Design Evaluation System, Design Evaluation Method, and Program

Abstract

Provided are a design evaluation system, a design evaluation method, and a program capable of objectively evaluating a design. 【Solution means】Based on a first learned model that has learned a first correspondence relationship between a material image showing the design of a material and a feature quantity in which the visual features of the design are digitized, the visual features of the design of the material shown by the material image input as an evaluation target are digitized, and a feature quantity digitization unit that generates feature quantity information indicating the feature quantity of the evaluation target; and based on a second learned model that has learned a second correspondence relationship between the feature quantity and a sensory quantity in which a sensory word representing the visual appearance of the design is digitized, the sensory word corresponding to the feature quantity of the evaluation target indicated by the input feature quantity information is digitized, and a sensory quantity digitization unit that generates sensory quantity information indicating the sensory quantity of the evaluation target. A design evaluation system comprising:

Description

Technical Field

[0001] The present invention relates to a design evaluation system, a design evaluation method, and a program.

Background Art

[0002] Conventionally, in the process of creating a design for building materials or the like, a business operator receives a request from a customer, creates a design, proposes the design to the customer, and receives an evaluation from the customer.

[0003] In relation to this, various technologies capable of automatically generating designs have been proposed. For example, Patent Document 1 below discloses a technology in which an image feature amount corresponding to a feeling such as an atmosphere or taste held by a human with respect to an original image is digitized, and a computer automatically generates other content having a similar feeling to the original image based on the image feature amount.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In the proposal of a design, in order to clearly convey the characteristics of the design, it may be preferable to evaluate the design from a perceptual point of view and convey the characteristics by a perceptual expression. However, since the evaluation of a design depends on an individual's perception, it has been difficult to make an objective evaluation.

[0006] In view of the above problems, an object of the present invention is to provide a design evaluation system, a design evaluation method, and a program capable of objectively evaluating a design.

Means for Solving the Problems

[0007] In order to solve the above problems, a design evaluation system according to one aspect of the present invention includes a feature quantification unit that quantifies the visual features of the design of a material shown in a material image input as an evaluation target based on a first learned model that has learned a first correspondence relationship between the material image showing the design of the material and a feature quantity in which the visual features of the design are digitized, and generates feature quantity information indicating the feature quantity of the evaluation target; and a sensory quantification unit that quantifies the sensory words representing the visual appearance of the design into a sensory quantity, and generates sensory quantity information indicating the sensory quantity of the evaluation target based on a second learned model that has learned a second correspondence relationship between the feature quantity and the sensory quantity. The feature quantity numerically indicates the degree to which the visual features of the design appear in the material image. The sensory word is a verbalization of what a person who has seen the material image feels about the visual appearance of the design. The sensory quantity numerically indicates the degree to which the sensory word applies to the visual appearance of the design. The first learned model is a model that has learned the first correspondence relationship using a first dataset in which the feature quantity indicated by a numerical value evaluated by a person is tagged as first label information for the material image. The second learned model is a model that has learned the second correspondence relationship using a second dataset in which the sensory quantity indicated by a numerical value evaluated by a person is tagged as second label information for the feature quantity information of the material image. It is a design evaluation system.

[0008] The design evaluation method according to one aspect of the present invention is based on a first learned model that has learned a first correspondence relationship between a material image showing the design of a material and a feature quantity in which the visual features of the design are digitized. A feature quantification process for quantifying the visual features of the design of the material shown by the material image input as an evaluation target and generating feature quantity information indicating the feature quantity of the evaluation target; and the feature quantity and the design Based on a second learned model that has learned a second correspondence relationship with a sentiment quantity in which a sentiment word representing the visual appearance is digitized, the sentiment word corresponding to the feature quantity of the evaluation target indicated by the input feature quantity information is digitized, and a sentiment quantification process for generating sentiment quantity information indicating the sentiment quantity of the evaluation target, wherein the feature quantity numerically indicates the degree to which the visual features of the design appear in the material image, and the sentiment word is a verbalization of what a person who has seen the material image feels about the visual appearance of the design, and the sentiment quantity numerically indicates the degree to which the sentiment word applies to the visual appearance of the design. The first learned model is a model that has learned the first correspondence relationship using a first dataset in which the feature quantity indicated by a numerical value evaluated by a human is tagged as first label information for the material image. The second learned model is a model that has learned the second correspondence relationship using a second dataset in which the sentiment quantity indicated by a numerical value evaluated by a human is tagged as second label information for the feature quantity information of the material image. It is a design evaluation method executed by a computer.

[0009] A program according to an aspect of the present invention causes a computer to digitize the visual features of the design of a material shown in a material image indicating the design of the material based on a first learned model that has learned a first correspondence relationship between the material image and a feature amount in which the visual features of the design are digitized, and generate feature amount information indicating the feature amount of the evaluation target; and digitize the perceptual word corresponding to the feature amount of the evaluation target indicated by the input feature amount information based on a second learned model that has learned a second correspondence relationship between the feature amount and a perceptual amount in which the perceptual word representing the visual appearance of the design is digitized, and generate perceptual amount information indicating the perceptual amount of the evaluation target, and function as: the feature amount numerically indicates the degree to which the visual features of the design appear in the material image; the perceptual word is a verbalization of what a human who has viewed the material image has felt about the visual appearance of the design; the perceptual amount numerically indicates the degree to which the perceptual word applies to the visual appearance of the design; the first learned model is a model that has learned the first correspondence relationship using a first dataset in which the feature amount indicated by a numerical value evaluated by a human is tagged as first label information for the material image; and the second learned model is a model that has learned the second correspondence relationship using a second dataset in which the perceptual amount indicated by a numerical value evaluated by a human is tagged as second label information for the feature amount information of the material image.

Advantages of the Invention

[0010] According to the present invention, the design can be objectively evaluated.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Mode for Carrying Out the Invention

[0012] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In this embodiment, a design evaluation system for evaluating the design of a material based on an image showing the design of the material (hereinafter also referred to as a "material image") will be described. The material is, for example, a material such as wood, stone, or fabric (cloth such as fabric or weave). The design to be evaluated is, for example, the appearance of the design of the material. The appearance of the design of the material is, for example, the pattern or motif on the surface of the material. In this case, the material image is an image showing the pattern or motif on the surface of the material. As an example, when the material is wood, the material image is, for example, an image showing the appearance of wood with a wood grain pattern on the surface. The design evaluation system evaluates the appearance of the design of the material shown in the material image input by the user, and outputs information indicating the evaluation result of the design (hereinafter also referred to as "evaluation information") to the user.

[0013] Hereinafter, the present embodiment will be described by taking an example of applying the design evaluation system to the building material design evaluation service. Building materials are materials and members used for interior and exterior decoration of buildings. Building materials include, for example, floor materials, decorative papers, wallpapers, interior decorative materials, etc. A decorative sheet is used for the surface decoration of building materials. Various designs can be printed on the surface of the decorative sheet. The above-described material image uses an image showing a design applicable to the design of building materials. Therefore, the user can apply the design of the material to the design of the building material by printing the material image on the decorative sheet and attaching the decorative sheet to the surface of the building material.

[0014] <1. Overview of Building Material Design Evaluation Service> Referring to FIG. 1, the overview of the building material design evaluation service according to the present embodiment will be described. FIG. 1 is a diagram showing the overview of the building material design evaluation service according to the present embodiment.

[0015] The building material design evaluation service SA shown in FIG. 1 is a service that evaluates the design of building materials handled by a user (for example, a business operator) and provides the user with evaluation information indicating the evaluation result. In the evaluation of the design, feature evaluation and sensory evaluation are performed.

[0016] In the feature evaluation, the visual features of the design of the material shown in the material image are evaluated. The visual features to be evaluated include, for example, layout, balance of square and board patterns, distribution of terry, impression of knots (including leaf knots), shading (intonation), finish feeling, color of conduit, impression of eye width, change amount of eye width, character, painting feeling, directionality of pattern, hardness feeling, roughness feeling, expression of texture, material feeling, etc. The result of the feature evaluation is shown by information in which the visual features of the design are quantified (hereinafter, also referred to as "feature quantity"). The feature quantity numerically indicates the degree to which the visual features of the design appear in the material image. The feature evaluation is performed based on the correspondence relationship (the first correspondence relationship) between the material image showing the appearance of the material design and the feature quantity in which the features of the appearance of the design are quantified. In this embodiment, the feature evaluation is performed by a feature quantification model (the first learned model) that has learned (machine learning) the correspondence relationship between the material image and the feature quantity. The feature quantification model takes as input a material image to be evaluated, quantifies the features of the appearance of the material shown by the material image, and generates and outputs information indicating the feature quantity of the evaluation target (hereinafter, also referred to as "feature quantity information"). The feature quantification model, for example, finely divides the input material image and classifies the feature quantity for each of the divided images by a deep learning model. Thereby, the feature quantification model quantifies the feature quantity of the appearance of the input material image.

[0017] In the sensory evaluation, the sensory words representing the appearance of the material design shown by the material image are evaluated. The sensory words are those in which what a person who has seen the material image feels about the appearance of the material shown by the material image is verbalized. The sensory words to be evaluated are, for example, gentle, elegant, flavorful, heavy, stylish, gorgeous, refined, lively, never boring, refreshing, healing, comfortable, calm, high-quality, natural, etc. The result of the sensory evaluation is indicated by information in which the sensory words for the appearance of the design are quantified (hereinafter, also referred to as "sensory quantity"). The sensory quantity numerically indicates the degree to which the sensory word applies to the appearance of the design. The sensory evaluation is performed based on the correspondence relationship (second correspondence relationship) between the visual feature amounts of the design shown in the material image and the sensory amounts of the visual appearance of the design shown in the material image. In the present embodiment, the sensory evaluation is performed by a sensory quantification model (second learned model) that has learned (machine learned) the correspondence relationship between the feature amount and the sensory amount. The sensory quantification model takes as input a material image to be evaluated and feature amount information, quantifies the sensory words corresponding to the feature amounts of the evaluation target indicated by the feature amount information, and generates and outputs information indicating the sensory amount of the evaluation target (hereinafter, also referred to as "sensory amount information").

[0018] As shown in FIG. 1, a user who uses the building material design evaluation service SA first inputs a material image MG1 to be evaluated into the feature quantification model MD1 (step S11). The feature quantification model MD1 performs a feature evaluation on the input material image MG1 and generates feature amount information. The material image MG1 is an image showing a wood grain design, for example, as shown in FIG. 1.

[0019] Next, the user inputs the feature amount information generated by the feature quantification model MD1 through the feature evaluation into the sensory quantification model MD2 (step S12). The sensory quantification model MD2 performs a sensory evaluation based on the input feature amount information and generates sensory amount information.

[0020] The sensory quantification model MD2 outputs the generated sensory amount information to the user as evaluation information (step S13). The user uses the sensory amount information to propose a design to be applied to the building material to the customer (step S14). The sensory amount indicated by the sensory amount information is the numerical value of the sensory words objectively evaluated by the sensory quantification model MD2. Therefore, by using the sensory amount information indicating an objective evaluation, the user can make a more persuasive proposal to the customer compared to the case of making a proposal using a subjective evaluation by the user himself / herself. Furthermore, by using the sensory amount information, the user can make an easy-to-understand proposal using sensory words compared to the case of making a proposal simply explaining the features of the design, which can also lead to an improvement in the number of customers acquired.

[0021] <2. Configuration of Design Evaluation System> As described above, the building material design evaluation service SA according to this embodiment has been explained. Subsequently, with reference to FIGS. 2 and 3, the configuration of the design evaluation system according to this embodiment will be explained. FIG. 2 is a block diagram showing an example of the configuration of the design evaluation system according to this embodiment. The design evaluation system 1 shown in FIG. 2 is a system for operating the building material design evaluation service SA whose outline was explained with reference to FIG. 1.

[0022] As shown in FIG. 2, the design evaluation system 1 includes an administrator terminal 10, a user terminal 20, a design learning device 30, and a design evaluation device 40. For the network NW, as a configuration for exchanging information, for example, LAN (Local Area Network), WAN (Wide Area Network), telephone network (mobile phone network, fixed phone network, etc.), regional IP (Internet Protocol) network, Internet, etc. are applicable.

[0023] (1) Administrator Terminal 10 The administrator terminal 10 is a terminal that an administrator (business operator) operates to manage the building material design evaluation service SA. The administrator terminal 10 is, for example, a smartphone, a tablet terminal, a PC (Personal Computer), etc. The administrator terminal 10 is communicably connected to the design learning device 30 and the design evaluation device 40 via the network NW.

[0024] In communication with the design learning device 30, the administrator terminal 10 transmits a plurality of material images, label information, etc. prepared for learning, and receives a feature quantification model MD1, a sensibility quantification model MD2, etc. The administrator operates the administrator terminal 10 to transmit the material images and label information for learning to the design learning device 30.

[0025] The learning material images are, for example, image data obtained by scanning pattern manuscript data prepared for building material designs (hereinafter also referred to as "scanned image data"), or image data showing designs created by designers or the like for building material designs (hereinafter also referred to as "design image data").

[0026] Label information is information tagged for learning based on the perspective of applying the material design to the building material design. There is label information prepared for feature learning (first label information) and label information prepared for sentiment learning (second label information).

[0027] The label information prepared for feature learning (hereinafter also referred to as "feature label information") is information indicating the visual feature amounts of the design shown in the learning material image. The feature label information includes information tagged to the learning material image by humans and information tagged to the learning material image by image recognition processing. The information tagged to the learning material image by humans is, for example, a feature amount indicated by a numerical value evaluated by humans. The information tagged to the learning material image by image recognition processing is, for example, a feature amount in which the features captured by image recognition of the learning material image are quantified. The person who evaluates the feature amounts is, for example, a person with knowledge about the material design (an expert). The expert is, for example, a designer. An expert such as a designer, in evaluating the feature amounts, for example, actually looks at the material image, captures the visual features of the design shown in the material image, and answers numerically the degree to which the features appear. Also, a human such as a designer may confirm whether or not each of the features of the design prepared in advance as evaluation items appears in the design of the material image and answer numerically the degree thereof.

[0028] The label information prepared for aesthetic learning (hereinafter also referred to as "aesthetic label information") is information indicating the aesthetic quantity of the appearance of the design shown in the learning material image. The aesthetic label information is information tagged by humans. The information tagged by humans is the aesthetic quantity indicated by a numerical value evaluated by humans. The humans who evaluate the aesthetic quantity are, for example, monitors. The monitors are ordinary people who answer questionnaires for aesthetic quantity evaluation. The persons serving as monitors are not particularly limited, and it is preferable that various persons can be evenly covered regardless of whether they are laypersons, experts, generations, genders, etc. In the questionnaire, for example, a material image to be evaluated and a plurality of aesthetic words are presented to the monitor, and the monitor is asked to answer the degree to which each aesthetic word fits the appearance of the design shown in the material image. The answer regarding the degree to which the aesthetic word fits can be in any method, such as a form in which the most appropriate numerical value is selected from a plurality of options, or a form in which a numerical value is freely input.

[0029] The feature quantification model MD1 and the aesthetic quantification model MD2 are generated by the design learning device 30 based on the learning material image and the label information transmitted from the administrator terminal 10.

[0030] In communication with the design evaluation device 40, the administrator terminal 10 transmits the feature quantification model MD1 and the aesthetic quantification model MD2. The feature quantification model MD1 and the aesthetic quantification model MD2 are those received by the administrator terminal 10 from the design learning device 30.

[0031] On the administrator terminal 10, various UIs (User Inteface) are displayed by an application (hereinafter also referred to as the "administrator application") for the administrator to manage the building material design evaluation service SA. The administrator can manage the building material design evaluation service SA by operating the UI displayed on the administrator terminal 10 by means of the administrator application. Note that the functions of the administrator app may be provided by installing the administrator app on the administrator terminal 10 (i.e., a native app), or may be provided by a web system (i.e., a web app). In the case of a web app, the administrator app is managed by the server, and its functions are provided via a web browser.

[0032] (2) User terminal 20 The user terminal 20 is a terminal that the user operates to use the building material design evaluation service SA. The user terminal 20 is, for example, a smartphone, a tablet terminal, a PC, etc. The user terminal 20 is communicably connected to the design evaluation device 40 via the network NW.

[0033] In communication with the design evaluation device 40, the user terminal 20 transmits a material image and receives evaluation information. The user operates the user terminal 20 to input a material image MG1 to be evaluated and transmits it to the design evaluation device 40. The evaluation information is perceptual quantity information generated by the design evaluation device 40 based on the material image transmitted from the user terminal 20, or information for display generated by processing the perceptual quantity information.

[0034] On the user terminal 20, various UIs are displayed by an application for the user to use the building material design evaluation service SA (hereinafter, also referred to as the "user app"). The user can use the building material design evaluation service SA by operating the UI displayed on the user terminal 20 by the user app. Note that the functions of the user app may be provided by installing the user app on the user terminal 20 (i.e., a native app), or may be provided by a web system (i.e., a web app). In the case of a web app, the user app is managed by the server, and its functions are provided via a web browser.

[0035] (3) Design learning device 30 The design learning device 30 is a device that generates a feature quantification model MD1 and a sensibility quantification model MD2. The design learning device 30 is constituted by, for example, one or more PCs, server devices (for example, cloud servers), and the like. The design learning device 30 is communicably connected to the administrator terminal 10 via the network NW.

[0036] In communication with the administrator terminal 10, the design learning device 30 receives a plurality of material images, label information, etc. prepared for learning, and transmits the feature quantification model MD1, the sensibility quantification model MD2, etc. The design learning device 30 creates a learning dataset based on the learning material images and label information received from the administrator terminal 10, and generates the feature quantification model MD1 and the sensibility quantification model MD2 using the learning dataset.

[0037] Here, referring to FIG. 3, the outline of the design learning according to the present embodiment will be described. FIG. 3 is a diagram showing the outline of the design learning according to the present embodiment. In the design learning, learning is performed on the correspondence relationship between the material image and the feature amount (the first correspondence relationship) and the correspondence relationship between the feature amount of the material image and the sensibility amount (the second correspondence relationship).

[0038] As shown in FIG. 3, in the building material design evaluation service SA, in advance, a learning material image MG2 is prepared by the administrator. The learning material image MG2 is, for example, scan image data or design image data. In the example shown in FIG. 3, as an example, a material image MG2-1 showing a wood grain design and a material image MG2-2 showing a wood grain design different from the material image MG2-1 are shown.

[0039] First, the administrator operates the administrator terminal 10 to tag the feature amount for the learning material image MG2 (step S21). For this purpose, the administrator inputs the feature label information into the administrator terminal 10 and transmits it to the design learning device 30. The feature label information input here is, for example, a feature amount indicated by a numerical value evaluated by a designer. When the design learning device 30 receives the feature label information, it creates a dataset (hereinafter also referred to as the "feature learning dataset") in which the feature label information is tagged to the corresponding learning material image MG2. For example, as shown in FIG. 3, for the material image MG2-1, the design learning device 30 tags the visual feature amount of the design of the material shown in the material image MG2-1 as label information. Also, for the material image MG2-2, the visual feature amount of the design of the material shown in the material image MG2-2 is tagged as label information. Thereby, the design learning device 30 creates a feature learning dataset DS1 in which the feature label information is tagged to each material image MG2.

[0040] Using the created feature learning dataset DS1, the design learning device 30 learns about the features of the material image (the correspondence between the material image and the feature amount), and generates a feature quantification model MD1 (step S22).

[0041] Next, the administrator operates the administrator terminal 10 to tag the emotional amount to the learning material image MG2 (step S23). For this purpose, the administrator inputs the emotional label information to the administrator terminal 10 and transmits it to the design learning device 30. The emotional label information input here is, for example, the emotional amount obtained by a questionnaire for the monitor. When the design learning device 30 receives the aesthetic label information, it creates a dataset (hereinafter also referred to as the "aesthetic learning dataset") in which the corresponding feature learning dataset DS1 is further tagged with the aesthetic label information. For example, as shown in FIG. 3, the design learning device 30 tags the feature learning dataset DS1 of the material image MG2-1 with the aesthetic quantity of the design of the material shown in the material image MG2-1 as label information. Also, for the feature learning dataset DS1 of the material image MG2-2, the design learning device 30 tags the aesthetic quantity of the design of the material shown in the material image MG2-2 as label information. Thereby, the design learning device 30 creates the aesthetic learning dataset DS2 in which the aesthetic label information is tagged for each feature learning dataset DS1.

[0042] Using the created aesthetic learning dataset DS2, the design learning device 30 learns about the aesthetics of the material image (the correspondence between the feature amount and the aesthetic amount of the material image) and generates the aesthetic quantization model MD2 (step S24).

[0043] (4) Design evaluation device 40 The design evaluation device 40 is a device that evaluates the design of the material image MG1 to be evaluated. The design evaluation device 40 is composed of, for example, one or more PCs or server devices (for example, a cloud server). The design evaluation device 40 is communicably connected to the administrator terminal 10 and the user terminal 20 via the network NW.

[0044] In communication with the administrator terminal 10, the design evaluation device 40 receives the feature quantization model MD1 and the aesthetic quantization model MD2. The design evaluation device 40 evaluates the design of the material image MG1 to be evaluated using the feature quantization model MD1 and the aesthetic quantization model MD2 received from the administrator terminal 10.

[0045] In communication with the user terminal 20, the design evaluation device 40 receives a material image and transmits evaluation information. The design evaluation device 40 inputs the material image received from the user terminal 20 into the feature quantification model MD1, and inputs the feature quantity information output from the feature quantification model MD1 into the sensibility quantification model MD2, thereby generating evaluation information.

[0046] <3. Functional Configuration of Design Learning Device 30> As described above, the configuration of the design evaluation system 1 according to the present embodiment has been described. Subsequently, with reference to FIG. 4, the functional configuration of the design learning device 30 according to the present embodiment will be described. FIG. 4 is a block diagram showing an example of the functional configuration of the design learning device 30 according to the present embodiment. As shown in FIG. 4, the design learning device 30 includes a communication unit 310, a storage unit 320, and a control unit 330.

[0047] (1) Communication Unit 310 The communication unit 310 has a function of transmitting and receiving various information. The communication unit 310 is communicably connected to the administrator terminal 10 via the network NW and transmits and receives various information. In communication with the administrator terminal 10, the communication unit 310 receives a plurality of material images, label information, etc. prepared for learning, and transmits the feature quantification model MD1, the sensibility quantification model MD2, etc.

[0048] (2) Storage Unit 320 The storage unit 320 has a function of storing various information. The storage unit 320 is configured by a storage medium included in the design learning device 30 as hardware, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), a RAM (Random Access read / write Memory), a ROM (Read Only Memory), or an arbitrary combination of these storage media. As shown in FIG. 4, the storage unit 320 includes a material image storage unit 321, a learning dataset storage unit 322, and a model storage unit 323.

[0049] (2-1) Material Image Storage Unit 321 The material image storage unit 321 has a function of storing material images. The material image storage unit 321 stores, for example, learning material images MG2 prepared in advance by an administrator and received by the communication unit 310 from the administrator terminal 10. The material image storage unit 321 stores a plurality of material images with different combinations of the appearance characteristics of materials and designs.

[0050] (2-2) Learning Dataset Storage Unit 322 The learning dataset storage unit 322 has a function of storing learning datasets. The learning dataset storage unit 322 stores, for example, learning datasets created by a data processing unit 332 described later. The learning dataset storage unit 322 stores, as learning datasets, a plurality of data with label information tagged for each material image stored in the material image storage unit 321.

[0051] (2-3) Model Storage Unit 323 The model storage unit 323 has a function of storing various models. As shown in FIG. 4, the model storage unit 323 stores a feature quantification model MD1 and a sensibility quantification model MD2.

[0052] (3) Control Unit 330 The control unit 330 has a function of controlling the overall operation of the design learning device 30. The control unit 330 is realized, for example, by causing a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit) included in the design learning device 30 as hardware to execute a program. As shown in FIG. 4, 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.

[0053] (3-1) Data acquisition unit 331 The data acquisition unit 331 has a function of acquiring various types of data. For example, the data acquisition unit 331 acquires the learning material image MG2 and label information received by the communication unit 310 from the administrator terminal 10. The label information acquired by the data acquisition unit 331 includes feature label information or sensory label information.

[0054] (3-2) Data processing unit 332 The data processing unit 332 has a function of performing various types of data processing. For example, the data processing unit 332 creates a learning dataset based on the learning material image MG2 and label information acquired by the data acquisition unit 331. When the data acquisition unit 331 acquires the learning material image MG2 and the feature label information, the data processing unit 332 creates a feature learning dataset DS1 in which the feature label information is tagged to the learning material image MG2. When the data acquisition unit 331 acquires the sensory label information, the data processing unit 332 creates a sensory learning dataset DS2 in which the sensory label information is further tagged to the feature learning dataset DS1.

[0055] (3-3) Learning unit 333 The learning unit 333 has a function of generating a learned model by machine learning. For example, the learning unit 333 learns various correspondence relationships using the learning dataset created by the data processing unit 332.

[0056] When the learning dataset is the feature learning dataset DS1, the learning unit 333 learns about the visual features of the design shown in the material image using the feature learning dataset DS1. In learning the features, the learning unit 333 learns the correspondence between the material image shown in the feature learning dataset DS1 and the feature amount. Through this learning, the learning unit 333 can generate a feature quantification model MD1 that, when a material image is input, quantifies the visual features of the design shown in the material image and generates and outputs feature amount information indicating the feature amount to be evaluated. That is, the feature quantification model MD1 is a model that has learned the correspondence between the material image and the feature amount using the feature learning dataset DS1 (the first dataset) in which the feature amount indicated by a numerical value evaluated by a human is tagged as feature label information (the first label information) for the material image.

[0057] Note that the feature label information according to this embodiment is information tagged for learning based on the perspective of applying the design of the material to the design of the building material. Also, the feature quantification model MD1 used for quantifying the features is a learned model learned using the feature learning dataset DS1 including the feature label information. Therefore, the feature amount information generated by the feature quantification model MD1 is information generated based on the perspective of applying the design of the material to the design of the building material.

[0058] In addition, the feature learning dataset DS1 used by the learning unit 333 for feature learning includes a dataset in which feature label information is tagged by a person having knowledge about the design of the material (for example, a designer), and a dataset in which feature label information is tagged by image recognition of the material image.

[0059] When the learning dataset is the aesthetic learning dataset DS2, the learning unit 333 learns about the visual aesthetics of the design shown in the material image using the aesthetic learning dataset DS2. In the aesthetic learning, the learning unit 333 learns the correspondence between the feature amount and the aesthetic amount of the material image shown in the aesthetic learning dataset DS2. Through this learning, the learning unit 333 can digitize the aesthetic words for the appearance of the design shown in the material image when the material image and the feature amount information are input, and generate and output the aesthetic amount information indicating the aesthetic amount of the evaluation target, thereby generating an aesthetic digitization model MD2. That is, the aesthetic digitization model MD2 is a model that learns the correspondence between the feature amount and the aesthetic amount of the material image using the aesthetic learning dataset DS2 (the second dataset) in which the aesthetic amount indicated by the numerical value evaluated by humans is tagged as aesthetic label information (the second label information) for the feature amount information of the material image.

[0060] Note that the aesthetic label information according to this embodiment is information tagged for learning based on the perspective when applying the design of the material to the design of the building material. Also, the aesthetic digitization model MD2 used for the digitization of aesthetics is a learned model learned using the aesthetic learning dataset DS2 including the aesthetic label information. Therefore, the aesthetic amount information generated by the aesthetic digitization model MD2 is information generated based on the perspective when applying the design of the material to the design of the building material.

[0061] (3-4) Output processing unit 334 The output processing unit 334 has a function of performing processing related to the output of various information. For example, the output processing unit 334 transmits the feature digitization model MD1 and the aesthetic digitization model MD2 generated by the learning unit 333 from the communication unit 310 to the administrator terminal 10.

[0062] <4. Functional configuration of the design evaluation device 40> The functional configuration of the design learning device 30 according to the present embodiment has been described above. Subsequently, with reference to FIG. 5, the functional configuration of the design evaluation device 40 according to the present embodiment will be described. FIG. 5 is a block diagram showing an example of the functional configuration of the design evaluation device 40 according to the present embodiment. As shown in FIG. 5, the design evaluation device 40 includes a communication unit 410, a storage unit 420, and a control unit 430.

[0063] (1) Communication unit 410 The communication unit 410 has a function of transmitting and receiving various information. The communication unit 410 is communicably connected to the administrator terminal 10 and the user terminal 20 via the network NW, and transmits and receives various information to and from each terminal. In communication with the administrator terminal 10, the communication unit 410 receives the feature quantification model MD1 and the sensibility quantification model MD2. In communication with the user terminal 20, the communication unit 410 receives the material image and transmits the evaluation information.

[0064] (2) Storage unit 420 The storage unit 420 has a function of storing various information. The storage unit 420 is configured by a storage medium included in the design evaluation device 40 as hardware, for example, an HDD, an SSD, a flash memory, an EEPROM, a RAM, a ROM, or any combination of these storage media. As shown in FIG. 5, the storage unit 420 includes a model storage unit 421 and a material image storage unit 422.

[0065] (2-1) Model storage unit 421 The model storage unit 421 has a function of storing various models. As shown in FIG. 5, the model storage unit 421 stores the feature quantification model MD1 and the sensibility quantification model MD2. These feature quantification model MD1 and sensibility quantification model MD2 are generated by the design learning device 30 and received by the communication unit 410 from the administrator terminal 10.

[0066] (2-2) Material image storage unit 422 The material image storage unit 422 has a function of storing material images. The material image storage unit 422 stores, for example, a material image MG1 that is prepared by the user and is the object to be evaluated received by the communication unit 410 from the user terminal 20.

[0067] (3) Control unit 430 The control unit 430 has a function of controlling the overall operation of the design evaluation device 40. The control unit 430 is realized, for example, by causing a CPU or GPU included in the design evaluation device 40 as hardware to execute a program. As shown in FIG. 5, the control unit 430 includes an image acquisition unit 431, a design evaluation unit 432, and an output processing unit 433.

[0068] (3-1) Image acquisition unit 431 The image acquisition unit 431 has a function of acquiring a material image MG1 to be evaluated. The image acquisition unit 431 acquires the material image MG1 received by the communication unit 410 from the user terminal 20 as the object to be evaluated.

[0069] (3-2) Design evaluation unit 432 The design evaluation unit 432 has a function of evaluating the design shown by the material image MG1 to be evaluated. As shown in FIG. 5, the design evaluation unit 432 includes a feature quantification unit 4321 and a sensibility quantification unit 4322.

[0070] (3-2-1) Feature quantification unit 4321 The feature quantification unit 4321 has a function of evaluating the design shown by the material image MG1 to be evaluated from the perspective of features. The feature quantification unit 4321 evaluates the visual features of the material shown in the material image MG1 to be evaluated, which is acquired by the image acquisition unit 431, based on the correspondence relationship between the learning material image MG2 and the feature quantity in which the visual features of the design shown in the material image MG2 are quantified. In the evaluation of the visual features, the feature quantification unit 4321 quantifies the visual features of the material shown in the material image MG1 to be evaluated, and generates feature quantity information indicating the feature quantity of the evaluation target.

[0071] The feature quantification unit 4321 generates the feature quantity information by inputting the material image MG1 to be evaluated into the feature quantification model MD1 stored in the model storage unit 421. By using the feature quantification model MD1, the feature quantification unit 4321 can objectively evaluate the features of the evaluation target.

[0072] (3-2-2) Aesthetic quantification unit 4322 The aesthetic quantification unit 4322 has a function of evaluating the design shown in the material image MG1 to be evaluated from the perspective of aesthetics. The aesthetic quantification unit 4322 evaluates the aesthetics of the appearance of the material shown in the material image MG1 to be evaluated, which is acquired by the image acquisition unit 431, based on the correspondence relationship between the feature quantity of the learning material image MG2 and the aesthetic quantity in which the aesthetic words representing the human aesthetics for the appearance of the design shown in the material image MG2 are quantified. In the evaluation of the aesthetics, the aesthetic quantification unit 4322 quantifies the aesthetic words corresponding to the feature quantity of the appearance of the material shown in the material image MG1 to be evaluated, and generates aesthetic quantity information indicating the aesthetic quantity of the appearance of the evaluation target.

[0073] The aesthetic quantification unit 4322 generates the aesthetic quantity information by inputting the feature quantity information of the material image MG1 to be evaluated into the aesthetic quantification model MD2 stored in the model storage unit 421. By using the aesthetic quantification model MD2, the aesthetic quantification unit 4322 can objectively evaluate the aesthetics of the evaluation target.

[0074] (3-3) Output processing unit 433 The output processing unit 433 has a function of performing processing related to the output of various information. For example, the output processing unit 433 transmits the perceptual quantity information generated by the perceptual quantification unit 4322 as evaluation information from the communication unit 410 to the user terminal 20 and causes it to be displayed on the user terminal 20.

[0075] <5. Flow of processing> As described above, the functional configuration of the design evaluation apparatus 40 according to the present embodiment has been described. Subsequently, with reference to FIGS. 6 and 7, the flow of processing according to the present embodiment will be described.

[0076] (1) Design learning processing With reference to FIG. 6, the flow of the design learning processing according to the present embodiment will be described. FIG. 6 is a flowchart showing an example of the flow of the design learning processing according to the present embodiment.

[0077] As shown in FIG. 6, first, the data acquisition unit 331 of the design learning apparatus 30 acquires a learning material image MG2 (step S101). Specifically, the data acquisition unit 331 acquires the learning material image MG2 that is transmitted from the administrator terminal 10 to the design learning apparatus 30 by the administrator and received by the communication unit 310.

[0078] Also, the data acquisition unit 331 acquires the feature label information corresponding to the acquired learning material image MG2 (step S102). Specifically, the data acquisition unit 331 acquires the feature label information that is transmitted from the administrator terminal 10 to the design learning apparatus 30 by the administrator and received by the communication unit 310.

[0079] Next, based on the learning material image MG2 and the feature label information acquired by the data acquisition unit 331, the data processing unit 332 of the design learning device 30 creates a feature learning dataset DS1 (step S103). Specifically, the data processing unit 332 tags each of the learning material images MG2 with the feature label information indicating the corresponding feature amount, thereby creating the feature learning dataset DS1.

[0080] Next, the learning unit 333 of the design learning device 30 uses the feature learning dataset DS1 created by the data processing unit 332 to learn the visual features of the design shown in the material image (step S104). Specifically, the learning unit 333 learns the correspondence between the learning material image MG2 and the feature amounts of the design appearance using the feature learning dataset DS1. Through this learning, the learning unit 333 generates a feature quantification model MD1 (step S105). The output processing unit 334 links the feature quantification model MD1 generated by the learning unit 333 to the design evaluation device 40 via the administrator terminal 10.

[0081] Next, the data acquisition unit 331 acquires the emotional label information corresponding to the acquired learning material image MG2 (step S106). Specifically, the data acquisition unit 331 acquires the emotional label information transmitted from the administrator terminal 10 to the design learning device 30 by the administrator and received by the communication unit 310.

[0082] Next, based on the feature learning dataset DS1 created by the data processing unit 332 and the emotional label information acquired by the data acquisition unit 331, the data processing unit 332 of the design learning device 30 creates an emotional learning dataset DS2 (step S107). Specifically, the data processing unit 332 tags each of the feature learning datasets DS1 with the emotional label information indicating the corresponding emotional amount, thereby creating the emotional learning dataset DS2.

[0083] Next, the learning unit 333 of the design learning device 30 learns the aesthetics of the design shown in the material image using the aesthetic learning dataset DS2 created by the data processing unit 332 (step S108). Specifically, the learning unit 333 learns the correspondence between the visual feature amount and the aesthetic amount of the design of the learning material image MG2 using the aesthetic learning dataset DS2. Through this learning, the learning unit 333 generates an aesthetic quantification model MD2 (step S109). The output processing unit 334 links the aesthetic quantification model MD2 generated by the learning unit 333 to the design evaluation device 40 via the administrator terminal 10.

[0084] (2) Design evaluation process Referring to FIG. 7, the flow of the design evaluation process according to the present embodiment will be described. FIG. 7 is a flowchart showing an example of the flow of the design evaluation process according to the present embodiment.

[0085] As shown in FIG. 7, first, the image acquisition unit 431 of the design evaluation device 40 acquires a material image MG1 to be evaluated (step S201). Specifically, the image acquisition unit 431 acquires the material image MG1 that is input to the user terminal 20 by the user, transmitted from the user terminal 20 to the design evaluation device 40, and received by the communication unit 410.

[0086] Next, the feature quantification unit 4321 of the design evaluation device 40 evaluates the visual features of the design shown in the material image to be evaluated (step S202). Specifically, the feature quantification unit 4321 generates feature amount information by inputting the material image acquired by the image acquisition unit 431 into the feature quantification model MD1 stored in the model storage unit 421.

[0087] Next, the aesthetic quantification unit 4322 of the design evaluation device 40 evaluates the aesthetic sense regarding the appearance of the design shown in the material image to be evaluated (step S203). Specifically, the aesthetic quantification unit 4322 inputs the feature quantity information generated by the feature quantification unit 4321 for the material image acquired by the image acquisition unit 431 into the aesthetic quantification model MD2 stored in the model storage unit 421, thereby generating aesthetic quantity information.

[0088] Next, the output processing unit 433 of the design evaluation device 40 outputs the evaluation information (step S204). Specifically, the output processing unit 433 transmits the aesthetic quantity information generated by the aesthetic quantification unit 4322 as evaluation information from the communication unit 410 to the user terminal 20 and causes it to be displayed on the user terminal 20.

[0089] The flow of the process according to the present embodiment has been described above. As described above, the design evaluation system 1 according to the present embodiment learns a first correspondence relationship between a material image indicating a design of a material and a feature quantity in which visual features of the design are digitized (a first learned model), and based on this, digitizes the visual features of the design of the material indicated by the material image input as an evaluation target, and generates feature quantity information indicating the feature quantity of the evaluation target. A feature quantity digitization unit 4321, and based on a perceptual digitization model (a second learned model) that learns a second correspondence relationship between the feature quantity and a perceptual quantity in which perceptual words representing the visual appearance of the design are digitized, digitizes the perceptual words corresponding to the feature quantity of the evaluation target indicated by the input feature quantity information, and generates perceptual quantity information indicating the perceptual quantity of the evaluation target. The feature quantity digitization unit 4322 is provided. The feature quantity numerically indicates the degree to which the visual features of the design appear in the material image. The perceptual word is a verbalization of what a person who has seen the material image feels about the visual appearance of the design. The perceptual quantity numerically indicates the degree to which the perceptual word fits the visual appearance of the design. The feature quantity digitization model is a model that learns the first correspondence relationship using a feature learning data set DS1 (a first data set) in which a feature quantity indicated by a numerical value evaluated by a person is tagged as feature label information (first label information) with respect to the material image. The perceptual digitization model is a model that learns the second correspondence relationship using a perceptual learning data set DS2 (a second data set) in which a perceptual quantity indicated by a numerical value evaluated by a person is tagged as perceptual label information (second label information) with respect to the feature quantity information of the material image.

[0090] With such a configuration, the design evaluation system 1 according to the present embodiment can evaluate a design without depending on an individual's sensibility. Therefore, the design evaluation system 1 according to the present embodiment enables objective evaluation of a design.

[0091] <6. Modification Example> The above describes the embodiments. Subsequently, modifications of the above-described embodiments will be described. Note that each of the modifications described below may be applied to the embodiments alone or in combination. Further, each modification may be applied in place of the configuration described in the embodiments, or may be additionally applied to the configuration described in the embodiments.

[0092] In the above-described embodiments, an example has been described in which the function of learning a design and the function of evaluating a design are realized by different devices (design learning device 30 and design evaluation device 40), but the present invention is not limited to such an example. For example, the function of learning a design and the function of evaluating a design may be realized by one device having both functions.

[0093] In the above-described embodiments, an example has been described in which the learning material image MG2 and the label information are each transmitted from the administrator terminal 10 to the design learning device 30, and the learning dataset DS is created by the design learning device 30, but the present invention is not limited to such an example. For example, the learning dataset DS may be created in advance by the administrator using the administrator terminal 10 and transmitted from the administrator terminal 10 to the design learning device 30.

[0094] In the above-described embodiments, an example has been described in which the feature quantification model MD1 and the sensibility quantification model MD2 generated by the design learning device 30 are registered in the design evaluation device 40 via the administrator terminal 10, but the present invention is not limited to such an example. For example, the design learning device 30 and the design evaluation device 40 may be able to communicate with each other via the network NW, and the feature quantification model MD1 and the sensibility quantification model MD2 may be directly transmitted from the design learning device 30 to the design evaluation device 40.

[0095] Also, in the above-described embodiments, an example in which the sensory quantity information is output as evaluation information has been described, but the present invention is not limited to such an example. For example, only the feature quantity information may be output as the evaluation information, or information combining the feature quantity information and the sensory quantity information may be output as the evaluation information.

[0096] The above describes the modification examples of the embodiments of the present invention. Note that a part or all of the design evaluation system 1, the administrator terminal 10, the user terminal 20, the design learning device 30, and the design evaluation device 40 in the above-described embodiments may be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed to realize it. Here, the “computer system” is assumed to include hardware such as an OS and peripheral devices. Also, the “computer-readable recording medium” refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, etc., and a storage device such as a hard disk built in a computer system. Further, the “computer-readable recording medium” refers to something that dynamically holds a program for a short time, like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, and may also include something that holds a program for a certain time, like a volatile memory inside a computer system serving as a server or a client in that case. Also, the above program may be for realizing a part of the above-described functions, and may further be realizable in combination with a program already recorded in a computer system for the above-described functions, or may be realized using a programmable logic device such as an FPGA (Field Programmable Gate Array).

[0097] The embodiments of the present invention have been described in detail with reference to the drawings above. However, the specific configuration is not limited to the above, and various design changes and the like can be made without departing from the gist of the present invention.

Explanation of Reference Signs

[0098] 1…Design evaluation system, 10…Administrator terminal, 20…User terminal, 30…Design learning device, 40…Design evaluation device, 310…Communication unit, 320…Storage unit, 321…Material image storage unit, 322…Learning dataset storage unit, 323…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…Model storage unit, 422…Material image storage unit, 430…Control unit, 431…Image acquisition unit, 432…Design evaluation unit, 4321…Feature quantification unit, 4322…Sensibility quantification unit, 433…Output processing unit, MD1…Feature quantification model, MD2…Sensibility quantification model, NW…Network, SA…Building material design evaluation service

Claims

1. a feature quantification unit that quantifies the appearance features of the design of a material represented by a material image input as an evaluation target based on a first trained model that has learned a first correspondence relationship between a material image representing the design of the material and a feature amount that quantifies the appearance features of the design, and generates feature amount information that indicates the feature amount of the evaluation target; a sensitivity quantification unit that quantifies the sensitivity words corresponding to the feature quantities of the evaluation object indicated by the input feature quantity information based on a second trained model that has learned a second correspondence relationship between the feature quantities and sensitivity quantities obtained by quantifying sensitivity words that represent the appearance of the design, and generates sensitivity quantity information indicating the sensitivity quantities of the evaluation object; Equipped with the feature amount is a numerical representation of the degree to which the visual feature of the design is expressed in the material image; The affective words are a verbalization of the feeling of a person who has viewed the material image about the appearance of the design, The affective amount is a numerical representation of the degree to which the affective word corresponds to the appearance of the design, The first trained model is a model that has learned about the first correspondence using a first dataset in which the feature amount indicated by a numerical value evaluated by a human is tagged as first label information for the material image, The second trained model is a model that has learned about the second correspondence using a second dataset in which the affective amount indicated by a numerical value evaluated by a human is tagged as second label information for the feature amount information of the material image. Design evaluation system.

2. The first dataset includes a dataset in which the material image is tagged with the first label information by image recognition. The design evaluation system according to claim 1 .

3. the feature amount information and the affective amount information are information generated based on a viewpoint in applying the design of the material to a design of a building material; The design evaluation system according to claim 1 .

4. a feature quantification process for quantifying the appearance characteristics of the design of a material represented by a material image input as an evaluation object based on a first trained model that has learned a first correspondence relationship between a material image representing the design of the material and a feature amount in which the appearance characteristics of the design have been quantified, and generating feature amount information indicative of the feature amount of the evaluation object; a sensitivity quantification process of quantifying the sensitivity words corresponding to the feature quantities of the evaluation object indicated by the input feature quantity information based on a second trained model that has learned a second correspondence relationship between the feature quantities and sensitivity quantities obtained by quantifying sensitivity words that represent the appearance of the design, and generating sensitivity quantity information indicating the sensitivity quantities of the evaluation object; Including, the feature amount is a numerical representation of the degree to which the visual feature of the design is expressed in the material image; The affective words are a verbalization of the feeling of a person who has viewed the material image about the appearance of the design, The affective amount is a numerical representation of the degree to which the affective word corresponds to the appearance of the design, The first trained model is a model that has learned about the first correspondence using a first dataset in which the feature amount indicated by a numerical value evaluated by a human is tagged as first label information for the material image, The second trained model is a model that has learned about the second correspondence using a second dataset in which the affective amount indicated by a numerical value evaluated by a human is tagged as second label information for the feature amount information of the material image. A computer implemented method for design evaluation.

5. Computer, a feature quantification means for quantifying the appearance features of the design of a material represented by a material image input as an evaluation object based on a first trained model that has learned a first correspondence relationship between a material image representing the design of the material and a feature amount that quantifies the appearance features of the design, and generating feature amount information that indicates the feature amount of the evaluation object; a sensitivity quantification means for quantifying the sensitivity words corresponding to the feature quantities of the evaluation object indicated by the input feature quantity information based on a second trained model that has learned a second correspondence relationship between the feature quantities and sensitivity quantities obtained by quantifying sensitivity words that represent the appearance of the design, and generating sensitivity quantity information indicating the sensitivity quantities of the evaluation object; Function as a the feature amount is a numerical representation of the degree to which the visual feature of the design is expressed in the material image; The affective words are a verbalization of the feeling of a person who has viewed the material image about the appearance of the design, The affective amount is a numerical representation of the degree to which the affective word corresponds to the appearance of the design, The first trained model is a model that has learned about the first correspondence using a first dataset in which the feature amount indicated by a numerical value evaluated by a human is tagged as first label information for the material image, The second trained model is a model that has learned about the second correspondence using a second dataset in which the affective amount indicated by a numerical value evaluated by a human is tagged as second label information for the feature amount information of the material image. program.

Images