Material image generation system, material image generation method, and program

Abstract

This invention provides a material image generation system, a material image generation method, and a program that can reduce the cost of design production. [Solution] A material image generation system comprising: a request information acquisition unit that acquires request information indicating the characteristics of the material pattern requested by the user; and an image generation unit that generates a second material image showing a pattern in accordance with the user's request by inputting the request information into a first trained model that has learned the correspondence between a first material image showing the material pattern and the characteristics of the material pattern shown by the first material image, wherein the first trained model learns the correspondence using a dataset in which label information is labeled to the first material image, the first material image is pattern manuscript data prepared for building material design, and the label information indicates the characteristics of the material pattern extracted from the pattern manuscript data based on the perspective of applying the material pattern to building material design.

Description

Technical Field

[0001] The present invention relates to a material image generation system, a material image generation method, and a program.

Background Art

[0002] Conventionally, in the process of creating designs for building materials and the like, businesses receive requests from customers, create designs, propose the designs to the customers, and receive evaluations from the customers. Usually, this process is often repeated multiple times, and both the businesses and the customers incur high costs until a design agreement is reached.

[0003] In relation to this, various technologies that can automatically generate designs have been proposed. For example, Patent Document 1 below discloses a technique in which a computer automatically generates a wood grain pattern that gives an impression closer to natural wood based on the pattern that appears on the cut surface when a three-dimensional tree model is cut by a predetermined plane.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the technique described in Patent Document 1 above, since the generated design cannot reflect the customer's requests, the business cannot obtain a design that satisfies the customer's requests. Therefore, with this technique, although the creation of the design can be automated, the time taken in the process of determining the design cannot be shortened, and the cost cannot be reduced.

[0006] In view of the above-mentioned problems, the object of the present invention is to provide a material image generation system, a material image generation method, and a program that can reduce the cost of design production. [Means for solving the problem]

[0007] To solve the above-mentioned problems, a material image generation system according to one aspect of the present invention comprises: a request information acquisition unit that acquires request information indicating the characteristics of the pattern of a material requested by a user; and an image generation unit that generates a second material image indicating a pattern generated in accordance with the user's request indicated by the request information by inputting the request information into a first material image indicating the pattern of a material and a first trained model that has learned the correspondence between the pattern of the material indicated by the first material image, wherein the first trained model is a model that has learned the correspondence using a dataset in which label information is labeled to the first material image, the first material image is image data obtained by scanning pattern manuscript data prepared for building material design, and the label information is information indicating the characteristics of the pattern of the material extracted from the pattern manuscript data based on the viewpoint of applying the pattern of the material to building material design.

[0008] A material image generation method according to one aspect of the present invention includes a request information acquisition process for acquiring request information indicating the characteristics of a material pattern requested by a user, and an image generation process for generating a second material image indicating a pattern that is generated in accordance with the user's request indicated by the request information by inputting the request information into a first material image indicating a material pattern and a first trained model that has learned the correspondence between the first material image and the characteristics of the material pattern indicated by the first material image, wherein the first trained model is a model that has learned the correspondence using a dataset in which label information is labeled to the first material image, the first material image is image data obtained by scanning pattern manuscript data prepared for building material design, and the label information is information indicating the characteristics of the material pattern extracted from the pattern manuscript data based on the viewpoint of applying the material pattern to building material design, and is a material image generation method executed by a computer.

[0009] A program according to one aspect of the present invention causes a computer to function as a request information acquisition means for acquiring request information indicating the characteristics of the pattern of a material requested by a user, and an image generation means for generating a second material image indicating a pattern that is generated in accordance with the user's request indicated by the request information by inputting the request information into a first pre-trained model that has learned the correspondence between a first material image indicating the pattern of a material and the characteristics of the pattern of the material indicated by the first material image, wherein the first pre-trained model is a model that has learned the correspondence using a dataset in which label information is labeled to the first material image, the first material image is image data obtained by scanning pattern manuscript data prepared for building material design, and the label information is information indicating the characteristics of the pattern of the material extracted from the pattern manuscript data based on the viewpoint of applying the pattern of the material to building material design. [Effects of the Invention]

[0010] According to the present invention, the cost of design production can be reduced. [Brief explanation of the drawing]

[0011] [Figure 1] This diagram shows an overview of the building materials design output service according to this embodiment. [Figure 2] This is a block diagram showing an example of the configuration of the material image generation system according to this embodiment. [Figure 3] This block diagram shows an example of the functional configuration of the design learning device according to this embodiment. [Figure 4] This is a block diagram showing an example of the functional configuration of the material image generation apparatus according to this embodiment. [Figure 5] This flowchart shows an example of the design learning process flow according to this embodiment. [Figure 6] This flowchart shows an example of the material image generation process according to this embodiment. [Modes for carrying out the invention]

[0012] Embodiments of the present invention will be described in detail below with reference to the drawings. This embodiment describes a material image generation system that generates images showing the design of a material (hereinafter also referred to as "material images"). Examples of materials include wood grain, stone, and fabric. Below, this embodiment will be described using an example of applying the material image generation system to a building materials design output service. A building materials design output service is a service in which businesses handling building materials provide designs for building materials to users (customers). Building materials include, for example, flooring, decorative paper, wallpaper, and interior decorative materials. Decorative sheets are used on building materials for surface finishing. The material image generation system generates material images showing designs applicable to building materials, in response to user requests, and outputs them to the user. The user can then print the material images output from the material image generation system onto decorative sheets and apply the generated material designs to the building materials by attaching these decorative sheets to the surface of the building materials.

[0013] <1. Overview of the Building Materials Design Output Service> Referring to Figure 1, an overview of the building materials design output service according to this embodiment will be described. Figure 1 is a diagram showing an overview of the building materials design output service according to this embodiment.

[0014] In the building materials design output service SA shown in Figure 1, a material image generation model MD is prepared in advance, which has been trained to generate building materials designs based on the training dataset DS (step S0). The training dataset DS is data in which label information is assigned to multiple material images MG1 (the first material image) prepared for training. For example, material image MG1-1 is labeled with the design features of the material represented by material image MG1-1 as label information. Similarly, material image MG1-2, which has a different design from material image MG1-1, is labeled with the design features of the material represented by material image MG1-2 as label information. The design features that are labeled as label information include, for example, color, surface treatment, and the presence or absence of knots.

[0015] A user who uses the building material design output service SA inputs a design request (step S1). The user can input information (hereinafter also referred to as "request information") indicating the characteristics of the design desired by the user, for example, by text, label, image, sketch, etc. In the case of input by text, the user verbalizes the characteristics of the desired design and inputs them as text. In the case of input by label, the user selects a label indicating the characteristics of the desired design. In the case of input by image, the user prepares and inputs an image indicating the characteristics of the desired design. In the case of input by sketch, the user prepares and inputs a sketch indicating the characteristics of the desired design.

[0016] Based on the request information input from the user, the material image generation model MD generates and outputs a material image MG2 (second material image) indicating a design along with the user's request indicated by the request information (step S2). The material image generation model MD generates and outputs an image showing a grain, for example, like the material image MG2 shown in FIG. 1.

[0017] The user evaluates the design of the output material image MG2 (step S3). As a result of the evaluation, if agreement on the output design is obtained (step S4), the user incorporates the design of the material image MG2 into the product design (step S5).

[0018] Note that the material image MG2 evaluated by the user is an image generated in accordance with the user's request and has a design with high appeal to the user. Therefore, it is a design that is likely to obtain agreement on the design. As a result, the frequency of regenerating the material image MG2 without obtaining agreement on the design can be reduced. On the other hand, if agreement on the design is not obtained, the user can also easily regenerate a material image MG2 with a design more in line with the request by adjusting the request to be input.

[0019] <2. Configuration of Material Image Generation System> The building materials design output service SA according to this embodiment has been described above. Next, the configuration of the material image generation system according to this embodiment will be described with reference to Figure 2. Figure 2 is a block diagram showing an example of the configuration of the material image generation system according to this embodiment. The material image generation system 1 shown in Figure 2 is a system for operating the building materials design output service SA, whose outline was explained with reference to Figure 1.

[0020] As shown in Figure 2, the material image generation system 1 comprises an administrator terminal 10, a user terminal 20, a design learning device 30, and a material image generation device 40. Networks (NW) can include various configurations for exchanging information, such as LANs (Local Area Networks), WANs (Wide Area Networks), telephone networks (mobile phone networks, fixed-line telephone networks, etc.), regional IP (Internet Protocol) networks, and the Internet.

[0021] (1) Administrator terminal 10 The administrator terminal 10 is a terminal operated by the administrator (business operator) to manage the building materials design output service SA. The administrator terminal 10 can be, for example, a smartphone, tablet, or PC (Personal Computer). The administrator terminal 10 is connected to the design learning device 30 and the material image generation device 40 via a network NW, enabling communication between them.

[0022] In communication with the design learning device 30, the administrator terminal 10 transmits multiple material images MG1 and label information prepared for learning, and receives the material image generation model MD. The administrator operates the administrator terminal 10 to transmit the material images MG1 and label information for learning to the design learning device 30.

[0023] The training material image MG1 is, for example, image data obtained by scanning pattern design data prepared for building materials design. Label information is information that is labeled to the learning material image MG1 based on the perspective of applying material design to building material design. This label information is prepared, for example, by a designer with experience in building material design, who extracts design features from each individual material image MG1. The material image generation model MD is generated by the design learning device 30 based on the training material image MG1 and label information transmitted from the administrator terminal 10.

[0024] In communication with the material image generation device 40, the administrator terminal 10 transmits the material image generation model MD. This material image generation model MD is received by the administrator terminal 10 from the design learning device 30.

[0025] On the administrator terminal 10, various UIs (User Interfaces) are displayed by an application (hereinafter also referred to as the "administrator app") for the administrator to manage the building materials design output service SA. The administrator can manage the building materials design output service SA by operating the UI displayed on the administrator terminal 10 using the administrator app. The administrator application's functionality may be provided by installing the administrator application on the administrator terminal 10 (i.e., as a native application), or by a web system (i.e., as a web application). In the case of a web application, the administrator application is managed on a server, and its functionality is provided via a web browser.

[0026] (2) User terminal 20 The user terminal 20 is a terminal that the user operates to utilize the building materials design output service SA. The user terminal 20 can be, for example, a smartphone, tablet, or PC. The user terminal 20 is connected to the material image generation device 40 via a network NW in a communicable manner.

[0027] In communication with the material image generation device 40, the user terminal 20 transmits request information and receives the material image MG2. The user operates the user terminal 20 to input request information indicating the desired design features and transmits it to the material image generation device 40. The material image MG2 is generated by the material image generation device 40 based on the request information transmitted from the user terminal 20.

[0028] The user terminal 20 displays various UI elements via an application (hereinafter also referred to as the "user app") that allows the user to utilize the building materials design output service SA. The user can use the building materials design output service SA by operating the UI displayed on the user terminal 20 via the user app. The functionality of the user application may be provided by installing the user application on the user terminal 20 (i.e., a native application), or by a web system (i.e., a web application). In the case of a web application, the user application is managed on a server, and its functionality is provided via a web browser.

[0029] (3) Design learning device 30 The design learning device 30 is a device that generates a material image generation model MD. The design learning device 30 is composed of, for example, one or more PCs or server devices (for example, a cloud server). The design learning device 30 is connected to the administrator terminal 10 via a network NW so as to be able to communicate.

[0030] In communication with the administrator terminal 10, the design learning device 30 receives multiple material images MG1 and label information prepared for learning, and transmits a material image generation model MD. Based on the learning material images MG1 and label information received from the administrator terminal 10, the design learning device 30 creates a learning dataset DS, and generates a material image generation model MD using the learning dataset DS.

[0031] (4) Material image generation device 40 The material image generation device 40 is a device that generates a material image MG2 of the design requested by the user. The material image generation device 40 is composed of, for example, one or more PCs or server devices (for example, a cloud server). The material image generation device 40 is connected to the administrator terminal 10 and the user terminal 20 via a network NW so as to be able to communicate with them.

[0032] During communication with the administrator terminal 10, the material image generation device 40 receives the material image generation model MD. The material image generation device 40 generates a material image MG2 using the material image generation model MD received from the administrator terminal 10.

[0033] In communication with the user terminal 20, the material image generation device 40 receives request information and transmits the material image MG2. The material image generation device 40 generates the material image MG2 by inputting the request information received from the user terminal 20 into the material image generation model MD.

[0034] <3. Functional configuration of the design learning device 30> The configuration of the material image generation system 1 according to this embodiment has been described above. Next, the functional configuration of the design learning device 30 according to this embodiment will be described with reference to Figure 3. Figure 3 is a block diagram showing an example of the functional configuration of the design learning device 30 according to this embodiment. As shown in Figure 3, the design learning device 30 comprises a communication unit 310, a storage unit 320, and a control unit 330.

[0035] (1) Communications Section 310 The communication unit 310 has the function of sending and receiving various types of information. The communication unit 310 is connected to the administrator terminal 10 via the network NW and sends and receives various types of information. In communication with the administrator terminal 10, the communication unit 310 receives multiple material images MG1 prepared for training, label information, etc., and transmits the material image generation model MD.

[0036] (2) Storage section 320 The memory unit 320 has the function of storing various types of information. The memory unit 320 is composed of storage media provided as hardware by the design learning device 30, such as an HDD (Hard Disk Drive), SSD (Solid State Drive), flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), RAM (Random Access read / write Memory), ROM (Read Only Memory), or any combination of these storage media. As shown in Figure 3, the storage unit 320 includes a material image storage unit 321, a training dataset storage unit 322, and a material image generation model storage unit 323.

[0037] (2-1) Material image storage unit 321 The material image storage unit 321 has the function of storing material images. The material image storage unit 321 stores, for example, a learning material image MG1 that is prepared in advance by the administrator and received by the communication unit 310 from the administrator terminal 10. The material image storage unit 321 stores multiple material images MG1 with different combinations of material and design features.

[0038] (2-2) Learning dataset storage unit 322 The training dataset storage unit 322 has the function of storing the training dataset DS. The training dataset storage unit 322 stores, for example, the training dataset DS created by the data processing unit 332, which will be described later. The training dataset storage unit 322 stores multiple data, each labeled with label information for each material image MG1 stored in the material image storage unit 321, as the training dataset DS.

[0039] (2-3) Material image generation model storage unit 323 The material image generation model storage unit 323 has the function of storing material image generation models MD. As shown in Figure 3, the material image generation model storage unit 323 stores a reference model MD1 (first trained model) and an additional model MD2 (second trained model) as material image generation models MD. Both the reference model MD1 and the additional model MD2 are trained models that have learned the correspondence between the training material image MG1 and the design features of the material shown by the training material image MG1. On the other hand, the reference model MD1 and the additional model MD2 use different training datasets DS for training.

[0040] The baseline model MD1 is a model that has learned the correspondence between the design features of the material shown in the training material image MG1 and the training dataset DS, which is labeled as label information for the training material image MG1. In other words, the baseline model MD1 is a model that has learned the design features comprehensively using all the prepared training datasets DS.

[0041] The additional model MD2 is a model that learns the correspondence between the material image MG1 and the design features of the material shown in the material image MG1, using a portion of the training dataset DS, where specific features are labeled as label information for the material image MG1. In other words, the additional model MD2 is a model that learns by focusing on specific design features using a portion of the prepared training dataset DS. The additional model MD2 can be implemented, for example, using LoRA (Low-Rank Adaptation).

[0042] (3) Control unit 330 The control unit 330 has the function of controlling the overall operation of the design learning device 30. The control unit 330 is realized, for example, by causing the CPU (Central Processing Unit) or GPU (Graphics Processing Unit) provided as hardware in the design learning device 30 to execute a program. As shown in Figure 3, the control unit 330 includes a data acquisition unit 331, a data processing unit 332, a learning unit 333, and an output processing unit 334.

[0043] (3-1) Data acquisition unit 331 The data acquisition unit 331 has the function of acquiring various types of data. For example, the data acquisition unit 331 acquires learning material images MG1 and label information received by the communication unit 310 from the administrator terminal 10.

[0044] (3-2) Data processing unit 332 The data processing unit 332 has the function of performing various data processing operations. For example, the data processing unit 332 creates a training dataset DS based on the training material image MG1 and label information acquired by the data acquisition unit 331.

[0045] (3-3) Learning Department 333 The learning unit 333 has the function of generating a trained model by machine learning. For example, the learning unit 333 uses the training dataset DS created by the data processing unit 332 to learn the correspondence between the training material image MG1 and the design features. Through this learning, the learning unit 333 generates a material image generation model MD that, when request information is input, can generate and output a material image that conforms to the user's request indicated by the request information.

[0046] When generating the reference model MD1, the learning unit 333 learns the correspondence between the training material images MG1 and the design features using, for example, all the training datasets DS stored in the training dataset storage unit 322.

[0047] When generating the additional model MD2, the learning unit 333 uses, for example, only the learning dataset DS stored in the learning dataset storage unit 322, in which specific features are labeled as label information, to learn the correspondence between the training material image MG1 and the design features. As an example, the learning unit 333 extracts from all training datasets DS that have "nodes present" labeled as a design feature for the training material image MG1. When training using this extracted training dataset DS, the learning unit 333 can generate an additional model MD2 that is more likely to output material image MG2 with a node-like design. In this way, the learning unit 333 generates and prepares additional models MD2 for multiple specific features in advance.

[0048] The learning unit 333 uses the high-resolution training material image MG1 without reduction to generate a material image generation model MD (reference model MD1 and additional model MD2) that has learned the correspondence relationship. As a result, the material image generation model MD can generate and output a high-resolution material image MG2.

[0049] (3-4) Output processing unit 334 The output processing unit 334 has the function of processing various types of information for output. For example, the output processing unit 334 transmits the material image generation model MD generated by the learning unit 333 from the communication unit 310 to the administrator terminal 10.

[0050] <4. Functional configuration of the material image generation device 40> The functional configuration of the design learning device 30 according to this embodiment has been described above. Next, the functional configuration of the material image generation device 40 according to this embodiment will be described with reference to Figure 4. Figure 4 is a block diagram showing an example of the functional configuration of the material image generation device 40 according to this embodiment. As shown in Figure 4, the material image generation device 40 comprises a communication unit 410, a storage unit 420, and a control unit 430.

[0051] (1) Communications Section 410 The communication unit 410 has the function of sending and receiving various types of information. The communication unit 410 is connected to the administrator terminal 10 and the user terminal 20 via the network NW, and sends and receives various types of information with each terminal. In communication with the administrator terminal 10, the communication unit 410 receives the material image generation model MD. In communication with the user terminal 20, the communication unit 410 receives request information and transmits the material image MG2.

[0052] (2) Storage section 420 The storage unit 420 has the function of storing various types of information. The storage unit 420 is composed of storage media provided as hardware by the material image generation device 40, such as an HDD, SSD, flash memory, EEPROM, RAM, ROM, or any combination of these storage media. As shown in Figure 4, the storage unit 420 comprises a material image generation model storage unit 421 and a material image storage unit 422.

[0053] (2-1) Material image generation model storage unit 421 The material image generation model storage unit 421 has the function of storing material image generation models MD. As shown in Figure 4, the material image generation model storage unit 421 stores a reference model MD1 and an additional model MD2 as material image generation models MD. These reference model MD1 and additional model MD2 are generated by the design learning device 30 and received by the communication unit 410 from the administrator terminal 10.

[0054] (2-2) Material image storage unit 422 The material image storage unit 422 has the function of storing material images. For example, the material image storage unit 422 stores the material image MG2 generated by the image generation unit 432, which will be described later.

[0055] (3) Control unit 430 The control unit 430 has the function of controlling the overall operation of the material image generation device 40. The control unit 430 is implemented, for example, by causing the CPU or GPU provided as hardware in the material image generation device 40 to execute a program. As shown in Figure 4, the control unit 430 includes a request information acquisition unit 431, an image generation unit 432, a model output control unit 433, and an output processing unit 434.

[0056] (3-1) Request Information Acquisition Unit 431 The request information acquisition unit 431 has the function of acquiring request information. The request information acquisition unit 431 acquires request information received by the communication unit 410 from the user terminal 20.

[0057] (3-2) Image generation unit 432 The image generation unit 432 has the function of generating material image MG2. Based on the correspondence between the training material image MG1 and the design features of the material shown by the training material image MG1, the image generation unit 432 generates material image MG2 that shows a design in line with the user's requests indicated by the request information acquired by the request information acquisition unit 431. The image generation unit 432 generates material image MG2 by inputting the request information into the reference model MD1 stored in the material image generation model storage unit 421. In this way, the image generation unit 432 can generate a highly designed material image MG2 that matches the user's image by using the requested information.

[0058] Furthermore, the image generation unit 432 generates a material image MG2 using a reference model MD1 that has been trained on the correspondence relationship using a high-resolution training material image MG1 without reduction. As a result, the image generation unit 432 can output a high-resolution material image MG2.

[0059] (3-3) Model output control unit 433 The model output control unit 433 has the function of controlling the output of the material image generation model MD. The model output control unit 433 controls the output of the reference model MD1 using an additional model MD2 that has learned the correspondence between specific features of the material design shown in the training material image MG1. For example, if the request information acquired by the request information acquisition unit 431 includes a specific feature, the model output control unit 433 uses the additional model MD2 to control the reference model MD1 to generate and output a material image MG2 with a design that is aligned with the specific feature. As a result, the model output control unit 433 can output a material image MG2 that matches the user's image with greater accuracy.

[0060] (3-4) Output processing unit 434 The output processing unit 434 has the function of processing various types of information for output. For example, the output processing unit 434 transmits the material image MG2 generated by the image generation unit 432 from the communication unit 410 to the user terminal 20 and displays it on the user terminal 20.

[0061] <5. Processing Flow> The functional configuration of the material image generation apparatus 40 according to this embodiment has been described above. Next, the processing flow according to this embodiment will be described with reference to Figures 5 and 6.

[0062] (1) Design learning process Referring to Figure 5, the flow of the design learning process according to this embodiment will be explained. Figure 5 is a flowchart showing an example of the flow of the design learning process according to this embodiment.

[0063] As shown in Figure 5, first, the data acquisition unit 331 of the design learning device 30 acquires the learning material image MG1 (step S101). Specifically, the data acquisition unit 331 acquires the learning material image MG1 that is transmitted from the administrator terminal 10 to the design learning device 30 by the administrator and received by the communication unit 310.

[0064] Furthermore, the data acquisition unit 331 acquires label information corresponding to the acquired learning material image MG1 (step S102). Specifically, the data acquisition unit 331 acquires label information that is transmitted from the administrator terminal 10 to the design learning device 30 by the administrator and received by the communication unit 310.

[0065] Next, the data processing unit 332 of the design learning device 30 creates a learning dataset DS based on the learning material images MG1 and label information acquired by the data acquisition unit 331 (step S103). Specifically, the data processing unit 332 creates the learning dataset DS by labeling each of the learning material images MG1 with the corresponding label information.

[0066] Next, the learning unit 333 of the design learning device 30 learns using the learning dataset DS created by the data processing unit 332 (step S104). Specifically, the learning unit 333 learns the correspondence between the learning material image MG1 and the design features using all of the learning dataset DS. Through this learning process, the learning unit 333 generates a reference model MD1 (step S105). The output processing unit 334 transmits the reference model MD1 generated by the learning unit 333 to the material image generation device 40 via the administrator terminal 10.

[0067] Next, the learning unit 333 extracts training datasets DS (step S106). Specifically, the learning unit 333 extracts from all training datasets DS that have been labeled with specific features, for example, those specified by the administrator, as label information.

[0068] Next, the learning unit 333 performs training using the extracted training dataset DS (step S107). Specifically, the learning unit 333 uses only the extracted training dataset DS to learn the correspondence between the training material image MG1 and the design features. Through this learning process, the learning unit 333 generates the additional model MD2 (step S108). The output processing unit 334 transmits the additional model MD2 generated by the learning unit 333 to the material image generation device 40 via the administrator terminal 10.

[0069] (2) Material image generation process Referring to Figure 6, the flow of the material image generation process according to this embodiment will be explained. Figure 6 is a flowchart showing an example of the flow of the material image generation process according to this embodiment.

[0070] As shown in Figure 6, first, the request information acquisition unit 431 of the material image generation device 40 acquires request information (step S201). Specifically, the request information acquisition unit 431 acquires request information that is input by the user to the user terminal 20, transmitted from the user terminal 20 to the material image generation device 40, and received by the communication unit 410.

[0071] Next, the model output control unit 433 of the material image generation device 40 checks whether the user's request, as indicated by the request information acquired by the request information acquisition unit 431, includes a specific feature (step S202). If the specific feature is included (step S202 / YES), the process proceeds to step S203. On the other hand, if the specific feature is not included (step S202 / NO), the process proceeds to step S205.

[0072] If the process proceeds to step S203, the model output control unit 433 selects an additional model MD2 to apply to the reference model MD1 (step S203). Specifically, the model output control unit 433 selects an additional model MD2 from among the available additional model MD2s that has learned specific features included in the user's request. The model output control unit 433 applies the selected additional model MD2 to the reference model MD1 (step S204). After application, the process proceeds to step S205.

[0073] If the process proceeds to step S205, the image generation unit 432 generates the material image MG2 (step S205). Specifically, the image generation unit 432 generates the material image MG2 by inputting the request information acquired by the request information acquisition unit 431 into the reference model MD1. Furthermore, if the user's request includes specific features, the image generation unit 432, based on control by the model output control unit 433, inputs the request information to the reference model MD1 to which the additional model MD2 has been applied, thereby generating a material image MG2 with a design tailored to those specific features.

[0074] Next, the output processing unit 434 of the material image generation device 40 transmits the material image MG2 generated by the image generation unit 432 from the communication unit 410 to the user terminal 20, and displays it on the user terminal 20 (step S206).

[0075] The processing flow according to this embodiment has been described above. As described above, the material image generation system 1 according to this embodiment includes a request information acquisition unit 431 that acquires request information indicating the design features requested by the user, and an image generation unit 432 that generates a second material image showing a design in line with the user's requests indicated by the request information, based on the correspondence between a first material image showing the material design and the design features of the material shown by the first material image.

[0076] With this configuration, the material image generation system 1 according to this embodiment can reflect the user's (customer's) requests in the generated design. As a result, businesses and customers can reduce the number of times they have to go through the cycle of production, proposal, and evaluation in the design production process, and shorten the time required for the design production process. Therefore, the material image generation system 1 according to this embodiment makes it possible to reduce the cost of design production.

[0077] Furthermore, the material image generation system 1 according to this embodiment can output a high-resolution material image MG2 by generating a material image MG2 using a reference model MD1 that has been trained using a high-resolution training material image MG1 without reduction. As a result, the user can obtain a high-resolution material image MG2 that is suitable for printing.

[0078] Furthermore, the material image generation system 1 according to this embodiment can generate a material image MG2 with high design quality (high aesthetic appeal) that matches the user's image by using the requested information. As a result, users can utilize high-aesthetic decorative sheets not only for building materials but also for products that use decorative sheets for surface decoration.

[0079] Furthermore, the material image generation system 1 according to this embodiment achieves a dramatic improvement in the efficiency of the design production process by shortening the time required for the design production process, and can support the expansion of the user's creativity and the realization of their ideal surface design or space. Furthermore, users can leverage the unique characteristics of the output material images (high resolution, sophisticated design, etc.) and utilize them in digital content.

[0080] <6. Variation> The embodiments have now been described. Next, modifications of the embodiments described above will be explained. Each modification described below may be applied to the embodiments individually or in combination. Furthermore, each modification may be applied in place of the configuration described in the embodiments or additionally to the configuration described in the embodiments.

[0081] The embodiments described above illustrate an example of applying the material image generation system to a building materials design output service, but the system is not limited to such examples. For example, the material image generation system may be applied to a service provided in a virtual space (metaverse space).

[0082] Furthermore, although the above-described embodiment explains an example in which the design learning function and the material image generation function are implemented by different devices (design learning device 30 and material image generation device 40), the invention is not limited to such an example. For example, the design learning function and the material image generation function may be implemented by a single device that has both functions.

[0083] Furthermore, in the above-described embodiment, an example was explained in which the learning material image MG1 and label information are transmitted from the administrator terminal 10 to the design learning device 30, and the learning dataset DS is created in the design learning device 30. However, the invention is not limited to this example. For example, the learning dataset DS may be created in advance by the administrator at the administrator terminal 10 and transmitted from the administrator terminal 10 to the design learning device 30.

[0084] Furthermore, although the above-described embodiment described an example in which the material image generation model MD generated by the design learning device 30 is registered with the material image generation device 40 via the administrator terminal 10, the invention is not limited to this example. For example, the design learning device 30 and the material image generation device 40 may be able to communicate with each other via a network NW, and the material image generation model MD may be transmitted directly from the design learning device 30 to the material image generation device 40.

[0085] The above describes some modified embodiments of the present invention. Furthermore, some or all of the material image generation system 1, administrator terminal 10, user terminal 20, design learning device 30, and material image generation device 40 in the above-described embodiment may be implemented using a computer. In that case, the program for implementing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be loaded into a computer system and executed. Hereinafter, "computer system" includes hardware such as the OS and peripheral devices. Furthermore, "computer-readable recording media" refers to portable media such as flexible disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computer systems. In addition, "computer-readable recording media" may also include those that dynamically hold programs for a short period of time, such as communication lines used when transmitting programs over networks such as the Internet or communication lines such as telephone lines, and those that hold programs for a certain period of time, such as volatile memory inside computer systems that act as servers or clients in such cases. Furthermore, the above program may be for the purpose of implementing some of the functions described above, or it may be for the purpose of implementing the above functions in combination with a program already recorded in the computer system, or it may be implemented using a programmable logic device such as an FPGA (Field Programmable Gate Array).

[0086] Although embodiments of this invention have been described in detail above with reference to the drawings, the specific configuration is not limited to those described above, and various design changes can be made without departing from the spirit of this invention. [Explanation of symbols]

[0087] 1…Material image generation system, 10…Administrator terminal, 20…User terminal, 30…Design learning device, 40…Material image generation device, 310…Communication unit, 320…Storage unit, 321…Material image storage unit, 322…Learning dataset storage unit, 323…Material image generation model storage unit, 330…Control unit, 331…Data acquisition unit, 332…Data processing unit, 333…Learning unit, 334…Output processing unit, 410…Communication unit, 420…Storage unit, 421…Material image generation model storage unit, 422…Material image storage unit, 430…Control unit, 431…Request information acquisition unit, 432…Image generation unit, 433…Model output control unit, 434…Output processing unit, MD…Material image generation model, MD1…Reference model, MD2…Additional model, NW…Network, SA…Building material design output service

Claims

1. A request information acquisition unit that acquires request information indicating the characteristics of the material pattern requested by the user, An image generation unit generates a second material image showing a pattern that is generated in accordance with the user's request, by inputting the request information into a first trained model that has learned the correspondence between a first material image showing the pattern of a material and the characteristics of the pattern of the material shown in the first material image. Equipped with, The first pre-trained model is a model that has learned the correspondence relationship using a dataset in which label information is assigned to the first material images. The first material image described above is image data obtained by scanning pattern design data prepared for building materials design. The label information is information that indicates the characteristics of the material's pattern, extracted from the pattern draft data, based on the perspective of applying the material's pattern to building material design. Material image generation system.

2. A model output control unit controls the output of the first trained model, using a second trained model which has learned the correspondence between the first trained model and the first material image, focusing on specific features of the material's pattern. The material image generation system according to claim 1, further comprising:

3. The model output control unit controls the system to generate and output the second material image of the pattern having the specific feature, using the second trained model, if the request information includes the specific feature. The material image generation system according to claim 2.

4. The request information acquisition unit acquires the request information for the newly generated material pattern during the production process of the material pattern to be printed on the decorative sheet used for surface finishing of the product. The material image generation system according to claim 1.

5. The process of obtaining request information involves acquiring request information that indicates the characteristics of the material pattern requested by the user, An image generation process that generates a second material image showing a pattern generated in accordance with the user's request, by inputting the request information into a first trained model that has learned the correspondence between a first material image showing the pattern of a material and the characteristics of the pattern of the material shown in the first material image, Includes, The first pre-trained model is a model that has learned the correspondence relationship using a dataset in which label information is assigned to the first material images. The first material image described above is image data obtained by scanning pattern design data prepared for building materials design. The label information is information that indicates the characteristics of the material's pattern, extracted from the pattern draft data, based on the perspective of applying the material's pattern to building material design. A method for generating material images performed by a computer.

6. Computers, A means for obtaining request information that indicates the characteristics of the pattern of the material requested by the user, Image generation means that generates a second material image showing a pattern generated in accordance with the user's request as indicated by the request information, by inputting the request information into a first trained model that has learned the correspondence between a first material image showing the pattern of a material and the characteristics of the pattern of the material shown in the first material image, To make it function as, The first pre-trained model is a model that has learned the correspondence relationship using a dataset in which label information is assigned to the first material images. The first material image described above is image data obtained by scanning pattern design data prepared for building materials design. The label information is information that indicates the characteristics of the material's pattern, extracted from the pattern draft data, based on the perspective of applying the material's pattern to building material design. program.

Images