KR102475823B1 - NFT-based metaverse clothing information generation system and method therefor - Google Patents

Abstract

The present invention relates to an NFT-based metaverse-wearing clothing information generation system and a method therefor, configured to generate metaverse clothing information by generating a 3D clothing model based on generated clothing design information and changing the generated 3D clothing model according to the size, body type, length, shape, and texture of a fitting target metaverse character so that clothing can be worn on the fitting target metaverse character, and configured to perform non-fungible tokenization by matching the generated metaverse clothing information with a blockchain-based cryptocurrency so as to prevent counterfeiting and enable transactions regardless of platforms. The NFT-based metaverse-wearing clothing information generation system can be used to: generate, with an operation, by a plurality of fashion companies, of sensing clothing produced thereby simply through a sensing device and selecting a desired metaverse character, the sensed clothing into a clothing item capable of being realized in the metaverse and enable the clothing item to be worn on the metaverse character; convert a clothing item into a non-fungible taken so that the clothing item can be traded; and provide functions of promoting actual clothing, providing information about the same, and enabling transactions of the same.

Description

NFT-based metaverse clothing information generation system and method {NFT-based metaverse clothing information generation system and method therefor}

The present invention relates to a technology for generating NFT-based metaverse wearing clothing information. More specifically, clothing design information is generated by sensing an actual clothing product, and a clothing 3D model is created based on the generated clothing design information. Change the clothing 3D model to fit the size, body type, length, shape, and texture of the metaverse character to be fitted to create metaverse clothing information that can be worn on the fitting metaverse character, prevent forgery, and correlate with the platform The purpose is to provide an NFT-based metaverse wearing information generation system and method that performs non-fungible tokenization by matching with blockchain-based cryptocurrencies to enable transactions without

Metaverse is a compound word of the English word 'Meta', which means 'virtual' and 'transcendence', and 'Universe', which means the universe. refers to the virtual world of the dimension. Metaverse is a concept that has evolved one step further from virtual reality (VR, a cutting-edge technology that allows people to experience the same as real life in a virtual world created by a computer). It is characterized by being able to do social and cultural activities that are the same as in real life.

In particular, the metaverse began to spread in the context of the commercialization of ultra-high-speed, hyper-connectivity, and ultra-low latency 5G and the COVID-19 pandemic that hit the world in 2020. In other words, with the commercialization of 5G, technologies capable of realizing virtual reality (VR), augmented reality (AR), and mixed reality (MR) have developed, and as non-face-to-face and online trends spread due to the COVID-19 crisis, the metaverse is attracting attention. There is.

Due to the growth of the metaverse, blockchain technology, and the development of computing power and network performance to support it, the scale and interest in item trading in the metaverse have recently been greatly expanded, and metahumans (virtual humans) who are active in the advertising world ) are active, and users are enthusiastic about the products worn and used by these metahumans.

Through the present invention, by sensing actual clothing and automatically generating NFT-based metaverse wearing clothing information, it is possible to automatically wear on real clothing-based metaverse characters without the help of engineers knowledgeable in blockchain and metaverse platforms. It is possible to create metaverse wearing clothing information, and furthermore, perform non-fungible tokenization for metaverse wearing clothing information and link it with virtual currency of the block chain, so that it can be freely traded outside as well as inside a specific metaverse platform. It is possible to implement a system that can provide not only clothes worn in the metaverse as an implemented item, but also information on actual clothes and a function to trade them.

According to an embodiment of the present invention, the NFT-based metaverse wearing clothing information generation system generates clothing design information that generates clothing design information including length, size, ratio, texture, and shape information by sensing shape data of actual clothing. wealth;

Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse; a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information; A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model; Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit; According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model; Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; The metaverse clothing information is generated by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information with at least one blockchain-based cryptocurrency to obtain NFT clothing information It may include an NFT clothing information generation unit that generates.

According to an embodiment of the present invention, the NFT-based metaverse wearing clothing information generation system generates clothing design information that generates clothing design information including length, size, ratio, texture, and shape information by sensing shape data of actual clothing. wealth; Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse; a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information; A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model; Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit; According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model; Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; The metaverse clothing information is generated by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information with at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generation unit that generates a feature point connection structure model generation unit, the artificial neural network-based feature point connection structure conversion model includes an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module. The area segmentation module receives a character 3D model and a 3D clothing model as input to an artificial neural network-based body part segmentation model, and divides each body part of the character 3D model and the clothing 3D model into a plurality of body areas, Among the divided body areas, the body area included in the clothing 3D model is selected as at least one feature point extraction target area to generate feature point extraction target area information, and the feature point extraction module extracts the character 3D model, the clothing 3D model, and the feature point After receiving target area information, a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area and designated as feature points to generate feature point information for each area for each 3D model, and the structure creation module , Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, to express the three-dimensional shape and volume. A character feature point connected structure model and a clothing feature point connected structure model may be created by forming a feature point connected structure capable of this.

According to an embodiment of the present invention, the NFT-based metaverse wearing clothing information generation system generates clothing design information that generates clothing design information including length, size, ratio, texture, and shape information by sensing shape data of actual clothing. wealth; Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse; a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information; A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model; Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit; According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model; Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; The metaverse clothing information is generated by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information with at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generation unit that generates a feature point connection structure model generation unit, the artificial neural network-based feature point connection structure conversion model includes an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module. The area segmentation module receives a character 3D model and a 3D clothing model as input to an artificial neural network-based body part segmentation model, and divides each body part of the character 3D model and the clothing 3D model into a plurality of body areas, Among the divided body areas, the body area included in the clothing 3D model is selected as at least one feature point extraction target area to generate feature point extraction target area information, and the feature point extraction module extracts the character 3D model, the clothing 3D model, and the feature point After receiving target area information, a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area and designated as feature points to generate feature point information for each area for each 3D model, and the structure creation module , Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, to express the three-dimensional shape and volume. A character feature point connected structure model and a clothing feature point connected structure model are formed by forming a feature point connected structure capable of this, and the wearing clothing 3D model generation unit determines the positions of a plurality of feature points included in the clothing feature point connected structure model. In each change, a vector value for the position of the changed feature point is calculated for each feature point, and the feature point connected structure of the clothing feature point connected structure model is on a straight line. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is Position change can be limited so that it is located on a curve with a constant curvature.

According to an embodiment of the present invention, the NFT-based metaverse wearing clothing information generation system generates clothing design information that generates clothing design information including length, size, ratio, texture, and shape information by sensing shape data of actual clothing. wealth; Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse; a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information; A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model; Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit; According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model; Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; The metaverse clothing information is generated by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information with at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generation unit that generates a feature point connection structure model generation unit, the artificial neural network-based feature point connection structure conversion model includes an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module. The area segmentation module receives a character 3D model and a 3D clothing model as input to an artificial neural network-based body part segmentation model, and divides each body part of the character 3D model and the clothing 3D model into a plurality of body areas, Among the divided body areas, the body area included in the clothing 3D model is selected as at least one feature point extraction target area to generate feature point extraction target area information, and the feature point extraction module extracts the character 3D model, the clothing 3D model, and the feature point After receiving target area information, a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area and designated as feature points to generate feature point information for each area for each 3D model, and the structure creation module , Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, to express the three-dimensional shape and volume. A character feature point connected structure model and a clothing feature point connected structure model are formed by forming a feature point connected structure capable of this, and the wearing clothing 3D model generation unit determines the positions of a plurality of feature points included in the clothing feature point connected structure model. In each change, a vector value for the position of the changed feature point is calculated for each feature point, and the feature point connected structure of the clothing feature point connected structure model is on a straight line. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is Position change is limited to be located on a curve with a constant curvature, and the feature point extraction module includes a first module, a second module, and a third module composed of a plurality of convolutional layers, and the first module is composed of a plurality of convolutional layers. The character 3D model, clothing 3D model, and feature point extraction target area information are input to the first deep learning network formed through the product layer, and a plurality of first feature points are output by deriving a preset number of feature points for the feature point extraction target area information, and the second generation module receives the character 3D model, clothing 3D model, and feature point extraction target area information to the second deep learning network formed through a plurality of convolutional layers, and applies the feature point extraction target area to the target area. A plurality of second feature point output information is generated by deriving a preset number of feature points, and the third generation module receives the first feature point output information and the second feature point output information and selects the first feature point output information and the second feature point output information. 2 A weight of 0 to 1 or less is applied to each of the feature point output information, and the first feature point output information and the second feature point output information to which the weight is applied are normalized to a value within a predetermined range, and the number preset for the feature point extraction target area as many feature points as possible.

According to an embodiment of the present invention, the NFT-based metaverse wearing clothing information generation system generates clothing design information that generates clothing design information including length, size, ratio, texture, and shape information by sensing shape data of actual clothing. wealth; Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse; a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information; A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model; Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit; According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model; Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; The metaverse clothing information is generated by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information with at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generation unit that generates, wherein the feature point connection structure model generation unit,

The artificial neural network-based feature point connection structure transformation model may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module, and the area segmentation module is a body part segmentation model based on an artificial neural network. Receives a character 3D model and a clothing 3D model, divides each body part of the character 3D model and clothing 3D model into a plurality of body areas, and extracts at least one feature point from the body area included in the clothing 3D model among the divided body areas Selecting a target area as a target area to generate feature point extraction target area information, and the feature point extraction module receives information of a character 3D model, a 3D clothing model, and a feature point extraction target area, and forms a body shape targeting the feature point extraction target area By deriving a preset number of vertices and designating them as feature points, the feature point information for each area is generated for each 3D model, and the structure generation module receives the character 3D model, clothing 3D model, and feature point information for each area, and creates a character 3D model A plurality of feature points derived from the clothing 3D model are connected to each other with a plurality of straight lines or curves to form a feature point connection structure that can express a three-dimensional shape and volume to form a character feature point connection structure model and a clothing feature point connection structure A structural model is generated, and the wearing clothing 3D model generating unit calculates a vector value for the location of the changed feature point for each feature point when changing the location of a plurality of feature points included in the clothing feature point connection structure model, respectively, and clothing Even after multiplying each vector value by the feature points located on a straight line in the feature point connected structure of the feature point connected structure model, each feature point is located on a straight line, and the feature point connected structure of the clothing feature point connected structure model is on a curve with a certain curvature. Even after multiplying the feature points located at by each vector value, the position change is limited so that each feature point is located on a curve with a constant curvature, and the feature point extraction module It includes a first module, a second module, and a third module composed of convolutional layers of , wherein the first module transmits the character 3D model, clothing 3D model, and feature points to a first deep learning network formed through a plurality of convolutional layers. Receives extraction target area information, derives a preset number of feature points from the feature point extraction target area, and generates a plurality of first feature point output information, and the second generation module generates second output information formed through a plurality of convolution layers The character 3D model, the clothing 3D model, and feature point extraction target area information are input to the deep learning network, and a plurality of second feature point output information is generated by deriving a preset number of feature points for the feature point extraction target area, and the second feature point output information is generated. 3 The generation module receives the first feature point output information and the second feature point output information, applies a weight of 0 to 1 or less, respectively, to the selected first feature point output information and the second feature point output information, The feature point output information and the second feature point output information are normalized to a value within a predetermined range to derive a preset number of feature points for the feature point extraction target area, and the NFT clothing information is obtained by using a blockchain-based smart contract. Actual clothing seller included in, actual clothing transaction amount, actual clothing size information. It may further include a clothing sales performer that provides information on actual clothing and a sales function using actual clothing image information.

Using the NFT-based metaverse wearing clothing information generation system implemented according to the embodiment of the present invention, a number of fashion companies can sense clothing produced by themselves through a simple sensing device and select a desired metaverse character. Sensed clothing can be created as a clothing item that can be implemented in the metaverse and worn on a character in the metaverse, and the clothing item can be converted into a non-fungible token to provide a trading function for the clothing item, furthermore real clothing There is an effect that can provide publicity, information provision, and even trading functions.

1 is a block diagram of an NFT-based metaverse wearing clothing information generation system implemented according to a first embodiment of the present invention.
2 is a diagram illustrating an internal data flow of a feature point connection structure conversion model based on an artificial neural network implemented according to an embodiment of the present invention.
3 is a diagram illustrating an internal data flow of a feature point extraction module implemented according to an embodiment of the present invention.
4 is a diagram showing a network structure of first, second, and third modules of a feature point extraction module implemented according to an embodiment of the present invention.
5 is a block diagram of an NFT-based metaverse wearing clothing information generation system implemented according to a second embodiment of the present invention.
6 is a flowchart of a method for generating NFT-based metaverse wearing clothing information according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily carry out the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein.

Terms used in the present invention are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise.

In the present invention, terms such as "comprise" or "having" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present invention, they should not be interpreted in an ideal or excessively formal meaning. don't

It will also be understood that combinations of each block of the drawings and flowchart drawings can be performed by computer program instructions, and these computer program instructions can be loaded into a processor of a general-purpose computer, special-purpose computer, or other programmable data processing equipment. Thus, those instructions executed by a processor of a computer or other programmable data processing equipment create means for performing the functions described in the flowchart block(s).

These computer program instructions may also be stored in a computer usable or computer readable memory that can be directed to a computer or other programmable data processing equipment to implement functionality in a particular way, such that the computer usable or computer readable memory The instructions stored in are also capable of producing an article of manufacture containing instruction means that perform the functions described in the flowchart block(s).

The computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operational steps are performed on the computer or other programmable data processing equipment to create a computer-executed process to generate computer or other programmable data processing equipment. Instructions for performing processing equipment may also provide steps for performing the functions described in the flowchart block(s).

Additionally, each block may represent a module, segment, or portion of code that includes one or more executable instructions for executing specified logical function(s).

And it should be noted that in some alternative embodiments it is also possible for the functions mentioned in the blocks to occur out of order. For example, two blocks shown in succession may in fact be executed substantially concurrently, or the blocks may sometimes be executed in reverse order depending on their function.

At this time, the term '~unit' used in this embodiment means software or a hardware component such as a field-programmable gate array (FPGA) or application specific integrated circuit (ASIC), and what role does '~unit' have? perform them

However, '~ part' is not limited to software or hardware. '~bu' may be configured to be in an addressable storage medium and may be configured to reproduce one or more processors.

Therefore, as an example, '~unit' refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, and procedures. , subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables. Functions provided within components and '~units' may be combined into smaller numbers of components and '~units' or further separated into additional components and '~units'. In addition, components and '~units' may be implemented to play one or more CPUs in a device or a secure multimedia card.

In describing the embodiments of the present invention in detail, an example of a specific system will be the main target, but the main subject matter to be claimed in this specification extends the scope disclosed herein to other communication systems and services having a similar technical background. It can be applied within a range that does not deviate greatly, and this will be possible with the judgment of those skilled in the art.

Hereinafter, a NFT-based metaverse wearing clothing information generation system and method according to an embodiment of the present invention will be described with reference to the drawings.

1 is a block diagram of an NFT-based metaverse wearing clothing information generation system implemented according to a first embodiment of the present invention.

Referring to FIG. 1, the NFT-based metaverse wearing clothing information generating system 1000 implemented according to the first embodiment of the present invention includes a clothing design information generating unit 100, a character appearance information generating unit 200, and a 3D model Generation unit 300, feature point connection structure model generation unit 400, feature point matching information generation unit 500, worn clothing 3D model generation unit 600, worn clothing meta information generation unit 700, NFT clothing information generation may include section 800 .

The clothing design information generation unit 100 may generate clothing design information including length, size, ratio, texture, and shape information by sensing shape data of actual clothing.

According to an embodiment of the present invention, three-dimensional numerical data of clothes such as length, size, ratio, texture, and shape information may be collected through a depth camera or a 3D scanner to sense shape data of actual clothes.

According to an embodiment of the present invention, a depth camera is used to photograph real clothes, and a pulse-modulated infrared (IR) beam is emitted to the target at the origin of the spherical coordinate system, so that the horizontal (pan, φ) and vertical (tilt) . 3D image information of a target in the background can be obtained based on time.

According to an embodiment of the present invention, a plurality of depth cameras may be installed based on a predetermined angle so as to photograph the 3D shape of at least one garment in all directions, and the image data and the image data may be preset based on time. Collected shape information may be generated by collecting transmission/reception time data for each part of the set clothing.

Here, the preset clothing part means a part of the clothing that can identify the characteristics of the clothing type, for example, sleeve, collar, wrist, elbow, shoulder, neck, hip, knee, and ankle. can be set around them.

According to an embodiment of the present invention, the clothing design information generating unit 100 may perform two-step depth measurement when collecting transmission/reception time data using a depth camera.

According to an embodiment of the present invention, two-step depth measurement may be divided into a first depth measurement step and a second depth measurement step.

The reason why the depth measurement is performed in two stages as in the above embodiment is that when a single frequency is used and depth measurement is performed using a low modulation frequency, the maximum range is widened but the measurement quality is degraded. , in the case of performing depth measurement using a high modulation frequency, the measurement quality can be improved, but there is a disadvantage that the maximum range is narrowed.

According to an embodiment of the present invention, in the first depth measurement, the depth may be measured using a low modulation frequency, and measurement may be performed with low measurement quality for a pre-set region of interest for each body part.

According to an embodiment of the present invention, the second depth measurement may increase measurement precision with high measurement quality by measuring the depth using a high modulation frequency based on the measurement result of the region of interest for each body part of the first depth measurement. have.

According to an embodiment of the present invention, the first depth measurement using a low modulation frequency to have a wide maximum depth range can measure an approximate depth targeting a wide depth range, and the measurement quality at this time is proportional to the modulation frequency. , the progress measured through the first depth measurement may provide low measurement quality for a wide area of interest.

According to an embodiment of the present invention, in the second depth measurement, a maximum depth range may be set based on the precision of the first depth measurement, and a relatively higher frequency is selected to provide high measurement quality for a narrow region of interest. An error in the first depth measurement result may be compensated for.

According to an embodiment of the present invention, a low modulation frequency may be selected based on Equation 1

는 낮은 변조 주파수를 의미하며,

는 광속,

는 최대 깊이 범위를 의미할 수 있다.here

denotes a low modulation frequency,

is the speed of light,

may mean a maximum depth range.

According to an embodiment of the present invention, a high modulation frequency may be selected as a value that is inversely proportional to the standard deviation measured when a low modulation frequency is used, and if the standard deviation is smaller than a preset limit value, this indicates a high signal-to-noise ratio. By determining, a relatively higher frequency may be selected than when the standard deviation is larger than a preset limit value.

)는 수학식 2를 기반으로 산출될 수 있다.where the standard deviation (

) can be calculated based on Equation 2.

는 산출하고자 하는 표준 편차를 의미할 수 있으며, dp는 제1 깊이 측정으로 통해 측정된 깊이를 의미할 수 있으며 μ는 관심 영역(RoI)에 대한 dp의 평균값을 의미하며 N은 관심 영역 내의 픽셀의 개수로서 자연수일 수 있다.here

May mean the standard deviation to be calculated, dp may mean the depth measured through the first depth measurement, μ means the average value of dp for the region of interest (RoI), and N is the number of pixels within the region of interest. The number may be a natural number.

According to an embodiment of the present invention, measurement accuracy may be increased through repeated measurement by performing a plurality of first depth measurement steps and second depth measurement steps.

According to an embodiment of the present invention, the clothing design information generating unit 100 inputs the image data for each depth camera included in the collected shape information to an artificial neural network-based shape information generation model, and outlines and outlines for each part of clothing. First shape information including angle information may be generated.

In addition, a plurality of consecutive depth data, velocity data, and amplitude data are generated by analyzing the transmission and reception time data of each depth camera included in the collected shape information, and a cumulative distribution function and a noise removal function according to the size of the amplitude using the amplitude data Second shape information may be generated by removing noise using , and averaging depth data, velocity data, and amplitude data for each part of clothing generated for each depth camera from which the noise is removed with a median value.

According to an embodiment of the present invention, the artificial neural network-based shape information generation model receives image data for each depth camera as input to an input layer and outputs outline and outline angle information for each pre-set garment part, which can be output for each garment. It may be learned to generate first shape information by grouping by part. According to an embodiment of the present invention, transmission/reception time data for each body part generated for each depth camera may be analyzed to generate continuous depth data, velocity data, and amplitude data for each body part over time.

Here, the depth data may refer to data capable of recognizing the three-dimensional position of the region of interest, and the speed data may refer to a movement speed calculated through a change in position over time based on the depth data. , The amplitude data may mean the light intensity of the surface caused by the reflection of the damaged light because the light intensity of the surface of the light transmitted by the depth camera is different depending on the degree of amplitude of the surface.

According to an embodiment of the present invention, noise can be removed using a cumulative distribution function and a noise function according to the amplitude of amplitude data.

According to the above embodiment, the amplitude data may represent how bright the light reflected by the depth camera is, and this is based on the fact that light intensity of the surface appears differently according to the degree of amplitude of the surface.

According to an embodiment of the present invention, filtering may be performed using a cumulative distribution function and a noise function to remove noise caused by a non-ideal waveform used in a depth camera and noise generated at a boundary of an object.

According to an embodiment of the present invention, the cumulative distribution function may be performed based on Equation 3.

Φ(x) represents a cumulative distribution function and NF may mean a noise function.

According to an embodiment of the present invention, the noise function NF may be performed based on Equation 4.

According to an embodiment of the present invention, the clothing design information generating unit 100 calculates an error by matching the first shape information and the second shape information based on time, and uses a continuous weighted median filter for the calculated error. It can be used to remove errors and create final action information.

According to an embodiment of the present invention, the first shape information and the second shape information are input to a final shape information generation model based on an artificial neural network, and the outline and shape of the clothing part included in the first shape information and the second shape information Contrasting line angle information calculates the difference as an error, applies a weight to each of the data where the error occurs, and averages the calculated values with the median value to remove the error. It is possible to create final shape information including

The character appearance information generation unit 200 may generate character appearance information by receiving size, body shape, length, shape, and texture information of a fitting target metaverse character from a metaverse server.

According to one embodiment of the present invention, character appearance information may mean modeling and numerical data for characters that can be implemented in the metaverse stored in the main server of the platform implementing the metaverse, and the size and body shape of the character. , length, shape, and texture can be used without any restrictions on the format.

The 3D model generation unit 300 may generate a character 3D model by modeling the appearance of the character in 3D based on the character appearance information, and generate a 3D clothing model by modeling the appearance of the clothing in 3D using clothing design information. have.

According to an embodiment of the present invention, a parametric technique of constructing a function expression inputting a plurality of independent parameters by analyzing character appearance information and clothing appearance information and performing 3D modeling based on a mathematical free curve using such a function expression Character 3D models and clothing 3D models can be created using

The feature point connection structure model generation unit 400 extracts a plurality of feature points for the appearance of each 3D model by inputting a character 3D model and a 3D costume model to an artificial neural network-based feature point connection structure conversion model, and extracts a plurality of extracted feature points. It is possible to create a character feature point connected structure model and a clothing feature point connected structure model converted into a feature point connected structure model expressed as a shape structure by connecting feature points.

According to an embodiment of the present invention, the feature point connection structure model generation unit 400 designs a feature point connection structure conversion model composed of a plurality of convolution operation layers, and the designed feature point connection structure conversion model is a character 3D model and clothes Learning may be performed by inputting learning data so that a 3D model is input and a character feature point connected structure model and a clothing feature point connected structure model are output.

According to one embodiment of the present invention, the feature point concatenation structure transformation model may be designed so that a plurality of convolution operation layers are connected as one group, but according to another embodiment, a plurality of groups are connected and each group is separately designed. It can be designed to exist in units of functional modules that can perform the functions of.

According to an embodiment of the present invention, the feature point connection structure model generation unit 400 uses an artificial neural network-based feature point connection structure transformation model, which includes an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module. can include

According to an embodiment of the present invention, a plurality of convolution operation layers are connected into a plurality of groups, and each group exists as a functional module unit capable of performing separate functions such as area segmentation, feature point extraction, and feature point connection structure creation. can be designed

According to an embodiment of the present invention, the area division module receives a character 3D model and a 3D clothing model to an artificial neural network-based body part division model, divides each body part of the character 3D model and the clothing 3D model into a plurality of body areas, Among the divided body areas, body areas included in the 3D clothing model may be selected as at least one feature point extraction target area to generate feature point extraction target area information.

According to an embodiment of the present invention, feature points are extracted from the character 3D model and the clothing 3D model, and the size, shape, and texture of the clothing 3D model are changed to match the size, shape, and texture of the character 3D model based on the extracted feature points In order to be properly fitted as if a real person is wearing clothes, extract the same number of feature points from the points that form the preset vertices in each 3D model to express the characteristics of the appearance, and match each feature point with each other. method can be used.

According to an embodiment of the present invention, when feature points are extracted from all parts of a character 3D model for more accurate fitting, calculation is performed even to parts where clothes are not worn, resulting in excessive use of computational resources and inaccurate results. Extraction of feature points can reduce accuracy. In order to solve this problem, each body part of the character 3D model and clothing 3D model is divided into a plurality of body areas, and at least one of the divided body areas is included in the clothing 3D model. By selecting the above feature point extraction target area, only the same body parts can be extracted by metabolic response.

According to the above embodiment, the feature point extraction module receives the character 3D model, the clothing 3D model, and the feature point extraction target area information, and identifies the vertex part constituting the shape of the body and clothing for the feature point extraction target area according to a preset criterion It is possible to generate feature point information for each area for each 3D model by labeling with a possible number and deriving a preset number of labeled vertices and designating them as feature points.

According to an embodiment of the present invention, the structure generation module receives information on the character 3D model, the clothing 3D model, and feature points for each area, and converts a plurality of feature points derived from the character 3D model and the clothing 3D model into a plurality of straight lines or curves, respectively. A character feature point connected structure model and a clothing feature point connected structure model may be created by forming a feature point connected structure capable of expressing a 3D shape and volume by connecting them to each other.

According to an embodiment of the present invention, the feature point connection structure model may mean that a plurality of feature points are modeled as a mesh structure capable of expressing a shape and a sense of volume by connecting a plurality of feature points to each other with a plurality of straight lines or curves.

The feature point matching information generation unit 500 matches a plurality of points included in the character feature point connected body structure and a plurality of points included in the costume feature point connected body structure model, respectively, and features point matching including information on each pair of matched feature points. information can be generated.

According to an embodiment of the present invention, the criterion for matching the plurality of points included in the character feature point connected structure with the plurality of points included in the costume feature point connected structure model is labeling for each feature point included in each feature point connected structure. Since there is an identifiable number, the feature point of the same number is defined as meaning the same body part and clothing part, and the location information of each feature point included in the character feature point connection structure and the same number included in the clothing feature point connection structure It is possible to generate feature point matching information including each location information and an identifiable number by using location information of each feature point as one feature point pair.

The wearing clothing 3D model generating unit 600 contacts and matches the feature points of the character feature point connection structure model matched with the feature points included in the clothing feature point connection structure model according to the information on each feature point pair of the feature point connection chain matching information. The location of a plurality of feature points included in the clothing feature point connection structure model is changed respectively to transform the shape of the clothing feature point connection structure model, and the clothing 3D model is converted based on the transformed clothing feature point connection structure model to fit. You can create a 3D model of clothes that can be worn on the target metaverse character.

According to an embodiment of the present invention, the wearing clothing 3D model generating unit 600 calculates a vector value for the position of the changed feature point for each feature point when changing the position of a plurality of feature points included in the clothing feature point connection structure model. And, even after multiplying the feature points located on a straight line in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is located on a straight line, and a certain curvature is applied to the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the feature points located on the curve with each vector value, the position change can be limited so that each feature point is located on the curve with a constant curvature.

According to an embodiment of the present invention, in order to calculate a vector value for the location of the changed feature point, the location information of the feature point included in the character feature point connection body structure corresponding to the identification number of the corresponding feature point included in the feature point matching information is connected to the clothing feature point. By using the location information of the feature points included in the body structure, calculations such as the dot product of the vector values that are changed so that the location information of the feature points included in the clothing feature point connected structure matches the feature point location information included in the character feature point connected structure. method can be used to calculate it.

The wearing clothing meta information generation unit 700 uses a blockchain-based smart contract to generate unique identifier information of the worn clothing 3D model, the type of clothing worn, the name of the worn clothing, the number of transaction records of the clothing, and the amount of the 3D model transaction amount of the worn clothing , wearing clothing 3D model creator, wearing clothing 3D model current owner, wearing clothing 3D model creation date, actual clothing seller, actual clothing transaction amount, actual clothing size information. Wearing clothing meta information including at least one item of actual clothing image information may be generated.

According to an embodiment of the present invention, items included in the meta-information on clothes to be worn can be added without limitation as long as information necessary to provide information on actual clothes and transaction functions as well as clothes implemented as items in the metaverse.

The NFT clothing information generation unit 800 generates metaverse clothing information by inserting the worn clothing meta information into the worn clothing 3D model, and matches the metaverse clothing information to at least one blockchain-based cryptocurrency to tokenize non-fungible It is possible to generate NFT clothing information by performing.

According to an embodiment of the present invention, a specific cryptocurrency based on a block chain is designated for metaverse clothing information including a 3D model of worn clothing and meta information of worn clothing, and the metaverse clothing information is included in the metadata of a specific cryptocurrency. In the future, it is possible to perform irreplaceable tokenization, and in the future, by using the smart contract function of the block chain, it is possible to provide trading functions without a separate trading platform.

2 is a diagram illustrating an internal data flow of a feature point connection structure conversion model based on an artificial neural network implemented according to an embodiment of the present invention.

Referring to FIG. 2, the internal data flow of the artificial neural network-based feature point connection structure conversion model implemented according to an embodiment of the present invention is shown, and the feature point connection structure conversion model includes an area segmentation module, a feature point extraction module, and structure generation. It can be designed to contain modules.

According to an embodiment of the present invention, the area segmentation module may be trained to receive a character 3D model and a 3D clothing model and output target area information for feature point extraction. It may be learned to receive a 3D model of clothing and output feature point information for each area.

According to an embodiment of the present invention, the structure generation module may be trained to receive area-specific feature point information, a character 3D model, and a clothing 3D model, and output a character feature point-connected structure model and a clothing feature point-connected structure model.

3 is a diagram illustrating an internal data flow of a feature point extraction module implemented according to an embodiment of the present invention.

Referring to FIG. 3, according to an embodiment of the present invention, the feature point extraction module may include a first module, a second module, and a third module composed of a plurality of convolutional layers, and each module may have the following characteristics. have.

According to an embodiment of the present invention, the first module receives information of a character 3D model, a 3D clothing model, and a feature point extraction target area into a first deep learning network formed through a plurality of convolutional layers, and performs a pre-targeting of a feature point extraction target area. A plurality of first feature point output information may be generated by deriving a set number of feature points.

According to an embodiment of the present invention, the second module receives the character 3D model, the clothing 3D model, and the feature point extraction target area information to the second deep learning network formed through a plurality of convolutional layers, and performs a pre-targeting of the feature point extraction target area. A plurality of second feature point output information may be generated by deriving a set number of feature points.

According to an embodiment of the present invention, the third module receives the first feature point output information and the second feature point output information, applies weights to the selected first feature point output information and the second feature point output information, respectively, and applies the weights. The first feature point output information and the second feature point output information are normalized to a value within a predetermined range and analyzed through a preset mathematical formula to derive a preset number of feature points for the feature point extraction target area.

The first, second, and third modules according to an embodiment of the present invention may be designed as different networks, and will be described in more detail with reference to FIG. 4 .

4 is a diagram showing a network structure of first, second, and third modules of a feature point extraction module implemented according to an embodiment of the present invention.

Referring to FIG. 4, the network structure of the first, second, and third modules of the feature point extraction module implemented according to an embodiment of the present invention is shown, and the network of each module can be designed as follows.

The first module according to an embodiment of the present invention may generate a plurality of first feature point output information by receiving a character 3D model, a 3D clothing model, and target area information for feature point extraction.

, 1번째 레이어 출력 데이터를

, 2번째 레이어 출력 데이터를

, L번째 레이어 출력 데이터를

으로 지칭할 수 있으며, 이때

는 아래 수학식 5와 같이 산출될 수 있다.According to an embodiment of the present invention, the first module may be composed of a plurality of calculation layers, and the input character 3D model and clothing 3D model, and feature point extraction target area information may be used to perform calculations through a plurality of calculation layers. At this time, the character 3D model, clothing 3D model, and feature point extraction target area information

, the first layer output data

, the second layer output data

, the Lth layer output data

can be referred to as, where

Can be calculated as in Equation 5 below.

The second module according to an embodiment of the present invention may generate a plurality of second feature point output information by receiving a character 3D model, a 3D clothing model, and target area information for feature point extraction.

, 1번째 레이어 출력 데이터를

, 2번째 레이어 출력 데이터를

, L번째 레이어 출력 데이터를

으로 지칭할 수 있으며, 이때

는 아래 수학식 6과 같이 산출될 수 있다.According to an embodiment of the present invention, the second module may be composed of a plurality of calculation layers, and calculation may be performed through a plurality of calculation layers by using the input character 3D model, clothing 3D model, and feature point extraction target area information. At this time, the character 3D model, clothing 3D model, and feature point extraction target area information

, the first layer output data

, the second layer output data

, the Lth layer output data

can be referred to as, where

Can be calculated as in Equation 6 below.

According to the above embodiment, the third module may receive the first feature point output information and the second feature point output information and learn the feature point extraction module to generate feature point information for each of a plurality of areas.

를 생성할 수 있고, 아래 수학식 7에 기반하여 복수의 에어리어 별 특징점 정보(

)를 생성할 수 있다.According to an embodiment of the present invention, the input first feature point output information and the second feature point output information are integrated

And based on Equation 7 below, feature point information for each area (

) can be created.

5 is a block diagram of an NFT-based metaverse wearing clothing information generation system implemented according to a second embodiment of the present invention.

Referring to FIG. 5, an NFT-based metaverse wearing clothing information generation system implemented according to the second embodiment of the present invention is shown, and NFT-based metaverse wearing clothing information generation implemented according to the second embodiment of the present invention The system may further include a clothing sales execution unit 900 in the metaverse wearing clothing information generating system of the first embodiment.

The clothing sales execution unit 900 uses a blockchain-based smart contract to obtain the actual clothing seller, actual clothing transaction price, and actual clothing size information included in the NFT clothing information. Using actual clothing image information, information on actual clothing and a sales function can be provided without a separate shopping platform and payment tool.

According to an embodiment of the present invention, the clothing sales execution unit 900 may be implemented through a smart contract function of a block chain including a cryptocurrency on which NFT conversion is performed.

6 is a flow chart of a method of generating NFT-based metaverse wearing clothing information according to an embodiment of the present invention.

Clothing design information including length, size, ratio, texture, and shape information is generated by sensing shape data of actual clothing (S10).

According to an embodiment of the present invention, clothing design information including length, size, ratio, texture, and shape information may be generated by sensing shape data of actual clothing.

According to an embodiment of the present invention, three-dimensional numerical data of clothes such as length, size, ratio, texture, and shape information may be collected through a depth camera or a 3D scanner to sense shape data of actual clothes.

According to an embodiment of the present invention, a depth camera is used to photograph real clothes, and a pulse-modulated infrared (IR) beam is emitted to the target at the origin of the spherical coordinate system, so that the horizontal (pan, φ) and vertical (tilt) . 3D image information of a target in the background can be acquired based on the transmission/reception time.

According to an embodiment of the present invention, a plurality of depth cameras may be installed based on a predetermined angle so as to photograph the 3D shape of at least one garment in all directions, and the image data and the image data may be preset based on time. Collected shape information may be generated by collecting transmission/reception time data for each part of the set clothing.

Here, the preset clothing part means a part of the clothing that can identify the characteristics of the clothing type, for example, sleeve, collar, wrist, elbow, shoulder, neck, hip, knee, and ankle. can be set around them.

According to an embodiment of the present invention, in collecting transmission/reception time data using a depth camera, two-step depth measurement may be performed.

According to an embodiment of the present invention, two-step depth measurement may be divided into a first depth measurement step and a second depth measurement step.

The reason why the depth measurement is performed in two stages as in the above embodiment is that when a single frequency is used and depth measurement is performed using a low modulation frequency, the maximum range is widened but the measurement quality is degraded. , in the case of performing depth measurement using a high modulation frequency, the measurement quality can be improved, but there is a disadvantage that the maximum range is narrowed.

According to an embodiment of the present invention, in the first depth measurement, the depth may be measured using a low modulation frequency, and measurement may be performed with low measurement quality for a pre-set region of interest for each body part.

According to an embodiment of the present invention, the second depth measurement may increase measurement precision with high measurement quality by measuring the depth using a high modulation frequency based on the measurement result of the region of interest for each body part of the first depth measurement. have.

According to an embodiment of the present invention, the first depth measurement using a low modulation frequency to have a wide maximum depth range can measure an approximate depth targeting a wide depth range, and the measurement quality at this time is proportional to the modulation frequency. , the progress measured through the first depth measurement may provide low measurement quality for a wide area of interest.

According to an embodiment of the present invention, in the second depth measurement, a maximum depth range may be set based on the precision of the first depth measurement, and a relatively higher frequency is selected to provide high measurement quality for a narrow region of interest. An error in the first depth measurement result may be compensated for.

According to an embodiment of the present invention, a low modulation frequency may be selected based on Equation 1.

According to an embodiment of the present invention, a high modulation frequency may be selected as a value that is inversely proportional to the standard deviation measured when a low modulation frequency is used, and if the standard deviation is smaller than a preset limit value, this indicates a high signal-to-noise ratio. By determining, a relatively higher frequency may be selected than when the standard deviation is larger than a preset limit value.

)는 수학식 2를 기반으로 산출될 수 있다.where the standard deviation (

) can be calculated based on Equation 2.

According to an embodiment of the present invention, measurement accuracy may be increased through repeated measurement by performing a plurality of first depth measurement steps and second depth measurement steps.

According to an embodiment of the present invention, the clothing design information generating unit 100 inputs the image data for each depth camera included in the collected shape information to an artificial neural network-based shape information generation model, and outlines and outlines for each part of clothing. First shape information including angle information may be generated.

In addition, a plurality of consecutive depth data, velocity data, and amplitude data are generated by analyzing the transmission and reception time data of each depth camera included in the collected shape information, and a cumulative distribution function and a noise removal function according to the size of the amplitude using the amplitude data Second shape information may be generated by removing noise using , and averaging depth data, velocity data, and amplitude data for each part of clothing generated for each depth camera from which the noise is removed with a median value.

According to an embodiment of the present invention, the artificial neural network-based shape information generation model receives image data for each depth camera as input to an input layer and outputs outline and outline angle information for each pre-set garment part, which can be output for each garment. It may be learned to generate first shape information by grouping by part. According to an embodiment of the present invention, transmission/reception time data for each body part generated for each depth camera may be analyzed to generate continuous depth data, velocity data, and amplitude data for each body part over time.

Here, the depth data may refer to data capable of recognizing the three-dimensional position of the region of interest, and the speed data may refer to a movement speed calculated through a change in position over time based on the depth data. , The amplitude data may mean the light intensity of the surface caused by the reflection of the damaged light because the light intensity of the surface of the light transmitted by the depth camera is different depending on the degree of amplitude of the surface.

According to an embodiment of the present invention, noise can be removed using a cumulative distribution function and a noise function according to the amplitude of amplitude data.

According to the above embodiment, the amplitude data may represent how bright the light reflected by the depth camera is, and this is based on the fact that light intensity of the surface appears differently according to the degree of amplitude of the surface.

According to an embodiment of the present invention, filtering may be performed using a cumulative distribution function and a noise function to remove noise caused by a non-ideal waveform used in a depth camera and noise generated at a boundary of an object.

According to an embodiment of the present invention, the cumulative distribution function may be performed based on Equation 3.

According to an embodiment of the present invention, the noise function NF may be performed based on Equation 4.

According to an embodiment of the present invention, an error is calculated by matching the first shape information and the second shape information based on time, and the calculated error is removed using a continuous weighted median filter to obtain final behavioral information. can be created with

According to an embodiment of the present invention, the first shape information and the second shape information are input to a final shape information generation model based on an artificial neural network, and the outline and shape of the clothing part included in the first shape information and the second shape information Contrasting line angle information calculates the difference as an error, applies a weight to each of the data where the error occurs, and averages the calculated values with the median value to remove the error. It is possible to create final shape information including

The size, body type, length, shape, and texture information of the fitting target metaverse character are received from the metaverse server to generate character appearance information (S20).

According to an embodiment of the present invention, information on the size, body shape, length, shape, and texture of a metaverse character to be fitted may be received from a server of the metaverse to generate character appearance information.

According to one embodiment of the present invention, character appearance information may mean modeling and numerical data for characters that can be implemented in the metaverse stored in the main server of the platform implementing the metaverse, and the size and body shape of the character. , length, shape, and texture can be used without any restrictions on the format.

Based on the character appearance information, the character's appearance is modeled in 3D to create a character 3D model, and the clothing design information is used to model the appearance of the clothing in 3D to create a clothing 3D model (S30).

According to an embodiment of the present invention, a character 3D model can be created by modeling the appearance of a character in 3D based on character appearance information, and a 3D model of clothing can be created by modeling the appearance of clothing in 3D using clothing design information. have.

According to an embodiment of the present invention, a parametric technique of constructing a function expression inputting a plurality of independent parameters by analyzing character appearance information and clothing appearance information and performing 3D modeling based on a mathematical free curve using such a function expression Character 3D models and clothing 3D models can be created using

Character 3D model and clothing 3D model are input to the feature point connection structure transformation model, a plurality of feature points for the appearance of each 3D model are extracted, and a plurality of extracted feature points are connected to character feature point connection structure structure model and clothing feature point connection body A structural model is created (S40).

According to an embodiment of the present invention, a plurality of feature points for the appearance of each 3D model are extracted by inputting a character 3D model and a 3D costume model to an artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected Then, a character feature point connected structure model and a clothing feature point connected structure model converted into a feature point connected structure model expressed as a shape structure can be created.

According to an embodiment of the present invention, a feature point connection structure conversion model composed of a plurality of convolutional operation layers is designed, and the designed feature point connection structure conversion model receives a character 3D model and a clothing 3D model to obtain a character feature point connection structure transformation model and learning may be performed by inputting learning data so as to output a structure model of a feature point connection structure of a garment.

According to one embodiment of the present invention, the feature point concatenation structure transformation model may be designed so that a plurality of convolution operation layers are connected as one group, but according to another embodiment, a plurality of groups are connected and each group is separately designed. It can be designed to exist in units of functional modules that can perform the functions of.

According to an embodiment of the present invention, the artificial neural network-based structure conversion model of the feature point connection may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module.

According to an embodiment of the present invention, a plurality of convolution operation layers are connected into a plurality of groups, and each group exists as a functional module unit capable of performing separate functions such as area segmentation, feature point extraction, and feature point connection structure generation. can be designed

According to an embodiment of the present invention, the area division module receives a character 3D model and a 3D clothing model to an artificial neural network-based body part division model, divides each body part of the character 3D model and the clothing 3D model into a plurality of body areas, Among the divided body areas, body areas included in the 3D clothing model may be selected as at least one feature point extraction target area to generate feature point extraction target area information.

According to an embodiment of the present invention, feature points are extracted from the character 3D model and the clothing 3D model, and the size, shape, and texture of the clothing 3D model are changed to match the size, shape, and texture of the character 3D model based on the extracted feature points In order to be properly fitted as if a real person is wearing clothes, extract the same number of feature points from the points that form the preset vertices in each 3D model to express the characteristics of the appearance, and match each feature point with each other. method can be used.

According to an embodiment of the present invention, when feature points are extracted from all parts of a character 3D model for more accurate fitting, calculation is performed even to parts where clothes are not worn, resulting in excessive use of computational resources and inaccurate results. Extraction of feature points can reduce accuracy. In order to solve this problem, each body part of the character 3D model and clothing 3D model is divided into a plurality of body areas, and at least one of the divided body areas is included in the clothing 3D model. By selecting the above feature point extraction target area, only the same body parts can be extracted by metabolic response.

According to the above embodiment, the feature point extraction module receives the character 3D model, the clothing 3D model, and the feature point extraction target area information, and identifies the vertex part constituting the shape of the body and clothing for the feature point extraction target area according to a preset criterion It is possible to generate feature point information for each area for each 3D model by labeling with a possible number and deriving a preset number of labeled vertices and designating them as feature points.

According to an embodiment of the present invention, the structure generation module receives information on the character 3D model, the clothing 3D model, and feature points for each area, and converts a plurality of feature points derived from the character 3D model and the clothing 3D model into a plurality of straight lines or curves, respectively. A character feature point connected structure model and a clothing feature point connected structure model may be created by forming a feature point connected structure capable of expressing a 3D shape and volume by connecting them to each other.

According to an embodiment of the present invention, the feature point connection structure model may mean that a plurality of feature points are modeled as a mesh structure capable of expressing a shape and a sense of volume by connecting a plurality of feature points to each other with a plurality of straight lines or curves.

According to an embodiment of the present invention, the feature point connection structure conversion model may be designed to include an area segmentation module, a feature point extraction module, and a structure generation module.

According to an embodiment of the present invention, the area segmentation module may be trained to receive a character 3D model and a 3D clothing model and output target area information for feature point extraction. It may be learned to receive a 3D model of clothing and output feature point information for each area.

According to an embodiment of the present invention, the structure generation module may be trained to receive area-specific feature point information, a character 3D model, and a clothing 3D model, and output a character feature point-connected structure model and a clothing feature point-connected structure model.

According to an embodiment of the present invention, the feature point extraction module may include a first module, a second module, and a third module composed of a plurality of convolution layers, and each module may have the following characteristics.

According to an embodiment of the present invention, the first module receives information of a character 3D model, a 3D clothing model, and a feature point extraction target area into a first deep learning network formed through a plurality of convolutional layers, and performs a pre-targeting of a feature point extraction target area. A plurality of first feature point output information may be generated by deriving a set number of feature points.

According to an embodiment of the present invention, the second module receives the character 3D model, the clothing 3D model, and the feature point extraction target area information to the second deep learning network formed through a plurality of convolutional layers, and performs a pre-targeting of the feature point extraction target area. A plurality of second feature point output information may be generated by deriving a set number of feature points.

According to an embodiment of the present invention, the third module receives the first feature point output information and the second feature point output information, applies weights to the selected first feature point output information and the second feature point output information, respectively, and applies the weights. The first feature point output information and the second feature point output information are normalized to a value within a predetermined range and analyzed through a preset mathematical formula to derive a preset number of feature points for the feature point extraction target area.

The network structure of the first, second, and third modules of the feature point extraction module implemented according to an embodiment of the present invention is shown, and the network of each module can be designed as follows.

The first module according to an embodiment of the present invention may generate a plurality of first feature point output information by receiving a character 3D model, a 3D clothing model, and target area information for feature point extraction.

, 1번째 레이어 출력 데이터를

, 2번째 레이어 출력 데이터를

, L번째 레이어 출력 데이터를

으로 지칭할 수 있으며, 이때

는 수학식 5와 같이 산출될 수 있다.According to an embodiment of the present invention, the first module may be composed of a plurality of calculation layers, and the input character 3D model and clothing 3D model, and feature point extraction target area information may be used to perform calculations through a plurality of calculation layers. At this time, the character 3D model, clothing 3D model, and feature point extraction target area information

, the first layer output data

, the second layer output data

, the Lth layer output data

can be referred to as, where

Can be calculated as in Equation 5.

The second module according to an embodiment of the present invention may generate a plurality of second feature point output information by receiving a character 3D model, a 3D clothing model, and target area information for feature point extraction.

, 1번째 레이어 출력 데이터를

, 2번째 레이어 출력 데이터를

, L번째 레이어 출력 데이터를

으로 지칭할 수 있으며, 이때

는 수학식 6과 같이 산출될 수 있다.According to an embodiment of the present invention, the second module may be composed of a plurality of calculation layers, and calculation may be performed through a plurality of calculation layers by using the input character 3D model, clothing 3D model, and feature point extraction target area information. At this time, the character 3D model, clothing 3D model, and feature point extraction target area information

, the first layer output data

, the second layer output data

, the Lth layer output data

can be referred to as, where

Can be calculated as in Equation 6.

According to an embodiment of the present invention, the third module may receive the first feature point output information and the second feature point output information and learn the feature point extraction module to generate feature point information for each of a plurality of areas.

를 생성할 수 있고, 수학식 7에 기반하여 복수의 에어리어 별 특징점 정보(

)를 생성할 수 있다.According to an embodiment of the present invention, the input first feature point output information and the second feature point output information are integrated

Can be generated, and based on Equation 7, feature point information for each of a plurality of areas (

) can be created.

Characteristic point matching information is generated by matching a plurality of points included in the character feature point connected structure with a plurality of points included in the costume feature point connected structure model (S50).

According to an embodiment of the present invention, feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. can create

According to an embodiment of the present invention, the criterion for matching the plurality of points included in the character feature point connected structure with the plurality of points included in the costume feature point connected structure model is labeling for each feature point included in each feature point connected structure. Since there is an identifiable number, the feature point of the same number is defined as meaning the same body part and clothing part, and the location information of each feature point included in the character feature point connection structure and the same number included in the clothing feature point connection structure It is possible to generate feature point matching information including each location information and an identifiable number by using location information of each feature point as one feature point pair.

According to information on each pair of feature points, positions of a plurality of feature points included in the clothing feature point connection structure model are respectively changed to create a wearable 3D model of clothing (S60).

According to an embodiment of the present invention, according to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connected structure model matched with the feature points included in the costume feature point connected structure model can be matched. The location of a plurality of feature points included in the clothing feature point connection structure model is changed to transform the shape of the clothing feature point connection structure model, and the clothing 3D model is converted based on the transformed clothing feature point connection structure model to form a fitting target meta A 3D model of clothes that can be worn on a bus character can be created.

According to an embodiment of the present invention, when changing the positions of a plurality of feature points included in the clothing feature point connected structure model, a vector value for the position of the changed feature point is calculated for each feature point, and the feature points of the clothing feature point connected structure model Even after the feature points located on a straight line in the connected body structure are multiplied by each vector value, each feature point is located on a straight line, and each vector is located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the values, the position change can be limited so that each feature point is located on a curve with a constant curvature.

According to an embodiment of the present invention, in order to calculate a vector value for the location of the changed feature point, the location information of the feature point included in the character feature point connection body structure corresponding to the identification number of the corresponding feature point included in the feature point matching information is connected to the clothing feature point. By using the location information of the feature points included in the body structure, calculations such as the dot product of the vector values that are changed so that the location information of the feature points included in the clothing feature point connected structure matches the feature point location information included in the character feature point connected structure. method can be used to calculate it.

Wearing clothing meta information is created using a blockchain-based smart contract (S70).

According to an embodiment of the present invention, by using a blockchain-based smart contract, unique identifier information of the 3D model of the worn clothing, the type of clothing worn and the name of the worn clothing, the number of transaction records of the clothing, the transaction amount of the 3D model of the worn clothing, and the wearing Clothing 3D model creator, worn clothing 3D model current owner, worn clothing 3D model creation date, actual clothing seller, actual clothing transaction amount, actual clothing size information. Wearing clothing meta information including at least one item of actual clothing image information may be generated.

According to an embodiment of the present invention, items included in the meta-information on clothes to be worn can be added without limitation as long as information necessary to provide information on actual clothes and transaction functions as well as clothes implemented as items in the metaverse.

Metaverse clothing information is created by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching with cryptocurrency to generate NFT clothing information (S80).

According to one embodiment of the present invention, metaverse clothing information is created by inserting worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is achieved by matching the metaverse clothing information to at least one blockchain-based cryptocurrency. It can be performed to generate NFT clothing information.

According to an embodiment of the present invention, a specific cryptocurrency based on a block chain is designated for metaverse clothing information including a 3D model of worn clothing and meta information of worn clothing, and the metaverse clothing information is included in the metadata of a specific cryptocurrency. In the future, it is possible to perform irreplaceable tokenization, and in the future, by using the smart contract function of the block chain, it is possible to provide trading functions without a separate trading platform.

According to an embodiment of the present invention, actual clothing seller, actual clothing transaction amount, and actual clothing size information included in the NFT clothing information using a blockchain-based smart contract. Using actual clothing image information, information on actual clothing and a sales function can be provided without a separate shopping platform and payment tool.

According to an embodiment of the present invention, it can be implemented through a smart contract function of a block chain containing a cryptocurrency in which NFT conversion has been performed.

Embodiments of the present invention are not implemented only through the devices and / or methods described above, and the embodiments of the present invention have been described in detail above, but the scope of the present invention is not limited thereto, and the following claims Various modifications and improvements of those skilled in the art using the basic concept of the present invention defined in , also belong to the scope of the present invention.

Claims (5)

a clothing design information generation unit configured to sense shape data of actual clothing and generate clothing design information including length, size, ratio, texture, and shape information;
Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse;
a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information;
A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model;
Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit;
According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model;
Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; and
Metaverse clothing information is created by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information to at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generator to generate,
The feature point connection structure model generation unit,
The artificial neural network-based feature point connection structural transformation model may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module,
The area division module,
The character 3D model and the clothing 3D model are input to the artificial neural network-based body part segmentation model, and each body part of the character 3D model and the clothing 3D model is divided into a plurality of body areas, and among the divided body areas, the clothing 3D model included in the 3D model Selecting a body area as at least one feature point extraction target area to generate feature point extraction target area information;
The feature point extraction module,
Character 3D model, clothing 3D model, and feature point extraction target area information are input, and a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area, and designated as feature points, and feature points for each area for each 3D model generate information,
The structure generation module,
Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, so that a three-dimensional shape and volume can be expressed. Possible feature point connection structure is formed to create a character feature point connection structure model and a clothing feature point connection structure model,
The wearing clothing 3D model generating unit,
In changing the position of a plurality of feature points included in the clothing feature point connected structure model, a vector value for the position of the changed feature point is calculated for each feature point, and a vector value is calculated on a straight line in the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is NFT-based metaverse wearing clothing information generation system, characterized in that to limit the position change to be located on a curve with a constant curvature. a clothing design information generation unit configured to sense shape data of actual clothing and generate clothing design information including length, size, ratio, texture, and shape information;
Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse;
a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information;
A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model;
Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit;
According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model;
Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; and
Metaverse clothing information is created by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information to at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generator to generate,
The feature point connection structure model generation unit,
The artificial neural network-based feature point connection structural transformation model may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module,
The area division module,
The character 3D model and the clothing 3D model are input to the artificial neural network-based body part segmentation model, and each body part of the character 3D model and the clothing 3D model is divided into a plurality of body areas, and among the divided body areas, the clothing 3D model included in the 3D model Selecting a body area as at least one feature point extraction target area to generate feature point extraction target area information;
The feature point extraction module,
Character 3D model, clothing 3D model, and feature point extraction target area information are input, and a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area, and designated as feature points, and feature points for each area for each 3D model generate information,
The structure generation module,
Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, so that a three-dimensional shape and volume can be expressed. Possible feature point connection structure is formed to create a character feature point connection structure model and a clothing feature point connection structure model,
The wearing clothing 3D model generating unit,
In changing the position of a plurality of feature points included in the clothing feature point connected structure model, a vector value for the position of the changed feature point is calculated for each feature point, and a vector value is calculated on a straight line in the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is Limit the position change to be located on a curve with a constant curvature,
The feature point extraction module includes a first module, a second module, and a third module composed of a plurality of convolutional layers,
The first module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the first deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. NFT-based metaverse wearing clothing information generation system, characterized in that for generating a plurality of first feature point output information by deriving. a clothing design information generation unit configured to sense shape data of actual clothing and generate clothing design information including length, size, ratio, texture, and shape information;
Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse;
a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information;
A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model;
Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit;
According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model;
Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; and
Metaverse clothing information is created by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information to at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generator to generate,
The feature point connection structure model generation unit,
The artificial neural network-based feature point connection structural transformation model may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module,
The area division module,
The character 3D model and the clothing 3D model are input to the artificial neural network-based body part segmentation model, and each body part of the character 3D model and the clothing 3D model is divided into a plurality of body areas, and among the divided body areas, the clothing 3D model included in the 3D model Selecting a body area as at least one feature point extraction target area to generate feature point extraction target area information;
The feature point extraction module,
Character 3D model, clothing 3D model, and feature point extraction target area information are input, and a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area, and designated as feature points, and feature points for each area for each 3D model generate information,
The structure generation module,
Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, so that a three-dimensional shape and volume can be expressed. Possible feature point connection structure is formed to create a character feature point connection structure model and a clothing feature point connection structure model,
The wearing clothing 3D model generating unit,
In changing the position of a plurality of feature points included in the clothing feature point connected structure model, a vector value for the position of the changed feature point is calculated for each feature point, and a vector value is calculated on a straight line in the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is Limit the position change to be located on a curve with a constant curvature,
The feature point extraction module includes a first module, a second module, and a third module composed of a plurality of convolutional layers,
The first module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the first deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. Deriving and generating a plurality of first feature point output information;
The second module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the second deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. NFT-based metaverse wearing clothing information generation system, characterized in that for generating a plurality of second feature point output information by deriving. a clothing design information generation unit configured to sense shape data of actual clothing and generate clothing design information including length, size, ratio, texture, and shape information;
Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse;
a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information;
A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model;
Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit;
According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model;
Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; and
Metaverse clothing information is created by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information to at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generator to generate,
The feature point connection structure model generation unit,
The artificial neural network-based feature point connection structural transformation model may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module,
The area division module,
The character 3D model and the clothing 3D model are input to the artificial neural network-based body part segmentation model, and each body part of the character 3D model and the clothing 3D model is divided into a plurality of body areas, and among the divided body areas, the clothing 3D model included in the 3D model Selecting a body area as at least one feature point extraction target area to generate feature point extraction target area information;
The feature point extraction module,
Character 3D model, clothing 3D model, and feature point extraction target area information are input, and a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area, and designated as feature points, and feature points for each area for each 3D model generate information,
The structure generation module,
Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, so that a three-dimensional shape and volume can be expressed. Possible feature point connection structure is formed to create a character feature point connection structure model and a clothing feature point connection structure model,
The wearing clothing 3D model generating unit,
In changing the position of a plurality of feature points included in the clothing feature point connected structure model, a vector value for the position of the changed feature point is calculated for each feature point, and a vector value is calculated on a straight line in the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is Limit the position change to be located on a curve with a constant curvature,
The feature point extraction module includes a first module, a second module, and a third module composed of a plurality of convolutional layers,
The first module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the first deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. Deriving and generating a plurality of first feature point output information;
The second module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the second deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. Deriving and generating a plurality of second feature point output information;
The third module receives the first feature point output information and the second feature point output information, applies a weight of 0 to 1 or less, respectively, to the selected first feature point output information and the second feature point output information, and NFT-based metaverse wearing clothing information generation system, characterized in that by normalizing 1 feature point output information and 2 feature point output information to a value within a predetermined range to derive a preset number of feature points for the feature point extraction target area. a clothing design information generation unit configured to sense shape data of actual clothing and generate clothing design information including length, size, ratio, texture, and shape information;
Character appearance information generation unit for generating character appearance information by receiving the size, body shape, length, shape, texture information of the fitting target metaverse character from the server of the metaverse;
a 3D model generation unit for generating a 3D character model by modeling the outer appearance of the character in 3D based on the outer appearance information of the character, and generating a 3D model of clothing by modeling the outer appearance of the clothing in 3D using the clothing design information;
A plurality of feature points for the appearance of each 3D model are extracted by inputting the character 3D model and the costume 3D model to the artificial neural network-based feature point connection structure transformation model, and the plurality of extracted feature points are connected to form a shape structure a characteristic point connected structure model generating unit for generating a character feature point connected structure model converted into a feature point connected structure model and a clothing feature point connected structure model;
Feature point matching information for generating feature point matching information including information on each matched feature point pair by matching a plurality of points included in the character feature point connected structure and a plurality of points included in the costume feature point connected structure model. generating unit;
According to the information on each feature point pair of the feature point connection matching information, the feature points of the character feature point connection structure model matched with the feature points included in the costume feature point connection structure model are matched to the costume feature point connection structure model. By changing the position of a plurality of included feature points, the shape of the clothing feature point connection structure model is converted, and the clothing 3D model is converted based on the converted clothing feature point connection structure model to wear clothing that can be worn on the fitting target metaverse character a wearing clothing 3D model generation unit that creates a 3D model;
Using a blockchain-based smart contract, the unique identifier information of the 3D model of the worn clothing, the type of clothing worn, the name of the worn clothing, the number of transaction history of the clothing, the transaction amount of the 3D model of the worn clothing, the creator of the 3D model of the worn clothing, and the 3D worn clothing The current owner of the model, the date of creation of the 3D model of the clothing worn, the actual clothing seller, the actual clothing transaction amount, and the actual clothing size information. a worn clothing meta information generating unit generating worn clothing meta information including at least one item of actual clothing image information; and
Metaverse clothing information is created by inserting the worn clothing meta information into the worn clothing 3D model, and non-fungible tokenization is performed by matching the metaverse clothing information to at least one blockchain-based cryptocurrency to obtain NFT clothing information Including an NFT clothing information generator to generate,
The feature point connection structure model generation unit,
The artificial neural network-based feature point connection structural transformation model may include an area segmentation module composed of a plurality of convolutional layers, a feature point extraction module, and a structure generation model module,
The area division module,
The character 3D model and the clothing 3D model are input to the artificial neural network-based body part segmentation model, and each body part of the character 3D model and the clothing 3D model is divided into a plurality of body areas, and among the divided body areas, the clothing 3D model included in the 3D model Selecting a body area as at least one feature point extraction target area to generate feature point extraction target area information;
The feature point extraction module,
Character 3D model, clothing 3D model, and feature point extraction target area information are input, and a preset number of vertices constituting the shape of the body are derived for the feature point extraction target area, and designated as feature points, and feature points for each area for each 3D model generate information,
The structure generation module,
Character 3D model, clothing 3D model, and feature point information for each area are input, and a plurality of feature points derived from the character 3D model and clothing 3D model are connected to each other with a plurality of straight lines or curves, respectively, so that a three-dimensional shape and volume can be expressed. Possible feature point connection structure is formed to create a character feature point connection structure model and a clothing feature point connection structure model,
The wearing clothing 3D model generating unit,
In changing the position of a plurality of feature points included in the clothing feature point connected structure model, a vector value for the position of the changed feature point is calculated for each feature point, and a vector value is calculated on a straight line in the feature point connected structure of the clothing feature point connected structure model. Even after multiplying the located feature points by each vector value, each feature point is located on a straight line, and after multiplying the feature points located on a curve with a constant curvature in the feature point connected structure of the clothing feature point connected structure model by each vector value, each feature point is Limit the position change to be located on a curve with a constant curvature,
The feature point extraction module includes a first module, a second module, and a third module composed of a plurality of convolutional layers,
The first module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the first deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. Deriving and generating a plurality of first feature point output information;
The second module receives the character 3D model, clothing 3D model, and feature point extraction target area information from the second deep learning network formed through a plurality of convolutional layers, and extracts feature points by a predetermined number targeting the feature point extraction target area. Deriving and generating a plurality of second feature point output information;
The third module receives the first feature point output information and the second feature point output information, applies a weight of 0 to 1 or less, respectively, to the selected first feature point output information and the second feature point output information, and 1 feature point output information and 2 feature point output information are normalized to a value within a predetermined range to derive a preset number of feature points for the feature point extraction target area;
The actual clothing seller, actual clothing transaction amount, and actual clothing size information included in the NFT clothing information using a blockchain-based smart contract. Using actual clothing image information, NFT-based metaverse wearing clothing information generation system further comprising a clothing sales execution unit that provides information and sales functions on actual clothing.

Images