Data processing method and apparatus

Abstract

The application discloses a data processing method and device. The method comprises: processing demand data about a target product input by a user by using a semantic recognition model to obtain a plurality of product demand parameters of the target product; obtaining a plurality of product physical parameters of the target product by using a physical structure model corresponding to the target product; at least some of the plurality of product physical parameters have production constraint conditions; processing the plurality of product demand parameters and the plurality of product physical parameters by using a parameter fusion model to obtain a target parameter set; the target parameter set comprises a first parameter and a second parameter; a parameter value of the first parameter is consistent with a parameter value of a corresponding parameter in the product demand parameters; a parameter value of the second parameter satisfies a production constraint condition corresponding to the second parameter in the physical structure model; and the target parameter set is used to produce the target product.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a data processing method and apparatus. Background Technology

[0002] Existing multimodal models can generate two-dimensional images and three-dimensional models based on user input, thereby enabling personalized visual content creation.

[0003] However, the images generated by this multimodal model only have visual effects. The product structure, size, assembly relationship and appearance parameters in the images do not meet the requirements of industrial processing and production, and do not have industrial manufacturability. Summary of the Invention

[0004] In view of the above, this application provides a data processing method and apparatus, as follows:

[0005] A data processing method, comprising:

[0006] Using a semantic recognition model, the user-input demand data for the target product is processed to obtain multiple product demand parameters for the target product;

[0007] Using the physical structure model corresponding to the target product, multiple physical parameters of the target product are obtained; at least some of the multiple physical parameters have production constraints.

[0008] Using a parameter fusion model, the multiple product requirement parameters and the multiple product physical parameters are processed to obtain a target parameter set;

[0009] The target parameter set includes a first parameter and a second parameter; the parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters; the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model; the target parameter set is used to produce the target product.

[0010] Optionally, the above method utilizes a parameter fusion model to process the multiple product requirement parameters and the multiple product physical parameters to obtain a target parameter set, including one of the following:

[0011] Using a parameter fusion model, the parameter value of the first parameter in the multiple product physical parameters is modified according to the parameter value of the first parameter in the multiple product requirement parameters to obtain the target parameter set;

[0012] Using a parameter fusion model, the parameter value of the second parameter in the multiple product requirement parameters is modified according to the parameter value of the second parameter in the multiple product physical parameters to obtain the target parameter set.

[0013] Optionally, after processing the multiple product requirement parameters and multiple product physical parameters using a parameter fusion model to obtain the target parameter set, the method further includes:

[0014] Using a parameter optimization model, adjust the color number of at least one product partition in the target parameter set to match the target color number;

[0015] The target color number is a color number in the set of color numbers used in the production of the product that meets the matching condition with the color number of the product partition.

[0016] Optionally, the matching condition in the above method includes: the color difference between the target color and the color of the product partition is the smallest in the color set.

[0017] Optionally, in the above method, the target parameter set is presented as a product manufacturing drawing;

[0018] Specifically, using a parameter optimization model, adjusting the color number of at least one product partition in the target parameter set to match the target color number includes:

[0019] Using a parameter optimization model, the product manufacturing drawings are divided into multiple product zones according to the color of the drawings; the color difference between adjacent product zones is greater than or equal to the difference threshold.

[0020] Using a parameter optimization model, the color number of each product partition is aligned with the corresponding target color number.

[0021] Optionally, after processing the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain the target parameter set, the above method further includes at least one of the following:

[0022] Using a parameter verification model, the parameter value of the second parameter in the target parameter set is verified according to the product production specification data corresponding to the target product, and the verification result is obtained; based on the verification result, the parameter value of the second parameter in the target parameter set is adjusted.

[0023] Using a parameter validation model, a corresponding parameter value is generated for at least one fourth parameter in the target parameter set according to the parameter value of at least one third parameter in the target parameter set; the third parameter is a parameter that is related to the fourth parameter.

[0024] Obtain production evaluation results for producing the target article using the target parameter set; adjust the parameter value of at least one parameter in the target parameter set based on the production evaluation results.

[0025] Optionally, after processing the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain the target parameter set, the above method further includes at least one of the following:

[0026] Using a parameter verification model, the parameter values ​​of the second parameter in the target parameter set are verified according to the industrial production specifications corresponding to the target product, and the verification results are obtained; based on the verification results, the parameter fusion model and the physical structure model corresponding to the target product are optimized.

[0027] Obtain the production evaluation results for producing the target product using the target parameter set; based on the production evaluation results, optimize the parameter fusion model and the physical structure model corresponding to the target product.

[0028] Optionally, the above method may further include:

[0029] Based on the industrial production specifications corresponding to the target product, a training sample is constructed;

[0030] In the training samples, at least one first parameter is not labeled with production constraints, and at least one second parameter is labeled with corresponding production constraints.

[0031] Using the training samples, a physical structure model corresponding to the target product is trained, such that the parameter value of the second parameter among the multiple physical parameters of the target product output by the physical structure model satisfies the corresponding production constraints.

[0032] Optionally, different types of industrial products are trained with different physical structure models as described above.

[0033] The method further includes at least one of the following:

[0034] Based on the product type of the target product in the multiple product requirement parameters, determine the physical structure model corresponding to the target product;

[0035] In response to the user's selection of a physical structure model, the physical structure model corresponding to the target product is determined.

[0036] A data processing apparatus, comprising:

[0037] The demand acquisition unit is used to process the demand data about the target product input by the user using a semantic recognition model, and obtain multiple product demand parameters of the target product.

[0038] The parameter acquisition unit is used to obtain multiple physical parameters of the target product using the physical structure model corresponding to the target product; at least some of the multiple physical parameters have production constraints.

[0039] The parameter fusion unit is used to process the multiple product requirement parameters and the multiple product physical parameters using the parameter fusion model to obtain a target parameter set.

[0040] The target parameter set includes a first parameter and a second parameter; the parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters; the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model; the target parameter set is used to produce the target product.

[0041] As can be seen from the above technical solutions, in the data processing method and apparatus disclosed in this application, compared with the situation where the 3D image output by the multimodal model only has visual effects and cannot be directly used for industrial production, this application identifies multiple demand parameters of the user for the target product through a semantic recognition model, and obtains multiple physical parameters of the target product using a physical structure model. This ensures from the source that the parameters meet the production constraints in industrial production. Then, the parameter fusion model is used to fuse the product demand parameters and the product physical parameters to obtain a target parameter set, which includes: parameters that meet the production constraints in the physical structure model and parameters that meet the user's needs but do not have production constraints. Thus, by combining the target parameter set obtained by the user's needs and the physical structure model of the product, this application not only retains the user's personalized needs and meets the user's required visual effects, but also meets the production constraints of industrial processing and production requirements, achieving compatibility between user needs and industrial production. In this way, when using the target parameter set to realize product production, it can improve the consistency of product production requirements, product qualification rate, and product manufacturability. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart illustrating a data processing method provided in an embodiment of this application;

[0044] Figure 2 Another flowchart of a data processing method provided in an embodiment of this application;

[0045] Figure 3 Another flowchart of a data processing method provided in an embodiment of this application;

[0046] Figure 4 Another flowchart of a data processing method provided in an embodiment of this application;

[0047] Figure 5 Another flowchart of a data processing method provided in an embodiment of this application;

[0048] Figure 6 Another flowchart of a data processing method provided in an embodiment of this application;

[0049] Figure 7 Another flowchart of a data processing method provided in an embodiment of this application;

[0050] Figure 8 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application;

[0051] Figure 9 This is another structural schematic diagram of a data processing apparatus provided in an embodiment of this application;

[0052] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0054] refer to Figure 1 The diagram shown is a flowchart illustrating the implementation of a data processing method provided in this application. This method can be applied to electronic devices that have deployed artificial intelligence models or that invoke artificial intelligence models through intelligent agents, such as mobile phones, laptops, computers, or servers. The technical solution in this embodiment is mainly used to improve the user experience.

[0055] Specifically, the method in this embodiment may include the following steps:

[0056] Step 101: Using a semantic recognition model, process the user-input demand data for the target product to obtain multiple product demand parameters for the target product.

[0057] The demand data refers to the data that users input into the semantic recognition model, representing their customization needs for the product, such as appearance, color, size, pattern, and material. Specifically, it can include: images uploaded by users (hand-drawn drawings or drawings created by software), text entered by users, and voice recordings by users.

[0058] Product requirements parameters can include parameters in multiple dimensions or types, such as appearance requirements parameters (e.g., color, pattern, texture, text, logo, gloss, transparency), size requirements parameters (e.g., length, width, height, diameter, thickness, outline dimensions), functional requirements parameters (e.g., material preference, load-bearing capacity, suitable objects, usage methods), style requirements parameters (e.g., shape, structural style, curved / right angle, simple / complex), and identification requirements parameters (e.g., numbering, engraving position, QR code, image features), etc.

[0059] In one implementation, the semantic recognition model can be an artificial intelligence model based on Natural Language Processing (NLP) and multimodal understanding, which can extract structured requirement parameters from requirement data such as text, images, and speech.

[0060] In another implementation, the semantic recognition model can be any one of the following: a large plain text model, an image semantic extraction model, or a speech semantic extraction model, or a combination of at least two models. Specifically, the large plain text model can recognize the text input by the user to extract parameters representing the user's product customization needs; the image semantic extraction model can extract image features, such as patterns, colors, and outlines, from the image uploaded by the user, and then convert these image features into parameters representing product customization needs; the speech semantic extraction model can convert the user's recorded speech into text, and then extract parameters representing product customization needs from the text.

[0061] Specifically, in an industrial production scenario, this embodiment provides a user interface for a semantic recognition model. The user inputs demand data into the interface, such as at least one of text, voice, and images. This demand data represents the user's product requirements for the target product to be produced. In response to the user inputting demand data into the interface, this embodiment inputs the demand data into the semantic recognition model, causing the semantic recognition model to output multiple product demand parameters.

[0062] The target product refers to the product that the user needs to produce, such as keycaps, stools, tables, Rubik's cubes, water cups, etc.

[0063] For example, in this embodiment, a semantic recognition model is pre-deployed. Users input their requirements through the interactive interface provided by the intelligent agent or the semantic recognition model, such as "I want to customize an ABS keycap with a height of 12mm, printed with a cat pattern, and a pink gradient" and an image with a cat pattern uploaded by the user. The semantic recognition model processes the requirements data to obtain the product requirements parameters, such as the product type "keycap", height "12mm", pattern "image with cat pattern", color "pink gradient", and material "ABS".

[0064] For example, in this embodiment, a semantic recognition model is pre-deployed. Users input their requirements through the interactive interface provided by the intelligent agent or the semantic recognition model. For instance, a user uploads a hand-drawn table pattern marked with desktop dimensions a*b and leg height c. The semantic recognition model processes the table pattern to obtain product requirement parameters, such as product type "table", desktop dimensions "a*b", and leg height "c".

[0065] Step 102: Using the physical structure model corresponding to the industrial product, obtain multiple physical parameters of the target product.

[0066] Among the various physical parameters of a product, at least some have production constraints. Product physical parameters refer to the inherent structural parameters that a target product must adhere to in industrial production, structural assembly, and physical form, output from the physical structure model corresponding to the target product. Product physical parameters define the manufacturable structural range, process boundaries, and assembly standards of the target product, and are core parameters ensuring that the target product can be industrially manufactured, assembled, and used normally. For example, when producing keycaps in a factory, at least the keycap size, color, and material parameters are required.

[0067] It should be noted that production constraints refer to mandatory value ranges, structural rules, or prohibitive requirements that are attached to some or all of the physical parameters of a product and are limited by industrial manufacturing processes, mold structures, assembly rules, material properties, dimensional tolerances, or industry standards. Production constraints are the bottom line for determining whether the required parameters of a product can be used in actual production; if they are not met, it cannot be manufactured or used normally. For example, production constraints include the target product's dimensions must be within a specific range, the minimum wall thickness must be greater than 1.0 mm, sharp corners cannot be machined, and positional deviations must not exceed a threshold.

[0068] The physical structure model, also known as the industrial characteristic structure model, is a model trained or constructed based on industrial production specifications such as the target product's industrial production drawings, standard structural specifications, assembly relationships, and manufacturing process requirements. The physical structure model of the target product records at least one production constraint condition corresponding to a physical parameter of the product based on the industrial production specifications. This ensures that the physical structure model can output physical parameters of the product that conform to the industrial production specifications, guaranteeing that the output physical parameters can be directly used for industrial manufacturing.

[0069] Industrial production standards are the collective term for mandatory standards, process rules, dimensional tolerances, structural requirements, material limitations, and quality requirements that a target product must adhere to throughout the entire process of industrial design, manufacturing, assembly, testing, and delivery. They serve as the baseline rules to ensure that products can be stably manufactured by industrial equipment such as mold processing, injection molding, stamping, and extrusion, and can be properly assembled and used. For example, industrial production standards can include specifications across multiple dimensions, such as dimensional tolerance specifications (allowable deviations for length / width / height, hole diameter, and shaft positions), structural process specifications (such as minimum wall thickness, draft angle, minimum fillet radius, no overhangs, and no interference), assembly matching specifications (such as compatibility with standard parts / general structures and meeting positional accuracy standards), material forming specifications (such as those applicable to injection molding, extrusion, stamping, and printing processes), industry standard specifications (such as general structural rules stipulated by national, industry, and enterprise standards), and quality inspection specifications (such as qualification standards for appearance, strength, stability, and safety).

[0070] It should be noted that different types of industrial products correspond to different physical structure models.

[0071] For example, the industrial product "keycap" has a corresponding physical structure model. In the physical structure model of the keycap, the parameter values ​​of some parameters of the keycap in industrial production are specified. For example, the size range of the inner hole of the keycap base is marked, and the total height of the keycap is between 8 mm and 15 mm, etc.

[0072] For example, the industrial product "table" has a corresponding physical structure model. The physical structure model of the table specifies some parameters of the table in industrial production, such as marking that the four legs of the table are of the same length.

[0073] Step 103: Using the parameter fusion model, process multiple product requirement parameters and multiple product physical parameters to obtain the target parameter set.

[0074] The target parameter set may include at least one first parameter and at least one second parameter. The parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters, while the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model.

[0075] Specifically, a parameter fusion model is an artificial intelligence model capable of matching identical parameters, verifying parameter values, and correcting or aligning parameter values ​​between product requirement parameters and product physical parameters. By fusing product requirement parameters and product physical parameters, the model ensures that a first parameter without production constraints matches the corresponding first parameter in the product requirement parameters, while simultaneously correcting a second parameter with production constraints to match the corresponding second parameter in the product physical parameters. This approach satisfies user requirements for the target product while ensuring its manufacturability.

[0076] For example, in this embodiment, the product requirement parameters set by the user for the keycap, "12mm, cat, pink gradient, ABS, height 20mm, wall thickness 1.0mm", are combined with the production constraints of the physical parameters of the keycap provided by the physical structure model, "height 8-15mm, wall thickness ≥1.2mm", to obtain the target parameter set "12mm, cat, pink gradient, ABS, height 15mm, wall thickness 1.2mm".

[0077] Specifically, the target parameter set is used to produce the target product. This target parameter set is provided to the product manufacturing factory, which then produces the target product according to the set, meeting both user needs and industrial production standards. For example, in this embodiment, the target parameter set is input into the keycap mold to produce a customized keycap with a height of 12mm, a pink gradient, a cat pattern, a wall thickness of 1.2mm, meeting industrial standards, and suitable for assembly and sale.

[0078] As can be seen from the above technical solutions, in the data processing method and apparatus provided in this application embodiment, compared to the situation where the 3D image output by the multimodal model only has visual effects and cannot be directly used for industrial production, this application identifies multiple demand parameters of the user for the target product through a semantic recognition model, and obtains multiple physical parameters of the target product using a physical structure model. This ensures from the source that the parameters meet the production constraints in industrial production. Then, the parameter fusion model is used to fuse the product demand parameters and the product physical parameters to obtain a target parameter set, which includes: parameters that meet the production constraints in the physical structure model and parameters that meet the user's needs but do not have production constraints. Thus, by combining the target parameter set obtained by the user's needs and the physical structure model of the product, this application retains the user's personalized needs and meets the user's required visual effects, while also meeting the production constraints of industrial processing and production requirements, achieving compatibility between user needs and industrial production. In this way, when using the target parameter set to realize product production, it can improve the consistency of product production requirements, product qualification rate, and product manufacturability.

[0079] In one implementation, step 103, when processing multiple product requirement parameters and multiple product physical parameters using the parameter fusion model, can be achieved in the following way:

[0080] By using a parameter fusion model, the parameter value of the first parameter in the physical parameters of multiple products is modified according to the parameter value of the first parameter in the multiple product requirement parameters, and the target parameter set is obtained.

[0081] Specifically, in this embodiment, based on multiple product requirement parameters, the parameter value of the first parameter among the multiple product physical parameters that does not have production constraints is modified, thereby modifying the parameter value of the first parameter in the target parameter set to be consistent with the parameter value of the corresponding first parameter in the product requirement parameters.

[0082] The parameter value (default value, such as the upper or lower limit of the range in the production constraint) of the second parameter in the target parameter set with production constraints remains unchanged; or, the parameter value of the second parameter in the target parameter set with production constraints is modified to be the closest to the parameter value of the corresponding second parameter in the product demand parameters while still satisfying the production constraints. Here, "closest to the parameter value of the corresponding second parameter in the product demand parameters" means that the difference between the modified second parameter value and the parameter value of the corresponding second parameter in the product demand parameters is minimized.

[0083] For example, in this embodiment, the product requirement parameters set by the user for the keycap, "12mm, cat, pink gradient, ABS, height 20mm, wall thickness 1.0mm", without production constraints, are taken as the standard. The product physical parameters of the keycap provided by the physical structure model, "height 10mm, no pattern, white, ABS" without production constraints, are modified to be consistent. Thus, the first parameter without production constraints in the product physical parameters is "12mm, cat, pink gradient, ABS". The product physical parameters with production constraints, "height, wall thickness", can be modified according to the product requirement parameters "height 20mm, wall thickness 1.0mm" and the production constraints "height 8–15mm, wall thickness ≥1.2mm" so that the product physical parameters with production constraints are modified to "height 15mm, wall thickness 1.2mm". Therefore, the modified physical parameters of the product are recorded as the target parameter set, namely "12mm, cat, pink gradient, ABS, height 15mm, wall thickness 1.2mm".

[0084] Therefore, this embodiment can clearly distinguish the modification logic of two types of parameters (first parameter and second parameter) of demand reservation and constraint enforcement, ensuring that personalized needs and industrial production specifications are met simultaneously.

[0085] In one implementation, step 103, when processing multiple product requirement parameters and multiple product physical parameters using the parameter fusion model, can be achieved in the following way:

[0086] By using a parameter fusion model, the parameter value of the second parameter in the multiple product requirement parameters is modified according to the parameter value of the second parameter in the multiple product physical parameters, thus obtaining the target parameter set.

[0087] Specifically, in this embodiment, based on multiple product physical parameters, the second parameter in the production constraint condition among the multiple product requirement parameters is modified. The parameter value of the second parameter in the resulting target parameter set is modified to be consistent with the parameter value of the corresponding second parameter in the product physical parameters, or modified to be the parameter value closest to the parameter value range of the corresponding second parameter in the product physical parameters.

[0088] For example, in this embodiment, based on the physical parameters of the keycap provided by the physical structure model, which have production constraints such as "height, wall thickness" and "height 8–15mm, wall thickness ≥ 1.2mm", the production constraint "height 20mm, wall thickness 1.0mm" in the user's product requirement parameters for the keycap, "12mm, cat, pink gradient, ABS, height 20mm, wall thickness 1.0mm", is modified to "height 15mm, wall thickness 1.2mm". Thus, the modified product requirement parameters are denoted as the target parameter set, namely "12mm, cat, pink gradient, ABS, height 15mm, wall thickness 1.2mm".

[0089] Therefore, this embodiment can clearly distinguish the modification logic of two types of parameters (first parameter and second parameter) of demand reservation and constraint enforcement, ensuring that personalized needs and industrial production specifications are met simultaneously.

[0090] Based on the above implementation scheme, in one implementation method, after obtaining the target parameter set in step 103 in this embodiment, the following processing can also be performed, such as... Figure 2 As shown:

[0091] Step 104: Using the parameter optimization model, adjust the color number of at least one product partition in the target parameter set to match the target color number.

[0092] Here, a color code is a standardized code that uniquely identifies a color to a user, such as a spray paint color code card or an injection molding color masterbatch number. The target color code is the color code in the color code set used in product manufacturing that matches the color code of the product's partition. The color code set refers to the collection of color codes used in industrial production; it is a list or library of all achievable color codes defined by production equipment, material suppliers, and industry standards. If any color code in the target product is not in the color code set used in product manufacturing, spray painting may be impossible, and the target product cannot be produced.

[0093] As can be seen, in this embodiment, after the parameter fusion model outputs the target parameter set, the parameter optimization model is used to further calibrate the color number of at least one product partition, so that the color number of the product partition is completely aligned with the industrially producible color number. This ensures that the final target product color can not only meet user needs, but also achieve industrial production.

[0094] Specifically, the parameter optimization model, also known as the texture refinement model, is a model used to finely correct or align the appearance parameters of colors in the target parameter set to ensure that the appearance of the target product in industrial production meets the color matching standards in printing, spraying or injection molding.

[0095] It should be noted that product zoning refers to dividing a target product into independent areas based on its structure, color, coating area, material, or process. Different product zoning areas have different color codes. Each product zoning area can have its own parameters such as color code, texture, or thickness. For example, product zoning can include: a red zone on the top surface, a green zone on the top surface, a yellow zone on the sides, etc.

[0096] Based on the above implementation, in this embodiment, a parameter optimization model can be used to replace the color number of at least one product partition in the target parameter set with the target color number that meets the matching conditions.

[0097] For example, if the target parameter set specifies a cherry blossom pink color for the top section of the keycap (RGB: 255, 165, 195), but this color is not found in the color set, a target color that meets the matching criteria, such as light pink (RGB: 255, 170, 200), is selected from the color set. Therefore, RGB: 255, 170, 200 is used to replace RGB: 255, 165, 195 in the target parameter set. This ensures successful coloring of the top section of the keycap in industrial production, guaranteeing the industrial production of the keycaps.

[0098] Specifically, matching conditions may include: the target color number has the smallest color number difference with the color number of the product partition in the color number set.

[0099] It should be noted that color difference refers to the numerical distance between two colors in a color space such as RGB. The smaller the color difference, the closer the two colors are. The smallest color difference means that the target color is visually closest to the color of the product's color section.

[0100] In other words, in this embodiment, the color number that is visually closest to the color number of the product zone is selected from all the color numbers that can be produced in industrial production as the target color number, thereby ensuring that the target product has a consistent appearance and is producible.

[0101] In one specific implementation, the calculation of color number difference can be based on the color space of the color number. For example, in the RGB (Red, Green, Blue) color space, the color number difference can be calculated using the Euclidean distance in the RGB space; similarly, in the HSV (Hue, Saturation, Value) color space, the color number difference can be calculated using the color difference in the HSV space; and so on.

[0102] As can be seen, in this embodiment, after the parameter fusion model outputs the target parameter set, the parameter optimization model is used to further calibrate at least one product partition according to the closest color number in the color number set, so that the color number of the product partition is completely aligned with the industrially producible color number. This ensures that the final target product color can not only guarantee the visual consistency of the target product to meet user needs, but also realize industrial production.

[0103] Based on the above implementation scheme, in one specific implementation method, the target parameter set is presented as a product manufacturing drawing. Based on this, in this embodiment, when using the parameter optimization model to adjust the color number of at least one product partition in the target parameter set to be consistent with the target color number, the parameter optimization model can be used to divide the product manufacturing drawing into multiple product partitions according to the drawing color; then, the parameter optimization model is used to align the color number of each product partition with the corresponding target color number.

[0104] Specifically, the color difference between adjacent product zones must be greater than or equal to a threshold. The color difference can be calculated based on the distance within the color space. For example, in the RGB color space, the color difference between adjacent product zones can be calculated using Euclidean distance. A color difference greater than or equal to the threshold ensures clear color boundaries between different product zones, allowing for independent color matching and separate processing.

[0105] It should be noted that product manufacturing drawings refer to 2D engineering drawings, 3D model drawings, assembly drawings, and dimensioned drawings generated from a set of target parameters and conforming to industrial manufacturing standards. Product manufacturing drawings can contain complete information that can be directly used for production, such as structure, dimensions, tolerances, colors, and layers; they are also known as industrial drawings.

[0106] Specifically, in this embodiment, a parameter fusion model is used to directly output 3D drawings and 2D engineering drawings of keycaps that can be directly used for industrial production, i.e., product production drawings, from the target parameter set (keycap height 12mm, wall thickness 1.2mm, pink gradient on top, white border, and cat pattern).

[0107] In one implementation, this embodiment can utilize a parameter optimization model to divide the product into zones according to the color of the product production drawing. For example, based on the RGB or HSV color space of different areas on the product production drawing, the product production drawing can be divided into multiple independent areas with distinct color characteristics. Each area is a product zone, which facilitates color separation printing, color separation spraying, color separation injection molding, or disassembled production.

[0108] For example, in this embodiment, the parameter optimization model is used to divide the keycap production drawings by color, resulting in three product partitions: partition 1 (pink gradient) for the top surface of the keycap, partition 2 (white) for the border of the keycap, and partition 3 (black outline of a cat) for the keycap pattern area.

[0109] Based on the above implementation, this embodiment uses a parameter optimization model to match the target color number with the smallest difference in the industrially producible color number set for each independent product partition, and replaces the color number of the product partition with the matched target color number, so as to ensure that the color of each product is not only producible, but also has the smallest color difference to achieve visual consistency.

[0110] For example, in this embodiment, a parameter optimization model is used to replace partition 1 (top pink) with the cherry blossom pink color with the smallest color difference in the color set, partition 2 (border white) with the standard milky white color with the smallest color difference in the color set, and partition 3 (cat black) with the high-gloss black color with the smallest color difference in the color set.

[0111] As can be seen, in this embodiment, when adjusting the color number in the parameter optimization model, the target parameter set is first presented in the form of industrial production drawings, and then the production drawings are automatically divided into multiple product zones according to the color differences in the drawings. Finally, the color number of each product zone is aligned with the target color number to achieve accurate color matching in multiple color zones and industrialized disassembly production.

[0112] In one implementation, the parameter optimization model can be a non-trained model. In this case, the parameter optimization model is a rule-driven model, specifically based on color space formulas, color difference threshold rules, standard color code matching algorithms, image segmentation algorithms, and color code alignment and replacement algorithms. For example, the parameter optimization model uses color difference threshold rules and image segmentation algorithms to divide the product manufacturing drawing into multiple product partitions, where the color difference between product partitions is greater than or equal to the difference threshold. The parameter optimization model calculates the color code difference between the color code of the product partition and the color codes in the color code set using color space formulas. The parameter optimization model uses the standard color code matching algorithm to traverse the color code set to the color code with the smallest color code difference between the product partition and the corresponding color code. The parameter optimization model uses the color code alignment and replacement algorithm to replace the color code of the product partition with the target color code, thus achieving color code alignment.

[0113] In another implementation, the parameter optimization model can be a trained model. In this case, the parameter optimization model can be built based on a Convolutional Neural Network (CNN), U-Net, or Transformer model, and trained on a large number of industrial color code drawings to achieve product partitioning, color code matching, and color code alignment.

[0114] In one implementation, after obtaining the target parameter set in step 103 of this embodiment, the following processing can also be performed: Figure 3 As shown:

[0115] Step 105: Using the parameter verification model, verify the parameter value of the second parameter in the target parameter set according to the product manufacturing specification data corresponding to the target product, and obtain the verification result; then, adjust the parameter value of the second parameter in the target parameter set according to the verification result.

[0116] Among them, the parameter verification model can be trained based on the industrial production specifications of the target product, so that the parameter verification model can verify whether the second parameter in the target parameter set meets the industrial production specifications, obtain the verification result, and adjust the second parameter accordingly. In this way, the accuracy of the second parameter in the target parameter set can be guaranteed before the industrial production of the target product, avoiding the situation where production is impossible.

[0117] Specifically, in this embodiment, a parameter verification model is used to compare the parameter range of each second parameter in the industrial production specification with the parameter value of the corresponding second parameter in the target parameter set, so as to verify whether the parameter value of the second parameter in the target parameter set meets the industrial production specification.

[0118] For example, in this embodiment, a parameter verification model is used to compare the parameter range (height 8–15mm) for “height” in the keycap production specification with the “height” of the keycap in the target parameter set. If the comparison does not match, then the “height 18mm” in the target parameter set is adjusted to match the parameter range for “height 8–15mm” in the production specification, such as changing it to “15mm”.

[0119] For example, in this embodiment, a parameter verification model is used to compare the parameter range (height 8–15mm) for the "height" setting in the keycap production specification with the "height" of the keycap in the target parameter set. If the comparison matches, such as the "height 10mm" of the keycap in the target parameter set, then the target parameter set is not adjusted.

[0120] In one implementation, after obtaining the target parameter set in step 103 of this embodiment, the following processing can also be performed: Figure 4 As shown:

[0121] Step 106: Obtain the production evaluation results for producing the target article using the target parameter set; then, based on the production evaluation results, adjust the parameter value of at least one parameter in the target parameter set.

[0122] The production evaluation result refers to the evaluation result after actually producing the target product using the target parameter set. The production evaluation result characterizes the manufacturability and product qualification level when generating the target product using the target parameter set. Specifically, the production evaluation result can be represented by a score; for example, a production evaluation result of 0 indicates that normal production is impossible, a production evaluation result of 1 indicates that manual fine-tuning is required, and a production evaluation result of 2 indicates that production is fully feasible. Based on this, this embodiment can utilize a parameter validation model to obtain the production evaluation result and use the production evaluation result to adjust the value of at least one parameter in the target parameters to optimize the target parameter set, facilitating better production of the target product.

[0123] In one implementation, after obtaining the target parameter set in step 103 of this embodiment, the following processing can also be performed: Figure 5 As shown:

[0124] Step 107: Using the parameter validation model, generate corresponding parameter values ​​for at least one fourth parameter in the target parameter set according to the parameter value of at least one third parameter in the target parameter set; the third parameter is a parameter that is related to the fourth parameter;

[0125] Among them, the parameter validation model can be trained based on the industrial production specifications of the target product, so that the parameter validation model can check whether there are missing or incorrect parameters in the target parameter set and optimize the parameters in the target parameter set accordingly.

[0126] Taking a table as an example, the table has four legs, but the target parameter set only contains the height of one leg. Alternatively, the height of one leg may be the height input by the user through the requirement data, while the heights of the other three legs are the default heights provided by the table's physical structure model. In this case, the parameter verification model, according to industrial production standards, determines that the heights of the other three legs do not meet the industrial production standards. Therefore, the parameter verification model can set the heights of the other three legs of the table according to the height of one leg input by the user in the requirement data, thereby ensuring the accuracy of the target parameter set and achieving manufacturability.

[0127] In one implementation, after obtaining the target parameter set in step 103 of this embodiment, the following processing can also be performed: Figure 6 As shown:

[0128] Step 108: Using the parameter verification model, verify the parameter value of the second parameter in the target parameter set according to the industrial production specifications corresponding to the target product, obtain the verification result, and then optimize the parameter fusion model and the physical structure model corresponding to the target product based on the verification result.

[0129] Among them, the parameter verification model can be trained based on the industrial production specifications of the target product, so that the parameter verification model can verify whether the second parameter in the target parameter set meets the industrial production specifications, so as to obtain the verification result and optimize the parameter fusion model and physical structure model of the target parameter set. This allows the parameter fusion model and physical structure model to provide a more accurate target parameter set, thereby ensuring the accuracy of the target parameter set and avoiding the situation of being unable to produce.

[0130] As can be seen, in this embodiment, the parameter fusion model and physical structure model can be automatically fine-tuned based on the verification results of the target parameter set to achieve model optimization.

[0131] In one implementation, after obtaining the target parameter set in step 103 of this embodiment, the following processing can also be performed: Figure 7 As shown:

[0132] Step 109: Obtain the production evaluation results of producing the target product using the target parameter set; then, based on the production evaluation results, optimize the parameter fusion model and the physical structure model corresponding to the target product.

[0133] The production evaluation result refers to the evaluation result after actually producing the target product using the target parameter set. The production evaluation result can be specifically represented by a score; for example, a production evaluation result of 0 indicates that normal production is impossible, 1 indicates that manual fine-tuning is required, and 2 indicates that production is fully feasible. Based on this, this embodiment can utilize a parameter validation model to obtain the production evaluation result, and use the production evaluation result to optimize and obtain the parameter fusion model and physical structure model of the target parameter set. This allows the parameter fusion model and physical structure model to provide a more accurate target parameter set, thereby ensuring the accuracy of the target parameter set and avoiding situations where production is impossible.

[0134] As can be seen, in this embodiment, the parameter fusion model and physical structure model can be automatically fine-tuned according to the specific production conditions of the target parameter set (such as whether production is possible and the production quality) to achieve model optimization.

[0135] In one implementation, this embodiment can construct training samples based on the industrial production specifications corresponding to the target product, and then use the training samples to train the physical structure model corresponding to the target product, so that the parameter value of the second parameter among the multiple physical parameters of the target product output by the physical structure model satisfies the corresponding production constraints.

[0136] In the training samples, at least one first parameter is not labeled with production constraints, and at least one second parameter is labeled with corresponding production constraints.

[0137] Specifically, in this embodiment, industrial production specifications such as the geometric dimensions, assembly relationships, and manufacturing processes of the industrial product (target product) can be used to construct training samples. These training samples include multiple production parameters, at least some of which have production constraints, and at least some of which do not. The production constraints are fixed parameter values ​​or ranges set for the corresponding parameters in the industrial production specifications. Furthermore, the production parameters are structured data.

[0138] For example, in this embodiment, training samples are constructed based on standard drawings for industrial production of keycaps, including length, width, height, inner length, width, and height. Based on these samples, a physical structure model of the keycaps is trained, ensuring that a set of target parameters for manufacturable keycaps can be generated regardless of the user's input requirements.

[0139] In one implementation, different types of industrial products (such as the target product) are trained with different physical structure models. Based on this, in this embodiment, the physical structure model corresponding to the target product can be determined according to the product type of the target product among multiple product requirement parameters; or, in this embodiment, the physical structure model corresponding to the target product can be determined in response to the user's selection operation of the physical structure model.

[0140] In other words, in this embodiment, the physical structure model corresponding to the target product is determined from the model library according to the product type of the target product represented by the user's input demand data; or, in this embodiment, it is determined according to the user's selection operation.

[0141] For example, in this embodiment, the physical structure model corresponding to the keycap is determined from the model library according to the product type "keycap" in the product requirement parameters corresponding to the user's input "I want to customize an ABS keycap with a height of 12mm, printed with a cat pattern, and a pink gradient".

[0142] For example, in this embodiment, the user selects the physical structure model of the keycap in the interactive interface to indicate that they want to customize a keycap.

[0143] refer to Figure 8 This is a schematic diagram of a data processing device provided in an embodiment of this application. The device can be installed in an electronic device that has an artificial intelligence model deployed or that invokes an artificial intelligence model through an intelligent agent, such as a mobile phone, laptop, computer, or server. The technical solution in this embodiment is mainly used to improve the user experience.

[0144] Specifically, the apparatus in this embodiment may include the following units:

[0145] The demand acquisition unit 801 is used to process the demand data about the target product input by the user using a semantic recognition model, and obtain multiple product demand parameters of the target product.

[0146] The parameter acquisition unit 802 is used to obtain multiple physical parameters of the target product using the physical structure model corresponding to the target product; at least some of the multiple physical parameters have production constraints.

[0147] The parameter fusion unit 803 is used to process the multiple product requirement parameters and the multiple product physical parameters using the parameter fusion model to obtain a target parameter set.

[0148] The target parameter set includes a first parameter and a second parameter; the parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters; the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model; the target parameter set is used to produce the target product.

[0149] As can be seen from the above technical solutions, in the data processing device provided in this application embodiment, compared with the situation where the 3D image output by the multimodal model only has visual effects and cannot be directly used for industrial production, this application identifies multiple demand parameters of the user for the target product through a semantic recognition model, and obtains multiple physical parameters of the target product using a physical structure model. This ensures from the source that the parameters meet the production constraints in industrial production. Then, the parameter fusion model is used to fuse the product demand parameters and the product physical parameters to obtain a target parameter set, which includes: parameters that meet the production constraints in the physical structure model and parameters that meet the user's needs but do not have production constraints. Thus, by combining the target parameter set obtained by the user's needs and the physical structure model of the product, this application not only retains the user's personalized needs and meets the user's required visual effects, but also meets the production constraints of industrial processing and production requirements, achieving compatibility between user needs and industrial production. In this way, when using the target parameter set to realize product production, it can improve the consistency of product production requirements, product qualification rate, and product manufacturability.

[0150] In one implementation, when the parameter fusion unit 803 processes the multiple product requirement parameters and the multiple product physical parameters using the parameter fusion model to obtain the target parameter set, it may include one of the following:

[0151] Using a parameter fusion model, the parameter value of the first parameter in the multiple product physical parameters is modified according to the parameter value of the first parameter in the multiple product requirement parameters to obtain the target parameter set;

[0152] Using a parameter fusion model, the parameter value of the second parameter in the multiple product requirement parameters is modified according to the parameter value of the second parameter in the multiple product physical parameters to obtain the target parameter set.

[0153] In one implementation, after the parameter fusion unit 803 processes the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain a target parameter set, it is further configured to: use a parameter optimization model to adjust the color number of at least one product partition in the target parameter set to be consistent with the target color number; wherein, the target color number is the color number in the color number set used for product production that meets the matching condition with the color number of the product partition.

[0154] The matching condition includes: in the color set, the color difference between the target color and the color of the product partition is the smallest.

[0155] Specifically, the set of target parameters is presented in the form of product manufacturing drawings;

[0156] Specifically, when the parameter fusion unit 803 uses the parameter optimization model to adjust the color number of at least one product partition in the target parameter set to be consistent with the target color number, it is used to: divide the product production drawing into multiple product partitions according to the drawing color using the parameter optimization model; the color difference between adjacent product partitions is greater than or equal to the difference threshold; and align the color number of each product partition with the corresponding target color number using the parameter optimization model.

[0157] In one implementation, after the parameter fusion unit 803 processes the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain a target parameter set, it is further configured to: use a parameter verification model to verify the parameter values ​​of the second parameter in the target parameter set according to the product production specification data corresponding to the target product, and obtain a verification result; adjust the parameter values ​​of the second parameter in the target parameter set according to the verification result; use the parameter verification model to generate corresponding parameter values ​​for at least one fourth parameter in the target parameter set according to the parameter values ​​of at least one third parameter in the target parameter set; the third parameter is a parameter that is related to the fourth parameter; obtain a production evaluation result for producing the target product using the target parameter set; and adjust the parameter values ​​of at least one parameter in the target parameter set according to the production evaluation result.

[0158] In one implementation, after the parameter fusion unit 803 processes the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain a target parameter set, it is further configured to: use a parameter verification model to verify the parameter values ​​of the second parameter in the target parameter set according to the industrial production specifications corresponding to the target product, and obtain a verification result; optimize the parameter fusion model and the physical structure model corresponding to the target product based on the verification result; obtain a production evaluation result for producing the target product using the target parameter set; and optimize the parameter fusion model and the physical structure model corresponding to the target product based on the production evaluation result.

[0159] In one implementation, the apparatus in this embodiment may further include the following units, such as... Figure 9 As shown:

[0160] The model processing unit 804 is used to construct training samples according to the industrial production specifications corresponding to the target product; wherein, in the training samples, at least one first parameter is not labeled with production constraints, and at least one second parameter is labeled with corresponding production constraints; using the training samples, the physical structure model corresponding to the target product is trained, such that the parameter value of the second parameter among the multiple physical parameters of the product output by the physical structure model for the target product satisfies the corresponding production constraints.

[0161] In one implementation, different types of industrial products are trained with different physical structure models;

[0162] The parameter acquisition unit 802 is also used to implement one of the following:

[0163] Based on the product type of the target product in the multiple product requirement parameters, determine the physical structure model corresponding to the target product;

[0164] In response to the user's selection of a physical structure model, the physical structure model corresponding to the target product is determined.

[0165] It should be noted that the specific implementation of each unit in this embodiment can be referred to the corresponding content above, and will not be described in detail here.

[0166] refer to Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include the following structure:

[0167] Memory 1001 is used to store computer programs and data generated during the execution of computer programs;

[0168] Processor 1002 is used to execute computer programs to perform the following processes:

[0169] Using a semantic recognition model, the user-input demand data for the target product is processed to obtain multiple product demand parameters for the target product;

[0170] Using the physical structure model corresponding to the target product, multiple physical parameters of the target product are obtained; at least some of the multiple physical parameters have production constraints.

[0171] Using a parameter fusion model, the multiple product requirement parameters and the multiple product physical parameters are processed to obtain a target parameter set;

[0172] The target parameter set includes a first parameter and a second parameter; the parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters; the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model; the target parameter set is used to produce the target product.

[0173] As can be seen from the above technical solutions, in the electronic device provided by the embodiments of this application, compared with the situation where the 3D image output by the multimodal model only has visual effects and cannot be directly used for industrial production, this application identifies multiple demand parameters of the user for the target product through a semantic recognition model, and obtains multiple physical parameters of the target product using a physical structure model. This ensures from the source that the parameters meet the production constraints in industrial production. Then, the parameter fusion model is used to fuse the product demand parameters and the product physical parameters to obtain a target parameter set, which includes: parameters that meet the production constraints in the physical structure model and parameters that meet the user's needs but do not have production constraints. Thus, by combining the target parameter set obtained by the user's needs and the physical structure model of the product, this application not only retains the user's personalized needs and meets the user's required visual effects, but also meets the production constraints of industrial processing and production requirements, achieving compatibility between user needs and industrial production. In this way, when using the target parameter set to realize product production, it can improve the consistency of product production requirements, product qualification rate, and product manufacturability.

[0174] Taking the industrial production of keycaps as an example, the technical solution of this application is illustrated below:

[0175] This application deploys an industrial drawing structure model (physical structure model), a personalized fusion model (parameter fusion model), a texture refinement model (parameter optimization model), a standard inspection model (parameter verification model), and a reward model (parameter verification model) for industrial products. This multi-level model architecture ensures that the target parameter set of the generated product (i.e., the product production drawings) is usable for industrial production, reduces model illusion, and guarantees the consistency of the final product. Specifically:

[0176] Industrial drawing structure model: Training is conducted based on the industrial structure of the corresponding product, i.e., the industrial production specifications, to ensure the basic industrial structure of the corresponding product.

[0177] Personalized Fusion Model: Based on the user's prompt (i.e., the product requirement parameters corresponding to the input demand data) and the data generated by the industrial drawing structure of the product selected by the user (i.e., the product physical parameters), a product production drawing (i.e., the target parameter set) for the user's personalized product is generated.

[0178] Texture Refinement Model: The personalized fusion model can generate corresponding 2D and 3D drawings that are suitable for industrial production. The texture refinement model breaks down the 3D drawings into parts according to different colors, and then aligns the color codes of the generated drawings with the color codes used in production to ensure the quality of the generated drawings.

[0179] Standard inspection model: Used to check whether the final product manufacturing drawings meet the bottom line standards for manufacturability (i.e., industrial production specifications) and adjust the target parameter set accordingly.

[0180] Reward Model: Used to score the produced products based on the production status (whether they can be produced in actual industrial production) (i.e., production evaluation results), and automatically fine-tune the detailed parameters of subsequent production, such as adjusting the target parameter set or optimizing the personalized fusion model, industrial drawing structure model, texture refinement model, etc.

[0181] For example, if a user wants to produce a custom keycap, they can train a model in advance based on the standard drawings of keycap industrial production specifications, using data such as length, width, height, and inner dimensions, to ensure that no matter how the user generates it, it will be a producible keycap.

[0182] When a user uploads a photo showing their customization requirements and selects to produce custom keycaps, the personalization fusion model will input the user's photo (product requirement parameters) and the physical parameters of the product output from the industrial drawing structure model of the custom keycaps into the personalization fusion model to generate keycap drawings based on the user's photo.

[0183] After the keycap drawings based on user photos are generated, they are input into the texture refinement model to correct the generated color codes. At the same time, the parts are disassembled according to different color codes to improve the quality of the finished product.

[0184] The refined keycap drawings based on user photos were then sent to a standard inspection model to confirm that the 3D structure fully met the manufacturing requirements.

[0185] The final keycap blueprints will be sent to the factory for industrial production. If a blueprint cannot be produced, the reward model will be given a score of 0. If manual fine-tuning is required, it will be given a score of 1. If it is fully producible, it will be given a score of 2. Blueprints that are given a score of 1 will have some data modified manually before and after the blueprint, and other models will be automatically fine-tuned to ensure that the model illusion is reduced.

[0186] As can be seen, by adopting the technical solution of this application, personalized industrial-grade product drawings that can be directly used in industrial production can be generated based on the user's prompt.

[0187] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0188] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0189] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0190] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data processing method, comprising: Using a semantic recognition model, the user-input demand data for the target product is processed to obtain multiple product demand parameters for the target product; Using the physical structure model corresponding to the target product, multiple physical parameters of the target product are obtained; at least some of the multiple physical parameters have production constraints. Using a parameter fusion model, the multiple product requirement parameters and the multiple product physical parameters are processed to obtain a target parameter set; The target parameter set includes a first parameter and a second parameter; the parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters; the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model; the target parameter set is used to produce the target product.

2. The method according to claim 1, wherein a parameter fusion model is used to process the multiple product requirement parameters and the multiple product physical parameters to obtain a target parameter set, including one of the following: Using a parameter fusion model, the parameter value of the first parameter in the multiple product physical parameters is modified according to the parameter value of the first parameter in the multiple product requirement parameters to obtain the target parameter set; Using a parameter fusion model, the parameter value of the second parameter in the multiple product requirement parameters is modified according to the parameter value of the second parameter in the multiple product physical parameters to obtain the target parameter set.

3. The method according to claim 1, after processing the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain a target parameter set, the method further includes: Using a parameter optimization model, adjust the color number of at least one product partition in the target parameter set to match the target color number; The target color number is a color number in the set of color numbers used in the production of the product that meets the matching condition with the color number of the product partition.

4. The method according to claim 3, wherein the matching conditions include: In the set of color codes, the color code difference between the target color code and the color code of the product partition is the smallest.

5. The method according to claim 3, wherein the target parameter set is presented as a product manufacturing drawing; in, Using a parameter optimization model, the color number of at least one product zone in the target parameter set is adjusted to match the target color number, including: Using a parameter optimization model, the product manufacturing drawings are divided into multiple product zones according to the color of the drawings; the color difference between adjacent product zones is greater than or equal to the difference threshold. Using a parameter optimization model, the color number of each product partition is aligned with the corresponding target color number.

6. The method according to claim 1, after processing the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain a target parameter set, the method further includes at least one of the following: Using a parameter verification model, the parameter value of the second parameter in the target parameter set is verified according to the product manufacturing specification data corresponding to the target product, and the verification result is obtained; based on the verification result, the parameter value of the second parameter in the target parameter set is adjusted. Using a parameter validation model, a corresponding parameter value is generated for at least one fourth parameter in the target parameter set according to the parameter value of at least one third parameter in the target parameter set; the third parameter is a parameter that is related to the fourth parameter. Obtain production evaluation results for producing the target article using the target parameter set; Based on the production assessment results, the parameter value of at least one parameter in the target parameter set is adjusted.

7. The method according to claim 1, after processing the multiple product requirement parameters and the multiple product physical parameters using a parameter fusion model to obtain a target parameter set, the method further includes at least one of the following: Using a parameter verification model, the parameter value of the second parameter in the target parameter set is verified according to the industrial production specifications corresponding to the target product, and the verification result is obtained. Based on the verification results, optimize the parameter fusion model and the physical structure model corresponding to the target product; Obtain production evaluation results for producing the target article using the target parameter set; Based on the production assessment results, optimize the parameter fusion model and the physical structure model corresponding to the target product.

8. The method according to claim 1, further comprising: Based on the industrial production specifications corresponding to the target product, a training sample is constructed; In the training samples, at least one first parameter is not labeled with production constraints, and at least one second parameter is labeled with corresponding production constraints. Using the training samples, a physical structure model corresponding to the target product is trained, such that the parameter value of the second parameter among the multiple physical parameters of the target product output by the physical structure model satisfies the corresponding production constraints.

9. The method according to claim 1, wherein different types of industrial products are trained with different physical structure models; in, The method further includes at least one of the following: Based on the product type of the target product in the multiple product requirement parameters, determine the physical structure model corresponding to the target product; In response to the user's selection of a physical structure model, the physical structure model corresponding to the target product is determined.

10. A data processing apparatus, comprising: The demand acquisition unit is used to process the demand data about the target product input by the user using a semantic recognition model, and obtain multiple product demand parameters of the target product. The parameter acquisition unit is used to obtain multiple physical parameters of the target product using the physical structure model corresponding to the target product; at least some of the multiple physical parameters have production constraints. The parameter fusion unit is used to process the multiple product requirement parameters and the multiple product physical parameters using the parameter fusion model to obtain a target parameter set. The target parameter set includes a first parameter and a second parameter; the parameter value of the first parameter is consistent with the parameter value of the corresponding parameter in the product requirement parameters; the parameter value of the second parameter satisfies the production constraint conditions corresponding to the second parameter in the physical structure model; the target parameter set is used to produce the target product.

Images