Information processing device, information processing system and information processing method

Abstract

To provide an information processing device, an information processing system, and an information processing method, which can assist in the selection of a desired fabric.SOLUTION: An information acquisition unit acquires material information indicating the characteristics of the material of the fabric and fabric information indicating the characteristics of the fabric produced from the material, and an image synthesis unit synthesizes an image of the fabric from the acquired material information and fabric information using an image synthesis model indicating the relationship between the material information and the fabric information and an image showing the three-dimensional form of the fabric. This embodiment can be achieved in any of the forms of an information processing device, an information processing system, and an information processing method.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an information processing device, an information processing system and an information processing method, and more particularly to the synthesis of images representing samples of fabrics made of synthetic materials. [Background technology]

[0002] When trading clothing fabrics and their materials, material manufacturers may create samples (samples) of the fabric and present the samples to traders such as fabric manufacturers and apparel manufacturers. The traders may refer to the presented samples to select samples or materials to be traded. Conventionally, attempts have been made to improve the efficiency of searching for samples having desired characteristics or for those characteristics.

[0003] For example, Patent Document 1 describes a fabric search system that stores product data including the product name and weight of a fabric, sensory evaluation value data that evaluates the fabric using visual and tactile sensory scales, and basic item data with the fabric type, material, and use as basic items, in association with each other, and allows the user to select whether to search for fabrics from the sensory evaluation value data or the basic item data.

[0004] Patent document 2 describes a fitting image creation system that creates a try-on image of a human body model having the user's body type wearing clothing based on the user's body type information and clothing information, displays it on a display device, and when an instruction is given to create a fabric sample, outputs the fabric, color, pattern, etc. as clothing information for the try-on image, prints the color and pattern on white fabric based on this information, and creates a fabric sample after going through the processes of post-processing, washing and drying. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2017-016378 A [Patent Document 2] JP 2002-024319 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, since sensitivities differ from person to person, some users may have difficulty finding the desired fabric when searching from the sensitivity scale described in Patent Document 1. In the method described in Patent Document 2, characteristics such as fabric, color, and pattern are identified based on clothing information in a fitting image, but the characteristics of the fabric itself are not identified.

[0007] The present invention has been made in view of the above points, and one of the objects of the present invention is to provide an information processing device, an information processing system, and an image processing method that can support the selection of a desired fabric. In addition, by providing an image showing the three-dimensional form of a fabric sample, it becomes unnecessary to produce a physical fabric sample, which improves the efficiency of fabric development and contributes to environmental protection by eliminating the waste of fabric samples. [Means for solving the problem]

[0008] (1) The present invention has been made to solve the above-mentioned problems, and one aspect of the present invention is an information processing system including an information acquisition unit that acquires material information indicating the characteristics of a fabric material and fabric information indicating the characteristics of fabric produced from the material, and an image synthesis unit that synthesizes an image of the fabric from the acquired material information and fabric information using an image synthesis model that indicates the relationship between the material information and fabric information, and an image representing the three-dimensional form of the fabric. (2) Another aspect of the present invention is an information processing system comprising an information acquisition unit that acquires an image representing a three-dimensional form of a fabric, and an image analysis unit that determines the material information and fabric information from the acquired image using a production condition model that indicates the relationship between the image, material information that indicates the characteristics of the material of the fabric, and fabric information that indicates the characteristics of the fabric produced from the material.

[0009] (3) Another aspect of the present invention is an information processing device comprising: an information acquisition unit that acquires material information indicating characteristics of a fabric material and fabric information indicating characteristics of fabric produced from the material; and an image synthesis unit that synthesizes an image of the fabric from the acquired material information and fabric information using an image synthesis model that indicates the relationship between the material information and fabric information and an image representing the three-dimensional form of the fabric. (4) Another aspect of the present invention is an information processing system comprising an information acquisition unit that acquires an image representing a three-dimensional form of a fabric, and an image analysis unit that determines the material information and fabric information from the acquired image using a production condition model that indicates the relationship between the image, material information that indicates the characteristics of the material of the fabric, and fabric information that indicates the characteristics of the fabric produced from the material.

[0010] (5) Another aspect of the present invention is an information processing method in an information processing device, in which the information processing device executes an information acquisition step of acquiring material information indicating characteristics of a fabric material and fabric information indicating characteristics of a fabric produced from the material, and an image synthesis step of synthesizing an image of the fabric from the acquired material information and fabric information using an image synthesis model indicating the relationship between the material information and fabric information, and an image representing the three-dimensional form of the fabric. (6) Another aspect of the present invention is an information processing method in an information processing device, in which the information processing device executes an information acquisition step of acquiring an image representing a three-dimensional form of a fabric, and an image analysis step of determining the material information and fabric information from the acquired image using a production condition model that indicates the relationship between the image, material information that indicates the characteristics of the material of the fabric, and fabric information that indicates the characteristics of the fabric produced from the material. Effect of the Invention

[0011] According to the present invention, it is possible to assist in the selection of a desired fabric. [Brief description of the drawings]

[0012] [Figure 1] 1 is a schematic block diagram showing a first example of the functional configuration of an information processing system according to the present embodiment. [Diagram 2] FIG. 2 is an explanatory diagram showing an example of learning and inference of the image synthesis model of the present embodiment. [Diagram 3] 11 is a table showing an example of production condition information according to the embodiment; [Figure 4] FIG. 2 is a diagram showing a first example of a fabric according to the present embodiment. [Diagram 5] FIG. 11 is a diagram showing a second example of a fabric according to the present embodiment. [Figure 6] FIG. 2 is a diagram showing a first example of a display screen according to the embodiment. [Figure 7] FIG. 11 is a diagram showing a second example of a display screen according to the embodiment. [Figure 8] FIG. 13 is a diagram showing an example of an order screen according to the embodiment; [Figure 9] FIG. 13 is a diagram showing an example of an evaluation screen according to the embodiment. [Figure 10] FIG. 4 is a schematic block diagram showing a second example of the functional configuration of the information processing system according to the present embodiment. [Figure 11] FIG. 2 is an explanatory diagram showing an example of learning and inference of an image analysis model according to the present embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. First, an overview of this embodiment will be described. FIG. 1 is a schematic block diagram showing a first example of a functional configuration of an information processing system 1 according to this embodiment. The information processing system 1 includes an information processing device 10 and a terminal device 20. The information processing device 10 and the terminal device 20 are connected to each other via a network so as to be able to transmit and receive various data wirelessly or wired. The network may be any one of the Internet, a public network, a local area network (LAN), a virtual private network (VPN), a dedicated line, and the like, or any combination of these.

[0014] The information processing device 10 acquires information indicating the production conditions of the material (sometimes referred to as "production condition information" in this application). The production condition information includes material information and fabric information. The material information is information indicating the characteristics of the material of the fabric. The fabric information is information indicating the characteristics of the fabric produced from the material. Specifically, the fabric is artificial fur (sometimes called synthetic fur, artificial fur, eco fur, fake fur, etc.). Artificial fur is generally a fabric with a raised texture formed by knitting or weaving fiber bundles made of chemical fibers on a base fabric (sometimes called a base fabric part). The fabric can be formed as a sample (sometimes called a specimen, sample, etc.). The material includes at least chemical fibers that are the raw materials of the fabric, and may also include natural fibers. The information processing device 10 can acquire the material information and fabric information from, for example, the terminal device 20. The information processing device 10 synthesizes an image of the fabric from the acquired material information and fabric information using an image synthesis model indicating the relationship between the material information and fabric information and an image representing the three-dimensional form of the fabric. The information processing device 10 transmits image data showing the synthesized image to the terminal device 20. The terminal device 20 displays an image showing the three-dimensional form of the fabric based on the image data received from the information processing device 10. A user of the terminal device 20 can visually confirm the displayed image and confirm whether the fabric shown in the image is the desired fabric.

[0015] The users of the terminal device 20 are mainly producers or sellers of fabrics, or designers, producers or sellers of clothing made from fabrics, general consumers, etc. The information processing device 10 acquires order information related to fabrics from the terminal device 20. The users of the information processing device 10 are mainly producers or sellers of materials, or businesses commissioned by them, designers, producers or sellers of clothing. Since the production of fabric samples can be reduced or omitted, the development of fabrics can be made more efficient. Furthermore, the users of the terminal device 20 can change the production condition information using the terminal device 20 and immediately confirm the desired fabric by visually checking the image data of the fabric sample, thereby saving the labor and time involved in the development of unnecessary fabric samples.

[0016] The information processing device 10 may be configured as a general-purpose device such as a personal computer (PC), a workstation, a server device, etc., or may be configured as a dedicated device. The terminal device 20 may be configured as a general-purpose device such as a PC, a tablet terminal device, a smartphone, etc., or may be configured as a dedicated device.

[0017] Next, a description will be given of an example of a functional configuration of the information processing device 10. The information processing device 10 includes a control unit 120, a storage unit 140, and an input / output unit 150. The control unit 120 includes a processor such as a CPU (Central Processing Unit). The processor executes processing indicated by instructions written in a predetermined program stored in advance in the storage unit 140, thereby fulfilling the functions of the information processing device 10. In this application, the execution of processing indicated by instructions written in a program may be referred to as "executing a program" or "running a program". The processor, for example, executes a predetermined program to fulfill the functions of the control unit 120. The control unit 120 may be realized using a dedicated member.

[0018] The storage unit 140 stores various data used in the processes executed by the control unit 120. The storage unit 140 stores various data acquired by the control unit 120. The storage unit 140 includes a storage medium such as a RAM (Random Access Memory) and a ROM (Read Only Memory). The input / output unit 150 inputs or outputs various data to or from other devices wirelessly or via wires. The input / output unit 150 may be connected to other devices via a network. The input / output unit 150 may be configured to include, for example, either an input / output interface, a communication interface, or both.

[0019] Next, a description will be given of the control unit 120. The control unit 120 includes an information acquisition unit 122, an image synthesis unit 124, a model learning unit 126, an order processing unit 128, and a user management unit 130. The information acquisition unit 122 acquires production condition information of the fabric. The information acquisition unit 122 receives the production condition information from the terminal device 20 using the input / output unit 150, and outputs the received production condition information to the image synthesis unit 124. Examples of parameters that are elements of the production condition information will be described later.

[0020] The image synthesis unit 124 uses a predetermined image synthesis model to synthesize an image showing the three-dimensional form of the fabric based on the production condition information input from the information acquisition unit 122, and generates three-dimensional image data showing the generated image (inference). The image synthesis unit 124 transmits the generated image data to the terminal device 20 using the input / output unit 150. The three-dimensional form is expressed by a three-dimensional shape and a distribution of colors forming a pattern applied to the surface. The color generally refers to one or both of a hue (color tone) and a gradation (shade). The three-dimensional shape and size may be set in advance. The three-dimensional shape shown in the image includes at least one surface and one cross section. As the three-dimensional image data, point cloud data, a set of two-dimensional image data and depth data, etc. may be used. As the image synthesis model, for example, a mathematical model such as a convolutional neural network (CNN) may be used. As a parameter group (also called hyperparameters, model parameters, etc.) for the calculation process using the image synthesis model, model parameters obtained by the model learning unit 126 are used. In the following description, a pixel, which is a unit for setting signal values ​​in image data and depth data, and a voxel, which is a unit for setting signal values ​​in point cloud data, may be collectively referred to as a "display unit."

[0021] The model learning unit 126 uses previously generated training data to calculate model parameters of an image synthesis model indicating the relationship between production condition information and images showing the three-dimensional form of the fabric (model learning). The training data includes a plurality of training sets. Each training set includes production condition information as input information and a set of images as output information. The model learning unit 126 recursively updates the model parameters so that, for example, for the entire training set, the difference between an estimated image based on the input information and an image forming the output information is reduced. The model parameters obtained by learning are used for inference in the image synthesis unit 124. Model learning and inference will be described later.

[0022] The order processing unit 128 receives order information related to the fabric whose three-dimensional form is shown in the image data generated by the image synthesis unit 124 from the terminal device 20 using the input / output unit 150. The order processing unit 128 can identify the production condition information shown in the received order information. The identified production condition information may be used for the production of the fabric. For example, the order processing unit 128 may transmit the identified production condition information to a fabric production facility. The order processing unit 128 can cause the production facility to produce the fabric according to the production conditions shown in the production condition information. In this embodiment, the product to be produced is mainly a fabric sample (hereinafter, sometimes referred to as a "fabric sample"). The order processing unit 128 may output the acquired order conditions to the user management unit 130 and have the user manage the order conditions for each user.

[0023] The user management unit 130 manages users who access using the terminal device 20. The user management unit 130 performs, for example, a login process. The storage unit 140 stores user registration data that includes user identification information and authentication information for each user and associates these information. For example, a user ID and a password are used as the user identification information and authentication information. The user management unit 130 acquires the user identification information and the authentication information from the terminal device 20. The user management unit 130 refers to the user registration data, compares the authentication information associated with the acquired user identification information with the acquired authentication information, and determines whether the user authentication is successful. The user management unit 130 permits access from the terminal device 20 that is the transmission source of the user identification information that has been successfully authenticated, and does not permit access from the terminal device 20 that is the transmission source of the user identification information that has failed to be authenticated. The user management unit 130 can identify the terminal device 20 that is the access source by using device identification information such as an IP address notified by the access from the terminal device 20 until a logout process is instructed from the terminal device 20.

[0024] Next, a description will be given of an example of the functional configuration of the terminal device 20. The terminal device 20 includes a control unit 220, a storage unit 240, an input / output unit 250, a display unit 260, and an operation unit 270. The functional configurations of the control unit 220, the storage unit 240, and the input / output unit 250 may be similar to the functional configurations of the control unit 120, the storage unit 140, and the input / output unit 150 of the information processing device 10. The functional configurations of the control unit 220, the storage unit 240, and the input / output unit 250 refer to the descriptions of the control unit 120, the storage unit 140, and the input / output unit 150, respectively.

[0025] The display unit 260 displays various information (for example, a display screen, etc.) according to display data input from the control unit 220. The display unit 260 includes a display such as a liquid crystal display or an organic electroluminescence display. The operation unit 270 receives a user's operation and generates an operation signal indicating various information (for example, coordinate values, characters, etc.) specified by the received operation. The operation unit 270 outputs the generated operation signal to the control unit 220. The operation unit 270 may include dedicated members such as buttons, levers, and knobs, or may include general-purpose members such as a touch sensor, a mouse, and a joystick. The touch sensor constituting the operation unit 270 may overlap with a display constituting the display unit 260 and be configured as a touch panel.

[0026] Next, a description will be given of the control unit 220. The control unit 220 includes an input processing unit 222 and a display processing unit 224. The input processing unit 222 cooperates with the display processing unit 224 to configure a graphic user interface (GUI). The input processing unit 222 identifies information instructed by an operation signal input to the operation unit 270 or the input / output unit 250. The identified information includes element information such as material information, fabric information, and order information, and an access request to the information processing device 10. The input processing unit 222 transmits the identified material information and fabric information to the information processing device 10. The input processing unit 222 may identify a display mode of an image based on the operation signal and notify the display processing unit 224 of the identified display mode. The input processing unit 222 identifies order information based on the operation information, and transmits the identified order information to the information processing device 10.

[0027] The display processing unit 224 configures various display screens and outputs display data showing the configured display screens to the display unit 260. The display screens may include items for specifying material information and fabric information, a display field for an image showing the three-dimensional form of the fabric, screen components for instructing the display mode of the image, and the like. The display screens may include items for specifying order conditions, evaluation items for the image, and the like. Specific examples of the display screens will be described later.

[0028] Next, learning and inference of the image synthesis model will be described. FIG. 2 is an explanatory diagram showing an example of learning and inference of the image synthesis model of this embodiment. The information processing device 10 is capable of executing a learning step (S02) and an inference step (S04). The learning step S02 is a process of determining model parameters for synthesizing an image representing a three-dimensional form of a fabric using a predetermined image synthesis model from input material information and fabric information. In the learning step S02, the model learning unit 126 executes supervised learning. In supervised learning, the model learning unit 126 uses a data set including known input values ​​and target values ​​as a training set, and recursively calculates model parameters so that an estimated value calculated according to the image synthesis model for the input value approximates the target value. In this embodiment, numerical values ​​indicating elements constituting each of the material information and fabric information are used as input values.

[0029] In the learning step S02, the model learning unit 126 uses numerical values ​​constituting material information and fabric information as input values, and uses signal values ​​for each display element constituting image data showing a fabric sample image as target values. The fabric sample image is an image that represents a fabric sample. In the learning step S02, the model learning unit 126 recursively calculates model parameters so that a loss function indicating the magnitude of the difference between an estimated value and a target value is minimized across multiple training sets. As a mathematical model used as an image synthesis model, for example, a convolutional neural network (CNN) can be used. Here, minimization means performing a calculation to estimate or search model parameters with the aim of minimizing as much as possible, in addition to absolutely minimizing. Therefore, the loss function does not necessarily decrease monotonically in the process of minimization, and may temporarily increase. As the loss function, for example, the Sum of Squared Differences (SSD), cross entropy error, etc. can be used. The method for determining the model parameters may be any method such as steepest descent, stochastic optimization, back-propagation, etc. When the loss function becomes equal to or less than a predetermined loss function threshold, the model learning unit 126 determines that the model has converged and stops the learning step S02. The model learning unit 126 stores in the storage unit 140 image synthesis model data indicating the model parameters calculated at that time.

[0030] The inference step S04 is a process of synthesizing an image showing the three-dimensional form of the fabric using a predetermined image synthesis model from the input material information and fabric information. The image synthesis unit 124 reads out the image synthesis model data stored in the storage unit 140, and develops the model parameters shown in the read image synthesis model data into the image synthesis model. In the inference step S04, the image synthesis unit 124 calculates an estimated value for each display element using the image synthesis model for the input values ​​showing the material information and fabric information, and generates image data showing the calculated estimated value for each display element. The generated image data expresses a three-dimensional image showing the three-dimensional form of the fabric sample. The three-dimensional shape of the fabric sample expressed at this point may be the basic shape of the fabric sample. The basic shape means a predetermined basic shape. The basic shape of the fabric sample is, for example, a flat rectangular parallelepiped. The shape of the main surface is a square or a rectangle. The thickness is usually smaller than the length of each of the two sides parallel to the main surface.

[0031] Next, CNN will be described as an example of a mathematical model used as an image synthesis model. CNN is a type of artificial neural network, and includes one input layer, multiple intermediate layers, and an output layer. Each layer has multiple nodes (also called nodes, neurons, etc.). Each node outputs a function value of a predetermined function for an input value as an output value. In the input layer, values ​​of each dimension representing material information and texture information are input to the node of the corresponding dimension, and are output to the corresponding node of the next layer. CNN includes one or more convolution layers and one or more pooling layers as intermediate layers.

[0032] A convolutional layer is a layer that calculates convolutional values ​​by performing convolutional calculations for each kernel on input values ​​input from the previous layer to each of multiple nodes. A kernel refers to a processing unit for calculating one output value at a time. One kernel consists of one or more parallel nodes. In the convolutional calculation, one or more input values ​​are input to each node from the previous layer, and independent convolution coefficients are used for each input value. A convolutional layer is a layer that calculates the function value of a specified activation function for a correction value obtained by adding the convolution value calculated for each kernel and a bias value as an output value, and outputs the calculated output value to the corresponding node in the next layer. The parameters of the convolutional coefficients, bias value, and activation function are part of one set of model parameters.

[0033] A pooling layer is a layer having nodes that determine one representative value from input values ​​input from multiple nodes in the previous layer and output the determined representative value as an output value to the next layer. The representative value is, for example, a value that statistically represents multiple input values, such as the maximum value, average value, or mode. In a pooling layer, the input value from the previous layer is downsampled to a lower dimension and the output value is provided to the next layer.

[0034] Note that the intermediate layers may include layers that perform other types of calculations, such as one or more of a normalization layer, a smoothing layer, and a fully connected layer. The normalization layer is a layer that normalizes the input values ​​input to each of multiple nodes within a predetermined value range using a common normalization parameter (batch normalization), and outputs the normalized values ​​as output values ​​to the next layer. A flattening layer is a layer that expands high-dimensional (e.g., three-dimensional) input values ​​into lower-dimensional (e.g., two-dimensional) output values ​​and outputs the expanded output values ​​to the next layer. The fully connected layer is a layer that calculates convolution values ​​by performing convolution operations on input values ​​input from the previous layer to each of multiple nodes, calculates an output value by adding the calculated convolution value and a bias value, and outputs the calculated output value to the next layer. The output layer may be composed of any one of a normalization layer, a flattening layer, and a fully connected layer. The output layer has a node for each display element, and outputs a signal value for each node as an output value.

[0035] The mathematical model used as the image synthesis model is not limited to CNN, and may be any of RNN (Recurrent Neural Network) and ResNet (Residual Network), or may be a mathematical model other than a neural network. The model learning unit 126 may obtain training data from another device and calculate the model parameters using the obtained training data.

[0036] Next, an example of production condition information will be described. FIG. 3 is a table showing an example of production condition information. The production condition information includes material information and fabric information. The material information is information indicating the characteristics of the material. In this embodiment, the material refers to the fibers that are the material of the fabric. The characteristics refer to properties, state, attributes, and the like. Elements of the material information include all or part of the brand, variety (type of fiber), fineness, fiber type, cross-sectional shape, gloss, color, potential shrinkage characteristics, shrinkage characteristics, fiber length, shrinkage characteristics, filament number, and total fineness of the filaments.

[0037] Brand refers to the name of a fiber. Product names are sometimes used as brands. Brands are often used to specify fabric specifications because they are easy for users to recognize. Brands are sometimes associated with fiber characteristics, such as cross-sectional shape and thickness, blending (color, luster), and shrinkage characteristics. Variety refers to the type of fiber. Fibers include filament (long fiber) and staple (short fiber). Filament refers to fibers that are relatively long in length (e.g., tens of meters to several kilometers, theoretically infinite). Staple refers to fibers that are relatively short in length (e.g., several centimeters to several tens of centimeters). Fiber types (common to filament and staple) are classified into, for example, natural fibers and chemical fibers. Natural fibers include plant fibers such as hemp and cotton, and animal fibers such as wool and cashmere. In addition to chemical fibers, natural fibers may also be used as a part of artificial fur materials. Chemical fibers include regenerated fibers such as rayon, animal fibers such as acetate, and synthetic fibers such as nylon, polyester, and acrylic. Acrylic fibers, acrylic-based fibers, and polyester fibers are often used as materials for artificial fur. An example of an acrylic-based fiber is modacrylic fiber. The variety is expressed, for example, by an integer value (variety code) that indicates the individual type.

[0038] Fineness refers to the weight per unit length of a single fiber. Units of fineness include decitex (dtex) and denier (den). 1 decitex is equivalent to the weight (g) per 10,000m. 1 denier is equivalent to the weight (g) per 9,000m. The fiber type is information indicating whether the fiber is classified as staple or filament. The cross-sectional shape is the shape of the cross section that appears when the fiber is cut perpendicular to the longitudinal direction. Cross-sectional shapes include round, flat, irregular flat (flat multi-lobed), Y-shaped, H-shaped, W-shaped, etc. For artificial fur, round, oval, flat, irregular flat, etc. are often used. Gloss refers to the degree of cloudiness in the appearance of the fiber, and is also called dullness. Dullness means that the fiber is cloudy, has poor transparency, and does not transmit light. There are various levels of dullness, including full dull, dull, semi-dull, and bright. Of these levels, full dull is the most dull, followed by decreasing levels of dullness.

[0039] Color refers to the color of the fiber. It corresponds to a combination of color tone and brightness. Color is expressed using color space values ​​expressed using a predetermined color space. For example, Lab can be used as the color space. The Lab color space is designed to approximate human vision. The Lab color space value is expressed as a three-dimensional vector value including L, a, and b as real number elements. L, a, and b correspond to the dimensions of lightness, the first complementary color, and the second complementary color, respectively. The value of a quantifies the color tone between red or magenta and green. The value of b quantifies the color tone between yellow and blue. In this embodiment, the Lab color space is not limited to the Lab color space, and the RGB color space may be used. Synthetic fibers may be colored not only during production (original dyeing), but also after production (dyeing).

[0040] The latent crimp property refers to, for example, the property of the spiral shape of a side-by-side type composite fiber. Fibers having such properties can obtain elasticity similar to that of wool, and are therefore sometimes used in the production of sheep-type artificial fur. Elements of the latent crimp property may include the presence or absence of latent crimp, the shape of the crimp (for example, the diameter and pitch of the spiral), and the like. The shrinkage characteristic is a characteristic that indicates the shrinkage rate when heat is applied to the fiber. As will be described later in the example of boa knit fabric in the step information, the tips of the fabric made from multiple types of raw cotton are cut to a certain length, and then a shrinking process (generally a heat treatment) is performed to form steps so that the length corresponds to the shrinkage rate. In this case, the shrinkage rate of each fiber is important in forming the steps.

[0041] All of the above items are common to both staple and filament, but the crimp property is an item specific to staple. The fiber length is not limited to one, but may be set to multiple. This is because the type of animal fur is characterized by a combination of multiple fiber lengths. This fact can be used as a clue to reproduce (simulate) the fur of a certain animal based on a combination of multiple fiber lengths. The crimping property refers to the property of the crimping of the staple. The crimping means the so-called shrinking, and is also called crimp. The crimping is a shape necessary to ensure the entanglement of the fibers when twisting multiple fibers together to form a thread. Elements of the crimping property include the number of crimps, the crimping rate, the residual crimping rate, and the crimping modulus. The definitions and test methods of these parameters are specified, for example, in the Japanese Industrial Standard JIS L1015:2010 "Testing Methods for Chemical Fiber Staples".

[0042] Items specific to filaments are the number of filaments and the total filament fineness. The number of filaments refers to the number of single yarns that make up one filament bundle. A filament bundle is usually made up of several tens of single yarns twisted together (multifilament). The total fineness of the filaments refers to the fineness of the entire bundle of filaments, which is calculated by multiplying the fineness of one single yarn by the number of filaments. In the information acquisition unit 122 of the information processing device 10, keywords representing trend information, value information, or usage information may be directly input in response to the operation of the user of the terminal device 20, or may be selected from a list of multiple keywords prepared in advance in response to the user's operation.

[0043] Fabric information is information that indicates the characteristics of fabric produced from materials. Generally, fabrics are produced by knitting or weaving methods. Artificial fur fabrics are mainly produced by knitting. Knitting methods used in the production of artificial fur include, for example, high pile knitting, boa knitting, raschel knitting, and other methods. Among them, high pile knitting (also called sliver knit) is generally produced using staples. Boa knitting can be realized by using worsted yarn made with staples or by directly using filaments. Raschel knitting can be realized by using worsted yarn made with staples (worsted raschel) or by directly using filaments.

[0044] Fabric information mainly refers to the properties, condition, attributes, etc. that appear in the fabric during the fabric production process. Elements of fabric information include animal type, composition ratio, pile length, discoloration processing state, basis weight, planting density, and total fineness. The type of animal means the type of animal that is simulated by the fabric. The type of animal is expressed by an integer value (brand code) that indicates the type of each animal. The "brand" of this "brand code" does not mean the "name of fiber" as described above, but the "type of animal". The types of animals include, for example, fox (Fig. 4), rabbit (Fig. 5), mink, raccoon, sable, sheep, wolf, and chinchilla (not shown). Depending on the type of animal, requirements such as the length of the pile (long, medium, short), the stepped structure (one-layer structure (plain), two-layer structure (consisting of guard hair and down hair), and three-layer structure (consisting of guard hair, down hair, and middle hair)), and whether or not the fiber is wavy, may differ. For example, a fox-type fabric has a long pile length and a three-layer structure. A sable-type fabric has a medium pile length and a two-layer structure. A Mongolian sheep-type fabric has a long pile length, a three-layer structure, and is wavy. The fabric type refers to the method of forming the fabric. The fabric type is not limited to the high pile, boa, and raschel mentioned above, but may also specify other weaves and other processing methods, such as flock processing.

[0045] The composition ratio is the composition ratio of the raw cotton used as a material. The composition ratio is given as a real value, for example, the weight ratio of each type of raw cotton. As a prerequisite, the material information may be defined for each type of raw cotton, or may be defined only for a limited type of raw cotton that is the main component. The above material information can be set for each type of raw cotton used as a material. The pile length indicates the length of the fiber bundle. The length of the fiber bundle corresponds to the length from the ground structure of the fabric to the tip of the hair. In this embodiment, the average value of the fiber bundle length may be given. In addition, the fabric information may include shearing information as information indicating whether the fiber bundle length is cut to a uniform length. The shearing information is expressed using a flag value indicating 1 or 0 depending on whether the fiber bundle length is cut to a uniform length. In the example of FIG. 5, the hair length is cut to a substantially uniform length, but in the example of FIG. 4, the fiber bundle length is not cut to a uniform length. If the fiber bundle length is not cut to a uniform length, the pile length does not need to be set.

[0046] The step information is information that indicates the characteristics of the step structure. The step information includes the presence or absence of a step structure, and for fabrics with a step structure, the length of the hair for each step (i.e., the length of the down hair and middle hair). There are various methods for creating steps. In high pile knitting, steps appear in fabrics made from multiple types of raw cotton with different fiber lengths. In boa knitted mink fabric, the tips of the hair of the fabric, which is made from multiple types of raw cotton with different shrinkage rates, are cut to a certain length and then shrinkage treatment (generally heat treatment) is performed, and steps appear so that the length corresponds to the shrinkage rate. In boa knitting, when the hair is cut for each loop, steps appear by changing the length of the loop for each loop.

[0047] The implantation pattern refers to the distribution pattern of the positions where the fibers of the standing pile portion are implanted into the stitches of the ground structure. Implantation patterns include implantation in each stitch, skip stitch, jacquard stitch, etc. Skip stitch means that the fibers are implanted by skipping stitches. Jacquard stitch means that the fibers are embedded in any stitch.

[0048] The discoloration processing state refers to the state of coloring or bleaching of a part of a discoloration processing for each hair. Discoloration processing is generally also called printing processing (coloring) and discharge printing processing (bleaching). The discoloration processing state is expressed by the color signal value of the color that appears by coloring or bleaching for each discoloration processing unit and the distribution of the discoloration processing units. For example, the root part, the middle part, and the tip part are each treated as a discoloration processing unit, and their length and color are expressed. The discoloration processing state may be expressed by the content and length of the discoloration processing treatment for each discoloration processing unit. The content of the discoloration processing treatment includes, for example, coloring, the color after coloring, and discharge printing. The basis weight is the weight per unit area. The larger the basis weight, the higher the fiber density tends to be. The planting density is the number of fibers per unit area.

[0049] Each artificial fur may be made of a plurality of types of fibers. In this case, the pile length, planting density, discoloration processing state, and total fineness may be set for each type of hair as a whole fabric of the artificial fur. Also, instead of each fiber, the artificial fur may have a fiber bundle forming a nap-raising structure on the base fabric. Each fiber bundle is formed by connecting the base ends of a plurality of fibers. In this case, the planting density is set for the fiber bundle, and does not have to be set for each type of fiber. The types of fibers include, for example, differential hair and fluff. Differential hair can be thicker and longer than fluff. For example, the fiber bundle is formed by bundling one differential hair and multiple (for example, 3 to 10) fluffs around it with their roots aligned, and is adhered to the base fabric. The configuration of the fiber bundle is not limited to this, and there are also ones that are made of only fluff or only differential hair. Also, a fiber bundle made of only fluff and a fiber bundle made of differential hair shavings may be arranged next to each other to form a series of nap-raising parts. Therefore, the fabric information may include fiber bundle configuration information regarding the configuration of the fiber bundle as element information. The fiber bundle configuration information may include any one of information indicating the presence or absence of a fiber bundle configuration, and if a fiber bundle is configured, information indicating the composition ratio (e.g., the ratio of the numbers) between fiber types in one fiber bundle, and information indicating the arrangement of multiple fiber types in one fiber bundle, or a combination thereof.

[0050] The elements of the fabric information may include information regarding the usage of the produced fabric, such as trend information, value information, and usage information, or a combination of these. Trend information is information that reflects the trends of fabrics. Trends mainly refer to fashion. Generally, clothing trends change with each season. Trend information is expressed by keywords that reflect, for example, the characteristics of the clothing, the characteristics of the fabric that caused the clothing to become fashionable, and the like.

[0051] The value information is information that reflects the values ​​that a user or end consumer desires for a fabric (e.g., "real," "gorgeous," "environmental protection," etc.) The value information is expressed, for example, by keywords that indicate the values. The use information is information that indicates the use of the fabric. Examples of uses of the fabric include use for clothing such as the outer surface of clothes and around the hood of a coat (also called trimming), use for home textiles such as cushion covers and sofa covers, and use for fashion accessories such as bags and shoes. The use information is expressed, for example, by keywords that indicate the use. In the information acquisition unit 122 of the information processing device 10, keywords representing trend information, value information, or usage information may be directly input in response to the operation of the user of the terminal device 20, or may be selected from a list of multiple keywords prepared in advance in response to the user's operation.

[0052] Next, the display screen will be described. Fig. 6 is a diagram showing a first example of the display screen. The display screen exemplified in Fig. 6 is a service screen for viewing the form of a fabric sample based on material information and fabric information specified by a user and ordering the fabric sample. The service screen has an image display field, a display condition specification field, a production condition specification field, and an order button bo. The image display field is a field for displaying a sample image showing the three-dimensional form of the fabric sample produced based on the specified material information and fabric information.

[0053] In FIG. 6, the image display field va is an area of ​​a certain size near the upper left corner of the display screen. The image display field va is located directly below the title "Eco fur sample". The line segments ax, ay, and az drawn in the image display field va respectively represent the x-axis, y-axis, and z-axis in a virtual three-dimensional space in which the fabric sample is displayed (hereinafter sometimes referred to as the "fabric sample space"). The fabric sample space is a virtual three-dimensional space whose origin is a representative point of the fabric sample (e.g., the center of gravity, center, etc.). When the fabric sample is not bent, the normal direction of its surface is oriented in the z-direction, and the normal direction of one cross section is oriented in the y-direction. The x-direction is perpendicular to both the y-direction and the z-direction. The direction of the line segments ax, ay, and az represents the direction of the viewpoint.

[0054] The production condition specification field is an area where screen components are arranged to determine production condition information according to operations. Directly below the title "Specify material", screen components are arranged to specify the material information elements, such as variety, fineness, fiber length, and color. Each element is specified by operating the screen components. The variety is selected from the variety candidates listed in the pull-down menu that appears when the input field mv is specified. The variety is not limited to one type, and multiple types can be specified. When multiple types of varieties are specified, other elements, that is, fineness, fiber length, and color, may be specified for each variety. Fineness and fiber length are specified by entering numerical values ​​in the input fields mf and ml, respectively. As for color, slider bars ml, ma, and mb are displayed to specify the brightness, redness, and blueness of the material by operations, respectively. Brightness, redness, and blueness indicate color elements. Brightness, redness, and blueness correspond to the L value (lightness), a value (one of the two complementary color axes), and b value (the other of the two complementary color axes) in the Lab color space, respectively.

[0055] Directly below the title "Fabric specification" are screen components for specifying the fabric information elements, such as animal type, composition ratio, pile length, planting density, basis weight, and total fineness. The animal type is selected from the animal type candidates listed in the pull-down menu that appears when input field cv is specified. The composition ratio is set using two or more partial areas and a slider bar cc that changes the area of ​​each partial area according to the operation. For each partial area, one type is selected from the raw cotton type candidates listed in the pull-down menu that appears when specified. The pile length, planting density, basis weight, and total fineness are specified by entering numerical values ​​in input fields cp, cd, cw, and cc, respectively.

[0056] The display condition specification field is an area for specifying parameters that form conditions for displaying (viewing) an image of a fabric sample in response to an operation signal. In the example of FIG. 6, the viewpoint direction, lighting conditions, and product type (product image) can be specified. As the viewpoint direction, dials dp and dt for specifying the azimuth angle (horizontal plane direction) and elevation angle are displayed. The input processing unit 222 can specify the azimuth angle and elevation angle according to the position of the operation point specified by dragging the dials dp and dt. As the lighting condition, slider bars ll, la, and lb for specifying the brightness, redness, and blueness of the illumination light by operation are displayed. The input processing unit 222 can change the position of the control point of each of the slider bars ll, la, and lb according to the position of the operation point specified by dragging, and specifies the L value, a value, and b value according to the position of the control point. As the product type, buttons se, sc, and sm for selecting one of a fabric sample (eco fur sample), an overcoat, and a muffler by operation are displayed. The input processing unit 222 waits for an operation on the buttons se, sc, and sm, and identifies a product corresponding to the button whose position is specified by any one of the operations. The input processing unit 222 transmits display condition information indicating the identified display condition to the information processing device 10.

[0057] The image synthesis unit 124 of the information processing device 10 manipulates (renders) the image by performing a predetermined graphics processing on the image synthesized using the image synthesis model in accordance with the display conditions indicated in the display condition information received from the input processing unit 222. When the specified product type is a product made of fabric (in this example, an overcoat or a muffler), the image synthesis unit 124 synthesizes an image showing the three-dimensional form of the specified product type using the fabric based on the image of the synthesized fabric sample (product type conversion). In the image synthesis unit 124, parameters showing the characteristics of typical three-dimensional forms for each product type are set as three-dimensional form characteristic parameters. The image synthesis unit 124 synthesizes an image showing the three-dimensional form of the product using the three-dimensional form characteristic parameters of the specified product type. The specified product type of object is not limited to an overcoat and a muffler, and may include other products produced using fabric as a material. However, when the product type of the object shown in the image generated using the image synthesis model is a fabric sample, the image synthesis unit 124 does not need to perform product type conversion processing.

[0058] If the specified product type is a product made of fabric (i.e., an overcoat or a scarf), the image synthesis unit 124 may synthesize an image showing the three-dimensional form of a person wearing the product. Therefore, parameters indicating the characteristics of the three-dimensional form when the product is worn by a person of a certain shape are set as the three-dimensional form characteristic parameters. The image synthesis unit 124 uses the three-dimensional form characteristic parameters of the specified product type to synthesize an image showing the three-dimensional form of the product, and synthesizes an image in which the three-dimensional form of the synthesized product is covered over the three-dimensional form of the person's body.

[0059] When a viewpoint direction is specified, the image synthesis unit 124 determines a position a predetermined distance away from the origin in the specified viewpoint direction as the viewpoint. The image synthesis unit 124 synthesizes an image showing the three-dimensional form of a synthesized object (or a texture sample) observed from the determined viewpoint and with its center of gravity set at the origin. When lighting conditions are specified, the image synthesis unit 124 converts the color distribution of the three-dimensional form of the synthesized object (or fabric sample) into a color distribution obtained by irradiating it with illumination light under the specified lighting conditions (color conversion).

[0060] The image synthesis unit 124 may transmit image data of an image synthesized based on the updated display conditions to the terminal device 20 each time an element of any of the display conditions is changed in response to an operation. The terminal device 20 receives image data according to the updated display conditions from the information processing device 10. The image synthesis unit 124 of the information processing device 10 may transmit image data of an image synthesized based on the updated display conditions to the terminal device 20 each time an element of any of the display conditions is changed in response to an operation.

[0061] The input processing unit 222 may transmit the production condition information to the information processing device 10 every time any of the production condition elements is updated in response to an operation. The image synthesis unit 124 synthesizes an image showing the three-dimensional form of the object based on the production conditions indicated in the production condition information updated in response to an operation, and transmits image data showing the synthesized image to the terminal device 20. The terminal device 20 receives the image data according to the updated production condition information from the information processing device 10. Therefore, the image showing the three-dimensional form is updated every time the production condition information or the display conditions are updated. The terminal device 20 can determine the production condition information and order a fabric sample while viewing the updated image on the display screen in response to an operation.

[0062] The order button bo is a screen component for instructing the ordering of a fabric sample by pressing it. When the display processing unit 224 detects the pressing of the order button bo, it erases the service screen being displayed at that time, and causes the display unit 260 to display the order screen. In this application, "pressing" includes not only the actual pressing, but also the input of an operation signal indicating a position within the display area, that is, an instruction given in response to an operation. An example of the order screen will be described later.

[0063] FIG. 8 is a diagram showing an example of an order screen. The order screen shown in FIG. 8 has an input field for inputting the order conditions of the fabric sample according to an operation, a confirmation button bc, and a previous screen button bb. As order conditions, input fields OP, OA, CM, DQ, and DS are arranged for the orderer's name, orderer's contact information, payment method, quantity, and size, respectively. The input fields MS and CS are input fields that allow material information and fabric information to be input according to an operation. The display processing unit 224 may display the material information and fabric information input using the service screen shown in FIG. 6 by arranging them in the input fields MS and CS, respectively. This allows the user to check the specified material information and fabric information and adjust them by input operations.

[0064] The Confirm button bc is a screen component for confirming an order by pressing it. When the display processing unit 224 detects the pressing of the Confirm button bc, it transmits to the information processing device 10 the order information indicating the input order conditions and the production condition information (i.e., material information and fabric information) in association with each other. The previous screen button bb is a screen item for instructing the display of the previously displayed service screen when pressed. When the display processing unit 224 detects the pressing of the previous screen button bb, it erases the order screen and causes the display unit 260 to display the previously displayed service screen.

[0065] The order processing unit 128 of the information processing device 10 receives order information from the terminal device 20. The order processing unit 128 issues a contract number for identifying each order. The contract number differs for each contract. For example, the order processing unit 128 determines a value that is incremented by a certain value each time new order information is received as the contract number.

[0066] Next, a second example of the display screen will be described. FIG. 7 is a diagram showing the second example of the display screen. The display screen exemplified in FIG. 7 is a service screen for viewing any one of a plurality of pieces of predefined production condition information and ordering a fabric sample. The service screen in FIG. 7 is arranged with an image display field, a display condition specification field, a sample specification field, a desired condition specification field, a fine adjustment button ba, and an order button bo. The image display field is a field for displaying a sample image showing the three-dimensional form of a fabric sample produced based on the specified material information and fabric information.

[0067] The sample designation field includes an input field sw for specifying a fabric sample associated with predetermined production condition information, and a display area for a sample candidate list sl. The input field sw is a field for inputting a characteristic word in response to an operation. The input processing unit 222 of the terminal device 20 specifies the word entered in the input field sw as a search word based on an operation signal, and transmits sample search request information indicating the search word to the information processing device 10 when a press of the search button is detected. The information acquisition unit 122 of the information processing device 10 specifies the search word indicated in the sample search request information received from the terminal device 20, and reads out sample identification information corresponding to a characteristic word matching the specified search word or a synonym thereof from the storage unit 140. The storage unit 140 stores a set of production condition information, image data, and characteristic words (keywords) that are associated with each sample identification information. The sample identification information is, for example, an identifier for uniquely identifying each piece of production condition information or a fabric sample. As the image data to be stored, image data indicating the three-dimensional form of the fabric sample generated by the image synthesis unit 124 based on the corresponding production condition information is used. As the characteristic word, a word expressing the characteristics of the three-dimensional form is used. The words and phrases used are not limited to objective information such as the type of animal (e.g., "fox"), but may include words and phrases expressing subjective information such as psychological impressions of the characteristics (e.g., "solid"), etc. Words and phrases expressing subjective information may include, for example, onomatopoeia (e.g., "fluffy"), etc.

[0068] The information acquiring unit 122 transmits the read sample identification information as a search result to the terminal device 20. The display processing unit 224 of the terminal device 20 arranges the list of sample identification information received from the information processing device 10 on the service screen as a sample candidate list sl. 7, the sample candidate list sl lists samples 01, 04, and 07. Each sample identification information is displayed on a button. The display processing unit 224 selects one of the sample identification information based on the operation signal, and transmits image request information to the information processing device 10 to request image data corresponding to the selected sample identification information. The information acquisition unit 122 of the information processing device 10 reads out from the storage unit 140 image data corresponding to the sample identification information instructed in the image request information received from the terminal device 20, and transmits it to the terminal device 20. The display processing unit 224 of the terminal device 20 arranges an image of the three-dimensional form of the fabric sample shown in the image data received from the information processing device 10 in the image display field of the service screen. In the example of FIG. 7, sample 01 is selected from the sample candidate list sl, and an image showing the three-dimensional form of a fabric sample generated based on the production condition information corresponding to sample 01 is displayed in the image display field.

[0069] The desired condition specification field is an input field for inputting desired conditions for the fabric sample according to operations. In the example of Fig. 7, the desired condition specification field includes input fields dq, ds, and sd for specifying quantity, size, and other conditions (detailed specifications). The order button bo is a button for instructing ordering of a fabric sample by pressing it based on the production condition information corresponding to the designated sample identification information and the desired condition information indicating the designated desired conditions. When the input processing unit 222 detects pressing of the order button bo, it transmits to the information processing device 10 the order information in which the selected sample identification information and the desired condition information indicating the input desired conditions are associated.

[0070] The fine adjustment button ba is a button for instructing, by pressing, fine adjustment of the production condition information corresponding to the selected sample identification information. For example, when detecting a press on the fine adjustment button ba, the display processing unit 224 causes the display unit 260 to display the above-mentioned production condition designation field (see FIG. 6). The user of the terminal device 20 can use the production condition designation field to specify each item of the production condition information by operation. Therefore, it is possible to specify production condition information that is more in line with the user's preferences, based on the production condition information corresponding to the selected sample.

[0071] The order processing unit 128 of the information processing device 10 reads out from the storage unit 140 production condition information corresponding to the sample identification information included in the order information received from the terminal device. The order processing unit 128 may determine a price as an element of the order conditions for the fabric sample based on the read production condition information and the desired condition information corresponding to the order information. The storage unit 140 stores, for example, price calculation data indicating the unit price per unit quantity and the basic fee for each size in association with the production condition information. The order processing unit 128 can refer to the price calculation data to identify the unit price and basic fee corresponding to the received production condition information, and calculate the price based on the identified unit price and basic fee, and the quantity and size indicated in the desired condition information.

[0072] Next, an example of an evaluation screen for a user to respond to evaluation information by operation will be described. FIG. 9 is a diagram showing an example of the evaluation screen. An image display field, an answer field, and an answer button ba are arranged on the evaluation screen. An image showing the three-dimensional form of a fabric sample synthesized based on production condition information associated with a contract number is displayed on the image display field. The answer field is arranged below the contract number and the name of the client. In the example of FIG. 9, color, realism, texture, usefulness, and overall evaluation are listed as answer items, and slider bars ec, er, eq, eu, and et are arranged for each answer item. An evaluation value indicating the subjective degree of goodness or badness for each answer item can be answered according to an operation. The degree of goodness or badness may be indicated by an integer in a predetermined value range (number of stages). The evaluation value may be defined so that the larger the value, the higher the evaluation. The input processing unit 222 is also provided with an answer field em in which any text can be input. The answer button ba is a screen component for instructing transmission of the answered evaluation information by pressing it. When detecting a press on the answer button ba, the input processing unit 222 transmits evaluation information indicating the answer content to the information processing device 10.

[0073] After the ordered fabric sample is produced, the user management unit 130 of the information processing device 10 may transmit guidance information indicating an address (e.g., URL: Uniform Resource Locator) indicating the location of image data synthesized based on the production condition information related to the order to the terminal device 20. The guidance information may be transmitted, for example, as an e-mail. The input processing unit 222 of the terminal device 20 can acquire image data located at the address notified in the guidance information received from the information processing device 10, and configure an evaluation screen based on an image based on the acquired image data, a contract number, and the name of the purchaser. The evaluation screen may include a statement encouraging a comparison with the actual fabric sample. In the example of FIG. 9, a message is written saying, "Please look at the sample you receive and give us your honest opinion on the image on the left." This makes it possible to encourage the purchaser to evaluate the synthesized image by comparing it with the fabric sample actually produced.

[0074] The user management unit 130 of the information processing device 10 stores the evaluation information received from the terminal device 20 in the storage unit 140 in association with the production condition information of the order. The user management unit 130 may select the evaluation value of any one item (e.g., overall evaluation) of the evaluation values ​​for each evaluation item shown in the evaluation information as a representative value, or may calculate a simple average or weighted average of the evaluation values ​​among the evaluation items as the representative value. The user management unit 130 may store the calculated representative value in the storage unit 140 by including it in the evaluation information.

[0075] The model learning unit 126 may configure a training set in which production condition information related to a shipped fabric sample is associated with image data showing a fabric sample image showing the three-dimensional form of the fabric sample. The model learning unit 126 may use the configured training set as part of the training data and execute the above-mentioned learning step to update the model parameters of the image synthesis model. In this learning step, the model learning unit 126 may use an unused newly configured training set to update the existing model parameters without using an existing training set (transfer learning). In addition, in the learning step, the model learning unit 126 may give more importance to a training set related to an image with a higher evaluation. More specifically, when calculating the update amount of the parameter set, the model learning unit 126 multiplies the difference between the signal value (target value) of the image for each training set and the estimated value calculated from the production condition information by a weighting coefficient to calculate a weighted difference, and calculates the update amount using the calculated weighted difference and the gradient of the weighting coefficient of the parameter set instead of a simple difference. Here, the model learning unit 126 may predefine a positive real number as the weighting coefficient, which increases with a representative value with a higher evaluation.

[0076] The fabric samples produced are usually formed by implanting hairs on a flat base fabric. Therefore, the general shape of the fabric sample shown in the fabric sample image used for each training set is a rectangular parallelepiped. In addition, the general shape of the fabric sample shown in the image synthesized by the image synthesis unit 124 is also a rectangular parallelepiped. The image synthesis unit 124 may use a known three-dimensional image processing (3D graphics) technique to deform a part of the shape of the fabric sample shown in the image synthesized in the inference step, and synthesize an image of the fabric sample having the deformed shape. For example, as shown in Figures 6 and 7, the model learning unit 126 may curve the middle part of one cross section of the base fabric forming the fabric sample so that it protrudes more than its surroundings. The position of the base end of each hair implanted on the fabric sample also moves in response to the deformation of the fabric, and the longitudinal direction of the hair changes to a direction intersecting the base fabric at the base end. Therefore, the distribution of colors and patterns of the observed fabric sample as a whole also changes. The model learning unit 126 may determine the shape of the surface of the base fabric using, for example, a predetermined mathematical model (for example, an equation of a quadric surface). In this case, the model learning unit 126 can calculate parameters of the mathematical model according to changes in the positions of predetermined control points, and determine the position of the entire surface using the determined parameters. The model learning unit 126 may adopt the position designated according to an operation as the control point.

[0077] The model learning unit 126 may synthesize an image in which the form and pattern (form) of the fabric sample change according to changes over time in one or both of the part and the amount of deformation of the base fabric. Here, the model learning unit 126 may synthesize an image showing the form of the fabric sample deformed according to the part and the amount of deformation of the base fabric specified in the deformation information received from the terminal device 20. Here, the input processing unit 222 of the terminal device 20 may determine the deformation information in response to an operation signal. Each time a new image is synthesized, the model learning unit 126 transmits image data showing the synthesized image to the terminal device 20. The user of the terminal device 20 can deform the form of the fabric sample image displayed on the service screen by operating it.

[0078] In the above description, the information processing device 10 mainly synthesizes an image showing the three-dimensional form of the fabric using an image synthesis model from production condition information including material information and fabric information, but this is not limited to the above. The information processing device 10 may be configured to determine production condition information including material information and fabric information from an image showing the three-dimensional form of the fabric using an image analysis model, as exemplified in FIG. 10. Next, a second example of the functional configuration of the information processing system 1 according to this embodiment will be described with reference to FIGS. 10 and 11. In the following description, differences from the above functional configuration example will be mainly described, and common reference numerals will be used to refer to the description of the common points.

[0079] FIG. 10 is a schematic block diagram showing a second example of the functional configuration of the information processing system 1 according to the present embodiment. The input processing unit 222 of the terminal device 20 acquires image data showing an image representing the three-dimensional form of the fabric, and transmits the acquired image data to the information processing device 10. The input processing unit 222 may acquire image data from a photographing unit (not shown) built into the terminal device 20 in response to a user's operation, or may acquire image data from other devices (e.g., a photographing device, an image synthesis device, etc.). The display processing unit 224 may display a display screen (e.g., Figs. 6 and 7) including guidance information for guiding the collection of such image data on the display unit 260. Such guidance information may include, for example, text showing a message such as "You can upload a three-dimensional image of your favorite sample and specify your desired conditions."

[0080] The information processing device 10 further includes an image analysis unit 134. In the example shown in Fig. 10, the information processing device 10 includes the image synthesis unit 124, but this may be omitted. The information acquisition unit 122 acquires image data indicating an image (fabric sample image) showing the three-dimensional form of the fabric from the terminal device 20 using the input / output unit 150. The information acquisition unit 122 outputs the acquired image data to the image analysis unit .

[0081] The image analysis unit 134 uses a predetermined image analysis model to determine production condition information including material information and fabric information based on the image shown in the image data input from the information acquisition unit 122. The image analysis unit 134 outputs the determined production condition information to the order processing unit 128. The production condition information can be used for receiving orders or producing fabrics as described above. When the image synthesis model is a reversible mathematical model, the input / output relationship of the image synthesis model can be applied inversely to function as an image analysis model. In that case, the image analysis unit 134 may be configured as a single model calculation unit integrated with the image synthesis unit 124. The model calculation unit can determine an output value indicating production condition information from an input value indicating an image shown in the acquired image data, using the image synthesis model.

[0082] However, if the image synthesis model is a non-reversible mathematical model, the model learning unit 126 needs to learn the image analysis model independently of the image synthesis model. Here, learning and inference of the image analysis model will be described. FIG. 11 is an explanatory diagram showing an example of learning and inference of the image analysis model of this embodiment. The information processing device 10 is capable of executing a learning step (S12) and an inference step (S14). The learning step S12 is a process of determining model parameters for determining material information and fabric information from an input image showing the three-dimensional form of the fabric. In the learning step S12, the model learning unit 126 uses a data set including known input values ​​and target values ​​as a training set, and recursively calculates model parameters so that an estimated value calculated according to the image analysis model for the input value approximates the target value.

[0083] In the learning step S12, the model learning unit 126 uses the signal value of each display element forming the image data as an input value, and uses the numerical value forming the material information and the fabric information as a target value. The fabric sample image is an image representing a fabric sample. In the learning step S12, the model learning unit 126 recursively calculates model parameters so that a loss function indicating the magnitude of the difference between the estimated value and the target value is minimized across multiple training sets. When the loss function becomes equal to or less than a threshold value of a predetermined loss function, the model learning unit 126 determines that the model has converged, and stops the learning step S02. The model learning unit 126 stores image analysis model data indicating the model parameters calculated at that time in the storage unit 140.

[0084] The inference step S14 is a process of calculating numerical values ​​indicating material information and texture information from the signal values ​​of each display element constituting the input image data, using a predetermined image analysis model. The image analysis unit 134 reads out the image analysis model data stored in the storage unit 140, and develops the model parameters indicated in the read image analysis model data into an image analysis model. In the inference step S14, the image analysis unit 134 can identify material information and fabric information based on estimated values ​​calculated using the image analysis model for input values ​​indicating the signal values ​​of each display element.

[0085] As described above, the information processing system 1 according to this embodiment includes an information acquisition unit 122 that acquires material information indicating the characteristics of a fabric material and fabric information indicating the characteristics of a fabric produced from the material, and an image synthesis unit 124 that synthesizes an image of the fabric from the acquired material information and fabric information using an image synthesis model that indicates the relationship between the material information, the fabric information, and an image showing the three-dimensional form of the fabric. The fabric may be artificial fur. According to this configuration, an image showing the three-dimensional form of the fabric is obtained based on the characteristics of the material itself and the characteristics of the fabric made from the material. By visually checking the obtained image, the three-dimensional form according to the characteristics of the material and the fabric can be grasped, which promotes the selection of the desired fabric.

[0086] The information processing system 1 according to this embodiment includes an information acquisition unit 122 that acquires an image showing the three-dimensional form of a fabric, and an image analysis unit 134 that determines the material information and fabric information from the acquired image using a production condition model that indicates the relationship between the image, material information that indicates the characteristics of the material of the fabric, and fabric information that indicates the characteristics of the fabric produced from the material. According to this configuration, material information showing the characteristics of the material itself and fabric information showing the characteristics of the fabric made from the material are specified based on the image showing the three-dimensional form of the fabric. Since the material information and fabric information are specified using the image showing the desired three-dimensional form, the requirements for producing fabric having the desired three-dimensional form can be efficiently understood.

[0087] The material information may include at least one of the brand, variety, fineness, fiber type, cross-sectional shape, luster, color, potential shrinkage characteristics, fiber length, shrinkage characteristics, filament length, total filament fineness, and shrinkage characteristics of the material, and the fabric information may include at least one of the animal type, fabric type, composition ratio between multiple materials, pile length, step information, discoloration processing state, planting density, fiber length, basis weight, planting pattern, trend information, value information, and usage information. With this configuration, at least one of the material brand, variety, fineness, fiber type, cross-sectional shape, luster, color, potential shrinkage characteristics, fiber length, shrinkage characteristics, filament length, total filament fineness, and shrinkage characteristics are used as material characteristics, and at least one of the animal species, fabric type, composition ratio between multiple materials, pile length, step information, discoloration processing state, planting density, fiber length, basis weight, planting pattern, trend information, value information, and usage information are used as fabric characteristics, so that an image can be obtained that shows a three-dimensional form with texture based on both the characteristics of the material itself and the characteristics of the process of processing it into fabric.

[0088] The surface of the fabric may be made of the material and covered with hair colored in a plurality of different colors in the length direction, and the color-changing finish state may show a distribution of the plurality of colors. The distribution of colors applied to the fur covering the fabric results in a highly realistic image of the fabric covered with fur.

[0089] The surface of the fabric may be implanted with fiber bundles, which are a set of multiple types of fibers made of the material, and the fabric information may include fiber bundle configuration information that indicates the configuration of the fiber bundles. The configuration of fiber bundles implanted in the fabric allows for a highly realistic image of the fabric with implanted fiber bundles.

[0090] The information acquisition unit 122 may acquire viewpoint information indicating a viewpoint in a three-dimensional space, and the image synthesis unit 124 may synthesize an image of the texture observed from the viewpoint. With this configuration, the three-dimensional shape of the material shown in the synthesized image can be viewed from various viewpoints, allowing the material to be selected more accurately than if the image were viewed only from a specific viewpoint.

[0091] The information acquisition unit 122 may acquire lighting information indicating lighting conditions, and the image synthesis unit 124 may synthesize an image of the texture represented by illuminating the light according to the lighting conditions. This configuration allows for the acquisition of images of the texture observed under the designated lighting conditions. By understanding the dependency of the three-dimensional shape on the lighting conditions, it is possible to select the texture more accurately than if it were observed under a specific lighting condition.

[0092] The image synthesis unit 124 may deform at least a portion of the fabric, and synthesize an image of the fabric having a shape and pattern that have changed in accordance with the deformation. According to this configuration, the image synthesis unit 124 can synthesize the image of the fabric by further considering either the deformation or the change in the pattern according to the deformation, or both. Therefore, by visually recognizing the three-dimensional form of the fabric shown in the synthesized image, a more accurate selection of the fabric is promoted.

[0093] The image synthesis unit 124 may synthesize images of fabrics whose shapes and patterns change over time. This configuration allows for the acquisition of images of fabrics whose shapes and patterns change over time, and by visually recognizing the dynamic changes in shapes and patterns shown in the synthesized images, the user is prompted to select a fabric more accurately.

[0094] The information acquisition unit 122 may acquire product type information indicating the product type, and the image synthesis unit 124 may synthesize an image of a product of the product type produced from the fabric. According to this configuration, an image of a product produced from the fabric and of the type indicated by the type information is synthesized. By visually checking the synthesized image, the impression of the product shown in the image can be taken into consideration, which encourages an appropriate selection of fabric.

[0095] The image synthesis unit 124 may synthesize an image of a person wearing the product. According to this configuration, the impression of the product when worn by a person and displayed in an image can be taken into consideration, which encourages accurate selection of fabrics.

[0096] Although the embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to the above, and various design changes, etc. are possible within the scope that does not deviate from the gist of the present invention.

[0097] For example, the set of element information constituting material information and the set of element information constituting fabric information are not limited to those described above and can be determined arbitrarily. Furthermore, the arrangement of screen parts on the display screen, the set of display items, the set of setting items, their arrangement, messages, shapes, colors, etc. are not limited to those described above and can be determined arbitrarily. The mathematical model used as the image synthesis model or image analysis model is not limited to CNN, and may be any of RNN (Recurrent Neural Network) and ResNet (Residual Network), or may be a mathematical model other than a neural network. The model learning unit 126 may obtain training data from another device and calculate the model parameters using the obtained training data.

[0098] The information processing device 10 may also include functional units corresponding to the input processing unit 222 and the display processing unit 224 of the terminal device 20. In that case, the transmission and reception of various information between the information processing device 10 and the terminal device 20 corresponds to input and output in the information processing device 10. Furthermore, the model learning unit 126 may be omitted in the information processing device 10. However, the image synthesis unit 124 is set with model parameters acquired in advance. A part of the processing of the image synthesis unit 124, for example, processing related to synthesis (rendering) of the texture sample image in response to any change in the lighting conditions, the viewpoint, or the texture sample shape, may be executed by the display processing unit 224 of the terminal device 20. One or both of the display unit 260 and the operation unit 270 may be omitted in the terminal device 20. The terminal device 20 only needs to be connected wirelessly or wired to a device that is an input source of an operation signal and a destination of display data so as to be able to transmit and receive various types of data.

[0099] In addition, a part or all of the information processing device 10 in the above-mentioned embodiment may be realized as an integrated circuit such as an LSI (Large Scale Integration). Each functional block of the information processing device 10 may be individually made into a processor, or a part or all of the functional blocks may be integrated into a processor. The integrated circuit method is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. In addition, when an integrated circuit technology that replaces LSI appears due to the progress of semiconductor technology, an integrated circuit based on that technology may be used. [Explanation of symbols]

[0100] 10...information processing device, 20...terminal device, 120...control unit, 122...information acquisition unit, 124...image synthesis unit, 126...model learning unit, 128...order processing unit, 130...user management unit, 140...storage unit, 150...input / output unit, 220...control unit, 240...storage unit, 250...input / output unit, 260...display unit, 270...operation unit

Claims

1. An information acquisition unit that acquires material information indicating characteristics of a material of a fabric and fabric information indicating characteristics of a fabric produced from the material; and an image synthesis unit that synthesizes an image of the fabric from the acquired material information and fabric information by using an image synthesis model that indicates a relationship between the material information and the fabric information, and an image that represents a three-dimensional form of the fabric. Information processing system.

2. The information acquisition unit Obtain viewpoint information indicating a viewpoint in three-dimensional space; The image synthesis unit includes: Synthesizing an image of the fabric as viewed from the viewpoint. The information processing system according to claim 1 .

3. The information acquisition unit obtaining lighting information indicative of lighting conditions; The image synthesis unit includes: Synthesizing an image of the texture illuminated according to the lighting conditions The information processing system according to claim 1 .

4. The image synthesis unit includes: Deforming at least a portion of the fabric; Synthesizing an image of the fabric having a shape and pattern that change in accordance with the deformation. The information processing system according to any one of claims 1 to 3.

5. The image synthesis unit includes: Synthesizing an image of the fabric in which the shape and pattern change over time 5. The information processing system according to claim 4.

6. The information acquisition unit Obtaining breed information indicating the breed, The image synthesis unit includes: Synthesize an image of the product of the variety produced from the fabric. The information processing system according to any one of claims 1 to 5.

7. The image synthesis unit includes: Synthesize an image of a person wearing the product 7. The information processing system according to claim 6.

8. an information acquisition unit for acquiring an image representing a three-dimensional form of the fabric; and an image analysis unit that determines the material information and the fabric information from the acquired image by using a production condition model that indicates a relationship between the image, material information that indicates characteristics of the material of the fabric, and fabric information that indicates characteristics of the fabric produced from the material. Information processing system.

9. The material information includes at least one of the brand, variety, fineness, fiber type, cross-sectional shape, gloss, color, potential shrinkage characteristics, shrinkage characteristics, shrinkage characteristics, fiber length, filament length, and total fineness of the material, The fabric information includes at least one of trend information, value information, use information, animal species, fabric type, composition ratio between a plurality of materials, pile length, step information, discoloration processing state, planting density, basis weight, and planting pattern. The information processing system according to any one of claims 1 to 8.

10. The surface of the fabric is made of the material and contains fibers colored in a plurality of different colors in the length direction, The discoloration processing state shows a distribution of the plurality of colors. The information processing system according to claim 9.

11. The surface of the fabric is implanted with fiber bundles which are a set of multiple types of fibers made of the material, The fabric information is Fiber bundle configuration information indicating the configuration of the fiber bundle The information processing system according to any one of claims 1 to 10.

12. The fabric is an artificial fur The information processing system according to any one of claims 1 to 11.

13. An information acquisition unit that acquires material information indicating characteristics of a material of a fabric and fabric information indicating characteristics of a fabric produced from the material; and an image synthesis unit that synthesizes an image of the fabric from the acquired material information and fabric information by using an image synthesis model that indicates a relationship between the material information and the fabric information, and an image that represents a three-dimensional form of the fabric. Information processing device.

14. an information acquisition unit for acquiring an image representing a three-dimensional form of the fabric; and an image analysis unit that determines the material information and the fabric information from the acquired image by using a production condition model that indicates a relationship between the image, material information that indicates characteristics of the material of the fabric, and fabric information that indicates characteristics of the fabric produced from the material. Information processing device.

15. A method in an information processing device, comprising: The information processing device, An information acquisition step of acquiring material information indicating characteristics of a material of a fabric and fabric information indicating characteristics of a fabric produced from the material; and an image synthesis step of synthesizing an image of the fabric from the acquired material information and fabric information by using an image synthesis model indicating a relationship between the material information and the fabric information and an image showing a three-dimensional form of the fabric. Information processing methods.

16. A method in an information processing device, comprising: The information processing device, An information acquisition step of acquiring an image representing a three-dimensional form of the fabric; and an image analysis step of determining the material information and the fabric information from the acquired image by using a production condition model that indicates a relationship between the image, material information that indicates the characteristics of the material of the fabric, and fabric information that indicates the characteristics of the fabric produced from the material. Information processing methods.

Images