Object design assistance device, control method of object design assistance device, and control program of object design assistance device

Abstract

The present invention provides an object design assistance device or the like with which it is possible to more easily design an object that can be actually produced. The object design assistance device includes: a storage unit for storing images of a plurality of actually present real objects in association with three-dimensional information of the real objects; a selection reception unit for receiving selection of at least one or more images as a reference image from the images of the real objects stored in the storage unit; an indication reception unit for receiving an indication for generation that designates a feature of the object to be generated; a generation unit for generating the image of a virtual object having the feature designated by the indication for generation on the basis of the reference image using a predetermined generation model; and a comparison unit for determining whether or not the result of comparing the image of the virtual object with the images of the plurality of real objects stored in the storage unit satisfies a predetermined condition. The generation unit repeats generation of the image of the virtual object until it is determined by the comparison unit that the predetermined condition is satisfied and, for the virtual object satisfying the predetermined condition, generates three-dimensional information of that virtual object.

Description

Object design support device, control method for object design support device, and control program for object design support device

[0001] The present invention relates to an object design support device, a control method for an object design support device, and a control program for an object design support device.

[0002] Conventionally, a virtual clothing generation system that generates 3D (3-dimensional) clothing model data, pattern data, and / or 3D fabric model data from 2D (2-dimensional) clothing image data using a machine learning model has been known (see, for example, Patent Literature 1). Furthermore, in recent years, there has been progress in the development of image generation AI (artificial intelligence) that generates images in response to sentences (so-called prompts) that specify features or images (see, for example, Non-Patent Literature 1).

[0003] JP 2023-110395 A

[0004] CANVA, "Free online AI image generator," [online], searched October 26, 2023, Internet <URL: https: / / www.canva.com / ja_jp / ai-image-generator / >

[0005] Using the above technology, it is possible to use image generation AI to design objects such as clothing as 2D images, then convert them into 3D model data and output them. However, further data processing is required before the clothing designed by image generation AI can be worn. In other words, a method for more easily designing producible objects is needed.

[0006] An object design support device according to one embodiment of the present invention includes a memory unit that stores images of multiple real objects that exist in reality in association with three-dimensional information about the real objects; a selection receiving unit that receives a selection of at least one image as a reference image from the images of the real objects stored in the memory unit; an instruction receiving unit that receives a generation instruction that specifies the characteristics of the object to be generated; a generation unit that generates an image of a virtual object having the characteristics specified in the generation instruction based on the reference image using a predetermined generative model; and a comparison unit that determines whether a result of comparing the image of the virtual object with images of the multiple real objects stored in the memory unit satisfies predetermined conditions. The generation unit repeats generating images of the virtual objects until the comparison unit determines that the predetermined conditions are satisfied, and generates three-dimensional information about the virtual object for virtual objects that satisfy the predetermined conditions.

[0007] In an object design support device according to an embodiment of the present invention, a creation instruction may include, as a feature of the object to be created, a designation of the shape of each element that constitutes the object.

[0008] In an object design support device according to an embodiment of the present invention, a creation instruction may include, as a feature of the object to be created, a specification of at least one of color, pattern, and material that indicates the aspect of the object.

[0009] In an object design support device according to an embodiment of the present invention, a generation instruction may include, as a feature of the object to be generated, a specification of the dimensions of each element that constitutes the object.

[0010] In the object design support device according to one embodiment of the present invention, the selection receiving unit may receive a reference image for each element that constitutes the object to be generated.

[0011] In an object design support device according to one embodiment of the present invention, the generation unit may generate an image of a virtual object for each of a plurality of real object images stored in the memory unit by using a generation model created by learning the association between the shape, dimensions, and appearance of each element constituting the real object as three-dimensional information of the real object.

[0012] In an object design support device according to one embodiment of the present invention, the comparison unit may determine that a predetermined condition is satisfied when the difference between the feature amount calculated from the feature of the virtual object generated by the generation unit and the feature amount calculated from the feature of a plurality of real objects stored in the memory unit is equal to or greater than a predetermined threshold value.

[0013] A control method for an object design support device according to one embodiment of the present invention includes the steps of: storing images of multiple real objects that exist in reality in a memory unit in association with three-dimensional information about the real objects; accepting the selection of at least one image as a reference image from the images of the real objects stored in the memory unit; accepting a generation instruction that specifies the characteristics of the object to be generated; generating an image of a virtual object having the characteristics specified in the generation instruction based on the reference image; and determining whether a result of comparing the image of the virtual object with the images of the multiple real objects stored in the memory unit satisfies a predetermined condition. The generating step repeats generation of images of the virtual object until the determining step determines that the predetermined condition is satisfied, and generates three-dimensional information about the virtual object for any virtual object that satisfies the predetermined condition.

[0014] A control program for an object design support device according to one embodiment of the present invention implements the following functions: storing images of multiple real objects that exist in reality in a memory unit in association with three-dimensional information about the real objects; accepting the selection of at least one image as a reference image from the images of the real objects stored in the memory unit; accepting a generation instruction that specifies the characteristics of the object to be generated; generating, using a predetermined generation model, an image of a virtual object having the characteristics specified in the generation instruction based on the reference image; and determining whether a result of comparing the image of the virtual object with images of the multiple real objects stored in the memory unit satisfies predetermined conditions. The generating function repeats generating images of the virtual object until the determining function determines that the predetermined conditions are satisfied, and generates three-dimensional information about the virtual object for virtual objects that satisfy the predetermined conditions.

[0015] According to one embodiment of the present invention, it is possible to provide an object design support device or the like that can more easily design producible objects.

[0016] Fig. 1 is a functional block diagram of an object design support device according to an embodiment of the present invention. Fig. 2 is an example of a data table stored in an object design support device according to an embodiment of the present invention. Fig. 3 is a flowchart showing an example of an information processing method according to an embodiment of the present invention. Fig. 4 is an example of the hardware configuration of an information processing device (computer) according to an embodiment of the present invention.

[0017] Hereinafter, an embodiment of the invention according to the present disclosure (also referred to as the present invention) will be described using the drawings. Note that the drawings are merely examples, and the present invention is not limited to those shown in the drawings. For example, the block diagram, flowchart, data table, etc. of the object design support device (information processing device) shown in the drawings are merely examples, and the present invention is not limited to these.

[0018] As used herein, an "object" refers to an object designed by a user, and includes, for example, "clothing," "furniture," "architecture," "tableware," "bags," "watches," "vehicles," "jewelry," "electrical appliances," "folk crafts," and "living things." Note that, as used herein, "design" refers not only to the creation of a two-dimensional design but also to the planning of the actual production of an object. Hereinafter, clothing will be described as an example of an object, but "clothing" may include pants, skirts, dresses, jumpsuits, shirts, T-shirts, jackets, coats, swimsuits, and the like, regardless of whether they are Western or Japanese clothing. Furthermore, "clothing" may include accessories such as hats, scarves, gloves, ties, and belts; ornaments such as necklaces, earrings, watches, glasses, and masks; footwear such as shoes, sandals, and geta; and sanitary masks.

[0019] <Overview of Object Design Support System> First, an object design support system using an object design support device will be described with reference to FIG. 1 . The object design support device 100 is connected to a user's information processing terminal (not shown) (hereinafter also referred to as a "user terminal"), generates an image of a virtual object in response to a prompt input at the user terminal, and outputs the image to the user terminal. The object design support device 100 generates an image of a virtual object using a generative model learned from an image of a real object that exists in reality, along with three-dimensional information about the real object associated with the image. The object design support device 100 may output a three-dimensional model of the virtual object. The three-dimensional model of the virtual object is information necessary to create a three-dimensional virtual object, and may be, for example, a polygon mesh model (mesh model) including a set of vertices, edges, and faces that define the shape of the virtual object's appearance (polyhedral object).

[0020] Images of real objects that exist in reality and three-dimensional information about the real objects are stored in a real object database (DB) 400. The three-dimensional information about the real objects may be information embedded in the images or information that can be identified from multiple images. For example, the images of the real objects may be multiple two-dimensional images of the real object captured from multiple directions, and the three-dimensional shape (stereoscopic shape) of the real object may be reconstructed from these multiple images. For example, the images of the real objects may be generated by photogrammetry based on multiple images of the object captured from multiple directions. Photogrammetry is a technique for capturing images of a subject from various angles and analyzing and integrating the digital images to create a three-dimensional 3DCG (3D computer graphics) three-dimensional model. The method for capturing images from multiple directions is not particularly limited. Multiple images may be captured by regularly arranging multiple (e.g., 200) camera units (cameras) around the subject, or by capturing images with cameras while rotating a turntable on which the subject is placed. Alternatively, multiple images may be taken by moving the camera while capturing images without moving the subject. Note that an existing method for generating a three-dimensional model from multiple images may be used.

[0021] Alternatively, the image and three-dimensional information of the object may be acquired by photographing the object with a 3D camera capable of capturing depth information.

[0022] The applicant has a large collection of rare vintage clothing, and in one embodiment of the present invention, the images and three-dimensional information of the real objects stored in the real object DB 400 may be obtained by photographing the clothing.

[0023] In the real object DB 400, images of real objects may be stored in association with information about the features of the real objects. FIG. 2 shows an example of a data table storing information about real objects when the real object is "clothing." Note that the figure is an example, and the stored information is not limited to that shown, and the stored information may be less than that shown. For example, in the figure, a "*" indicates that information is stored, and a "-" indicates that information is not stored. Note that the information described below may be associated by manual annotation or by machine learning.

[0024] The data table TB10 may store information relating to the real object's category, attributes, parts, decorations, age, etc., in association with an identifier that uniquely identifies the real object. A category is information indicating the type of real object, and may be, for example, "skirt," "shirt," "jacket," "pullover," etc. Categories may also be divided into major and minor categories. The real object DB 400 may also store information relating to the attributes of real objects. The attributes of real objects are information indicating the characteristics of the real object, and may be, for example, information such as color, pattern, material, etc.

[0025] Furthermore, in the real object DB 400, an image of a real object may be associated with information about the shape of each element that constitutes the real object. Each element that constitutes an object is a part that forms the object. For example, in the case of clothing, it may be a pattern or clothing accessory that is sewn together with thread, such as a collar, sleeves, shoulders, length, pockets, buttons, etc. Information about the shape of each element that constitutes the object is information about the shape and type of each element. For example, if the element is a "collar," information about the shape may include information about the shape, such as a regular collar, roll collar, shawl collar, bow tie, open collar, etc. Furthermore, if the element is a "sleeve," information about the length, such as short sleeves or long sleeves, as well as information about the shape, such as dolman sleeves, French sleeves, sleeveless, puff sleeves, etc.

[0026] 2, information on the dimensions of each element constituting the real object may be associated with an image of the real object in the real object DB 400. For example, if an element is a "sleeve," the dimensions of the sleeve length and cuff may be associated with the image.

[0027] Furthermore, information regarding decorations applied to the real object and the year in which the real object was manufactured may be further associated with the image of the real object in the real object DB 400. Decorations may be, for example, studs, cutouts, piping, fly-fronts, fringes, etc.

[0028] <Hardware Configuration> Next, the hardware configuration of the object design support device (information processing device) 100 according to one embodiment of the present invention will be described with reference to FIG.

[0029] The object design support device 100 is configured as an information processing device, and may typically be a server. The object design support device 100 realizes the functions and methods described in this embodiment through cooperation of a processor 101, a memory 102, a storage 103, an input / output interface (I / F) 104, and a communication I / F 105. For example, the functions or methods disclosed herein are realized by the processor 101 executing instructions included in a program loaded into the memory 102.

[0030] The processor 101 executes functions and / or methods realized by codes or instructions contained in a program stored in the storage 103, and includes, for example, a central processing unit (CPU), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), etc., and realizes each process disclosed in each embodiment by a logic circuit (hardware) formed in an integrated circuit, etc., or a dedicated circuit.

[0031] The memory 102 temporarily stores programs loaded from the storage 103 and provides a working area for the processor 101. The memory 102 includes, for example, a random access memory (RAM) and a read only memory (ROM).

[0032] The storage 103 stores a program for causing the information processing device to function as an object design support device. The storage 103 includes, for example, a hard disk drive (HDD), a solid state drive (SSD), a flash memory, and the like.

[0033] The communication I / F 105 is implemented as hardware such as a network adapter, communication software, or a combination of these, and transmits and receives various data to and from a user terminal (not shown) via a wireless or wired network.

[0034] The input / output I / F 104 includes an input device for inputting various operations to the object design support device 100, and an output device for outputting processing results processed by the object design support device 100. The input / output I / F 104 may be an integrated input device and output device, or may be separate input device and output device.

[0035] <Functional Configuration> Next, the functional configuration of an object design support device according to one embodiment of the present invention will be described using Figure 1. The object design support device 100 includes a reception unit 111, a generation unit 114, a comparison unit 115, an output unit 116, and a learning unit 117 as functions realized by a control unit 110, and the reception unit 111 also includes an instruction reception unit 112 and a selection reception unit 113. Note that in Figure 1, functional units that are not essential to the embodiments described below may be omitted. Furthermore, the functions or processes of each functional unit may be realized by machine learning or AI to the extent feasible.

[0036] Hereinafter, one embodiment of the present invention will be described with reference to the processing of each functional unit of the object design support device 100 and the flowchart of Fig. 3. Fig. 3 is an example of a flowchart showing processing by the object design support device 100 according to one embodiment of the present invention.

[0037] First, as described above, images of multiple real objects that actually exist are stored in association with three-dimensional information about the real objects (step S11). This information is stored in the real object DB 400. Note that, although the real object DB 400 is shown in FIG. 1 as being separate from the object design support device 100, the real object DB 400 may be stored within the storage of the object design support device 100. Furthermore, as described above, information about the features of the real objects may be stored in the images of the real objects.

[0038] When a user starts designing an object, the object design support device 100 transmits an image of the real object to the user terminal to prompt the user to select an image of the real object on the user terminal. The user can select an image 12 that resembles the object they want to design from multiple real object images. The selection receiving unit 113 receives a selection of at least one image as a reference image 12 from the images of real objects stored in the real object DB 400 (step S12). The instruction receiving unit 112 also receives a generation instruction 11 from the user terminal that specifies the characteristics of the object to be generated (step S13). The generation instruction may be a so-called prompt that specifies the characteristics of the object to be generated.

[0039] The generation unit 114 generates an image of a virtual object having the characteristics specified in the generation instruction 11 based on the reference image 12 using a predetermined generation model (step S14). The generation model may be a learning model that learns images and characteristics of real objects stored in the real object DB 400 and generates an image of a virtual object based on a prompt. Note that the learning unit 117 may create a generation model by machine learning, for each real object image stored in the real object DB 400, the shape and dimensions of each element constituting the real object and associations with the aspect of the real object as three-dimensional information of the real object.

[0040] Alternatively, the generator 114 may generate a virtual object using an algorithm based on a deep neural network, such as a variational autoencoder or a generative adversarial network (GAN).

[0041] The comparison unit 115 determines whether the result of comparing the image of the virtual object with the images of a plurality of real objects stored in the real object DB 400 satisfies a predetermined condition (step S15).

[0042] Here, the predetermined conditions will be described. In one embodiment of the present invention, the generation unit 114 generates a virtual object using a generative model that has learned images of existing real objects, etc. However, if the real object is a product designed by someone else, it is undesirable to generate an imitation. Therefore, it is desirable to create a new design while referring to the excellent design of the real object. According to one embodiment of the present invention, the design of the virtual object may be determined using, as the predetermined condition, a condition that can determine that the virtual object is not similar to the real object and has a new design.

[0043] For example, the comparison unit 115 may determine that a predetermined condition is satisfied if the difference between the feature amount calculated from the feature of the virtual object generated by the generation unit 114 and the feature amount calculated from the feature of multiple real objects stored in the real object DB 400 is equal to or greater than a predetermined threshold. For example, the feature amount of an object (including a real object and a virtual object) may be shape, color, pattern, dimensions (width, depth), a combination of elements, a combination of materials, etc. The comparison unit 115 may express these feature amounts numerically and determine that a predetermined condition is satisfied if the difference between the numerical value of the feature amount of the real object and the numerical value of the feature amount of the virtual object is equal to or greater than a threshold at which the two objects can be said to be dissimilar. The comparison may be performed for each feature amount or may be performed on the sum of the feature amounts. This allows for the output of a new design that is dissimilar to the real object.

[0044] If the comparison unit 115 determines that the predetermined condition is not satisfied (NO in step S15), the process returns to step S14, and the generation unit 114 may repeat the process of generating a virtual object based on the prompt. If the comparison unit 115 determines that the predetermined condition is satisfied (YES in step S15), the output unit 116 generates three-dimensional information about the virtual object for the virtual object that satisfies the predetermined condition and outputs an image of the virtual object (step S16). In one embodiment of the present invention, the generation unit 114 generates a virtual object using a generative model that has learned three-dimensional information about a real object, so that the generation can also include three-dimensional information. The output unit 116 may output the virtual object in a manner that allows it to be rotated 360 degrees on the user terminal. This may be achieved, for example, by using a lightweight JavaScript library for rendering 3D images in a browser, such as Three.js. The file format of the virtual object output by the output unit 116 may be, but is not limited to, the obj format, the fdx format, the unitypackage format, or the like.

[0045] The instruction receiving unit 112 may receive, as a generation instruction, a designation of the shape of each element constituting the object, which is a characteristic of the object to be generated. For example, if the object is a shirt, the generation instruction may include the shapes of the collar, sleeves, body, etc. constituting the shirt. This allows an object with a design that matches the user's image to be generated.

[0046] Furthermore, the generation instruction may include a specification of at least one of the color, pattern, and material that indicates the appearance of the object as a feature of the object to be generated. For example, if the object is a jacket, the generation instruction may include specifications such as green, tartan check, and wool. Note that the appearance of the entire object may be specified, or the appearance may be specified for each element that makes up the object. For example, the color and material may be specified separately for the sleeves and the body. This allows an object with a design that matches the user's image.

[0047] Furthermore, the generation instructions may include specifications for the dimensions of each element that makes up the object as characteristics of the object to be generated. For example, prompts may be used to input numerical values ​​for the cuffs and neckline, or specifications such as "long bare hands all the way to the fingertips" or "three-quarter length, XX cm shorter from the wrist." This allows for the generation of an object that more closely matches the user's body type and preferences.

[0048] The selection receiving unit 113 may receive a reference image for each element constituting the object to be generated. For example, a reference image may be selectable for each element, such as "Image A" for the "collar," "Image B" for the "sleeves," and "Image C" for the "pattern."

[0049] The output unit 116 may output information for realizing a virtual object in the real world. For example, the output unit 116 may output design data for a 3D printer, a pattern, a design drawing, etc., depending on the object to be created. Furthermore, the generation unit 114 may generate multiple virtual objects, and the output unit 116 may output the multiple virtual objects to a user terminal. The user can select a virtual object to be created and have the output unit 116 output information for creation.

[0050] Although the present invention has been described based on the drawings and examples, it should be noted that those skilled in the art would easily be able to make various modifications and alterations based on this disclosure. Therefore, it should be noted that these modifications and alterations are within the scope of the present invention. For example, the functions included in each component, step, etc. can be rearranged so as not to cause logical inconsistencies, and multiple components, steps, etc. can be combined into one or divided. Furthermore, the configurations shown in the above embodiments may be combined as appropriate.

[0051] 1 shows only one object design support device 100, the present invention is not limited to this and multiple object design support devices may exist. Furthermore, the object design support device 100 is a distributed server that cooperates by communicating via a network, and may include an edge server or a so-called cloud server. In other words, the object design support device 100 is not limited to a physical server, but may also include a virtual server.

[0052] Additionally, when generating a generative model, the model may learn, for example, the association between recommended weight and height for a certain piece of clothing, and the dimensions of the automatically generated clothing may be determined by the user including their height and weight in the prompt.

[0053] In the above description, clothing has been used as an example of an object. However, as described above, furniture, buildings, tableware, bags, watches, vehicles, jewelry, electrical appliances, folk crafts, living things, and the like may also be target objects. In the case of furniture, for example, the legs, tabletop, and the like may be specified as the elements constituting a table, along with their material and color, and the legs, backrest, seat, and the like may be specified as the elements constituting a chair, along with their material and color. In the case of a building, for example, characteristics may be specified regarding the exterior wall color, exterior wall pattern, balcony position, number of floors, doors, windows, roof, and the like. As real object data for a building, the appearance and dimensions of the building may be obtained by image processing from satellite images provided by existing services.

[0054] In addition, in the program of the present disclosure, a two-dimensional virtual object image (2D image) temporarily generated by the generation unit 114 may also be acquired as a product of the generation process. That is, the output unit 116 may be capable of outputting a 2D image of the virtual object generated by the generation unit 114. Furthermore, the output unit 116 may be capable of outputting the virtual object generated by the generation unit 114 by superimposing it on another object. For example, the output unit 116 may output an image of a clothing model worn by a person.

[0055] The programs of the embodiments of the present disclosure may be provided in a state stored in a storage medium readable by an information processing device. The storage medium may store the programs in a "non-transitory tangible medium." The programs include, for example, software programs and control programs. When the functional units of the object design support device are realized by software, the object design support device functions as a receiving unit 111, an instruction receiving unit 112, a selection receiving unit 113, a generating unit 114, a comparing unit 115, an output unit 116, and a learning unit 117 by a processor executing a program loaded into memory.

[0056] Furthermore, the program of the present disclosure may be provided to the object design support device via any transmission medium (such as a communication network or broadcast waves) capable of transmitting the program.

[0057] Furthermore, each embodiment of the present disclosure may be realized in the form of a data signal embedded in a carrier wave, in which the program is embodied by electronic transmission. Note that the program of the present disclosure may be implemented using, for example, a scripting language such as JavaScript (registered trademark) or Python, or the C language, Go language, Swift, Koltin, Java (registered trademark), or the like.

[0058] REFERENCE SIGNS LIST 100 Object design support device (information processing device) 110 Control unit 111 Receiving unit 112 Instruction receiving unit 113 Selection receiving unit 114 Generation unit 115 Comparison unit 116 Output unit 117 Learning unit 10 User input 11 Generation instruction 12 Reference image 20 Output image

Claims

1. An object design support device comprising: a memory unit that stores images of multiple real objects that exist in reality in association with three-dimensional information of the real objects; a selection receiving unit that receives a selection of at least one image as a reference image from the images of the real objects stored in the memory unit; an instruction receiving unit that receives a generation instruction that specifies the characteristics of an object to be generated; a generation unit that generates an image of a virtual object having the characteristics specified in the generation instruction based on the reference image using a predetermined generative model; and a comparison unit that determines whether a result of comparing the image of the virtual object with images of the multiple real objects stored in the memory unit satisfies a predetermined condition, wherein the generation unit repeats generation of images of the virtual objects until the comparison unit determines that the predetermined condition is satisfied, and generates three-dimensional information of the virtual object for any virtual object that satisfies the predetermined condition.

2. The object design support device according to claim 1, wherein the generation instructions include, as features of the object to be generated, a designation of the shape of each element constituting the object.

3. An object design support device according to claim 1 or 2, wherein the generation instructions include designation of at least one of a color, a pattern, and a material that represents the aspect of the object to be generated as a feature of the object to be generated.

4. An object design support device according to claim 1 or 2, wherein the generation instructions include, as features of the object to be generated, designation of dimensions of each element constituting the object.

5. The object design support device according to claim 1, wherein the selection receiving unit receives a reference image for each element that constitutes the object to be created.

6. The object design support device according to claim 3, wherein the generation unit generates an image of the virtual object for each of the images of the multiple real objects stored in the memory unit by using the generation model created by learning the association between the shape and dimensions of each element constituting the real object and the appearance of the real object as three-dimensional information of the real object.

7. The object design support device according to claim 1, wherein the comparison unit determines that the specified condition is satisfied when a difference between a feature amount calculated from the features of the virtual object generated by the generation unit and a feature amount calculated from the features of the multiple real objects stored in the memory unit is equal to or greater than a specified threshold value.

8. A control method for an object design support device comprising the steps of: storing images of a plurality of real objects that exist in reality in a memory unit in association with three-dimensional information of the real objects; accepting a selection of at least one image as a reference image from the images of the real objects stored in the memory unit; accepting a generation instruction that specifies characteristics of an object to be generated; generating an image of a virtual object having characteristics specified in the generation instruction based on the reference image; and judging whether a result of comparing the image of the virtual object with images of the plurality of real objects stored in the memory unit satisfies a predetermined condition, wherein the generating step repeats generation of images of the virtual objects until it is determined by the judging step that the predetermined condition is satisfied, and generates three-dimensional information of the virtual object for any virtual object that satisfies the predetermined condition.

9. A control program for an object design support device which provides an information processing device with the following functions: a function for storing images of multiple real objects that exist in reality in a memory unit in association with three-dimensional information of the real objects; a function for accepting the selection of at least one image as a reference image from the images of the real objects stored in the memory unit; a function for accepting a generation instruction that specifies the characteristics of an object to be generated; a function for generating, using a predetermined generation model, an image of a virtual object having the characteristics specified in the generation instruction based on the reference image; and a function for determining whether a result of comparing the image of the virtual object with images of the multiple real objects stored in the memory unit satisfies a predetermined condition, wherein the generating function repeats generation of an image of the virtual object until the judging function determines that the predetermined condition is satisfied, and generates three-dimensional information of the virtual object for any virtual object that satisfies the predetermined condition.

Images