Design requirement data processing method, device, storage medium and processor

Abstract

The present invention discloses a method, device, storage medium and processor for processing design requirement data. The method comprises: obtaining the design requirement of an object to be designed, and collecting material images based on the design requirement; using a product classification model to analyze the material images, and obtaining product image tags for marking the design elements of the object to be designed; based on the product image tags, obtaining a three-dimensional product model associated with the object to be designed. The present invention solves the technical problem of low processing efficiency of design requirement data.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method, device, storage medium and processor for processing design requirement data. Background Art

[0002] At present, when processing data related to design needs, designers usually need to browse a large number of fashion trend portals to collect popular materials of the season, and then abstract popular concepts based on their own understanding, and finally complete the final material presentation of product design with design manuscripts or physical samples.

[0003] However, the above method is limited by the number of design materials and the differences in artistic qualities of designers. Their understanding of popular elements is often subjective and one-sided, and cannot truly reflect the overall market trends. In addition, due to information differences between industries, communication costs are caused, and the production of physical products brings time and raw material costs, which greatly delays designers' ability to respond quickly to the market.

[0004] Therefore, in the prior art, the design requirements for the target object are completed through manual design, which leads to low efficiency in the digital conversion process of the design content, thereby causing a technical problem of low efficiency in processing the design requirement data.

[0005] With respect to the above-mentioned technical problem of low efficiency in processing design requirement data, no effective solution has been proposed yet. Summary of the invention

[0006] Embodiments of the present invention provide a method, device, storage medium and processor for processing design requirement data, so as to at least solve the technical problem of low processing efficiency of design requirement data.

[0007] According to one aspect of an embodiment of the present invention, a method for processing design requirement data is provided. The method may include: obtaining the design requirement of an object to be designed, and collecting material images based on the design requirement; using a product classification model to analyze the material images, and obtaining product image tags for marking design elements of the object to be designed; based on the product image tags, obtaining a three-dimensional product model associated with the object to be designed.

[0008] According to one aspect of an embodiment of the present invention, another method for processing design requirement data is provided. The method may include: obtaining the design requirement of the object to be designed by calling a first interface, and collecting material images based on the design requirement, wherein the first interface includes: a first parameter, and the parameter value of the first parameter is the design requirement; using a product classification model to analyze the material image, and obtaining a product image label for marking the design element of the object to be designed; based on the product image label, obtaining a three-dimensional product model associated with the object to be designed; outputting the three-dimensional product model by calling a second interface, wherein the second interface includes: a second parameter, and the parameter value of the second parameter is the three-dimensional product model.

[0009] According to one aspect of an embodiment of the present invention, another method for processing design requirement data is provided. The method may include: obtaining design requirements of an object to be designed on an e-commerce platform, and collecting e-commerce material images based on the design requirements; using a product classification model to analyze the e-commerce material images, and obtaining e-commerce image tags for marking design elements of the object to be designed; based on the e-commerce image tags, obtaining a three-dimensional e-commerce model associated with the object to be designed.

[0010] According to another aspect of an embodiment of the present invention, a device for processing design requirement data is also provided. The device may include: a first acquisition unit, used to acquire the design requirement of the object to be designed, and collect material images based on the design requirement; a second acquisition unit, used to analyze the material images using a product classification model, and acquire product image tags for marking the design elements of the object to be designed; and a third acquisition unit, used to acquire a three-dimensional product model associated with the object to be designed based on the product image tags.

[0011] According to another aspect of the present invention, a computer-readable storage medium is provided, which includes a stored program, wherein when the program is executed by a processor, the device where the computer-readable storage medium is located is controlled to execute the method for processing design requirement data of the present invention.

[0012] According to another aspect of an embodiment of the present invention, a processor is provided, which is used to run a program, wherein when the program is run by the processor, the method for processing design requirement data of an embodiment of the present invention is executed.

[0013] According to another aspect of an embodiment of the present invention, a system for processing design requirement data is also provided. The system includes: a processor; a memory connected to the processor, and used to provide the processor with instructions for processing the following processing steps: obtaining the design requirements of the object to be designed, and collecting material images based on the design requirements; using a product classification model to analyze the material images, and obtaining product image tags for marking the design elements of the object to be designed; based on the product image tags, obtaining a three-dimensional product model associated with the object to be designed.

[0014] In the method for processing design requirement data of this embodiment, the design requirement of the object to be designed is obtained, and material images are collected based on the design requirement; the material images are analyzed by using a product classification model to obtain product image tags for marking the design elements of the object to be designed; and based on the product image tags, a three-dimensional product model associated with the object to be designed is obtained. In other words, the present application uses a product classification model to analyze the material images searched based on the design requirement, obtains product image tags for marking the design elements of the object to be designed, and then achieves the purpose of obtaining the three-dimensional product model associated with the object to be designed based on the product image tags, which helps designers to have a more accurate control of the design elements and provide guarantees for the design scheme of the product object. Moreover, obtaining the three-dimensional product model associated with the object to be designed based on the product image tags can help designers quickly see the three-dimensional product model, reduce the time and raw material cost of physical proofing, reduce communication costs, and accelerate designers to respond quickly to market trends, thereby solving the technical problem of low processing efficiency of design requirement data and achieving the technical effect of improving the processing efficiency of design requirement data. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0016] Figure 1 It is a hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for processing design requirement data according to an embodiment of the present invention;

[0017] Figure 2 is a flow chart of a method for processing design requirement data according to an embodiment of the present invention;

[0018] Figure 3 is a schematic diagram of a three-dimensional clothing generation solution based on intelligent fashion trend analysis according to an embodiment of the present invention;

[0019] Figure 4 is a schematic diagram of an image completed by a self-developed image annotation website according to an embodiment of the present invention;

[0020] Figure 5 is a schematic diagram of image data visualization in a color space according to an embodiment of the present invention;

[0021] Figure 6 is a schematic diagram of color space clothing image browsing according to an embodiment of the present invention;

[0022] Figure 7 is a schematic diagram of a component stitching algorithm process according to an embodiment of the present invention;

[0023] Figure 8 is a schematic diagram of a three-dimensional model retrieval and component stitching result based on clothing image tags according to an embodiment of the present invention;

[0024] Fig. 9 is a schematic diagram of three-dimensional clothing model generation and visual rendering according to an embodiment of the present invention;

[0025] Fig.10 is a schematic diagram of a device for processing design requirement data according to an embodiment of the present invention;

[0026] Fig.11 It is a structural block diagram of a computer terminal according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations:

[0030] 3D garments are virtual garments created through 3D modeling technology, usually in the form of triangular meshes and quadrilateral meshes, which need to be visualized through rendering algorithms;

[0031] Clothing blocks are parts of clothing with semantic information, usually obtained by stitching two-dimensional patterns. They are the smallest units for replacement, reorganization, and deformation in the clothing generation algorithm of this scheme.

[0032] Style, the overall cut shape of a garment;

[0033] Fabric, the material that clothing is made of, which defines the intrinsic physical properties of three-dimensional clothing, such as cotton, linen, silk, canvas, etc.

[0034] Accessories: materials that play a decorative role in clothing design, such as zippers, buttons, linings, etc.

[0035] Texture is an important visual attribute of three-dimensional clothing, which describes the color, surface pattern, and design of clothing;

[0036] Material refers to the material and texture of fabrics on three-dimensional clothing. It describes the detailed properties of the fabric surface, such as color, texture, bumps, and light perception.

[0037] Example 1

[0038] According to an embodiment of the present invention, an embodiment of a method for processing design requirement data is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0039] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 1 is a hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for processing design requirement data according to an embodiment of the present invention. Figure 1 As shown, the computer terminal 10 (or mobile device 10) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission module 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0040] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0041] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the method for processing design requirement data in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the method for processing design requirement data of the above-mentioned application program is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0042] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0043] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0044] It should be noted that, in some optional embodiments, the above Figure 1 The computer device (or mobile device) shown may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of hardware elements and software elements. It should be noted that Figure 1This is merely one example of a particular embodiment and is intended to illustrate the types of components that may be present in the above-described computer device (or mobile device).

[0045] exist Figure 1 In the operating environment shown, this application provides Figure 2 It should be noted that the method for processing images in the video of this embodiment can be Figure 1 The illustrated embodiment is executed by a mobile terminal.

[0046] Figure 2 FIG. 1 is a flow chart of a method for processing design requirement data according to an embodiment of the present invention. Figure 2 As shown, the method for processing the design requirement data may include the following steps:

[0047] Step S202, obtaining design requirements of the object to be designed, and collecting material images based on the design requirements.

[0048] In the technical solution provided in the above step S202 of the present invention, the object to be designed may be clothing to be designed in the clothing industry, and the design requirements of the object to be designed are obtained. The design requirements are also the elements that need to be met when designing the object. Based on the design requirements, a large number of material images related to the design requirements can be searched. The material images may be popular material data, which are used as materials for data analysis, and may include popular image data and three-dimensional (3D) clothing model data, wherein the popular image data is also the image data resource, and may include clothing image data.

[0049] Optionally, when collecting material images based on design requirements, this embodiment may use a web crawler to collect material images based on design requirements, and may use a web crawler to collect material images on fashion shows, fashion magazines, social networks, fashion blogs and other websites in a targeted manner, thereby obtaining a large amount of popular image data; this embodiment may also obtain 3D clothing model data from cooperative customers and designers, which may include most common categories of men's and women's clothing (including T-shirts, hoodies, pants, skirts, dresses, jackets, etc.).

[0050] Optionally, this embodiment may use a cross-platform computer vision and machine learning software library (opencv) to complete the reading, storage, resizing and format conversion of the above-mentioned material images.

[0051] Step S204: Analyze the material image using the product classification model to obtain product image tags for marking design elements of the object to be designed.

[0052] In the technical solution provided in the above step S204 of the present invention, after collecting material images based on design requirements, a product classification model can be used to analyze the material images to obtain product image labels for marking design elements of the object to be designed, wherein the design elements may include at least one of the following: color material, material material and pattern material.

[0053] In this embodiment, the product classification model can be a product classification network, which is used to analyze the acquired material images to obtain product image labels for marking the design elements of the objects to be designed. The product image labels are used to describe the product objects in multiple dimensions. They can be popular element labels or clothing component labels (clothing image labels). For example, product image labels can be used to describe the clothing in multiple dimensions, such as basic clothing attributes such as style, pattern, and category, product attributes such as brand and season, and sales attributes such as transaction volume. Optionally, this embodiment may extract design elements from a large amount of material images in a data-driven manner, and the design elements may be popular design elements. The product classification model may focus on color space and perform primary color analysis on the material images. The acquired material images may be analyzed based on clothing segmentation technology of a deep learning segmentation network (UNet) to obtain relatively pure clothing areas in the material images, thereby avoiding as much as possible the noise influence of the background of the material image and human skin color on the primary color analysis of clothing. The product classification model of this embodiment may also perform pixel color statistics on the clothing area in the material image through a bar graph, and obtain the top-ranked colors as the primary color tones of the clothing, for example, the top 5 colors (Top5) are used as the primary color tones of the clothing.

[0054] Step S206: acquiring a three-dimensional product model associated with the object to be designed based on the product image tag.

[0055] In the technical solution provided in the above step S206 of the present invention, after obtaining the product image tags for marking the design elements of the object to be designed, the three-dimensional product model associated with the object to be designed can be obtained based on the product image tags.

[0056] In this embodiment, when the product object itself has the characteristics of diverse textures and multiple wrinkles, it is very difficult and unstable to directly generate a three-dimensional product model from the material image, which will bring uncertain noise effects to the estimation of the three-dimensional information of the product object, thereby causing inaccurate depth estimation. In order to solve the above problems, this embodiment can use a three-dimensional product model retrieval and generation algorithm based on product image tags to achieve the generation of a three-dimensional product model, and can retrieve and generate a three-dimensional product model associated with the object to be designed based on the product image tag. The three-dimensional product model can be a three-dimensional clothing model.

[0057] Optionally, this embodiment can generate a new 3D product model by splicing the 3D product component models based on the 3D product model generation algorithm of the component model through the iterative closest point algorithm (ICP), the closest point farthest distance sampling and Poisson reconstruction. This algorithm can make the 3D product model have a diverse generation capability, which is equivalent to providing a self-expansion method for a self-owned model database.

[0058] Optionally, when the product object is clothing and the three-dimensional product model is a three-dimensional clothing model, 1,000 common three-dimensional clothing models can be accumulated, and the association relationship between the material image and the three-dimensional clothing model can be established through product image tag retrieval, thereby achieving the purpose of retrieving three-dimensional clothing models with material images, and further realizing the generation of new three-dimensional clothing models through Poisson reconstruction.

[0059] This application obtains the design requirements of the object to be designed through the above steps S202 to S206, and collects material images based on the design requirements; uses the product classification model to analyze the material images, and obtains product image tags for marking the design elements of the object to be designed; based on the product image tags, obtains the three-dimensional product model associated with the object to be designed. In other words, this application uses the product classification model to analyze the material images searched based on the design requirements, obtains the product image tags for marking the design elements of the object to be designed, and then achieves the purpose of obtaining the three-dimensional product model associated with the object to be designed based on the product image tags, which helps designers to have a more accurate control of the design elements and provide guarantees for the design scheme of the product object. In addition, obtaining the three-dimensional product model associated with the object to be designed based on the product image tags can help designers quickly see the three-dimensional product model, reduce the time and raw material cost of physical proofing, reduce communication costs, and accelerate designers to respond quickly to market trends, solve the technical problem of low processing efficiency of design demand data, and achieve the technical effect of improving the processing efficiency of design demand data.

[0060] The following is a further introduction to the method for processing the above-mentioned design requirement data of this embodiment.

[0061] As an optional implementation, step S206, based on the product image label, obtains the three-dimensional product model associated with the object to be designed, including: calculating the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database; based on the calculation result, obtains the three-dimensional product model with the highest similarity from the three-dimensional product model database, wherein the three-dimensional product model with the highest similarity is the three-dimensional product model associated with the object to be designed.

[0062] In this embodiment, when obtaining the three-dimensional product model associated with the object to be designed based on the product image label, the product image label and the label of each three-dimensional product model in the three-dimensional product model database can be respectively calculated for similarity, and the cosine similarity of the product image label and the label of each three-dimensional product model in the three-dimensional product model database can be respectively calculated to obtain a calculation result. For example, if the product image label is represented by a vector I=(I1, I2, ...In), and the label of each three-dimensional product model in the three-dimensional product model database is represented by a vector M=(M1, M2, ...Mn), then the similarity can be calculated as Sim=I·M / ||I||||M||', where "·" is used to represent the inner product of the vector, and ||·|| is used to represent the modulus of the vector. After obtaining the above calculation result, the three-dimensional product model with the highest similarity between the label and the product image label can be obtained from the three-dimensional product model database, and then the three-dimensional product model with the highest similarity can be determined as the three-dimensional product model associated with the object to be designed.

[0063] Optionally, the labels of each 3D product model in the 3D product model database have different descriptions of the characteristics of the product object. For example, if the product object is clothing, gender, category, length, etc. describe the significant characteristics of the clothing, while texture, fabric, etc. describe the detailed characteristics of the clothing. Therefore, this embodiment defines different weights w=(w1, w2, ...wn) for the labels of each 3D product model in the 3D product model database, and ∑ k∈[1，n] ω k =1, the weight can be substituted into the above cosine similarity formula, and the weighted cosine similarity formula is Sim = (wI·wM) / ||wI||||wM||°

[0064] As an optional implementation, before calculating the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database, the method also includes: creating a three-dimensional product model database; associating the label in the three-dimensional product model database with the pre-stored product image label; dividing each three-dimensional product model in the three-dimensional product model database into components according to the label to obtain multiple component models; and marking the stitchable points of the multiple component models, wherein each component serves as the smallest unit for retrieval and reorganization of the three-dimensional product model.

[0065] In this embodiment, before calculating the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database, a three-dimensional product model database is created. The three-dimensional product model database includes labels of multiple three-dimensional product models, which can be a three-dimensional clothing model database. In this embodiment, the label in the three-dimensional product model database is associated with the pre-stored product image label. The pre-stored product image label can be a pre-stored two-dimensional product image label, for example, a two-dimensional clothing image label. The label in the three-dimensional clothing model database can be associated with the two-dimensional clothing image label, that is, the three-dimensional clothing model has the same label attribute as the two-dimensional clothing image, and therefore has label comparability.

[0066] After associating the labels in the 3D product model database with the pre-stored product image labels, each 3D product model in the 3D product model database can be segmented into components according to the labels, thereby obtaining multiple component models, which will become the minimum primitives for retrieval and reorganization of the 3D product model, and then the stitchable points of these component models are marked. Among them, the component model can be a clothing component, which can also be called a 3D clothing component model.

[0067] When a designer selects an inspirational design image, the algorithm obtains the product image label through the deep learning network (DeepFashion network), and retrieves the relevant 3D product models and component models from the 3D product model database based on the content of the product image label. It then selects the 3D product model with the highest similarity to the product image label and returns it to the designer.

[0068] As an optional implementation, before marking the stitchable points of multiple component models, the method also includes: obtaining a coordinate transformation matrix of each component model, wherein the coordinate transformation matrix of each component model constitutes a re-gridded three-dimensional product component model; transforming each component to a predetermined position and obtaining the coordinate position of each component model; and determining the stitchable points between any two or more components based on the coordinate position of each component model.

[0069] In this embodiment, before marking the stitchable points of multiple component models, the coordinate transformation matrix of each component model can be obtained first. The coordinate transformation matrix of each component model can be obtained by ICP calculation. The coordinate transformation matrix of each component model can be used to construct a re-grid three-dimensional product component model. The coordinate transformation matrix can be used to transform each component model to a predetermined position, that is, to a suitable position, and then the coordinate position of each component model is obtained. Based on the coordinate position of each component model, the stitchable points between any two or more components are determined. Since there may be inconsistencies in the number of stitching points between component models, this embodiment samples the stitching points by finding the farthest distance to the nearest stitchable point, thereby ensuring the consistency of the number of stitchable points and establishing a corresponding relationship between stitching points. Optionally, this embodiment gives a stitching line L = {v0, v1, ...v m} and L′={v0′, v1′, … v n '}, and m<n. For the suture point v i ∈L, the distance v can be found on L′ i The distance is less than the set A of other stitching points on L = {vi′, v j ′,…v p ′}, according to the natural sequence of the stitching points on L′, select the point with the largest sequence number as the sampling point on L′.

[0070] In order to ensure the uniformity of the grid at the fusion boundary of the component model, this embodiment adopts the method of regridding the grid of the resampled component model. This embodiment can adopt an incremental regridding algorithm, which can take the three-dimensional grid of the component model as input and execute the following steps: 1) set the side length e of the target grid of the component model and mark the edge at the stitching point as a fixed edge; 2) in the grid, all the edges greater than 4 / 3e inserts a new vertex at the edge center and splits the triangles associated with the edge; 3) deletes all triangles in the mesh that are smaller than 4 / 5 e edge and merge the triangles related to the edge; 4) delete the vertices with degree greater than 6 in the mesh; 5) in order to ensure the smoothness of the mesh after resampling, the position coordinates of all vertices can be updated on the tangent plane along the normal direction to obtain the new coordinate p′=q+nn T (pq), where q is the cosine Laplace coordinate of vertex p, n is the normal vector of the point, and n T is the transpose of n; 6) repeat the above steps 2)-5) multiple times until the mesh is smooth enough. Optionally, based on experience, this embodiment repeats the above algorithm 5 times to obtain a re-grid component model.

[0071] Optionally, this embodiment has component model A and component model B, and the corresponding stitching points are V A=(v1, v2, ..., v n ) and V B =(v1, v2, ..., v n ), then we can solve the linear equation arg min V′ ||L(V′)-δ||+||V A -V B || to complete the stitching of component model A and component model B, where V′ is used to represent the vertex position coordinates of component model A after reconstruction, L(·) is used to represent the vertex coordinate Laplace operator, and δ is used to represent the target vertex Laplace coordinates.

[0072] As an optional implementation, the stitching between any two components is performed using Poisson reconstruction, so that the stitching has smooth features.

[0073] Optionally, the stitching between any two component models of this embodiment may adopt Poisson reconstruction to ensure the smoothness of the stitching.

[0074] As an optional implementation, in step S202, after collecting material images based on design requirements, the method further includes: performing latent spatial encoding on the main color tone of the material image to reduce the display dimension of the material image; based on the reduced display dimension, visualizing the material image in the screen space, wherein the image data of the material image after the display dimension is reduced is clustered in the screen space according to the data amount and color distribution.

[0075] In this embodiment, in order to facilitate designers to freely browse related designs in the color space, after collecting material images based on design requirements, the main color of the material images can be determined first, and the main color of the material images can be encoded in a hidden space to reduce the display dimension of the material images. For example, the main color of the material images can be encoded from the 15-dimensional space R 15 Reduced to 2D space R 2 , based on the reduced display dimension, the material image is visualized in the screen space.

[0076] For example, each image has 5 main colors, denoted by c i∈[1，5] =(r i , g i , b i ), so it can be written as a 15-dimensional vector v 15 =(c1, c2, c3, c4, c5). By principal component analysis (PCA), we can get v 2 = PCA(v 15 ), and v 2=(x, y) represents the screen space coordinates and also represents a popular image in the high-dimensional space, thereby achieving the purpose of visualizing the material image in the screen space.

[0077] In this embodiment, the designer can select the area of ​​interest in the screen space and browse the popular material images freely according to the main color, which can greatly save the designer's time cost of aimlessly searching for design materials in the whole space. It should be noted that the material image data after PCA dimensionality reduction will show clustering characteristics in the screen space according to the amount of data and the diversity of color distribution in the image data, thereby explicitly providing the designer with popular color information in the global data.

[0078] As an optional implementation, in step S204, before using the product classification model to analyze the material image, the method also includes: obtaining product sample material, wherein the product sample material includes: product image and product design model; labeling the product sample material with attribute labels, wherein the attribute labels include at least one of the following: product attributes and transaction attributes; training the product image and the corresponding attribute labels with a neural network model to generate a product classification model, wherein the product classification model is used to label the corresponding attribute labels for the object to be analyzed.

[0079] In this embodiment, before using the product classification model to analyze the material image, a product sample material can be obtained, and the product sample material can include: product image and product design model, wherein the product image can be popular image data, and the product design model can be a clothing model. This embodiment uses attribute tags to mark the product sample material, and the attribute tags include at least one of the following: product attributes and transaction attributes, wherein the product attributes can be the basic attributes of the product, and the transaction attributes can be the commodity attributes and sales attributes of the product. Optionally, this embodiment can define attribute tags such as the basic attributes, sales attributes, and scene information of the product from the attribute source and scene dimension of the product sample material. Among them, the basic attribute is the objective descriptive information of the product as a physical object itself; the attribute tag of the commodity attribute is the relevant strategy information of the product as a commodity in the sales scene; the attribute tag of the sales information describes the dynamic data related to marketing and the market generated by the product during the sales process.

[0080] For example, the attribute tag of this embodiment is a clothing fashion tag, which can be shown in Table 1.

[0081] Table 1 Clothing fashion label list

[0082]

[0083] The above-mentioned fashion label table of the embodiment is defined for the product image being a clothing image and the product design model being a three-dimensional clothing model. It should be noted that the three-dimensional clothing model of the embodiment may only have product attributes.

[0084] Optionally, when the product image is a clothing image, this embodiment may define 17 refined attribute tags based on product attributes, commodity attributes, and sales attributes of the clothing to describe the clothing image in multiple dimensions.

[0085] Optionally, for the scalability of the algorithm, this embodiment can first complete the collection of product images and corresponding attribute labels through a self-developed image annotation website, and then use a neural network model to train the product images and the corresponding attribute labels. For example, the product images and the corresponding attribute labels are trained using the DeepFashion2 network to generate the above-mentioned product classification model, which is used to annotate the corresponding attribute labels of the objects to be analyzed, that is, to process the annotation work of new data.

[0086] As an optional implementation, in step S206, after obtaining the three-dimensional product model associated with the object to be designed, the method also includes: selecting a texture with the highest similarity to the product image label from a pre-stored texture database as the texture of the three-dimensional product model; and visually rendering the three-dimensional product model based on the texture of the three-dimensional product model.

[0087] In this embodiment, the pre-stored texture database, that is, the self-owned texture database, may have textures of different three-dimensional product models. After obtaining the three-dimensional product model associated with the object to be designed, this embodiment may deduct a portion of the image from the material image as a reference based on the inspiration image selected by the designer, and may select the texture with the highest similarity to the product image label of the object to be designed from the pre-stored texture database in the form of an image search as the texture of the three-dimensional product model, and then perform visual rendering on the three-dimensional product model based on the texture of the three-dimensional product model, and may further perform visual rendering on the three-dimensional product model in combination with the apparent information such as the material. Optionally, this embodiment completes the visual rendering of the three-dimensional product model through the threejs plug-in, thereby enhancing the designer's physical perception of the design concept, assisting them in more accurately grasping the real demands of the market and customers, speeding up the product design process, reducing time and raw material costs, and achieving the goal of improving the efficiency of product digital design.

[0088] As an optional example, the design requirements of the object to be designed are obtained by calling a first interface, and material images are collected based on the design requirements, wherein the first interface includes: a first parameter, and the parameter value of the first parameter is the design requirement; a product classification model is used to analyze the material image, and a product image label for marking the design elements of the object to be designed is obtained; based on the product image label, a three-dimensional product model associated with the object to be designed is obtained; and the three-dimensional product model is output by calling a second interface, wherein the second interface includes: a second parameter, and the parameter value of the second parameter is the three-dimensional product model.

[0089] In this embodiment, the above-mentioned first interface can be an interface for data interaction between the server and the client. The client can pass the design requirement data of the object to be designed into the first interface as a parameter of the first interface to achieve the purpose of uploading the design requirement data of the object to be designed to the server.

[0090] In this embodiment, the server obtains the design requirements of the object to be designed through the first interface. The design requirements are elements that need to be met when designing the object. The server can search for a large amount of material images related to the design requirements based on the design requirements.

[0091] Optionally, when the server of this embodiment collects material images based on design requirements, it can use a web crawler to collect material images based on design requirements. Optionally, this embodiment can use cross-platform computer vision and opencv to complete the reading, storage, size scaling and format conversion of the above material images.

[0092] After the server collects material images based on design requirements, the product classification model can be used to analyze the material images to obtain product image tags for marking design elements of the object to be designed, wherein the design elements may include at least one of the following: color materials, material materials, and pattern materials. The product classification model may be a product classification network, which is used to analyze the acquired material images to obtain product image tags for marking design elements of the object to be designed, and the product image tags are used to describe the product object in multiple dimensions. Optionally, this embodiment may extract design elements from a large amount of material images in a data-driven manner, and the design elements may be popular design elements. The product classification model may focus on color space and perform primary color analysis on the material images.

[0093] After the server obtains the product image tag used to mark the design elements of the object to be designed, the 3D product model associated with the object to be designed can be obtained based on the product image tag. This embodiment can use a 3D product model retrieval and generation algorithm based on product image tags to achieve the generation of the 3D product model, and can retrieve and generate the 3D product model associated with the object to be designed based on the product image tag. The 3D product model can be a 3D clothing model.

[0094] Optionally, the server of this embodiment may generate a new three-dimensional product model by splicing the three-dimensional product component models based on the three-dimensional product model generation algorithm of the component model through ICP, closest point farthest distance sampling and Poisson reconstruction.

[0095] The second interface of this embodiment can be an interface for data interaction between the server and the client. The server can transmit the three-dimensional product model data to the second interface as a parameter of the second interface to achieve the purpose of the server sending the three-dimensional product model to the client.

[0096] The following is an introduction to the method for processing design requirement data in an embodiment of the present invention from the perspective of an e-commerce scenario.

[0097] As another optional example, design requirements of the object to be designed are obtained on the e-commerce platform, and e-commerce material images are collected based on the design requirements; the e-commerce material images are analyzed using a product classification model to obtain e-commerce image tags for marking design elements of the object to be designed; based on the e-commerce image tags, a three-dimensional e-commerce model associated with the object to be designed is obtained.

[0098] In this embodiment, the user can input the design requirements of the object to be designed on the e-commerce platform, so that the design requirements of the object to be designed can be obtained in the e-commerce platform, wherein the e-commerce platform can be a comprehensive e-commerce platform or an e-commerce platform of vertical categories. The object to be designed can be an object to be designed in the e-commerce platform, and the design requirements of the object to be designed are obtained. The design requirements are also the elements that need to be met when designing the object. Based on the design requirements, a large number of e-commerce material images related to the design requirements can be searched, and the e-commerce material images can be popular e-commerce material image data, which can be e-commerce pictures.

[0099] Optionally, when collecting material images based on design requirements, this embodiment may use a web crawler to collect the above-mentioned e-commerce material images based on design requirements, thereby obtaining a large amount of popular e-commerce material image data; this embodiment may also obtain e-commerce material images from cooperative customers and designers.

[0100] Optionally, this embodiment may use a cross-platform computer vision and machine learning software library (opencv) to complete the reading, storage, resizing and format conversion of the above-mentioned e-commerce material images.

[0101] After collecting e-commerce material images based on design requirements, a product classification model can be used to analyze the e-commerce material images to obtain e-commerce image tags for marking design elements of the object to be designed, wherein the design elements may include at least one of the following: color material, material material, and pattern material.

[0102] In this embodiment, the product classification model can be a product classification network, which is used to analyze the acquired e-commerce material images to obtain e-commerce image tags for marking the design elements of the object to be designed. The e-commerce image tags are used to describe the e-commerce objects in multiple dimensions and can be popular e-commerce element tags. Optionally, this embodiment can extract design elements from a large amount of e-commerce material images in a data-driven manner. The design elements can be popular e-commerce design elements. The product classification model can focus on color space and perform primary color analysis on the e-commerce material images.

[0103] After obtaining the e-commerce image tags used to mark the design elements of the object to be designed, the three-dimensional e-commerce model associated with the object to be designed can be obtained based on the e-commerce image tags. This embodiment can use a three-dimensional e-commerce model retrieval and generation algorithm based on e-commerce image tags to achieve the generation of the three-dimensional e-commerce model, and can retrieve and generate the three-dimensional e-commerce model associated with the object to be designed based on the e-commerce image tags.

[0104] Optionally, this embodiment can generate a new three-dimensional e-commerce model by splicing the three-dimensional product component model based on the three-dimensional e-commerce model generation algorithm of the component model through ICP, closest point maximum distance sampling and Poisson reconstruction. This algorithm can make the three-dimensional e-commerce model have a diverse generation capability, which is equivalent to providing a self-expansion method for a self-owned model database.

[0105] The method for processing design requirement data of this embodiment is an algorithm for generating three-dimensional product models based on intelligent fashion trend analysis. Fashion trends have always been the weather vane guiding product design solutions. The algorithm framework of this embodiment can extract popular design elements from massive material images in a data-driven manner, and can understand material images through deep learning algorithms, assisting designers to freely browse the entire domain of material images purposefully, and obtain relevant fashion trend data, which helps designers to have a more accurate control of popular design elements and provide guarantees for the fashion and trendiness of design solutions; in addition, this embodiment's three-dimensional product model retrieval generation algorithm based on product image tags can help designers quickly see samples of the objects to be designed, reduce the time and raw material costs of physical proofing, reduce communication costs, and accelerate designers' rapid response to market trends; further, This embodiment is an innovative process framework that combines image visual understanding and three-dimensional product model generation, and has capabilities that existing trend analysis portals and three-dimensional design software do not have; the three-dimensional product model generation algorithm based on component models in this embodiment can be to achieve splicing of component models through ICP, closest point longest distance sampling and Poisson reconstruction, so as to generate a new three-dimensional product model. This algorithm can be used as a generation capability for the diversity of three-dimensional product models, and provides a self-expansion method for an own model database. Therefore, this embodiment solves the technical problem of low processing efficiency of design requirement data through the above method, and achieves the technical effect of improving the processing efficiency of design requirement data.

[0106] Example 2

[0107] The technical solution of the present invention is illustrated below in conjunction with preferred implementation modes, specifically taking clothing as an example of the object to be designed.

[0108] The advent of the 5G era has greatly accelerated the development of traditional manufacturing industries empowered by artificial intelligence (AI) towards a more efficient and intelligent direction. Following the tide of the times, the apparel industry has also begun to explore how to make full use of the powerful data analysis and computing capabilities of deep learning to achieve the digital transformation of the entire industry, thereby achieving a rapid response to the ever-changing market supply and demand. As the source of the entire fast-response link, whether clothing design can accurately grasp the demands of the times and generate design solutions efficiently and at low cost really plays a pivotal role.

[0109] At present, clothing design is in a stage where digitalization is not yet mature. Designers usually need to browse a large number of popular trend portals to collect popular materials of the season, and then abstract popular concepts based on their own understanding, and finally complete the final materialization of clothing design with design manuscripts or physical samples. However, due to the incompleteness of designers' manual collection of design materials and the differences in designers' own design literacy, everyone's understanding and interpretation of popular trends are highly subjective, so they cannot truly reflect the overall trend of the market and the real demands of customers. In addition, due to the differences in information between industries, communication costs are caused, and the production of physical samples brings time costs and raw material costs. The cycle from concept design to final design of a piece of clothing usually takes about 30 days, which is very unfriendly to the low-cost embodiment of design concepts, thus greatly delaying the clothing industry's ability to respond quickly to the market from the source.

[0110] 3D clothing design software for professional designers aims to quickly display designers' design concepts in the form of 3D clothing, thereby reducing the production cost and time cost of sample production. They support designers to freely specify the texture and fabric material of each pattern of 3D clothing, so as to intuitively feedback the performance of different textures and fabric materials on 3D clothing. Some software also provides 3D body scanning and real fabric parameter measurement functions, aiming to provide 3D clothing fit customization, realistic physical simulation and rendering functions.

[0111] The online clothing customization platform for non-professionals aims to provide users with direct visual feedback in the form of 3D clothing. Currently, it supports users to perform personalized editing operations such as replacing parts, personalized logo mapping and texture material replacement on some daily 3D clothing.

[0112] The fashion trend portal aims to provide designers with a portal platform for clothing design resources and materials, covering independent designer works, fashion week shows and other high-definition pictures and fashion magazines and books, providing designers with current manifold materials in terms of color, fabric, style, texture and printing to inspire their design inspiration.

[0113] Related technologies have also collected nearly 4,000 fashion show images, analyzed their colors and organized their color plates, providing designers with a convenient service to browse fashion designs according to color space and draw design inspiration. However, this solution only provides image browsing functions, and the image labels only include a few dimensions such as clothing brands, designers, and years. It does not understand the details of clothing, and does not have the ability to generate three-dimensional clothing models. Therefore, the role it can play in accelerating the design process is also very limited.

[0114] The above-mentioned related technologies for existing trend analysis solutions often only serve to list design materials, but do not conduct in-depth data mining to generate specific popular element labels; clothing design software focuses on providing designers with modeling and generation capabilities from two-dimensional clothing pieces to three-dimensional clothing, and has not yet involved how to integrate popular elements into the design process itself, but instead handed over the interpretation of popular elements to designers.

[0115] Therefore, this embodiment proposes a 3D clothing design and generation scheme based on intelligent fashion trend analysis, aiming to provide designers with an automated process for intelligently generating 3D clothing samples from abstract fashion trend labels. Taking the massive fashion images on the Internet as data analysis materials, this embodiment can first use the DeepFashion deep learning network to analyze the material images as clothing images to obtain clothing image labels of design elements; then construct the color space through clothing color scheme analysis supplemented by PCA dimensionality reduction, and use color as the main direction of trend analysis; then associate the clothing image labels with the 3D clothing models, complete the ability to retrieve 3D clothing models and component models based on the clothing image labels, and realize the generation of 3D clothing models through the splicing and reorganization of clothing components, supplemented by texture, material and other surface information, and complete the visualization of 3D clothing models through rendering, thereby enhancing designers' perception of design concepts, assisting them to more accurately grasp the real demands of the market and customers, speeding up the clothing design process, reducing time and raw material costs, and achieving the goal of improving the efficiency of digital clothing design.

[0116] This embodiment can take color as the main fashion trend in fashion trend analysis to understand the clothing itself. With the help of DeepFashion's deep learning capability, it can use 16 refined tags to describe the clothing in multiple dimensions from basic clothing attributes such as version, color, category, product attributes such as brand, season, and sales attributes such as transaction volume, and establish an association between clothing image tags and 3D clothing model tags, thereby realizing the ability to search 3D clothing models and components with images and generate designed 3D models. These are capabilities that existing trend websites do not have.

[0117] The above method of this embodiment is further introduced below.

[0118] Figure 3 is a schematic diagram of a three-dimensional clothing generation solution based on intelligent fashion trend analysis according to an embodiment of the present invention. Figure 3As shown, a large amount of clothing image data can be obtained first, that is, popular material data can be obtained, and then popular element labels that need to be analyzed can be formulated and main color extraction can be performed, wherein popular element labels can include women's clothing, dresses, lady style, short style, slim fit, solid color, spring and summer labels; then color latent space dimensionality reduction visualization can be performed, that is, popular color version latent space encoding can be performed; image three-dimensional clothing model retrieval and component splicing generation can be performed; finally, three-dimensional clothing model visualization rendering can be performed.

[0119] The following is an introduction to the acquisition of popular material data of this embodiment.

[0120] The material images used in this embodiment may include two categories: popular image data and 3D clothing model data. First, web crawlers can be used to collect data on fashion shows, fashion magazines, social networks, fashion blogs and other websites in a targeted manner to obtain massive image data resources; 3D clothing model data can be obtained from cooperative customers and designers, covering most common categories of men's and women's clothing (including T-shirts, sweatshirts, pants, skirts, dresses, jackets, etc.).

[0121] The analysis material images of this embodiment are introduced below.

[0122] The analysis material image of this embodiment can be a popular element analysis, and the popular element labels that need to be analyzed can be set for the material image. In terms of popular elements, this embodiment can focus on color space and perform main color analysis on the material image. Among them, the key technologies that can be used can be: 1) UNet-based clothing segmentation technology, which aims to obtain a relatively pure clothing area in the material image, and avoid the noise influence of background and human skin color on the main color analysis of clothing as much as possible; 2) For the clothing area, pixel color statistics are performed through a bar graph, and the Top5 colors are obtained as the main color of the clothing.

[0123] This embodiment can define the basic attributes, sales attributes and scene information of clothing from the attribute source of the material image and the scene dimension. Among them, the basic attributes are the objective descriptive information of clothing as a physical object itself; the commodity attributes are the relevant strategy information of clothing as a commodity in the sales scene; and the sales information describes the dynamic data related to marketing and the market generated by clothing during the sales process. This embodiment defines the description tags shown in Table 1 for clothing images and 3D clothing models. It should be noted that the 3D clothing model only has basic attribute tags.

[0124] For the scalability of the algorithm, this embodiment can first complete the collection of basic attribute labels of the image through a self-developed image annotation website. Figure 4 FIG. 1 is a schematic diagram of an image completed by a self-developed image annotation website according to an embodiment of the present invention. Figure 4As shown, the collection of basic attribute labels of the image is completed through the self-developed image annotation website, such as gender, category, style, collar type, length, sleeve length, tightness, texture, pattern, season, style, details, fabric and other basic attribute labels. Optionally, this embodiment can set the number of tags to start reading on the self-developed image annotation website, and then determine that the read tags are the tags that need to be collected.

[0125] After completing the collection of basic attribute labels for the images, the DeepFashion2 network can be used to train the images and labels to obtain a classification network, which can be used to process the labeling of new data.

[0126] The popular color version hidden space coding of this embodiment is introduced below.

[0127] Figure 5 is a schematic diagram of a color space image data visualization according to an embodiment of the present invention. Figure 5 As shown in the figure, in order to facilitate designers to browse related designs freely in the color space, the main color of the image data can be encoded in the latent space, that is, it is converted from the 15-dimensional space R 15 Reduced to 2D space R 2 , so as to visualize it in screen space. Optionally, each image has 5 main colors which can be denoted by c i∈[1，5] =(r i , g i , b i ), so it can be written as a 15-dimensional vector v 15 =(c1, c2, c3, c4, c5). Through PCA dimensionality reduction, we get v 2 = PCA(v 15 ), and v 2 =(x, y) represents the screen space coordinates and also represents a popular image in high-dimensional space.

[0128] Figure 6 FIG. 1 is a schematic diagram of a color space clothing image browsing according to an embodiment of the present invention. Figure 6 As shown, designers can freely browse popular image data by main color by selecting the area of ​​interest in the screen space, thereby greatly saving the designer's time cost of aimlessly searching for design materials in the entire space. It should be noted that the image data after PCA dimensionality reduction in this embodiment will show clustering characteristics in the screen space according to the amount of data and the diversity of color distribution in the image data, thereby explicitly providing designers with popular color information in the global data.

[0129] The following is an introduction to the image-based 3D clothing model retrieval and generation of this embodiment.

[0130] Generating a three-dimensional clothing model directly from a material image is very difficult and unstable. The main reason is that the clothing itself has the characteristics of diverse textures and multiple wrinkles, which will bring uncertain noise effects to the estimation of clothing three-dimensional information, thereby causing inaccurate depth estimation. In order to solve the above problems, this embodiment adopts a three-dimensional clothing model retrieval and generation algorithm based on clothing image labels to realize the generation of three-dimensional clothing models based on images. Specifically, first, the two-dimensional clothing image label is associated with the label in the own three-dimensional clothing model database, that is, the three-dimensional clothing model has the same label attribute as the two-dimensional clothing image, and therefore has label comparability; then, this embodiment divides the three-dimensional clothing model into component models according to the label, and marks the stitchable points of the component model. These component models will become the smallest primitives for retrieval and reorganization of three-dimensional clothing models. When the designer selects an inspiration design image, the algorithm will obtain the clothing image label through the DeepFashion network, and retrieve the relevant models and component models in the three-dimensional clothing model database based on the content of the clothing image label, and select the most similar three-dimensional clothing model from it and return it to the designer.

[0131] The similarity between the three-dimensional clothing model and the image in this embodiment can be quantified by the matching degree of the labels. The solution can adopt cosine similarity. If the clothing image label is represented by a vector I = (I1, I2, ..., I n ), the 3D clothing model label is represented by a vector M = (M1, M2, ..., M n ), then the similarity can be calculated as Sim = I·M / ||I||||M||, where “·” is used to represent the inner product of the vector and “||·||” is used to represent the modulus of the vector. However, the degree of description of clothing features of each 3D clothing model label is different. For example, gender, category, length, etc. describe the significant features of clothing, while texture, fabric, etc. describe the detailed features of clothing. Therefore, the designer defines different weights ω = (ω1, ω2, …, ω n ), and∑ k∈[1，n] ω k = 1. This embodiment can bring the weight into the above cosine similarity formula, and obtain the weighted cosine similarity Sim = (wI·wM) / ||wI||||wM||°

[0132] In addition, this embodiment also supports the retrieval and splicing of clothing component models. Figure 7 FIG. 1 is a schematic diagram of a component stitching algorithm process according to an embodiment of the present invention. Figure 7 As shown, (a) is used to represent the clothing component model. In this embodiment, the coordinate transformation matrix of the clothing component model can be first obtained by ICP calculation, and the component can be transformed to a suitable position. Figure 7 (b). However, there may be inconsistencies in the number of stitching points between components. In this case, this embodiment samples stitching points by finding the farthest distance to the nearest stitchable point, thereby ensuring the consistency of the number of stitchable points and then establishing a corresponding relationship between stitching points, such as Figure 7 (c). Optionally, the wound is a given suture line L = {v0, v1, ...v m} and L′={v0′, v1′, … v n '}, and m<n. For point v i ∈L, the distance v can be found on L′ i The set A whose distance is less than that to other points on L = {v i ′,v j ′,…v p '}, select the point with the largest number as the sampling point on L' according to the natural number of the point on L'. In order to ensure the uniformity of the grid at the component fusion boundary, this embodiment can use the method of re-gridding the resampled component grid, such as Figure 7 (d).

[0133] Optionally, this embodiment can use an incremental re-meshing algorithm, with the three-dimensional mesh of the clothing component model as input, and can be executed according to the following six steps: 1) set the side length e of the clothing component model mesh target, and mark the edge on the stitching point as a fixed edge; 2) all the meshes larger than 4 / 3 e) Insert a new vertex at the edge center and split the triangles associated with the edge; 3) Delete all triangles in the mesh that are smaller than 4 / 5 e edge and merge the triangles related to the edge; 4) delete the vertices with degree greater than 6 in the mesh; 5) in order to ensure the smoothness of the mesh after resampling, update the position coordinates of all vertices on the tangent plane along the normal direction to obtain the new coordinate p′=q+nn T (pq), where q is the cosine Laplace coordinate of vertex p, n is the normal vector of the point, and n T is the transpose of n; 6) The above steps 2)-5) can be repeated multiple times until the mesh is smooth enough. Based on experience, this scheme repeats the above algorithm 5 times to obtain a re-grid three-dimensional clothing component model. The stitching between clothing component models uses Poisson reconstruction to ensure the smoothness of the stitching, such as Figure 7 (e) Then, the clothing component models are fused to obtain the component fusion results, for example, Figure 7 (f).

[0134] For example, if there are clothing component models A and B, and the corresponding seam points are V A =(v1, v2, ..., v n ) and V B =(v1, v2, ..., vn ), then we can solve the linear equation arg min v′ ||L(V′)-δ||+||V A -V B ||Complete the stitching of clothing component models A and B, where V′ is the vertex position coordinate of the reconstructed clothing component model A, L(·) is the vertex coordinate Laplace operator, and δ is the target vertex Laplace coordinate.

[0135] Figure 8 FIG. 1 is a schematic diagram of a 3D model retrieval and component splicing result based on clothing image tags according to an embodiment of the present invention. Figure 8 As shown, the algorithm for generating 3D clothing model retrieval and component splicing results based on clothing image labels can help designers quickly see virtual 3D samples, reduce the time and raw material costs of physical proofing, reduce communication costs, and accelerate designers' rapid response to market trends.

[0136] The clothing texture retrieval method of this embodiment is further introduced below.

[0137] Fig. 9 FIG. 1 is a schematic diagram of a three-dimensional clothing model generation and visual rendering according to an embodiment of the present invention. Fig. 9 As shown, this embodiment can extract a part of the image from the clothing image as a reference based on the inspiration image selected by the designer, and use the image search method to select the most similar texture from the own texture database as the texture of the 3D clothing model. Combined with the own material library, the 3D clothing model visualization rendering is completed through threejs.

[0138] Compared with simple image color analysis, this embodiment focuses more on understanding more clothing attributes, and uses three-dimensional clothing model generation as the output and focus of the algorithm. Optionally, this embodiment defines multiple refined tags from the basic attributes, commodity attributes and sales attributes of clothing to describe clothing images in multiple dimensions. In addition, this embodiment can also accumulate nearly 1,000 common three-dimensional clothing models, and establish an association between images and three-dimensional clothing models through basic attribute tags, thereby realizing image retrieval of three-dimensional clothing models and generating new three-dimensional clothing models through Poisson reconstruction. The above method is not achievable with existing trends.

[0139] The method of this embodiment is a three-dimensional clothing generation algorithm framework based on intelligent fashion trend analysis. Fashion trends have always been the vane guiding clothing design solutions. The algorithm framework proposed in this embodiment can extract popular elements (taking color space as an example) from massive popular clothing images in a data-driven manner, understand clothing images through deep learning algorithms, assist designers to browse freely in the entire domain of clothing images with purpose, and obtain relevant fashion trend data, which helps designers to have a more accurate control of popular elements and provide guarantees for the fashion and trendiness of design solutions; in addition, the three-dimensional clothing model retrieval generation algorithm based on image tags in this embodiment can help designers quickly see virtual three-dimensional samples, reduce the time and raw material costs of physical proofing, reduce communication costs, and accelerate designers' rapid response to market trends. Furthermore, this embodiment is an innovative process framework that combines image visual understanding and three-dimensional clothing model generation, and has capabilities that existing trend analysis portals and three-dimensional design software do not have.

[0140] The three-dimensional clothing model generation algorithm based on the component model of this embodiment realizes the splicing of the three-dimensional clothing component models through ICP, closest point maximum distance sampling and Poisson reconstruction, thereby generating a new three-dimensional clothing model. The algorithm can be used as a generation capability for the diversity of three-dimensional clothing models, and provides a self-expansion method for a self-owned model database.

[0141] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0142] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0143] Example 3

[0144] According to an embodiment of the present invention, a device for processing design requirement data for implementing the above-mentioned method for processing design requirement data is also provided.

[0145] Fig.10 is a schematic diagram of a device for processing design requirement data according to an embodiment of the present invention. Fig.10 As shown, the design requirement data processing device 100 may include: a first acquisition unit 101 , a second acquisition unit 102 and a third acquisition unit 103 .

[0146] The first acquisition unit 101 is used to acquire the design requirements of the object to be designed and collect material images based on the design requirements.

[0147] The second acquisition unit 102 is used to analyze the material image using the product classification model to acquire product image tags for marking design elements of the object to be designed.

[0148] The third acquisition unit 103 is used to acquire a three-dimensional product model associated with the object to be designed based on the product image tag.

[0149] It should be noted that the first acquisition unit 101, the second acquisition unit 102 and the third acquisition unit 103 correspond to steps S202 to S206 of Example 1, respectively. The three units and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in Example 1. It should be noted that the above units, as part of the device, can be run in the computer terminal 10 provided in Example 1.

[0150] In the design requirement data processing device of this embodiment, the product classification model is used to analyze the material images searched based on the design requirements, and product image tags used to mark the design elements of the object to be designed are obtained, and then the purpose of obtaining the three-dimensional product model associated with the object to be designed is achieved based on the product image tags. This helps designers to have a more accurate control over the design elements and provide guarantees for the design plans of the product objects. In addition, obtaining the three-dimensional product model associated with the object to be designed based on the product image tags can help designers quickly see the three-dimensional product model, reduce the time and raw material costs of physical proofing, reduce communication costs, and accelerate designers' rapid response to market trends, thereby solving the technical problem of low efficiency in processing design requirement data and achieving the technical effect of improving the efficiency of processing design requirement data.

[0151] Example 4

[0152] The embodiment of the present invention can provide a design requirement data processing, the design requirement data processing includes a computer terminal, and the computer terminal can be any computer terminal device in a computer terminal group. Optionally, in this embodiment, the above-mentioned computer terminal can also be replaced by a terminal device such as a mobile terminal.

[0153] Optionally, in this embodiment, the computer terminal may be located in at least one network device among a plurality of network devices of the computer network.

[0154] In this embodiment, the above-mentioned computer terminal can execute the program code of the following steps in the method for processing design requirement data of an application: obtaining the design requirements of the object to be designed, and collecting material images based on the design requirements; using a product classification model to analyze the material images, and obtaining product image tags for marking the design elements of the object to be designed; based on the product image tags, obtaining a three-dimensional product model associated with the object to be designed.

[0155] Optionally, Fig.11 is a structural block diagram of a computer terminal according to an embodiment of the present invention. Fig.11 As shown, the computer terminal A may include: one or more (only one is shown in the figure) processors 112 , a memory 114 , and a transmission device 116 .

[0156] Wherein, the transmission device is used to obtain the design requirements of the object to be designed. Wherein, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the method and device for processing design requirement data in the embodiment of the present invention, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method for processing design requirement data. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal A via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0157] The processor can call the information and application programs stored in the memory through the transmission device to execute the following steps: obtain the design requirements of the object to be designed, and collect material images based on the design requirements; use the product classification model to analyze the material images, and obtain product image tags for marking the design elements of the object to be designed; based on the product image tags, obtain a three-dimensional product model associated with the object to be designed.

[0158] Optionally, the processor may also execute the program code of the following steps: performing similarity calculations between the product image label and the labels of each three-dimensional product model in the three-dimensional product model database; based on the calculation results, obtaining the three-dimensional product model with the highest similarity from the three-dimensional product model database, wherein the three-dimensional product model with the highest similarity is the three-dimensional product model associated with the object to be designed.

[0159] Optionally, the processor may also execute program code for the following steps: creating a three-dimensional product model database before calculating the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database; associating the labels in the three-dimensional product model database with pre-stored product image labels; dividing each three-dimensional product model in the three-dimensional product model database into components according to the labels to obtain multiple component models; and marking the stitchable points of the multiple component models, wherein each component serves as the smallest unit for retrieval and reorganization of the three-dimensional product model.

[0160] Optionally, the processor may also execute program code for the following steps: before marking the stitchable points of multiple component models, obtaining the coordinate transformation matrix of each component model, wherein the coordinate transformation matrix of each component model constitutes a re-gridded three-dimensional product component model; transforming each component to a predetermined position and obtaining the coordinate position of each component model; and determining the stitchable points between any two or more components based on the coordinate position of each component model.

[0161] Optionally, the processor may also execute the program code of the following steps: after collecting material images based on design requirements, performing latent spatial encoding on the main color tone of the material image to reduce the display dimension of the material image; based on the reduced display dimension, visualizing the material image in the screen space, wherein the image data of the material image after reducing the display dimension is clustered in the screen space according to the data amount and color distribution.

[0162] Optionally, the processor may also execute program code for the following steps: obtaining product sample material before analyzing the material image using the product classification model, wherein the product sample material includes: product image and product design model; labeling the product sample material using attribute labels, wherein the attribute labels include at least one of the following: product attributes and transaction attributes; training the product image and the corresponding attribute labels using a neural network model to generate a product classification model, wherein the product classification model is used to label the object to be analyzed with corresponding attribute labels.

[0163] Optionally, the processor may also execute the program code of the following steps: after obtaining the three-dimensional product model associated with the object to be designed, selecting a texture with the highest similarity to the product image label from a pre-stored texture database as the texture of the three-dimensional product model; and visually rendering the three-dimensional product model based on the texture of the three-dimensional product model.

[0164] As an optional example, the processor can call the information and application stored in the memory through the transmission device to perform the following steps: obtain the design requirements of the object to be designed by calling the first interface, and collect material images based on the design requirements, wherein the first interface includes: a first parameter, and the parameter value of the first parameter is the design requirement; use the product classification model to analyze the material image to obtain a product image label for marking the design elements of the object to be designed; based on the product image label, obtain a three-dimensional product model associated with the object to be designed; output the three-dimensional product model by calling the second interface, wherein the second interface includes: a second parameter, and the parameter value of the second parameter is the three-dimensional product model.

[0165] As another optional example, the processor may call the information and application programs stored in the memory through the transmission device to perform the following steps: obtaining the design requirements of the object to be designed on the e-commerce platform, and collecting e-commerce material images based on the design requirements; using the product classification model to analyze the e-commerce material images, and obtaining e-commerce image tags for marking the design elements of the object to be designed; based on the e-commerce image tags, obtaining a three-dimensional e-commerce model associated with the object to be designed.

[0166] According to an embodiment of the present invention, a method for processing design requirement data is provided. The design requirements of the object to be designed are obtained, and material images are collected based on the design requirements; the material images are analyzed using a product classification model to obtain product image tags for marking the design elements of the object to be designed; and the three-dimensional product model associated with the object to be designed is obtained based on the product image tags. That is to say, the present application uses a product classification model to analyze the material images searched based on the design requirements, obtains product image tags for marking the design elements of the object to be designed, and then achieves the purpose of obtaining the three-dimensional product model associated with the object to be designed based on the product image tags, which helps designers to have a more accurate control of the design elements and provide guarantees for the design scheme of the product object. In addition, obtaining the three-dimensional product model associated with the object to be designed based on the product image tags can help designers quickly see the three-dimensional product model, reduce the time and raw material cost of physical proofing, reduce communication costs, and accelerate designers to respond quickly to market trends, thereby solving the technical problem of low processing efficiency of design requirement data and achieving the technical effect of improving the processing efficiency of design requirement data.

[0167] It can be understood by those skilled in the art that Fig.11The structure shown is for illustration only, and the computer terminal A may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a handheld computer, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Fig.11 It does not limit the structure of the above-mentioned computer terminal A. For example, the computer terminal A may also include Fig.11 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Fig.11 Different configurations are shown.

[0168] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0169] Example 5

[0170] The embodiment of the present invention further provides a computer-readable storage medium. Optionally, in this embodiment, the computer-readable storage medium can be used to store the program code executed by the method for processing design requirement data provided in the first embodiment.

[0171] Optionally, in this embodiment, the computer-readable storage medium may be located in any one of the computer terminals in a computer terminal group in a computer network, or in any one of the mobile terminals in a mobile terminal group.

[0172] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: obtaining design requirements of the object to be designed, and collecting material images based on the design requirements; using a product classification model to analyze the material images, and obtaining product image labels for annotating the design elements of the object to be designed; based on the product image labels, obtaining a three-dimensional product model associated with the object to be designed.

[0173] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: calculating the similarity between the product image label and the labels of each three-dimensional product model in the three-dimensional product model database; based on the calculation results, obtaining the three-dimensional product model with the highest similarity from the three-dimensional product model database, wherein the three-dimensional product model with the highest similarity is the three-dimensional product model associated with the object to be designed.

[0174] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: creating a three-dimensional product model database before calculating the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database; associating the labels in the three-dimensional product model database with pre-stored product image labels; dividing each three-dimensional product model in the three-dimensional product model database into components according to the labels to obtain multiple component models; and marking the stitchable points of the multiple component models, wherein each component serves as the smallest unit for retrieval and reorganization of the three-dimensional product model.

[0175] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: before marking the stitchable points of multiple component models, obtaining a coordinate transformation matrix of each component model, wherein the coordinate transformation matrix of each component model constitutes a re-gridded three-dimensional product component model; transforming each component to a predetermined position and obtaining the coordinate position of each component model; and determining the stitchable points between any two or more components based on the coordinate position of each component model.

[0176] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: after collecting material images based on design requirements, implicitly encoding the main color tones of the material images to reduce the display dimension of the material images; based on the reduced display dimension, visualizing the material images in the screen space, wherein the image data of the material images after the display dimension is reduced are clustered in the screen space according to the data amount and color distribution.

[0177] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: before using a product classification model to analyze a material image, obtaining product sample material, wherein the product sample material includes: a product image and a product design model; labeling the product sample material with attribute labels, wherein the attribute labels include at least one of the following: product attributes and transaction attributes; training the product images and corresponding attribute labels with a neural network model to generate a product classification model, wherein the product classification model is used to label the object to be analyzed with corresponding attribute labels.

[0178] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: after obtaining a three-dimensional product model associated with the object to be designed, selecting a texture with the highest similarity to the product image label from a pre-stored texture database as the texture of the three-dimensional product model; and visually rendering the three-dimensional product model based on the texture of the three-dimensional product model.

[0179] As an optional example, a computer-readable storage medium is configured to store program code for performing the following steps: obtaining design requirements of an object to be designed by calling a first interface, and collecting material images based on the design requirements, wherein the first interface includes: a first parameter, and a parameter value of the first parameter is the design requirement; using a product classification model to analyze the material image, and obtaining a product image label for marking the design elements of the object to be designed; based on the product image label, obtaining a three-dimensional product model associated with the object to be designed; and outputting the three-dimensional product model by calling a second interface, wherein the second interface includes: a second parameter, and a parameter value of the second parameter is the three-dimensional product model.

[0180] As another optional example, a computer-readable storage medium is configured to store program code for executing the following steps: obtaining design requirements for an object to be designed on an e-commerce platform, and collecting e-commerce material images based on the design requirements; using a product classification model to analyze the e-commerce material images, and obtaining e-commerce image tags for annotating design elements of the object to be designed; based on the e-commerce image tags, obtaining a three-dimensional e-commerce model associated with the object to be designed.

[0181] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0182] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0183] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0184] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0185] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0186] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0187] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for processing design requirement data, characterized in that: include: Obtaining design requirements of the object to be designed, and collecting material images based on the design requirements; Analyzing the material image using a product classification model to obtain product image labels for marking design elements of the object to be designed; Based on the product image tag, obtaining a three-dimensional product model associated with the object to be designed; Wherein, based on the product image tag, a three-dimensional product model associated with the object to be designed is obtained, including: based on a three-dimensional product model database, a plurality of component models are determined, wherein the stitchable points of the component models are determined by the coordinate transformation matrix and coordinate position of each of the component models; based on the product image tag and the plurality of component models, the three-dimensional product model is determined.

2. The method according to claim 1, characterized in that Based on the product image tag, obtaining a three-dimensional product model associated with the object to be designed includes: Calculate the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database; Based on the calculation result, the three-dimensional product model with the highest similarity is obtained from the three-dimensional product model database, wherein the three-dimensional product model with the highest similarity is the three-dimensional product model associated with the object to be designed.

3. The method according to claim 2, characterized in that Before respectively calculating the similarity between the product image label and the label of each three-dimensional product model in the three-dimensional product model database, the method further includes: Creating the three-dimensional product model database; Associating the labels in the three-dimensional product model database with pre-stored product image labels; Segment each 3D product model in the 3D product model database into components according to the labels to obtain a plurality of component models; The stitchable points of the multiple component models are marked, wherein each component is used as the smallest unit for retrieval and reorganization of the three-dimensional product model.

4. The method according to claim 3, characterized in that Before marking the stitchable points of the plurality of component models, the method further includes: Acquire the coordinate transformation matrix of each component model, wherein the coordinate transformation matrix of each component model constitutes a re-grid three-dimensional product component model; Transforming each of the components to a predetermined position, and obtaining the coordinate position of each component model; The sewable points between any two or more components are determined based on the coordinate position of each component model.

5. The method according to claim 4, characterized in that The stitching between any two components is reconstructed using Poisson reconstruction, which makes the stitching smooth.

6. The method according to claim 1, characterized in that After collecting material images based on the design requirements, the method further includes: Performing latent spatial encoding on the main color tone of the material image to reduce the display dimension of the material image; Based on the reduced display dimension, the material image is visualized in the screen space, wherein the image data of the material image after reducing the display dimension is clustered in the screen space according to data volume and color distribution.

7. The method according to any one of claims 1 to 6, characterized in that Before analyzing the material image using the product classification model, the method further includes: Acquire product sample materials, wherein the product sample materials include: product images and product design models; The product sample material is labeled using attribute tags, wherein the attribute tags include at least one of the following: product attributes and transaction attributes; The product images and corresponding attribute labels are trained using a neural network model to generate the product classification model, wherein the product classification model is used to annotate the corresponding attribute labels for the objects to be analyzed.

8. The method according to claim 1, characterized in that After acquiring the three-dimensional product model associated with the object to be designed, the method further includes: Selecting a texture having the highest similarity to the product image label from a pre-stored texture database as the texture of the three-dimensional product model; Based on the texture of the three-dimensional product model, the three-dimensional product model is visually rendered.

9. A method for processing design requirement data, characterized in that: include: The design requirements of the object to be designed are obtained by calling a first interface, and material images are collected based on the design requirements, wherein the first interface includes: a first parameter, and a parameter value of the first parameter is the design requirement; Analyzing the material image using a product classification model to obtain product image labels for marking design elements of the object to be designed; Based on the product image tag, obtaining a three-dimensional product model associated with the object to be designed; Outputting the three-dimensional product model by calling a second interface, wherein the second interface includes: a second parameter, a parameter value of the second parameter is the three-dimensional product model; Wherein, based on the product image tag, a three-dimensional product model associated with the object to be designed is obtained, including: based on a three-dimensional product model database, a plurality of component models are determined, wherein the stitchable points of the component models are determined by the coordinate transformation matrix and coordinate position of each of the component models; based on the product image tag and the plurality of component models, the three-dimensional product model is determined.

10. A method for processing design requirement data, characterized in that: include: Obtaining design requirements of the object to be designed on the e-commerce platform, and collecting e-commerce material images based on the design requirements; Analyzing the e-commerce material image using a product classification model to obtain e-commerce image tags for marking design elements of the object to be designed; Based on the e-commerce image tag, obtaining a three-dimensional e-commerce model associated with the object to be designed; Among them, based on the e-commerce image tag, a three-dimensional e-commerce model associated with the object to be designed is obtained, including: based on a three-dimensional product model database, a plurality of component models are determined, wherein the stitchable points of the component models are determined by the coordinate transformation matrix and coordinate position of each of the component models; based on the e-commerce image tag and the plurality of component models, the three-dimensional e-commerce model is determined.

11. A device for processing design requirement data, characterized in that: include: A first acquisition unit, used to acquire design requirements of the object to be designed, and collect material images based on the design requirements; A second acquisition unit is used to analyze the material image using a product classification model to acquire a product image label for marking the design elements of the object to be designed; A third acquisition unit, configured to acquire a three-dimensional product model associated with the object to be designed based on the product image tag; Among them, the third acquisition unit is also used to determine multiple component models based on the three-dimensional product model database, wherein the stitchable points of the component models are used to determine the coordinate transformation matrix and coordinate position of each component model; and is also used to determine the three-dimensional product model based on the product image label and multiple component models.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed by a processor, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 10.

13. A processor, characterized in that: The processor is used to run a program, wherein the program executes the method according to any one of claims 1 to 10 when being run by the processor.

14. A system for processing design requirement data, characterized in that: include: processor; A memory, connected to the processor, for providing the processor with instructions for processing the following processing steps: obtaining design requirements of the object to be designed, and collecting material images based on the design requirements; The material image is analyzed using a product classification model to obtain a product image label for marking the design elements of the object to be designed; based on the product image label, a three-dimensional product model associated with the object to be designed is obtained; wherein, based on the product image label, the three-dimensional product model associated with the object to be designed is obtained, including: based on a three-dimensional product model database, a plurality of component models are determined, wherein the stitchable points of the component models are determined by the coordinate transformation matrix and the coordinate position of each of the component models; and the three-dimensional product model is determined based on the product image label and the plurality of component models.

Images