that defines the relationship between the two.

[0028] The avatar may be either a two-dimensional image or a three-dimensional image, and may be not only a still image but also computer graphics that perform a predetermined action according to action information. Although the avatar is an image of a person imitating a target person, it is not limited to a human body and may be an image imitating an animal or an imaginary creature. As mentioned above, the avatar candidate generation unit 4 may be configured to generate a semi-finished avatar. In this case, the avatar generation unit 22 may be omitted or a component that realizes only some of the functions of the avatar generation unit 22 may be added.

[0029] The avatar candidate selection unit 5 is configured to select an avatar candidate to be used for the avatar output by the avatar output unit 6 from among the multiple avatar candidates generated by the avatar candidate generation unit 4. The avatar candidate selection unit 5 has a function of visually displaying the multiple avatar candidates generated by the avatar candidate generation unit 4 in an array, and selecting one avatar candidate according to an instruction from a selector (e.g., a person who will use the avatar, i.e., a person associated with the avatar). Note that the specific configuration of the avatar candidate selection unit 5 is not limited to a configuration in which the avatar candidates are selected by a specific selector, but may also be a configuration in which each avatar candidate is made public on the Internet for a certain period of time, and an avatar candidate is automatically selected depending on the number of views, comments, citations (such as the number of times the avatar is linked elsewhere), etc.

[0030] The avatar output unit 6 is for outputting an avatar using the avatar candidate selected by the avatar candidate selection unit 5. Specifically, if the selected avatar candidate is a finished avatar, the avatar output unit 6 has a function of outputting the avatar as is, and if the selected avatar candidate is a semi-finished avatar, the avatar output unit 6 has a function of outputting the avatar processed into the finished product. The output mode of the avatar output unit 6 is basically output to the outside in the form of electronic data, but may also be output in a form stored in a recording medium, or in a form visually displayed on paper or a display. In addition to the configuration of outputting to the outside, if the avatar output device according to the first embodiment constitutes part of a system such as a game or SNS, it may also be configured to output to other components within the system.

[0031] The storage unit 7 stores information about an avatar candidate selected by the avatar candidate selection unit 5 from among the generated avatar candidates, in association with feature point information, which is information about feature points of the original image used to generate the avatar candidate, and attribute information about the person using the avatar. Specifically, the storage unit 7 has a function of storing, for an avatar candidate previously selected as avatar candidate information, information about the design of each part constituting the avatar candidate, information about the positional distribution of feature points extracted from the original image (in the first embodiment, feature point information is information about the positions of feature points and the meaning of each feature point (e.g., corresponding to a knee joint), and attribute information, which is information about the attributes of a person associated with the avatar shown in the original image, in association with each other. The information stored in the storage unit 7 is output to the avatar candidate generation unit 4 according to instructions from the related avatar candidate information provision unit 8 and the attribute avatar candidate information provision unit 9, which will be described later, and is used in the partial design extraction process and / or the overall design construction process.

[0032] The related avatar candidate information providing unit 8 is configured to provide the avatar candidate generation unit 4 with related avatar candidate information, which is information on avatar candidates associated with feature point information common to feature points of the original image, when an original image is input by the original image input unit 1. Specifically, when a new original image is input, the related avatar candidate information providing unit 8 has a function to extract, based on the feature point information extracted from the original image by the feature point extraction unit 3, avatar candidate information on avatar candidates associated with feature point information common to the feature points of the original image, from the avatar candidate information stored in the storage unit 7, and provide the extracted avatar candidate information to the avatar candidate generation unit 4 as related avatar candidate information.

[0033] In the context of minutiae information, "being in common" or "having commonality" most simply means that the positions (normalized) of minutiae with the same meaning all match, but this is not limiting and it is also possible to "have commonality" when the distance between minutiae with the same meaning is equal to or less than a certain value. Furthermore, instead of comparing all minutiae, it is also possible to determine whether or not there is commonality by comparing only some of the minutiae.

[0034] The related avatar candidate information providing unit 8 may extract information about the avatar candidate with the highest commonality in the feature point information, or may extract information about multiple avatar candidates with a certain level of commonality or more.Furthermore, it may be configured to preferentially extract information about avatar candidates that have been selected as related avatar candidate information in the past from among the avatar candidates with commonality.

[0035] The related avatar candidate information provided by the related avatar candidate information providing unit 8 may include image data of all or part of the design of the extracted avatar candidate, but preferably includes all or part of information specifying the design of each part corresponding to the design of the avatar candidate and stored in the part design storage unit 18. More preferably, the related avatar candidate information providing unit 8 also includes in the related avatar candidate information specific content indicating the strength of commonality in the feature point information (such as the positions of all common points being the same, the distance between common points having the same meaning being within a certain range, some feature points satisfying the commonality condition, etc.).

[0036] When generating a new avatar candidate, the avatar candidate generation unit 4, which has received the related avatar candidate information, references design information included in the related avatar candidate information, and extracts partial designs using the partial design extraction unit 19, and the overall design construction unit 21 determines the combination of partial designs. The simplest configuration may be to generate an avatar candidate that is identical to the related avatar candidate. Alternatively, the partial design extraction unit 19 may extract all of the partial designs specified in the related avatar candidate information as all or part of the design candidate, and the overall design construction unit 21 may construct an overall design consisting of the same combination of partial designs as the combination of the partial designs specified in the related avatar candidate information as all or part of the avatar design. As another configuration of the avatar candidate generation unit 4 in the first embodiment, the partial design extraction unit 19 may extract only some of the partial designs specified in the related avatar candidate information, or the overall design construction unit 21 may construct an overall design in which each partial design is combined differently.

[0037] Furthermore, the avatar candidate generation unit 4 may construct an overall design using, instead of the partial design itself included in the related avatar candidate information, another partial design that has a certain relationship to the partial design (for example, a partial design that is visually similar to a partial design included in the related avatar candidate information, or a partial design that falls into the same type (group) when the partial designs are categorized (grouped) based on predetermined classification criteria). In the first embodiment, the partial design storage unit 18 is configured to store similarity information associated with the partial designs, and the partial design extraction unit 19 is configured to extract similar partial designs based on this similarity information. However, other configurations may also be used, for example, to perform a similar image search process each time a partial design is extracted, to extract other similar partial designs.

[0038] The attribute avatar candidate information providing unit 9 is configured to, when an original image is input via the original image input unit 1, output attribute avatar candidate information including information on some or all of the designs of past avatar candidates that share attribute information with a person associated with the avatar through the original image to the avatar candidate generation unit 4. Specifically, when a new original image is input, the attribute avatar candidate information providing unit 9 has a function to, based on the attribute information input via the attribute information input unit 2 in association with the original image, extract avatar candidate information that shares attribute information from the information stored in the storage unit 7, and output the extracted avatar candidate information as attribute avatar candidate information.

[0039] In the case of attribute information, "being in common" or "having commonality" most simply means that the information content matches for all items of the attribute information, but this does not have to be limited to this. Commonality can be considered to exist if the items are within a similar range (for example, an age difference of less than 10 years), and it can also be considered to be "being in common" or "having commonality" when commonality exists for only some items rather than the entire attribute information.

[0040] The attribute avatar candidate information providing unit 9 may be configured to output, as the attribute avatar candidate information, not only information on a single avatar candidate with the highest commonality in terms of attribute information, but also information on multiple avatar candidates with a certain level of commonality or higher, or may be configured to preferentially extract information on avatar candidates that are commonly selected in multiple cases with commonality as the attribute avatar candidate information. Furthermore, the attribute avatar candidate information may be configured to include image data of all or part of the design of the extracted avatar candidate, but preferably includes all or part of information identifying the partial designs that make up the design of the avatar candidate. The attribute avatar candidate information may also include information regarding the strength of commonality in the attribute information.

[0041] When the avatar candidate generation unit 4 receives the attribute avatar candidate information and generates a new avatar candidate, it references information about the designs included in the attribute avatar candidate information, extracts partial designs in the partial design extraction unit 19, and determines the combination of partial designs in the overall design construction unit 21. The simplest configuration is to generate an avatar candidate using the designs included in the attribute avatar candidate information as is, or to extract only some of the partial designs included in the attribute avatar candidate information and construct an overall design that uses only some of the combinations. Furthermore, instead of extracting and combining the designs included in the attribute avatar candidate information themselves, the overall design may be constructed by extracting partial designs that are visually similar to the design in question or partial designs that belong to the same type (group) when the partial designs are categorized (grouped) based on predetermined classification criteria. In this embodiment 1, a configuration is adopted in which the partial design storage unit 18 stores similarity information associated with the partial designs, and a configuration is adopted in which the partial design extraction unit 19 extracts similar partial designs based on this similarity information.However, in addition to this configuration, it is also possible to extract other similar partial designs, for example by performing a similar image search process each time a partial design is extracted.

[0042] The token generation unit 10 is for generating avatar tokens, which are non-fungible tokens corresponding to the avatars output by the avatar output unit 6. "Non-fungible tokens" refer to so-called NFTs (Non-Fungible Tokens), that is, tokens that have unique data and thus cannot be substituted for other tokens, and are issued, for example, based on ERC721, a standard of Ethereum (registered trademark). The avatar tokens in the first embodiment are issued based on ERC721 or other predetermined standards, and are configured such that information about the user of the corresponding avatar and information such as feature point information extracted from an original image of the corresponding avatar are stored on the blockchain.

[0043] "Blockchain" is a technology that uses encryption technology to synchronize data between multiple computers that make up a distributed network. Specifically, each block is made up of a collection of information about tokens, such as agreed-upon transaction records, and information for connecting to other blocks (information from the previous block), and a blockchain is made up of multiple blocks linked together. Even if data is tampered with on one of the multiple computers, the correct data is selected by majority vote among the other computers, making it extremely difficult to destroy or tamper with data.

[0044] The correspondence between the avatar token and the avatar is a one-to-one link between the avatar token identifier and the avatar's identification information (identification information may be attached to each avatar, or identification information (such as name) related to the person the avatar is for may be used). More directly, the avatar token identifier and the avatar's identification information may be the same, or a correspondence may be established between them composed of different character strings, etc. Alternatively, the avatar token identifier and the avatar's identification information may be linked to URL information on the Internet where the data that constitutes the avatar is saved.

[0045] In addition, in the first embodiment, the blocks constituting the distributed ledger corresponding to the generated avatar tokens are assumed to record transaction records relating to the transfer of ownership of the avatar tokens, as well as information relating to the user of the avatar and information relating to feature point information extracted from the original image that served as the basis for the avatar. These pieces of information are converted into transactions, which are information output to the distributed network by the transaction generation unit 11.

[0046] The avatar token may be generated by the token generation unit 10 itself, or may be generated by an external system that is directly or indirectly connected to the token generation unit 10 by issuing a predetermined command to the external system. The specific format of the avatar token is not limited to one that complies with ERC721, the standard of Ethereum (registered trademark), and any format may be used as long as it is non-fungible and the transaction history is stored on the blockchain.

[0047] The transaction generation unit 11 has a function of generating transactions to be stored in blocks in the distributed ledger, and corresponds to an example of the information generation means in the claims. Specifically, the transaction generation unit 11 generates information including, as the content of the transaction, information on the user of the corresponding avatar and feature point information of the original image extracted by the feature point extraction unit 3, in addition to information on the change of the name holding the avatar token.

[0048] The electronic signature generation unit 12 has a function of generating an electronic signature, which is data that proves that the creation and output of information included in a transaction complies with the intentions of the avatar token holder. Specifically, the electronic signature generation unit 12 has a function of generating a hash value of the transaction generated by the transaction generation unit 11 and encrypting the hash value with a private key held by the avatar token holder to generate an electronic signature. Note that a "hash value" is a fixed-length value obtained by applying a specific calculation procedure to the original data. Because the calculation procedure is irreversible, it is considered impossible to restore the original data from the hash value. Furthermore, a "private key" is a sequence of numbers used in the encryption process that can be decrypted using a corresponding "public key." Note that, in the first embodiment, the transaction generation unit 11 and the electronic signature generation unit 12 are configured to generate the transaction and electronic signature, which are one form of ownership change information, by themselves. However, this is not a limitation, and the transaction generation unit 11 and the electronic signature generation unit 12 may simply have a function of instructing an external device to generate the transaction and electronic signature.

[0049] The output unit 13 instructs the transaction generated by the transaction generation unit 11 and the electronic signature generated by the electronic signature generation unit 12 to be output to a distributed network associated with the avatar token. The output unit 13 functions as one aspect of the output instruction means in the claims. In this embodiment, the output unit 13 is configured to directly output the transaction and the electronic signature, but may be configured to only instruct other components (including those provided outside the avatar output device) to output the transaction. In this embodiment, the output unit 13 is directly or indirectly connected to the distributed network and is configured to be able to output predetermined data to the distributed network. The distributed network generates a hash value of the output transaction and compares it with data obtained by decrypting the electronic signature using a public key (= the hash value of the transaction). If the two values ​​differ, the distributed network determines that the transaction is not made by the owner of the tangible object and rejects registration. If the two values ​​match, the distributed network determines that the transaction is made by the owner and stores the information composed of the transaction in a block constituting the distributed ledger. The transaction output by the output unit 13 includes information regarding the change of ownership, and by outputting the transaction to the distributed network, this information is stored in the blocks that make up the distributed ledger established on the distributed network.

[0050] The feature point information acquisition unit 14 accesses the distributed network to acquire feature point information included in a transaction. Specifically, the feature point information acquisition unit 14 acquires feature point information related to the original image of the avatar as a comparison target when verifying the identity of the person using the avatar. The feature point information acquisition unit 14 outputs the acquired feature point information to the identity authentication unit 15, and the identity authentication unit 15 uses the feature point information to perform authentication processing for the avatar user.

[0051] The personal authentication unit 15 is for performing the authentication process based on the acquired feature point information. The personal authentication unit 15 performs a feature point extraction process on the image of the person using the avatar and performs the personal authentication process by comparing the extracted feature points with the feature points of the original image provided by the feature point information acquisition unit 14. The image of the person using the avatar may be obtained through a camera function provided in the information terminal used by the person, or may be obtained by directly photographing the person in real life at a predetermined location with a camera or the like. The feature point extraction process of the person's image is preferably performed using the same process as that of the feature point extraction unit 3; for example, the person's image may be output from the personal authentication unit 15 to the feature point extraction unit 3, and the information extracted by the feature point extraction unit 3 may then be used.

[0052] Next, the process up to avatar output in the avatar output device according to the first embodiment will be described. As shown in Fig. 2, the feature point extraction unit 3 extracts feature point information based on the input original image (step S101). Based on the extracted feature point information, the related avatar candidate information providing unit 8 selects, from previously selected avatar candidates, a past avatar candidate associated with feature point information of the original image from which the past avatar candidate was generated that has commonality with the feature point information extracted in step S101 (step S102). The avatar candidate information on the selected avatar candidate is provided to the avatar candidate generation unit 4 as related avatar candidate information (step S103). Based on information on the design of the avatar candidate included in the related avatar candidate information, the avatar candidate generation unit 4 extracts a partial design stored in the partial design storage unit 18, and generates a new avatar candidate after processing the extracted partial design as necessary (step S104).

[0053] Thereafter, based on the attribute information input via the attribute information input unit 2, the attribute avatar candidate information providing unit 9 selects an avatar candidate associated with attribute information that has commonality with the input attribute information from among previously created avatar candidates (step S105), and provides avatar candidate information on the selected avatar candidate as attribute avatar candidate information to the avatar candidate generation unit 4 (step S106).The avatar candidate generation unit 4 extracts partial designs stored in the partial design storage unit 18 based on information on partial designs and their combinations included in the attribute avatar candidate information, and generates avatar candidates after processing the extracted partial designs as necessary (step S107).

[0054] Thereafter, the avatar candidate generation unit 4 outputs a plurality of avatar candidates, including the avatar candidates generated in steps S104 and S107, to the avatar candidate selection unit 5, which then selects one avatar candidate from the plurality of avatar candidates (step S108). An avatar using the selected avatar candidate is output by the avatar output unit 6 (step S109), and avatar candidate information about the selected avatar candidate is associated with feature point information about the original image and person attribute information and then stored in the storage unit 7 (step S110). The avatar candidate information stored in the storage unit 7 is used when extracting past avatar candidate information in steps S101 and S105.

[0055] Next, advantages of the avatar output device according to the first embodiment will be described. First, the avatar output device according to the first embodiment employs a configuration in which the avatar candidate generation unit 4 generates new avatar candidates after receiving related avatar candidate information including information about the designs of avatar candidates associated with feature point information that has commonality with feature points of the input original image from among previously selected avatar candidates. By employing such a configuration, the avatar output device according to the first embodiment has the advantage that, when generating a new avatar candidate, it can generate an avatar candidate using the design of another avatar candidate that has commonality in feature points of the original image and has been selected as an avatar candidate in the past. When the feature points of the original images are common, reusing the design of a previous avatar candidate for a new avatar candidate does not create a sense of incongruity in design, and an avatar candidate with a natural design for the user can be generated. Furthermore, reusing the design of a previously selected avatar candidate makes the newly generated avatar candidate attractive and likely to be selected by the user, thereby advantageously presenting appropriate candidates for avatar selection.

[0056] Furthermore, the avatar output device according to the first embodiment is configured such that the avatar candidate generation unit 4 generates an avatar design after receiving attribute avatar candidate information, which is information about avatar designs that have commonality in attribute information of a person associated with the avatar to be generated from among previously selected avatar designs. By adopting such a configuration, the avatar output device according to the first embodiment can utilize avatar candidate designs previously selected for people with common attributes when generating a new avatar design. Since people with common attributes tend to have common design preferences compared to people without common attributes, generating new avatar candidates based on attribute avatar candidate information has the advantage of being able to present avatar candidates that are in line with the design preferences of the person associated with the avatar, i.e., appropriate avatar candidates that are likely to be selected by that person.

[0057] Furthermore, the avatar device according to the first embodiment employs a configuration in which personal authentication processing is performed using feature point information extracted from an original image. By employing such a configuration, it is possible to realize authentication processing without separately preparing identification data for personal authentication. Furthermore, by employing a configuration in which information relating to the feature point information extracted from an original image is included in a transaction and stored on a distributed network related to an avatar token that corresponds to the avatar, it is possible to prevent tampering with the feature point information, thereby advantageously preventing impersonation during personal authentication.

[0058] (Embodiment 2) Next, an avatar output device according to embodiment 2 will be described. In embodiment 2, components with the same names and symbols as those in embodiment 1 will have the same functions as those in embodiment 1, unless otherwise specified.

[0059] 3 is a schematic diagram showing the configuration of an avatar output device according to embodiment 2. As shown in Fig. 3, the avatar output device according to embodiment 2 includes, in addition to the same components as those of the avatar output device according to embodiment 1, an avatar candidate generation unit 23 that generates avatar candidates by changing the positions of feature points extracted from an original image rather than generating avatar designs by extracting and combining partial designs stored in a partial design storage unit, an associated avatar candidate information providing unit 24 that provides the avatar candidate generation unit 23 with information including position information of all or some of the feature points of the avatar candidates as associated avatar candidate information, an attribute avatar candidate information providing unit 25 that provides the avatar candidate generation unit 23 with information including position information of all or some of the feature points of the avatar candidates as attribute avatar candidate information, and a storage unit 26 that stores information including position information of feature points on the previously selected avatar designs extracted by the associated avatar candidate information providing unit 24 and the attribute avatar candidate information providing unit 25.

[0060] In the second embodiment, the avatar candidate generation unit 23 does not select a pre-stored partial design itself, but generates avatar candidates by appropriately adjusting the positions of feature points based on position information of feature points extracted from the original image. More specifically, the avatar candidate generation unit 23 includes a reference image storage unit 27 that stores information about feature points of one or more reference images different from the original image, a feature point position adjustment unit 28 that sets the positions of feature points in the partial designs to positions obtained by dividing the positions of feature points in the reference image and the positions of feature points in the original image at a certain ratio, a partial design generation unit 29 that generates partial designs based on the feature point positions adjusted by the feature point position adjustment unit 28, and an overall design construction unit 30 that combines the generated partial designs to construct an overall avatar design. In the second embodiment, a reference image is separately prepared as an index, and feature points are positioned at positions obtained by dividing the positions of feature points in the reference image and the positions of corresponding feature points in the original image at a certain ratio, thereby making it possible to generate partial designs and avatar designs that reflect the features of the reference image while maintaining the features of the original image.

[0061] Even when such a configuration is adopted, as in embodiment 1, the avatar output device of embodiment 2 can provide appropriate avatar designs that reflect the characteristics of the original image to a certain extent as selection candidates, and also has the advantage of being able to provide avatar designs that suit the design preferences of people with specific attributes.

[0062] Although the present invention has been described above in connection with the first and second embodiments, the technical scope of the present invention should not be construed as being limited to the specific configurations described in the embodiments. As long as the functions of the present invention can be realized, various modifications and applications of the above-described embodiments also fall within the technical scope of the present invention.

[0063] For example, the avatar candidate generation unit may generate a partial design for a specific part using avatar candidate generation unit 23, select existing partial designs for other parts using avatar candidate generation unit 4, and then combine the two to generate an overall avatar design. Also, the avatar candidate generation unit 23 may omit the reference image storage unit 27 and adopt a configuration in which the positions of feature points in the original image are arbitrarily changed by feature point position adjustment unit 28.

[0064] Furthermore, the avatar candidate generation units 4 and 23 may not only generate the entire avatar design, but may also generate a partial design, such as only the face or the entire head. In this case, the avatar output unit 6 may use an existing design for the other parts and then generate a single avatar. [Industrial Applicability]

[0065] The present invention can be used as a technique for generating an avatar of an avatar user after presenting multiple candidates based on an original image related to the user. [Explanation of symbols]

[0066] 1. Original image input section 2. Attribute information input section 3. Feature point extraction section 4, 23 Avatar candidate generation unit 5 Avatar candidate selection section 6 Avatar output section 7, 26 Memory section 8, 24 Related Avatar Candidate Information Department 9, 25 Attribute Avatar Candidate Information Providing Department 10 Token Generation Unit 11 Transaction Generation Unit 12 Electronic signature generation section 13 Output section 14. Feature point information acquisition unit 15 Personal authentication unit 18 Partial Design Memory 19 Partial Design Extraction Unit 20 Partial Design Processing Department 21, 30 Overall Design and Construction Department 22 Avatar Generation Unit 27 Reference image storage unit 28 Feature point position adjustment unit 29 Partial Design Generation Unit

Claims

1. An avatar output device that outputs an avatar using one avatar candidate selected from a plurality of avatar candidates that are semi-finished or finished avatars, feature point extraction means for extracting feature points of an original image; an avatar candidate generation means for generating a plurality of avatar candidates, which are semi-finished or finished avatars, from the original image; an avatar output means for outputting an avatar using one avatar candidate selected from a plurality of avatar candidates; a storage means for storing avatar candidate information, which is information about the selected avatar candidate, in association with feature point information, which is information about the feature points extracted by the feature point extraction means, and attribute information, which is information about the attributes of a person associated with the generated avatar through an original image; a related avatar candidate information providing means for extracting, from the avatar candidate information stored in the storage means, related avatar candidate information which is avatar candidate information associated with feature point information having commonality with the feature points extracted by the feature point extracting means, and providing avatar candidates based on the extracted related avatar candidate information to the avatar candidate generating means; an attribute avatar candidate information providing means for providing the avatar candidate generating means with attribute avatar candidate information, which is avatar candidate information associated with attribute information having commonality with attributes of a person associated with a newly generated avatar through an original image, from among the avatar candidate information stored in the storage means; Equipped with The avatar output device is characterized in that the avatar candidate generation means generates avatar candidates including designs that are identical and / or similar to all or part of the designs of avatar candidates corresponding to the related avatar candidate information provided by the avatar candidate information providing means, and avatar candidates including designs that are identical and / or similar to all or part of the designs of avatar candidates corresponding to the attribute avatar candidate information provided by the attribute avatar candidate information providing means.

2. An avatar output method for outputting an avatar using one avatar candidate selected from a plurality of avatar candidates that are semi-finished or finished avatars, comprising: a feature point extraction step of extracting feature points of the original image; an avatar candidate generation step of generating a plurality of avatar candidates, which are semi-finished or finished avatars, from the original image; an avatar output step of outputting an avatar using one avatar candidate selected from the plurality of avatar candidates; a storage step of storing avatar candidate information, which is information about the selected avatar candidate, in association with feature point information, which is information about the feature points extracted in the feature point extraction step, and attribute information, which is information about attributes of a person associated with the generated avatar through an original image; a related avatar candidate information providing step of extracting related avatar candidate information, which is avatar candidate information associated with feature point information having commonality with the feature points extracted in the feature point extracting step, from the avatar candidate information stored in the storing step, and providing avatar candidates based on the extracted related avatar candidate information to the avatar candidate generating step; an attribute avatar candidate information providing step of providing attribute avatar candidate information, which is avatar candidate information associated with attribute information having commonality with attributes of a person associated with a newly generated avatar through an original image, from among the avatar candidate information stored in the storing step, to the avatar candidate generating step; Including, The avatar output method is characterized in that, in the avatar candidate generation step, avatar candidates including designs identical and / or similar to all or part of the designs of avatar candidates corresponding to the related avatar candidate information provided in the avatar candidate information providing step and avatar candidates including designs identical and / or similar to all or part of the designs of avatar candidates corresponding to the attribute avatar candidate information provided in the attribute avatar candidate information providing step are generated.

3. An avatar output program that causes a computer to output an avatar using one avatar candidate selected from a plurality of avatar candidates that are semi-finished or finished avatars, to the computer, A feature point extraction function that extracts feature points from the original image; an avatar candidate generation function for generating a plurality of avatar candidates, which are semi-finished or finished avatars, from the original image; an avatar output function for outputting an avatar using one avatar candidate selected from a plurality of avatar candidates; a storage function for storing avatar candidate information, which is information about the selected avatar candidate, in association with feature point information, which is information about the feature points extracted by the feature point extraction function, and attribute information, which is information about the attributes of a person associated with the generated avatar through an original image; and a related avatar candidate information providing function that extracts related avatar candidate information, which is avatar candidate information associated with feature point information having commonality with the feature points extracted by the feature point extracting function, from the avatar candidate information stored in the storage function, and provides avatar candidates based on the extracted related avatar candidate information to the avatar candidate generating function; an attribute avatar candidate information providing function that provides the avatar candidate generation function with attribute avatar candidate information, which is avatar candidate information associated with attribute information having commonality with attributes of a person associated with a newly generated avatar through an original image, from among the avatar candidate information stored in the storage function; Execute The avatar candidate generation function generates avatar candidates including designs identical and / or similar to all or part of the designs of avatar candidates corresponding to the related avatar candidate information provided from the avatar candidate information providing function, and avatar candidates including designs identical and / or similar to all or part of the designs of avatar candidates corresponding to the attribute avatar candidate information provided from the attribute avatar candidate information providing function.

Images