Texture map generation method and device

Abstract

The invention relates to a texture map generation method and device. The method comprises the following steps: receiving input data related to a target fabric from a user; classifying the input data into mapping data, wherein the mapping data is used for generating a texture map corresponding to the target fabric information; inputting the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse reflection map; and generating a synthetic texture map corresponding to the target fabric based on the normal map or the diffuse reflection map.

Description

Technical Field

[0001] The following embodiments relate to a texture map generation method and apparatus. Background Art

[0002] To simulate virtual clothing naturally in a 3D virtual space, it's necessary to naturally represent fabric characteristics and patterns. When representing patterns on fabric, simply synthesizing colors and shapes onto the fabric for simulation can make it difficult to naturally reproduce the texture of the patterned fabric. Therefore, when representing virtual clothing in a 3D virtual space, it's necessary to generate high-quality texture maps that can represent the fabric's unique texture. Summary of the Invention

[0003] Technical methods to solve problems

[0004] According to one embodiment, a texture map generation method may include the following steps: receiving input data related to a target fabric from a user; classifying the input data into mapping data, wherein the mapping data is used to generate a texture map corresponding to the target fabric information; inputting the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse map; and generating a synthetic texture map corresponding to the target fabric based on the normal map or the diffuse map.

[0005] The step of receiving the input data may include the step of detecting the target fabric information based on the text data when text data input by the user is received.

[0006] The step of receiving the input data may include the step of detecting the target fabric information included in the image data when the image data input by the user is received.

[0007] The step of classifying the target fabric information into the mapping data comprises the following steps: classifying the target fabric information into normal map information, diffuse map information or other information, wherein the normal map information comprises information related to the fabric type, and the diffuse map information comprises information related to the fabric pattern presentation.

[0008] The artificial neural network model may include a generative artificial neural network model that generates the texture map as a seamless map.

[0009] The artificial neural network model may include a machine learning model that generates a preset by learning a correlation between the fabric type and the normal map.

[0010] The step of classifying into the mapping data may include the step of generating the mapping data based on the preset setting when the input data matches the preset setting.

[0011] The classifying into the mapping data may include outputting mapping data related to the target fabric based on the input data when the input data does not match the preset setting.

[0012] The step of outputting mapping data related to the target fabric may include, when the mapping data does not match the preset setting and the mapping data is normal map information, outputting a preset setting candidate group similar to the normal map information.

[0013] The texture map generation step may include the following steps: processing the normal map into a filter image, and using the filter image to transform the diffuse map; and mixing the transformed diffuse map and the filter image.

[0014] The blending step may comprise the step of blending the filtered image and the transformed diffuse map by image compositing.

[0015] The texture map generation method of the present invention may also include the following steps: inputting the mapping data into the artificial neural network model to generate other maps, and the step of generating the synthetic texture map includes the following steps: generating a final synthetic texture map corresponding to the target fabric based on the normal map, the diffuse reflection map and at least one of the other maps, and the other maps are generated based on the normal map information and other information.

[0016] The texture map generation method of the present invention may further include the following steps: simulating the synthesized texture map to a three-dimensional virtual garment, thereby displaying the three-dimensional virtual garment to which the synthesized texture map is applied through a user interface.

[0017] The texture map generation method of the present invention may further include the following steps: outputting a feedback message to the user so that the user can adjust the input data in real time.

[0018] According to one embodiment, an electronic device for simulating virtual clothing may include: a memory for storing instructions; and at least one processor, and when the at least one processor executes the instructions, the electronic device performs the following steps: receiving input data related to a target fabric from a user; classifying the input data into mapping data, the mapping data being used to generate a texture map corresponding to the target fabric information; inputting the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse map; and generating a synthetic texture map corresponding to the target fabric based on the normal map or the diffuse map. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flowchart of a texture map generation method according to an embodiment.

[0020] Figure 2 is a flowchart of a method for generating a synthetic texture map according to an embodiment.

[0021] Figure 3 is an example diagram of a normal map, a diffuse map, and a composite texture map according to one embodiment.

[0022] Figure 4 FIG. 4 is a schematic diagram of steps for generating a synthetic texture map by an electronic device according to an embodiment.

[0023] Figure 5 FIG. 4 is a schematic diagram of steps for generating a synthetic texture map by an electronic device according to an embodiment.

[0024] Figure 6 FIG. 4 is a schematic diagram of a process of obtaining a normal map using an artificial neural network model according to an embodiment.

[0025] Figures 7 to 8D is a schematic diagram illustrating a user interface according to an embodiment.

[0026] Figure 9 is a block diagram of an electronic device according to an embodiment. DETAILED DESCRIPTION

[0027] The description of the specific structures or functions of the embodiments is for illustrative purposes only, and the embodiments may be implemented in various forms. The actual implementation is not limited to the specific embodiments disclosed, and the scope of this specification includes all modifications, equivalents, and even substitutes within the technical concepts described by the embodiments.

[0028] Terms such as first or second can be used to describe different components, but are only used to distinguish one component from other components. For example, a first component can be named a second component, and similarly, a second component can also be named a first component.

[0029] When it is stated that one component is “connected to” another component, it can be directly connected to or in contact with the other component; however, it can also be understood that there are other components between them.

[0030] Unless otherwise specified in the content, a single expression includes multiple meanings. In this specification, the terms "including" or "having" are used to express the presence of the features, numbers, steps, operations, components, accessories, or combinations thereof described in the specification, and do not exclude the possibility of the presence or addition of at least one other feature, number, step, operation, component, accessory, or combination thereof.

[0031] In the content, each of the phrases "A or B", "at least one of A and B", "at least one of A or B", "A, B or C", "at least one of A, B and C", "at least one of A, B or C", etc. may include any or all possible combinations of the items listed in the phrase.

[0032] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the ordinary meanings understood by those skilled in the art. Commonly used terms as defined in dictionaries should be understood as having the meanings in the relevant technical context and should not be interpreted as idealized or overly formalized unless otherwise defined in this specification.

[0033] In the following, the embodiments are described in detail with reference to the accompanying drawings. In the description of the embodiments with reference to the accompanying drawings, the same reference numerals are used for the same components and repeated description is omitted.

[0034] Figure 1 is a flowchart of a texture map generation method according to an embodiment.

[0035] For ease of explanation, use Figure 9 The electronic device 900 is shown to perform steps 110 to 140. However, these steps (steps 110 to 140) may be performed via any other suitable electronic device and in any suitable system.

[0036] In addition, you can follow Figure 1 The operations are performed in the order and manner shown, but the order of some operations may be changed or some operations may be omitted without departing from the spirit and scope of the illustrated embodiments. Figure 1 The multiple operations shown may be executed concurrently or in parallel.

[0037] According to an embodiment, the electronic device 900 needs to simultaneously display the fabric type and fabric pattern of the virtual garment to simulate the 3D virtual garment. Therefore, the electronic device 900 needs to generate a texture map similar to the actual fabric of the 3D virtual garment and apply it to the simulation process.

[0038] A texture map is a collection of images or data used to visually represent fabric surfaces, colors, and patterns in 3D modeling and computer graphics. Texture maps provide visual effects for the appearance of objects in 3D simulations. Texture maps can take many forms, including diffuse maps, normal maps, displacement maps, specular maps, height maps, roughness maps, and ambient occlusion maps. Furthermore, multiple texture map types can be combined to create realistic surface representations.

[0039] When fabrics are represented using images that only depict patterns in color and resolution, the resulting virtual garments during simulation differ significantly from reality. In the described embodiments, a synthetic texture map is one that simultaneously represents both the fabric pattern and the material. This allows for high-quality synthetic texture maps that can even simulate pattern distortion caused by the unique nature of the material. The resulting synthetic texture map is identical or similar to the target fabric, and providing it to the user allows the user to generate the fabric that will be applied to the garment they desire, based on their intent.

[0040] The following will describe how the electronic device 900 uses an artificial neural network model (eg, Figure 4 Artificial neural network model 420, Figure 5 The artificial neural network model 520 in Figure 6 A method for generating texture maps using an artificial neural network model 602 in FIG.

[0041] In step 110, the electronic device 900 may receive input data related to a target fabric from the user. Fabric can refer to a flexible material made using thread or fiber through weaving, knitting, crocheting, or other methods. Fabric has a variety of uses, including making clothing, home textiles (such as curtains and bedding), and industrial materials. Fabric properties can be determined by fiber type (such as cotton, wool, silk, polyester, etc.), production method (such as weaving, knitting), and processing method (such as dyeing, printing, coating). This diversity allows fabrics to have different textures, weights, strengths, stretchability, and transparency, and different fabrics can be selected for different applications. The target fabric refers to the fabric that reflects visual or physical properties that the user wants to generate using the electronic device 900. It is the fabric that the user will use when creating a virtual garment in the virtual space. For example, if the user wants to generate a "cotton" patterned fabric, the target fabric can be "cotton with a pattern."

[0042] The electronic device 900 provides a user interface (eg, the user interface 700 in FIG. 8 ) for a user to input input data, or provides a service through a server so that a user can use the user interface through a user terminal.

[0043] A terminal is a device that can install and run server-related applications and provide an interface to the user. This interface can be provided by the terminal itself. For example, it can be provided by the terminal's operating system (OS) or by an application installed on the terminal. Alternatively, the interface can be provided by the server, or the terminal can simply receive and display the interface provided by the server.

[0044] The user interface may include an input area for the user to input text or upload an image. In addition, the user interface may include touch screen operation, sliding menus, etc. to provide diverse and convenient user input. The electronic device 900 has different data processing methods depending on the type and format of the input data received by the user interface.

[0045] When receiving text data (or a message) input by a user, the electronic device 900 according to an embodiment can detect target fabric information based on the text data. The target fabric information refers to the type and details of the data related to the target fabric.

[0046] The target fabric information may include fabric type, fabric texture, fabric pattern, fabric color, fabric thickness and weight, and fabric stretchability, etc. The electronic device 900 may detect and output fabric type and composition, etc.

[0047] Examples of fabric types include 'Boucle', 'Canvas', 'Challis', 'Chambray' / 'Oxford', 'Chiffon', 'Clipjacquard', 'Corduroy', 'Crepe' / 'CDC', 'Crepeknit', 'Crochet', 'Denim', 'Dewspo', 'Dobby', 'Dobbymesh', 'Doubleknit' / 'Interlock', 'Doubleweave',

[0048] 'Eyelet', 'Flannel', 'Flatbackrib', 'Fleece', 'French Terry' terry', 'Gauze' / 'Double gauze', 'Georgette', 'ITY' / 'Matte jersey', 'Jacquard' / 'Brocade', 'Jacquard knit', 'Jersey', 'Lace', 'Loopterry', 'Lowgauge knit', 'Melton' / 'Boiled', 'Memory', 'Mesh' / 'Tulle', 'Neoprene' / 'Scuba', 'Organza', 'Ottoman', 'PVC',

[0049] 'Pique', 'Plaid', 'Plain', 'Pointelle', 'Polarfleece', 'Ponte', 'Poplin', 'Quiltedknit', 'Rib', 'Ripstop', 'Satin', 'Seersucker', 'Sherpa', 'TRS', 'Taffeta' The target fabric types include 'Taffeta', 'Tricot', 'Tweed', 'Twill', 'Tyvek', 'Veganfur', 'Veganleather', 'Vegansuede', 'Velour', 'Velvet', 'Velvet' / 'Velveteen', 'Voile', 'Waffle', etc. However, the target fabric types listed are only examples, and the embodiment is not limited thereto.

[0050] Examples of fabric patterns include 'Argyle', 'Art Deco', 'Batik', 'Baroque', 'Brocade', 'Camouflage', 'Chevron', 'Chintz', 'Damask', 'Ditsy Print', 'Floral', 'Fleur de Lis', 'Gingham', 'Greek Key', 'Herringbone', 'Houndstooth',

[0051] 'Ikat', 'Jacquard', 'Lattice', 'Leopard Print', 'Moroccan Tile', 'Ogee', 'Paisley', 'Plaid (Tartan)', 'Polka Dot', 'Quatrefoil', 'Seersucker', 'Stripes', 'Toile de Jouy', 'Zebra Print', etc. However, the target fabric patterns listed are only examples, and the embodiment is not limited thereto.

[0052] For example, assuming that the user inputs "Fleur De Lis Patterns" and "Challis Type" through the user interface, the electronic device 900 can detect "Fleur De Lis Patterns" as target fabric pattern information and "Challis Type" as target fabric type information from the text message.

[0053] As another example, the user can enter a text message in the input area without explicitly specifying the target fabric type. For example, suppose the user enters the text message "wool fabric with lily pattern" in the input area. In this case, the electronic device 900 can detect "lily pattern" as target fabric pattern information and "wool fabric" as target fabric type information from the text message.

[0054] If the target fabric information cannot be accurately detected from the text message input by the user, or the user wants to specify a desired target fabric, the electronic device 900 may guide the user to use an interactive user interface to enable the user to input more text messages.

[0055] When receiving image data input by a user, the electronic device 900 according to one embodiment can detect the target fabric information included in the image data. For example, when the user inputs an image of a "warbler tail flower pattern", the electronic device 900 can extract features from the image and detect the pattern information of the target fabric.

[0056] The electronic device 900 can receive image data and text messages simultaneously or sequentially. For example, the electronic device 900 can simultaneously receive an image of "Oryx pattern", a text data message of "printed silk wool type", or an input preset setting. At this time, the electronic device 900 can detect that the pattern information of the target fabric is "Oryx pattern", and detect that the type information of the target fabric is "printed silk wool type". Preset setting input can refer to a method in which the electronic device 900 provides data that has been previously learned and stored in a database through a user interface, so that the user can input by clicking or sliding a menu, etc. For example, the preset setting input is a method in which the user enters a setting from a menu. Figures 8A to 8C The first texture generation interface 710 includes a fabric type option 711 , a content type option 712 , a pattern size option 713 , and other inputs selected from the options pre-set in the user interface.

[0057] In step 120, the electronic device 900 may classify the input data into mapping data for generating a texture map corresponding to the target fabric information. The mapping data is classified as data for generating different types of texture maps according to the type of the input data by analyzing the input data.

[0058] According to one embodiment, the electronic device 900 may classify target fabric information into normal map information, diffuse map information, or other information. The normal map information includes information related to the fabric type, and the diffuse map information includes information related to the fabric pattern. According to one embodiment, the artificial neural network model includes a machine learning model that generates a preset by learning the correlation between the fabric type and the normal map.

[0059] When the preset setting matches the input data, the electronic device 900 according to an embodiment may generate mapping data based on the preset setting.

[0060] When the input data does not match the preset setting, the electronic device 900 according to an embodiment may output mapping data related to the target fabric based on the input data.

[0061] The user interface according to one embodiment may provide a list of preset settings that have been learned and stored in a database in a manner such as a sliding menu for the user to input the preset settings. When the user inputs information related to a normal map through the interactive user interface and the mapping data does not match the preset settings, the electronic device 900 may provide a candidate group of normal maps related to the mapping data in the preset settings.

[0062] When the mapping data does not match the preset settings, and the mapping data is normal map information, the electronic device 900 according to one embodiment may output a preset setting candidate group similar to the normal map information. When the artificial neural network model generates and outputs a diffuse map and a normal map, whether to use the generated artificial neural network model or the machine learning model that has learned the preset settings is determined based on user settings or the texture map generation environment.

[0063] The normal map may be a texture map that presents the specific texture and wrinkles of the fabric surface according to the fabric type. Therefore, the normal map information related to the fabric type may correspond to the target fabric type information detected from the input data input by the user.

[0064] The diffuse reflectance map can be a texture map that defines the color or pattern of a fabric surface. Therefore, the diffuse reflectance map information related to the fabric pattern representation can correspond to the target fabric pattern information detected from the input data input by the user. The target fabric pattern information can include not only the pattern information presented by the target fabric, but also color information.

[0065] Other maps can be any maps other than the Normal map and Diffuse map such as the Displacement map, Specular map, Height map, Roughness map and Ambient Occlusion map.

[0066] The displacement map can be a texture map that presents the actual physical depth and details of the fabric by reflecting the geometric features of the fabric surface. When the electronic device 900 applies the displacement map to the 3D virtual clothing, the real depth of the fabric can be simulated and provided to the user interface.

[0067] The specular map can be a texture map that reflects the reflectivity and glossiness of the fabric surface by reflecting light reflected from the fabric surface. When the electronic device 900 applies the specular map to the 3D virtual garment, certain parts of the fabric may have or may not have gloss, thereby simulating various textures of the fabric and providing them to the user interface.

[0068] The height map can be a texture map that represents the height of the fabric surface through a grayscale image, thereby presenting texture depth and wrinkles. When the electronic device 900 applies the height map to the 3D virtual clothing, the real texture of the specific height changes of the fabric surface can be simulated and provided to the user interface.

[0069] The roughness map can be a texture map that represents the roughness of the fabric surface. When the electronic device 900 applies the roughness map to the 3D virtual clothing, the roughness variation of the fabric surface can be realistically presented to simulate the characteristics of various fabric materials and provide the user interface.

[0070] An ambient occlusion map can be a texture map that renders shadows of parts of the surrounding environment that are not touched by light. When the electronic device 900 applies the ambient occlusion map to a 3D virtual garment, it can provide a sense of depth and realism through the user interface by simulating soft shadows in the gaps, corners, and small holes of the 3D virtual garment.

[0071] Texture maps other than the normal map and the diffuse map are not limited to the above embodiments, and may include texture maps such as an opacity map commonly used by those skilled in the art.

[0072] According to an embodiment, the electronic device 900 may receive input data related to a target fabric, analyze the input data, and then generate other maps.

[0073] According to one embodiment, the electronic device 900 can analyze input data to classify mapping data corresponding to target fabric information. The remaining information in the mapping data, excluding normal map information and diffuse map information, can be classified as "other information." This other information can include physical properties or composition information related to the target fabric, such as fabric thickness and weight, fabric roughness, and fabric stretchability.

[0074] According to one embodiment, the electronic device 900 may generate a normal map by inputting mapping data into an artificial neural network model. The electronic device 900 may generate a normal map based on normal map information or other information. According to one embodiment, the electronic device 900 may generate other maps by inputting mapping data into an artificial neural network model. The electronic device 900 may generate other maps based on normal map information or other information.

[0075] According to one embodiment, other information may be input into the artificial neural network model and used to generate other maps such as displacement maps, specular maps, height maps, roughness maps, and ambient occlusion maps. For example, if the other information is information related to roughness (e.g., softness), the artificial neural network model may generate a roughness map that represents the surface of a soft fabric.

[0076] In step 130, the electronic device 900 may input the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse map of the target fabric. The artificial neural network model may include at least one of a Generative Adversarial Networks (GANs) and a machine learning model.

[0077] The artificial neural network model according to an embodiment includes a generative artificial neural network model that generates a texture map as a seamless map. A seamless map may refer to a texture map in which the fabric surface of the texture map is smoothly connected without any visible boundaries or seams. The seamless map may be designed such that the texture repeats infinitely in any direction in the texture map. The generative artificial neural network model may generate a normal map, a diffuse map, and other maps as seamless maps.

[0078] For example, when the user inputs a text message of "iris pattern", the electronic device 900 may generate a diffuse map related to the "iris pattern" as a seamless map in the artificial neural network model. Also for example, when the user inputs a non-seamless image of "iris pattern", the electronic device 900 may generate a diffuse map related to the "iris pattern" as a seamless map through the artificial neural network model. This also applies to the normal map and other maps, so the artificial neural network model may generate each map as a seamless map according to the input.

[0079] Since the normal map is a texture map generated using the type information of the fabric, if the artificial neural network model has learned the correlation between the normal map and the fabric type, the normal map may be generated from the fabric type information. As described above, the preset settings may refer to data that pairs various information such as fabric type, fabric pattern, fabric texture, and fabric thickness that can define the fabric with the fabric. The artificial neural network model that has learned the correlation between the preset settings and the actual fabric characteristics may compare the user input with the mapping data and output the normal map.

[0080] If the mapping data corresponds to a preset setting learned by the artificial neural network model, the artificial neural network model can output a texture map based on the learned correlation. For example, if the artificial neural network model has learned "Orysalis pattern" and "Printed silk wool material" and saved them as preset settings, when a user inputs "Orysalis pattern + Printed silk wool material type" through the user interface, the artificial neural network model searches for the preset settings corresponding to "Orysalis pattern" and "Printed silk wool material type" and outputs a diffuse map and normal map that are identical or similar to the preset settings. Furthermore, because the artificial neural network model includes a generative artificial neural network model, a new diffuse map related to the "Orysalis pattern" can be generated.

[0081] On the other hand, if the mapping data does not correspond to the preset settings learned by the artificial neural network model or does not exist, the artificial neural network model can analyze the mapping data and generate an appropriate texture map.

[0082] For example, suppose a user enters "wrinkled fabric with a blue dot pattern" in the input area. The electronic device 900 may detect "blue dot pattern" as mapping data associated with the target fabric pattern, and "wrinkled" as mapping data associated with the target fabric type. The artificial neural network model may receive the detected mapping data as input. The artificial neural network model includes a generative artificial neural network model, and thus may generate a new diffuse reflectance map for "blue dot pattern" or a new diffuse reflectance map based on a pattern learned in preset settings, such as "coin-shaped blue dots."

[0083] According to one embodiment, the electronic device 900 can generate a normal map by directly reflecting the "wrinkles" input by the user. For example, the electronic device 900 can select a normal map representing "wrinkles" from various normal map examples generated and provided by the artificial neural network model, which can be a directly generated result without relying on existing preset settings. When the electronic device 900 selects a normal map representing "wrinkles", it provides various normal map examples generated by the artificial neural network model on the user interface to request additional information from the user through the user interface, or request the user to select additional information required for the normal map, thereby ultimately generating a normal map.

[0084] According to one embodiment, an artificial neural network model can learn the correlation between inputs describing fabric characteristics, such as "wrinkles," "rough texture," and "smoothness," and normal maps. For example, the artificial neural network model can learn the correlation between inputs describing fabric characteristics, such as "wrinkles," and generating a "corduroy normal map."

[0085] In addition, the artificial neural network model learns the correlation between inputs describing fabric characteristics such as "wrinkles", "rough texture" and "smoothness" and fabric types, and generates a normal map corresponding to the corresponding fabric characteristics based on the fabric types detected during learning. For example, after learning the correlation of detecting the "corduroy" fabric type based on the input describing fabric characteristics, the artificial neural network model can generate a normal map corresponding to "corduroy" based on the detected fabric type. Generally, for diffuse reflection maps, even if a virtual pattern that does not exist in reality is generated based on user input, it can be realized in reality. However, for normal maps, the user can use an interactive user interface instead of the preset settings provided in the user interface, which does not require the input of the fabric type name, but instead interactively inputs the characteristics or features of the fabric type to generate the target fabric.

[0086] Reference together Figure 8D Even if the user does not input the target fabric type-related input that matches the learned preset setting name, the artificial neural network model will compare the user's input with the preset setting to output a preset setting candidate group similar to the user input, and the electronic device 900 can provide the preset setting candidate group to the user interface to guide the user to select a normal map.

[0087] The electronic device 900 can display the generated diffuse reflection map and normal map as intermediate results through the user interface. If the user inputs an input in the user interface that the intermediate result is not the expected result, the electronic device 900 can generate another intermediate result through the artificial neural network model. In addition, the electronic device 900 can interactively guide the user to input additional information about the input data through the user interface to generate a texture map that matches the user's intention. However, the user interface of the electronic device 900 is not limited to an interactive form, and can adopt a method of displaying various intermediate results and allowing the user to select input by sliding a menu or clicking on a candidate group.

[0088] In step 140, the electronic device 900 may generate a synthetic texture map corresponding to the target fabric based on the normal map or the diffuse map. The synthetic texture map may include a texture map generated by synthesizing texture maps (eg, normal map, diffuse map, other maps).

[0089] According to an embodiment, the electronic device 900 may generate a final synthesized texture map corresponding to the target fabric based on at least one of a normal map, a diffuse map, and other maps.

[0090] According to one embodiment, the electronic device 900 generates an ND composite texture map based on the normal map N and the diffuse map D, and then further applies other maps to the ND composite texture map to generate a final composite texture map. In this case, after the electronic device 900 generates the ND composite texture map by synthesizing the normal map N and the diffuse map D according to user input, it then applies other maps, such as a specular map or a roughness map, to generate a final composite texture map that presents additional features such as surface gloss and roughness. This allows physical properties or features to be added to the basic texture, color, and pattern of the target fabric the user desires to generate.

[0091] According to one embodiment, the electronic device 900 generates a composite NE texture map based on a normal map N and other maps E. The composite NE texture map is then further applied with a diffuse map to generate a final composite texture map. For example, the electronic device 900 combines the generated normal map and roughness map to generate a composite NE texture map that reflects a basic texture. Based on user input, the diffuse map is then further applied to generate a final texture map with the same texture and multiple colors and patterns, which is then presented to the user interface. This allows the user to add a variety of colors and patterns to the desired target fabric while maintaining the same texture and physical properties.

[0092] According to one embodiment, the electronic device 900 can generate a final synthetic texture map by simultaneously reflecting the generated normal map, diffuse map, and / or other maps. Alternatively, the electronic device 900 can simultaneously input the normal map information, diffuse map information, and other map information included in the data input by the user into an artificial neural network model to generate a final synthetic texture map that integrates the information of each map. In this way, the user can obtain a final synthetic texture map that comprehensively reflects the texture of the normal map of the target fabric to be generated, the color and pattern of the diffuse map, and the additional physical properties or characteristics of the other maps.

[0093] The described embodiments are only some of the methods for generating the final synthesized texture map. The method and order for generating the final synthesized texture map are not limited to the described embodiments. The electronic device 900 can generate the final synthesized texture map by reflecting the normal map, diffuse map, and other maps in various combinations and orders based on user requests and input data.

[0094] Figure 2 is a flowchart of a method for generating a synthetic texture map according to an embodiment.

[0095] refer to Figure 1 The description can also be applied to Figure 2 , and the repeated content is omitted here.

[0096] In step 140 , the electronic device 900 may execute steps 141 to 143 to generate a synthetic texture map corresponding to the target fabric.

[0097] In step 141, the electronic device 900 may adjust the size of the normal map and the diffuse map. For example, assuming that the size of the diffuse map to be used is 512x512 and the size of the normal map is 1822x602. To synthesize the normal map and the diffuse map, the electronic device 900 may adjust the size of the diffuse map to 3584x512 and the size of the normal map to 3644x602 so that the two maps have similar sizes. However, adjusting the width and height of the map to an integer ratio is for ease of explanation, and the sizes of the normal map and the diffuse map can be adjusted to be similar or the same.

[0098] In step 142, the electronic device 900 may process the normal map into a filtered image and use the filtered image to transform the diffuse map. For example, the electronic device 900 may convert the normal map into a 1-channel grayscale image to emphasize fabric patterns represented in the normal map. Furthermore, the gamma value in the normal map image may be adjusted to emphasize patterns such as fabric wrinkles.

[0099] A 1-channel grayscale image represents image color information using a single channel. These images contain only brightness information, not color intensity, and pixel values ​​typically range from 0 (black) to 255 (white). Each pixel value represents the brightness of that pixel, and intermediate values ​​can represent various shades of gray.

[0100] 1-channel grayscale images can be used in situations where color information is not required, or to reduce the complexity of texture analysis, pattern recognition, and image processing. In addition, storage space and processing time can be saved, and the efficiency of data processing can be improved. Since the electronic device 900 requires information such as fabric wrinkles presented in the normal map, the normal map can be converted into a 1-channel grayscale image (1-channel gray scale). However, the 1-channel grayscale method is an example of a method for the electronic device 900 to highlight wrinkles in the normal map, and other methods for highlighting wrinkles can be used.

[0101] The electronic device 900 according to an embodiment may transform the diffuse map using the filter image to apply fabric wrinkles and the like represented in the normal map to the diffuse map.

[0102] For example, assume that the dimensions of the normal map and diffuse map are width x and height y. The three-dimensional RGB color value of each pixel that constitutes the diffuse map (here, it is assumed that the RGB color value ranges from 0 to 255) is input[u][v] (where u ranges from 0 to less than x, and v ranges from 0 to less than y). Assume that the 1-channel grayscale color value of the filtered image (here, it is assumed that the grayscale color value ranges from 0 to 255) is filter[u][v]. At this time, the three-dimensional RGB color value of each pixel of the transformed diffuse map is output[u'][v'], where the transformed diffuse map is the result image output after applying the diffuse map to the filtered image. Use the following formula 1 to calculate output[u'][v'].

[0103] Formula 1

[0104]

[0105] In formula 1, Represents the integer part of n, ɑ and β are the distortion strengths; larger values ​​result in greater distortion of the diffuse map. ɑ refers to the horizontal distortion strength of the image, and β refers to the vertical distortion strength of the image. The range of each factor is unlimited; a negative number (minus, -) results in a reverse wrinkle appearance. The electronic device 900 can use a fixed value, but can also modify the value as needed.

[0106] In step 143, the electronic device 900 may blend the transformed diffuse reflectance map and the filtered image. Figures 3 to 5 The electronic device 900 can mix the transformed diffuse reflectance map and the filtered image to generate a composite texture map as a final result desired by the user. The electronic device 900 can mix the filtered image and the transformed diffuse reflectance map through image compositing. Finally, the electronic device 900 can generate a composite texture map.

[0107] Image compositing (or image compositing mode) refers to the process of combining multiple images or elements in image editing software to generate a new image. The electronic device 900 can use image compositing to appropriately arrange the elements and utilize blending modes to combine different elements to achieve the desired visual effect. In image compositing mode, the electronic device 900 primarily performs tasks such as image combination, layer adjustment, masking, and tonal and lighting adjustments.

[0108] Image composition is the process of combining multiple images to generate a new image. This can be used to create scenes that composite people or objects onto background images or combine multiple photos.

[0109] Adjustment layers are the process of stacking each image or element into layers, and then adjusting the transparency or blending mode of each layer to achieve the desired effect. This allows you to naturally coordinate image elements or emphasize specific parts.

[0110] Masking is the process of using masks to obscure or reveal portions of an image, allowing only the desired portion to be displayed. This allows for manipulation of parts of an image or emphasis on specific areas. Tonal and lighting adjustments are the process of improving the overall visual quality by adjusting the tonality or brightness of an image or adding lighting effects.

[0111] For example, the electronic device 900 may use a soft light method in an image synthesis method to blend the filtered image and the transformed diffuse reflection map.

[0112] Soft Lighting is a method of compositing two images based on a single image. You can use the color values ​​of one image to enhance or reduce the brightness of the other. Soft Lighting can be used to combine two images to create a new image. Blending based on Soft Lighting simulates lighting effects, adding soft illumination to an image.

[0113] The Soft Light mode is essentially determined by two factors: the brightness of the object being blended and the brightness of the blended color. In the Soft Light mode, light and dark areas are treated differently. Light areas are adjusted to a lighter hue in the layer that is the Soft Light object, while dark areas are adjusted to a darker hue. Images blended using the Soft Light mode can produce a soft, softly lit effect.

[0114] Soft Light allows you to adjust the color, brightness, and lighting of an image. Blending images with Soft Light can achieve a variety of effects, such as brightening a poorly lit photo or adding soft shadows to an image.

[0115] That is to say, the electronic device 900 can use a soft light method to make the relationship between each pixel of the transformed diffuse reflection map clearer based on the filtered image. When it is determined that the fabric texture or wrinkles are not well presented in the normal map due to insufficient brightness of the transformed diffuse reflection map, the electronic device 900 can repeatedly apply the soft light method. However, the image synthesis method of the described embodiment is not limited to the soft light method, and those skilled in the art can apply other applicable image synthesis methods such as normal light, overlay light, hard light, etc.

[0116] According to an embodiment, the electronic device 900 may generate other maps by inputting other information related to the target fabric included in the input data into the artificial neural network model, and may synthesize the other maps with the synthesized texture map to update the synthesized texture map.

[0117] For example, if the additional information is information about the stretchability of the target fabric, the artificial neural network model can generate an additional map as a texture map related to stretchability. However, the additional map related to stretchability may be difficult to display in two dimensions within a user interface. Nevertheless, if the user applies a synthesized texture map that actually incorporates the additional map to a 3D virtual garment, the stretchability data can be displayed in the simulation results, or physical forces can be applied to the 3D virtual garment to simulate the garment, thereby simultaneously incorporating stretchability into the simulation.

[0118] Figure 3 is an example diagram of a normal map, a diffuse map, and a composite texture map according to one embodiment.

[0119] Reference Figure 1 and Figure 2 The same description applies to Figure 3 , repeated description is omitted here.

[0120] Reference Figure 3 According to an embodiment, the electronic device 900 may generate a synthesized texture map 350 by synthesizing the diffuse map 310 and the normal map 330. The diffuse map 310 may be an example diffuse map generated by the electronic device 900 by receiving "a warbler's tail pattern" in text or image form. The normal map 330 may be an example normal map generated by the electronic device 900 by receiving "quilted knit fabric (88% polyester, 9% viscose, 3% elastane / spandex)" in text or image form. The electronic device 900 may synthesize the diffuse map 310 and the normal map 330 to create a synthesized texture map 350 of "quilted knit fabric" with a "a warbler's tail pattern." The electronic device 900 may display the generated synthesized texture map 350 through a user interface.

[0121] Figure 4 FIG. 4 is a schematic diagram of steps for generating a synthetic texture map by an electronic device according to an embodiment.

[0122] Reference Figure 1 and Figure 2 The description can also be applied to Figure 4 , repeated description is omitted here.

[0123] Reference Figure 4 , the electronic device 900 can receive input related to the target fabric. The user can input different normal maps and diffuse maps according to the type of user interface. For example, if the user interface is a texture generation interface (e.g., a texture generation interface in which normal map related information is input through an input area based on a preset setting and diffuse map related information is input through an input area in an input prompt manner), Figure 8A), the user selects "canvas" in the preset settings and enters "coin-shaped blue dot" in the prompt. As another example, if the user interface is a user interface with interactive input areas for inputting normal maps and diffuse maps by means of prompts (e.g., Figure 8D In the second texture generation interface 810), the user inputs "canvas with coin-shaped blue dots".

[0124] The electronic device 900 can classify the received input 410 and detect mapping data. For example, since the preset setting is "canvas", the electronic device 900 detects "canvas" as information related to the target fabric type, and the user enters "coin-shaped blue dots" in the input area, and can detect "coin-shaped blue dots" as information related to the target fabric pattern. For another example, the electronic device 900 detects "canvas" as information related to the target fabric type and "coin-shaped blue dots" as information related to the target fabric pattern from the user input "canvas with coin-shaped blue dots". Based on this, the electronic device 900 can classify "coin-shaped blue dots" as information related to the diffuse map, and classify "canvas" as information related to the normal map. Then, the electronic device 900 can classify "coin-shaped blue dots, canvas" as mapping data 410.

[0125] Electronic device 900 can input mapping data 410, which contains information related to the classified diffuse and normal maps, into artificial neural network model 420. For example, artificial neural network model 420 can utilize a generative artificial neural network model and a machine learning model to generate a "diffuse map 431 of a coin-shaped blue dot" from the diffuse map information and a "normal map 432 of the canvas" from the normal map information. In this case, the artificial neural network model generates the "diffuse map 431 of a coin-shaped blue dot" and the "normal map 432 of the canvas" as a seamless map. This allows the pattern or type to be presented continuously even when each map is connected and expanded.

[0126] The electronic device 900 can use the synthesis algorithm 440 to generate a synthetic texture map for the diffuse map and normal map generated by the artificial neural network model. For example, the artificial neural network model can use the synthesis algorithm 440 to generate the "diffuse map 431 of coin-shaped blue dots" and the "normal map 432 of the canvas" into the "texture map 450 of the canvas with coin-shaped blue dots". The synthesis algorithm 440 can be used to generate the synthetic texture map for the diffuse map and normal map generated by the artificial neural network model. Figure 2 Describes the algorithm of the method.

[0127] The electronic device 900 may simultaneously display the “texture map 450 of the canvas with coin-shaped blue dots” and the “normal map 432 of the canvas” generated in response to the user input in the result window 460 of the user interface.

[0128] Figure 5 FIG. 4 is a schematic diagram of steps for generating a synthetic texture map by an electronic device according to an embodiment.

[0129] Reference Figure 1 and Figure 2 The description can also be applied to Figure 5 , repeated description is omitted here.

[0130] Reference Figure 5 , the electronic device 900 can receive input related to the target fabric. For example, suppose a user uploads a target image 511 related to a desired target fabric pattern and enters "canvas 512" as the target fabric type. The electronic device 900 can receive the target image 511 and "canvas 512" as input related to the target fabric. In this case, "canvas 512," which serves as normal map-related information, can be selected by the user from a preset setting provided in the user interface, or can be normal map-related mapping data detected based on the information entered by the user in the input area.

[0131] Electronic device 900 can classify the received input and detect it as mapping data. For example, electronic device 900 can detect information related to the pattern of the target fabric from target image 511, which is user input. In this case, electronic device 900 can extract features of target image 511 and detect a "coin-shaped blue dot" pattern. "Canvas 512" can be information related to the type of the target fabric. Therefore, electronic device 900 can classify target image 511 as information related to a diffuse map and "canvas 512" as information related to a normal map.

[0132] The electronic device 900 can input mapping data 510 into the artificial neural network model 520, where the mapping data 510 includes information related to the classified diffuse map and information related to the normal map. For example, the artificial neural network model 520 uses a generative artificial neural network model and a machine learning model to generate a "diffuse map 531 of coin-shaped blue dots" from the diffuse map information and a "normal map 532 of canvas" from the normal map information. In this case, the artificial neural network model generates the "diffuse map 531 of coin-shaped blue dots" and the "normal map 532 of canvas" into a seamless map. This allows the pattern or type to be presented continuously without interruption even when each map is connected and expanded. Since the artificial neural network model 520 receives the target image 511 from the user, it can generate only the seamless map without generating the target image 511.

[0133] When generating a seamless map, the electronic device 900 can utilize inpainting technology. Image inpainting can refer to a technology that naturally restores missing or damaged areas when processing an image. Image inpainting can fill the missing area based on the data around the missing part, thereby completely presenting the image. When using image inpainting technology in the process of generating a seamless map, the edges of the texture can be adjusted so that they are well connected to the center, thereby avoiding seams when the image is repeated. In other words, even if the texture map is assumed to extend infinitely, the same pattern can be repeated without generating seams.

[0134] The electronic device 900 can use the synthesis algorithm 540 to generate the diffuse map and normal map generated by the artificial neural network model into a synthetic texture map. For example, the artificial neural network model can use the synthesis algorithm 540 to generate the "diffuse map 531 of coin-shaped blue dots" and the "normal map 532 of canvas" into the "texture map 550 of canvas with coin-shaped blue dots". The synthesis algorithm 540 can be used to generate the "texture map 550 of canvas with coin-shaped blue dots". Figure 2 Describes the algorithm of the method.

[0135] The electronic device 900 may simultaneously display the “problem map 550 of the canvas with coin-shaped blue dots” and the “normal map 532 of the canvas” generated in response to the user input in the result window 560 of the user interface.

[0136] Figure 6 FIG. 4 is a schematic diagram of a process of obtaining a normal map using an artificial neural network model according to an embodiment.

[0137] Reference Figures 1 to 5 The same description applies to Figure 6 , and the repeated content is omitted here.

[0138] According to an embodiment, the electronic device 900 may include an artificial neural network model capable of generating a normal map. The artificial neural network model 602 may learn from a machine learning device 601 to generate a normal map based on learning data 610. The machine learning device 601 includes a first encoder 621 capable of outputting a first latent vector 622 from the learning data 610 (e.g., the preset settings described above) and a first decoder 623 capable of outputting a learning result 640 from the first latent vector 622.

[0139] For example, refer to Figure 6 The machine learning device 601 can receive learning data 610, which includes data pairing (or labeling) the names of multiple fabric types (e.g., canvas 611, corduroy 612, terry cloth 613, etc.) with normal maps of the fabric types.

[0140] The machine learning device 601 can extract a first latent vector 622 from the received learning data 610. The first latent vector 622 is a concept used in machine learning and deep learning models and is a vector representing data features. The first latent vector 622 can compress or summarize the features of the original data and simply represent complex data structures. Therefore, the first encoder 621 extracts important information from data such as images and, from this, extracts the first latent vector 622, enabling the first decoder 623 of the machine learning device to learn the data features.

[0141] The machine learning device 601 can use the first decoder 623 to extract the learning result 640 from the first latent vector 622. The first decoder 623 can be learned (or updated) to become the second decoder 634 to output a normal map suitable for the fabric type input by the user based on the first latent vector 622, the learning result 640, and the learning data 610. In this process, the first latent vector can also be learned (or updated) to become the second latent vector 633 to include more appropriate image features.

[0142] The artificial neural network model 602 can be modeled to include a second decoder 634 and a second latent vector 633 learned from the machine learning device 601. The artificial neural network model 602 includes a second encoder 631 capable of performing one-hot encoding, a multi-layer perceptron (MLP) 632 that outputs the second latent vector 633, and a second decoder 634 that outputs a normal map corresponding to the user input from the second latent vector 633.

[0143] One-hot encoding can be a data preprocessing technique used to input categorical data into a machine learning model. For example, assuming the categories representing fabric types are ["canvas 611", "corduroy 612", "terry cloth 613"], as in learning data 610, the categorical data can be digitized by encoding "canvas 611" as [1, 0, 0], "corduroy 612" as [0, 1, 0], and "terry cloth 613" as [0, 0, 1]. In this case, if the user enters "terry cloth", encoder 631 can use one-hot encoding to encode the user input as [0, 0, 1].

[0144] The multilayer perceptron 632 may receive the encoded user input and extract a latent vector 633. For example, the encoded user input [0, 0, 1] may be received and extracted as a latent vector (633) of "terrycloth 613".

[0145] Decoder 634 can use latent vector 633 of “terry cloth 613” to output a learning result, “normal map of terry cloth 613.” As a result, electronic device 900 can display the “normal map of terry cloth 613” generated by artificial neural network model 602 on the user interface.

[0146] Figures 7 to 8D is a schematic diagram illustrating a user interface according to an embodiment.

[0147] Reference Figures 1 to 6 The same description applies to Figures 7 to 8D , repeated description is omitted here.

[0148] Reference Figure 7 In step 150 , the electronic device 900 simulates the synthesized texture map onto the three-dimensional virtual garment, and displays the three-dimensional virtual garment to which the synthesized texture map is applied through the user interface 700 .

[0149] According to one embodiment, the user interface 700 includes an avatar 720, a simulation area 701 as a virtual clothing display area, and a texture generation interface (a first texture generation interface 710 or a second texture generation interface 810). The user can use the texture generation interface 710 to instruct the electronic device 900 to simulate fabrics.

[0150] Reference Figures 8A to 8C The first texture generation interface 710 includes a fabric type option 711, a content type option 712, a pattern size option 713, a prompt 714, and a synthesized texture map area 715 representing a synthesized texture map.

[0151] The fabric type option 711 provides preset setting options of normal maps that have been learned by the electronic device 900. The user can select a desired fabric type from the fabric type option 711. Based on this, the electronic device 900 can generate a normal map based on the correlation between the learned preset setting and the normal map.

[0152] The option selected in the content type option 712 may be an option regarding other information to be attached to the synthesized texture map in addition to the information regarding the diffuse map and the information regarding the normal map.

[0153] The option selected in the pattern size option 713 is an option for the size of the pattern to be displayed in the synthesized texture map.

[0154] Prompt 714 can be an area for the user to input fabric pattern information. In the described embodiment, for ease of explanation, the input content is text, but as mentioned above, the user can also input an image.

[0155] like Figure 8D As shown, when the user specifies the fabric type option 711 as "canvas", enters the text "coin-shaped blue dots" in the prompt 714 and clicks the generate button, the electronic device 900 can receive the user input of the first texture generation interface 710 and generate a synthetic texture map of "canvas with coin-shaped blue dots" in the synthetic texture map area 715.

[0156] Reference Figure 8C When the user clicks the "Apply" button on the first texture generation interface 710, the electronic device 900 receives the user input and simulates the application of the virtual clothing 721 of the synthesized texture map on the avatar 720. At this time, the user can modify the properties of the synthesized texture map through the property window interface 730.

[0157] Reference Figure 8D The second texture generation interface 810 can be implemented as an interactive interface. The user can enter text to generate a texture map 811 representing a canvas material with "coin-shaped blue dots." Alternatively, the user can enter "tough and rough texture" instead of the name "canvas."

[0158] Upon receiving the user input, the electronic device 900 may output the synthesized texture map 814 without going through the intermediate steps, but may initially display the canvas's normal map 812 based on user settings or commands. In this case, the electronic device 900 may output feedback information to the user (e.g., whether it is the desired target fabric type, or output a preset candidate group to guide the user's selection) so that the user can adjust the input data in real time. Upon receiving the user's input 813 confirming the canvas's normal map, the electronic device 900 may display the synthesized texture map 814 of "canvas with coin-shaped blue dots" on the second texture generation interface 810.

[0159] However, even in this case, if the user does not expect to obtain other synthetic texture maps, or the user wants to further input other information (such as information related to fabric properties such as stretchability and softness) to obtain a synthetic problem map with more information added, additional operations can be further performed to generate the final synthetic texture map.

[0160] The electronic device 900 may receive a user's command or message to apply the synthetic texture map 814 and simulate the virtual clothing 721 of the avatar 720 with the synthetic texture map 814. Figures 8A to 8C, a property window interface 730 of the applicable synthesized texture map 814 may be displayed to allow the user to modify the properties of the synthesized texture map.

[0161] Figures 8A to 8D The illustrated embodiment shows an avatar 720 wearing a three-dimensional virtual garment 721 , but embodiments of the present invention are not limited thereto, and the synthetic texture map may also be represented by a two-dimensional paper pattern.

[0162] Figure 9 is a block diagram of an electronic device according to an embodiment.

[0163] Reference Figure 9 According to an embodiment, the electronic device 900 may include a processor 930, a memory 950, and an output device 970 (e.g., a display). The processor 930, the memory 950, and the output device 970 may be connected to each other via a communication bus 905. In the above process, for ease of explanation, the electronic device 900 may include a processor 930 for executing at least one method or an algorithm corresponding to the at least one method.

[0164] The output device 970 may display the user interface 700 for receiving user input regarding states and operations related to generating texture maps provided by the processor 930 .

[0165] Memory 950 can store the texture map generation algorithm executed by processor 930, as well as data obtained from sensors and the communication unit. Furthermore, memory 950 can store various information generated during the processing of processor 930. Furthermore, memory 950 can store various data and programs. Memory 950 can include volatile memory or non-volatile memory. Memory 950 includes a large-capacity storage medium such as a hard disk, which stores various data.

[0166] In addition, the processor 930 may execute the same Figure 1At least one method or algorithm corresponding to at least one method described in Figure 8. The processor 930 can be a data processing device embodied by hardware, wherein the hardware includes a circuit having a physical structure for performing a target operation (desired operations). For example, the target operation may include a code (code) or instructions (instructions) included in a program. For example, the processor 930 can be configured as a central processing unit (CPU), a graphics processing unit (GPU), or a neural network processing unit (NPU). For example, the electronic device 900 implemented by hardware may include a microprocessor (microprocessor), a central processing unit (central processing unit), a processor core (processor core), a multi-core processor (multi-core processor), a multiprocessor (multiprocessor), an application-specific integrated circuit (ASIC), and a field programmable gate array (FPGA).

[0167] The processor 930 may execute a program and control the electronic device 900. The program code executed by the processor 930 may be stored in the memory 950.

[0168] The processor 930 may receive input data related to a target fabric from a user; classify the input data into mapping data, the mapping data being used to generate a texture map corresponding to the target fabric information; input the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse map; and generate a synthetic texture map corresponding to the target fabric based on the normal map or the diffuse map.

[0169] The above learning process and step algorithm may be executed on a server including the above electronic device 900 and communicate with the user terminal to implement simulation through the user interface 700 displayed by the user terminal.

[0170] For example, a server operates a service platform that provides virtual clothing simulation, and a user can use the service platform to generate a synthetic texture map. Therefore, the user uses the user interface of the service platform to send input data through the user terminal, and the server sends the simulation result of the synthetic texture map generated based on the user input received through the electronic device 900 through the server and the service platform, so that the user terminal can display the virtual clothing.

[0171] Alternatively, the user terminal may be the electronic device 900 that executes the above-mentioned steps, algorithms and methods.

[0172] The devices described above can be implemented by hardware components, software components, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments can be implemented using one or more general-purpose computers or special-purpose computers such as a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor (DSP), a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device can execute an operating system (OS) and one or more application software executed in the operating system. Furthermore, the processing device responds to the execution of the software to access, store, operate, process, and generate data. For ease of understanding, a method with only one processing device is described, but those skilled in the art should understand that the processing device can include multiple processing elements and / or multiple types of processing elements. For example, the processing device can include multiple processors or a processor and a controller. Furthermore, other processing configurations such as parallel processors can also be included.

[0173] Software can include a computer program, code, instructions, or a combination of more than one of these, that causes a processing device to operate in a desired manner or, individually or collectively, instructs a processing device. Software and / or data can be permanently or temporarily embodied in any type of device, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave for interpretation by the processing device or to provide commands or data to the processing device. Software is distributed across computer systems connected via a network and can be stored or executed in a distributed manner. Software and data can be stored in one or more computer-readable storage media.

[0174] The method according to one embodiment is embodied in the form of program commands that can be executed by various computer means and recorded in a computer-readable medium. The computer-readable medium can include program commands, data files, data structures, etc. in a single or combined form. The program instructions recorded on the medium can be instructions specially designed and constructed to implement the embodiment, or instructions that can be used based on known knowledge by a person of ordinary skill in the field of computer software. Computer-readable recording media can include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical media similar to CD-ROMs, DVDs, etc.; magneto-optical media similar to floppy disks, and hardware devices specially constructed for storing and executing program commands similar to read-only memories (ROMs), random access memories (RAMs), flash memories, etc. Examples of program instructions include not only machine language codes generated by compilers, but also high-level language codes that can be executed by computers using interpreters, etc.

[0175] To perform the operations of the embodiments, the hardware device may be configured in such a manner that the operations are implemented with one or more software modules, and vice versa.

[0176] In summary, the embodiments have been described using limited drawings, and persons skilled in the art will be able to make various modifications and variations based on the description. For example, the described techniques may be performed in a different order than the described methods, and / or the described systems, structures, devices, circuits, and other components may be combined or combined in a different manner than the described methods, or replaced or substituted with other components or equivalents, and appropriate results may still be achieved.

[0177] Accordingly, other embodiments, other examples, and their equivalents are within the scope of the appended claims.

Claims

1. A texture map generation method, characterized in that: The following steps are involved: receiving input data related to a target fabric from a user; Classifying the input data into mapping data, wherein the mapping data is used to generate a texture map corresponding to the target fabric information; inputting the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse map; and A synthetic texture map corresponding to the target fabric is generated based on the normal map or the diffuse map.

2. The texture map generation method according to claim 1, characterized in that: The step of receiving the input data comprises the following steps: When the text data input by the user is received, the target fabric information is detected based on the text data.

3. The texture map generation method according to claim 1, characterized in that: The step of receiving the input data comprises the following steps: When the image data input by the user is received, the target fabric information included in the image data is detected.

4. The texture map generation method according to claim 1, characterized in that: The step of classifying the mapping data comprises the following steps: Classify the target fabric information into normal map information, diffuse map information or other information, The normal map information includes information related to the fabric type, and the diffuse map information includes information related to the fabric pattern presentation.

5. The texture map generation method according to claim 1, characterized in that: The artificial neural network model includes a generative artificial neural network model that generates the texture map as a seamless map.

6. The texture map generation method according to claim 1, characterized in that: The artificial neural network model includes a machine learning model, which generates preset settings by learning the correlation between the fabric type and the normal map.

7. The texture map generation method according to claim 6, characterized in that: The step of classifying the mapping data comprises the following steps: When the input data matches the preset settings, the mapping data is generated based on the preset settings.

8. The texture map generation method according to claim 6, characterized in that: The step of classifying the mapping data comprises the following steps: When the input data does not match the preset setting, mapping data related to the target fabric is output based on the input data.

9. The texture map generation method according to claim 8, characterized in that: The step of outputting mapping data related to the target fabric comprises the following steps: When the mapping data does not match the preset setting, and the mapping data is normal map information, a preset setting candidate group similar to the normal map information is output.

10. The texture map generation method according to claim 1, characterized in that: The texture map generation step comprises the following steps: processing the normal map into a filtered image, and using the filtered image to transform the diffuse map; and The transformed diffuse map and the filtered image are blended.

11. The texture map generation method according to claim 9, characterized in that: The mixing step comprises the following steps: The filtered image and the transformed diffuse map are blended by image compositing.

12. The texture map generation method according to claim 1, characterized in that: The following steps are also included: Inputting the mapping data into the artificial neural network model to generate other maps, The step of generating the synthetic texture map comprises the following steps: generating a final synthetic texture map corresponding to the target fabric based on at least one of the normal map, the diffuse map and the other maps, The other maps are generated based on the normal map information and other information.

13. The texture map generation method according to claim 1, characterized in that: The following steps are also included: The synthesized texture map is simulated to a three-dimensional virtual garment, thereby displaying the three-dimensional virtual garment to which the synthesized texture map is applied through a user interface.

14. The texture map generation method according to claim 1, characterized in that: The following steps are also included: A feedback message is output to the user so that the user can adjust the input data in real time.

15. An electronic device for simulating virtual clothing, characterized in that: include: a memory for storing instructions; and at least one processor, When the at least one processor executes the instructions, the electronic device performs the following steps: receiving input data related to a target fabric from a user; Classifying the input data into mapping data, wherein the mapping data is used to generate a texture map corresponding to the target fabric information; inputting the mapping data into an artificial neural network model to generate at least one of a normal map and a diffuse map; and A synthetic texture map corresponding to the target fabric is generated based on the normal map or the diffuse map.

Images