Costume design assisting method and system based on artificial intelligence

Abstract

The invention discloses a costume design assisting method and system based on artificial intelligence, and belongs to the field of costume intelligent design, and the method comprises the steps: extracting multi-dimensional style feature information from a pre-constructed costume image data set and a fashion trend database through employing a convolutional neural network, and constructing a style vector space model; vector similarity matching is carried out, and a costume design sketch is generated by adopting a graph neural network; based on the design sketch, generating a structured process data packet for proofing; the design sketch and the structural data are input into a three-dimensional human body modeling and simulation module, and wearing effect simulation of virtual clothes is realized based on human body motion capture and physical cloth simulation technologies; and constructing a human-computer interaction interface, and quickly responding to a modification request based on the generative adversarial network model. Structured information such as the garment pattern structure, the sewing sequence, the material specification and the process route is automatically generated through the reinforcement learning algorithm, and the garment proofing efficiency and accuracy are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent clothing design, and more specifically, relates to an artificial intelligence-based clothing design auxiliary method and system. Background Art

[0002] With the advancement of information technology, fashion design has gradually evolved from traditional manual drawing and empirical proofing to digital and intelligent design. Currently, technologies such as computer-aided design, virtual fitting, and 3D human modeling have been adopted to a certain extent in the apparel industry. However, existing apparel design processes still suffer from common shortcomings: Traditional design processes rely on the experience and judgment of professional designers, requiring manual operations from sketching to pattern making. This is inefficient and difficult to adapt to rapidly changing market demands. Current intelligent design systems lack effective human-computer interaction mechanisms, making it difficult for users to customize styles or make minor modifications through intuitive methods such as natural language and voice, creating a high barrier to entry for personalized customization. While some AI-assisted tools can generate patterns or recommend colors, they lack in-depth analysis and reasoning of the resulting structural rationality, sewing logic, and material feasibility, and are unable to provide complete, manufacturable process solutions. Existing tools often separate garment design from fitting simulation and construction processes, lacking a coherent process based on a common semantic intent. This makes it difficult to quickly implement design results and leads to lengthy verification cycles.

[0003] Therefore, there is an urgent need for an intelligent design assistance method that integrates artificial intelligence algorithms, human-computer interaction technology and clothing industry processes, which can achieve integrated design support from user intention analysis, style sketch generation, structure deduction to virtual fitting, and improve the efficiency, accuracy and user participation of clothing customization. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above or existing problems of artificial intelligence-based clothing design auxiliary methods and systems, the present invention is proposed.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] An embodiment of the present invention provides an artificial intelligence-based clothing design assistance method, including: parsing the design requirements input by the user through a natural language processing model, extracting multi-dimensional style feature information from a pre-constructed clothing image data set and fashion trend database using a convolutional neural network, and constructing a style vector space model; performing vector similarity matching between the design vector and the style vector space model, and generating a clothing design sketch using a graph neural network; based on the design sketch, calling a reinforcement learning algorithm to automatically deduce the clothing pattern structure, sewing sequence, material specifications and processing route, and generating a structured process data packet for proofing; inputting the design sketch and structure data into a three-dimensional human body modeling and simulation module, and realizing the simulation of the wearing effect of virtual clothing based on human motion capture and physical fabric simulation technology; constructing a human-computer interaction interface to support users to make modification suggestions for local style, color or structure through voice, and quickly respond to modification requests based on a generative adversarial network model.

[0008] As a preferred solution of the artificial intelligence-based clothing design assistance method of the present invention, the design requirements input by the user are parsed through a natural language processing model, including:

[0009] The system collects user demand information about clothing design through a graphical interface, voice recognition module or text input box, uses a pre-trained language model to encode text semantics, uses a deep learning model for intent classification and slot filling, and converts design parameters into a unified design vector representation.

[0010] As a preferred embodiment of the artificial intelligence-based clothing design auxiliary method of the present invention, a convolutional neural network is used to extract multi-dimensional style feature information from a pre-built clothing image dataset and a fashion trend database, and a style vector space model is constructed, including:

[0011] The image's style vectors form a high-dimensional style space. Principal component analysis is used for visualization and dimensionality reduction. A clustering algorithm is used to divide the style space into several style clusters. Metric learning methods are used to enhance the distinguishability between vectors. An indexing mechanism is then constructed to enable subsequent matching.

[0012] The constructed style vector space is connected to the user intention vector to support subsequent semantic query and design generation.

[0013] As a preferred solution of the artificial intelligence-based clothing design auxiliary method of the present invention, wherein: the design vector is matched with the style vector space model for vector similarity, and a graph neural network is used to generate a clothing design sketch, including:

[0014] For the selected k image samples, a graph structure G = (V, E) is constructed, where the node set V = {v T ,hi1 ,...,h ik}, the edge set E constructs edge weights based on the similarity between style vectors;

[0015] Use graph attention network or graph convolution network to aggregate the information in the graph into a generated vector z T :

[0016] z T =GNN(G)

[0017] Among them, GNN (G) is the forward propagation of the graph neural network, z T It is the global feature that is ultimately used to generate the sketch;

[0018] Based on z T Input, use the generative adversarial network to generate sketch images, and add sketch structure guidance in the generation process.

[0019] As a preferred embodiment of the artificial intelligence-based clothing design assistance method of the present invention, the method includes: based on the design sketch, calling a reinforcement learning algorithm to automatically deduce the clothing pattern structure, sewing sequence, material specifications and processing route, and generating a structured process data package for proofing, including:

[0020] Use the image structure analysis network to perform image semantic segmentation on the sketch, extract clothing feature areas, and use the policy gradient method in reinforcement learning to optimize the design strategy:

[0021]

[0022] Among them, θ is the policy network parameter, γ is the reward discount factor, r t A score based on the comparison between the generated solution and the target;

[0023] Design reward function:

[0024] r t =α1r 结构相符 +α2r 缝制合理 +α3r 材料利用 +α4r 工艺可行性

[0025] Among them, α i is the weight coefficient of the corresponding sub-reward, r 结构相符 is the structural consistency score, r 缝制合理 is the sewing rationality score, r 材料利用 is the material utilization score, r 工艺可行性 Score the process feasibility.

[0026] As a preferred embodiment of the artificial intelligence-based clothing design assistance method of the present invention, the design sketch and structural data are input into a three-dimensional human body modeling and simulation module, and the wearing effect simulation of virtual clothing is realized based on human motion capture and physical cloth simulation technology, including:

[0027] Use 3D human body modeling tools to build an interactive high-precision human body mesh model H(x,y,z). Use human motion capture equipment to collect user static and dynamic posture data, and drive the 3D human body model to perform posture animation reconstruction:

[0028] H t =f pose (H0,M t )

[0029] Among them, H0 is the initial human body model, M t is the posture data of time step t, H t is the 3D human body model after the action, f pose is the posture transformation function;

[0030] The 2D pattern structure data is converted into a corresponding 3D mesh of patches, where each patch represents a pattern unit. Sewing operations are performed on each mesh boundary point according to the sewing sequence defined in the structure data to construct an initial 3D garment model. Based on the mechanical properties of the fabric, physical parameters are assigned to each garment mesh, initializing the physics solver in the fabric simulation engine.

[0031] The clothing model is bound to the 3D human body model, and the dynamic deformation of the clothing as the human body moves is simulated through collision detection and constraint solving algorithms; the simulation results are rendered in real time as a 3D animation try-on effect, and the user can choose different body shapes and postures to observe the wearing fit and wrinkle changes of the clothing.

[0032] As a preferred solution of the artificial intelligence-based clothing design assistance method of the present invention, a human-computer interaction interface is constructed to support users to make modification suggestions for local styles, colors or structures through voice, and a generative adversarial network model is used to quickly respond to modification requests, including:

[0033] A conditional generative adversarial network is constructed for clothing design. The input is the current design sketch and a semantic condition vector. The condition vector is encoded as a latent vector by the semantic parsing result. The original design drawing and the semantic modification vector are input into the generator. The generator performs target style transfer or image redrawing on local areas of the image according to different types of modification instructions. For color matching modifications, the style coloring module is used to adjust the overall or local area color matrix. The generated clothing sketch is rendered back to the human-computer interface in real time. The user can zoom in, rotate or overlay historical comparison images on the interface to confirm whether the modification is as expected. If the user is not satisfied, he or she can initiate a voice command again to enter a new round of modification cycle.

[0034] An artificial intelligence-based clothing design assistance system includes: a feature extraction module for parsing user input design requirements through a natural language processing model, extracting multi-dimensional style feature information from a pre-built clothing image dataset and fashion trend database using a convolutional neural network, and constructing a style vector space model;

[0035] The matching generation module is used to perform vector similarity matching between the design vector and the style vector space model, and generate clothing design sketches using a graph neural network;

[0036] The auxiliary generation module is used to call the reinforcement learning algorithm to automatically deduce the garment pattern structure, sewing sequence, material specifications and processing route based on the design sketch, and generate a structured process data package for proofing;

[0037] The fitting simulation module is used to input design sketches and structural data into the 3D human body modeling and simulation module. Based on human motion capture and physical fabric simulation technology, it can simulate the wearing effect of virtual clothing.

[0038] The interactive correction module is used to build a human-computer interaction interface, allowing users to make modification suggestions for local styles, colors or structures through voice, and quickly respond to modification requests based on the generative adversarial network model.

[0039] A computing device, comprising:

[0040] at least one processor, memory, and input-output unit;

[0041] The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the steps of the artificial intelligence-based clothing design assistance method.

[0042] A computer-readable storage medium includes instructions that, when executed on a computer, enable the computer to execute the steps of a clothing design assistance method based on artificial intelligence.

[0043] The beneficial effects of the present invention are as follows: by introducing a natural language processing model, the present invention can accurately extract information such as design requirements, style preferences, functional characteristics, etc. from the text or voice input by the user, establish a semantic-driven clothing design process, and enhance user participation and interactive intuitiveness. Based on the clothing image database and fashion trend corpus, a convolutional neural network is used to construct a style vector space to achieve modeling and matching of multi-dimensional style features. The input design vector is combined with a graph neural network to perform style mapping and style sketch generation, thereby improving the personalization and diversity of the design response. The reinforcement learning algorithm is used to automatically generate structured information such as clothing pattern structure, sewing sequence, material specifications, and process routes, and can quickly output data packets that can be used for proofing or industrial production, significantly improving the efficiency and accuracy of clothing proofing. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 A flowchart of an artificial intelligence-based clothing design assistance method provided in an embodiment of the present invention.

[0046] Figure 2 A schematic diagram of the structure of an artificial intelligence-based clothing design assistance system provided in an embodiment of the present invention.

[0047] Figure 3 The figure schematically shows a structural diagram of a medium according to an embodiment of the present invention.

[0048] Figure 4 The figure schematically shows a structural diagram of a computing device according to an embodiment of the present invention.

[0049] In the drawings, the same or corresponding reference numerals denote the same or corresponding parts. DETAILED DESCRIPTION

[0050] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0053] Example

[0054] Reference below Figure 1 , Figure 1 This is a flow chart of an artificial intelligence-based clothing design assistance method provided by an embodiment of the present invention. It should be noted that the embodiments of the present invention can be applied to any applicable scenario.

[0055] Figure 1 The process of the artificial intelligence-based clothing design assistance method provided by one embodiment of the present invention includes:

[0056] S1: The natural language processing model is used to parse the design requirements input by the user. From the pre-built clothing image dataset and fashion trend database, a convolutional neural network is used to extract multi-dimensional style feature information and construct a style vector space model.

[0057] Preferably, the system collects user demand information about clothing design through a graphical interface, voice recognition module or text input box, uses a pre-trained language model to encode text semantics, uses a deep learning model to perform intent classification and slot filling, and converts design parameters into a unified design vector representation.

[0058] Preferably, the image style vector set forms a high-dimensional style space, which is visualized and reduced using principal component analysis. A clustering algorithm is used to divide the style space into several style clusters. Metric learning methods are used to enhance the distinguishability between vectors, and an indexing mechanism is constructed to achieve subsequent matching.

[0059] The constructed style vector space is connected to the user intention vector to support subsequent semantic query and design generation.

[0060] Furthermore, we extract deep style features from pre-built high-quality clothing image datasets (such as DeepFashion and Lookbook datasets) using pre-trained ResNet, EfficientNet, or Vision Transformer models to obtain the vector corresponding to each image. The set of all image style features F = {f1, f2, ..., f N}Construct a high-dimensional style space and use PCA (principal component analysis) for dimensionality reduction visualization: Use unsupervised clustering methods such as K-means and DBSCAN to classify the style vectors after dimensionality reduction and obtain several style clusters C1, C2, ..., C k ; Use metric learning methods such as Triplet Loss to train the discriminant network to enhance the discriminability between different style vectors:

[0061]

[0062] Among them, f a is the anchor image style vector, f p is a positive sample, f n is a negative sample, ∈ is an interval;

[0063] Build a vector retrieval structure for fast matching and retrieval of style vectors, shortening the response time for subsequent design generation;

[0064] The design vector v design Similarity matching with the style vector space, using cosine similarity or Euclidean distance:

[0065]

[0066] Among them, f i The style vector of the i-th style image;

[0067] Representative images from the Top-K most similar style clusters are selected as references to improve user satisfaction and style consistency of the generated images; the matching results are used as input reference samples for the graph neural network generation module, and combined with the design intent vector to jointly generate clothing sketch images with consistent style and reasonable structure.

[0068] S2: Perform vector similarity matching between the design vector and the style vector space model, and use graph neural network to generate clothing design sketches.

[0069] Preferably, for the selected k image samples, a graph structure G = (V, E) is constructed, where the node set V = {v T ,h i1 ,...,h ik}, the edge set E constructs edge weights based on the similarity between style vectors;

[0070] Use graph attention network or graph convolution network to aggregate the information in the graph into a generated vector z T :

[0071] z T =GNN(G)

[0072] Among them, GNN (G) is the forward propagation of the graph neural network, zT It is the global feature that is ultimately used to generate the sketch;

[0073] Based on z T Input, use the generative adversarial network to generate sketch images, and add sketch structure guidance in the generation process.

[0074] Furthermore, metric learning or similarity matching algorithm is used to retrieve the style vector space of clothing images that are similar to v T The most similar Top-K image samples, the node set V contains the design vector and k candidate image style vectors, G is input into the graph neural network model, and semantic aggregation processing is performed on the information of each node in the graph. For each node, the attention coefficient is calculated based on the features of its neighboring nodes and the edge weight. Through the multi-layer graph neural network, the context information of each node in the graph is integrated to finally obtain the global design representation vector. The conditional GAN ​​based on structure guidance is used, and the input is the global feature vector z T and sketch structure prior S prior , the discriminator receives the generated image and the design vector as conditional input, and judges the authenticity of the image and its style consistency with the intent vector. The overall GAN ​​loss function is:

[0075]

[0076] in, is the structure preservation loss, and λ is the weight coefficient of the structure constraint term.

[0077] S3: Based on the design sketch, the reinforcement learning algorithm is called to automatically deduce the garment pattern structure, sewing sequence, material specifications and processing route, and generate a structured process data package for proofing.

[0078] Preferably, an image structure parsing network is used to perform semantic segmentation on the sketch, extract clothing feature areas, and optimize the design strategy using the policy gradient method in reinforcement learning:

[0079]

[0080] Among them, θ is the policy network parameter, γ is the reward discount factor, r t A score based on the comparison between the generated solution and the target;

[0081] Design reward function:

[0082] r t =α1r 结构相符 +α2r 缝制合理 +α3r 材料利用 +α4r 工艺可行性

[0083] Among them, α iis the weight coefficient of the corresponding sub-reward, r 结构相符 is the structural consistency score, r 缝制合理 is the sewing rationality score, r 材料利用 is the material utilization score, r 工艺可行性 Score the process feasibility.

[0084] Furthermore, sketch semantic structure mask:

[0085] "collar":{"position":[x1,y1,x2,y2],"category":"collar","curve":[...]},

[0086] "sleeve":{"position":[x3,y3,x4,y4],"category":"sleeve","style":"puff"},

[0087] "hem":{"position":[x5,y5,x6,y6],"category":"hem","symmetry":0.98}, ...

[0089] }

[0090] State vector s t :

[0091] s_t=[s_vec,m_vec,p_vec]

[0092] in:

[0093] -s_vec∈R 64 : Structural features extracted from segmented regions (such as aspect ratio, complexity, symmetry, etc.)

[0094] -m_vec∈R 16 : Feature vectors of selected fabric type, elasticity, weight, and color

[0095] -p_vec∈R 32 : Sewing sequence, sewing method, processing temperature and other process parameters;

[0096] Update the policy network parameters using the policy gradient method:

[0097]

[0098] Among them, Πθ is the probability of the policy network generating the current action, R t is the discounted cumulative reward, γ is the reward discount factor, and strategy optimization algorithms such as REINFORCE are used to update θ.

[0099] S4: Input the design sketch and structural data into the 3D human body modeling and simulation module, and simulate the wearing effect of virtual clothing based on human motion capture and physical fabric simulation technology.

[0100] Preferably, a 3D human body modeling tool is used to build an interactive high-precision human body mesh model H(x, y, z), a human motion capture device is used to collect static and dynamic posture data of the user, and the 3D human body model is driven to perform posture animation reconstruction:

[0101] H t =f pose (H0,M t )

[0102] Among them, H0 is the initial human body model, M t is the posture data of time step t, H t is the 3D human body model after the action, f pose is the posture transformation function;

[0103] The 2D pattern structure data is converted into a corresponding 3D mesh of patches, where each patch represents a pattern unit. Sewing operations are performed on each mesh boundary point according to the sewing sequence defined in the structure data to construct an initial 3D garment model. Based on the mechanical properties of the fabric, physical parameters are assigned to each garment mesh, initializing the physics solver in the fabric simulation engine.

[0104] The clothing model is bound to the 3D human body model, and the dynamic deformation of the clothing as the human body moves is simulated through collision detection and constraint solving algorithms; the simulation results are rendered in real time as a 3D animation try-on effect, and the user can choose different body shapes and postures to observe the wearing fit and wrinkle changes of the clothing.

[0105] Furthermore, the user action sequence is collected by the motion capture device: action frame M t ={j 1t ,j 2t ,...,j mt}, represents the posture of each bone node, and uses the bone skinning algorithm to map the bone action to the mesh model to achieve deformation-driven animation;

[0106] Each pattern piece P i Construct a two-dimensional patch mesh Mi(u,v); let A_i and A_j be two paper pattern boundary segments (vertex sequence is and ); adopt the boundary point pairing stitching strategy, define the stitching function, perform stitching operations sequentially, and generate the overall three-dimensional clothing mesh model F0(x,y,z).

[0107] Call the physical simulation engine, configure the collision detection module and the physical solver, establish the binding relationship between the clothing F0 and the human body H0, and for each frame action t: use H t =f_pose(H0,M t ) Update the human body model; execute the physics solver to calculate the dynamic response of the clothing mesh as the action changes F t Detect clothing penetration and deformation information, adjust fabric tension, collision rebound; output the coordinate set of clothing vertices after deformation F t ={v'1,v'2,...,v'n}.

[0108] S5: Build a human-computer interaction interface that allows users to make modification suggestions for local styles, colors, or structures through voice, and quickly respond to modification requests based on a generative adversarial network model.

[0109] Preferably, a conditional generative adversarial network is constructed for clothing design, with the current design sketch and a semantic condition vector as input. The condition vector is encoded as a latent vector by the result of semantic parsing. The original design drawing and the semantic modification vector are input into the generator, and the generator performs target style transfer or image redrawing on the local area of ​​the image according to different types of modification instructions; for color matching modifications, the style coloring module is used to adjust the overall or local area color matrix; the generated clothing sketch is rendered back to the human-computer interface in real time, and the user zooms in, rotates or overlays historical comparison images on the interface to confirm whether the modification is as expected; if the user is not satisfied, he or she can initiate a voice command again to enter a new round of modification cycle.

[0110] Furthermore, the input fusion module performs early fusion of the original image I0 with the conditional vector Zc. The conditional vector indicates the areas requiring modification, generating an attention heatmap A(x,y). The network performs style transfer or redrawing only in areas where A(x,y) > 0.5. The inputs are (I0,C,G(I0,C)) and the ground truth image (I0,C,I_target). The discriminator not only distinguishes authenticity from artifacts but also verifies semantic consistency. When the conditional type is "Color / Color Matching," the style colorization network ColorNet is activated. This processing includes: overall hue transfer, using AdaIN or ColorHistogram Matching to transfer the hue of the target style image to the original image; and localized color recoloring, using the semantic mask provided by the conditional vector C_area to perform Lab color space adjustments on the local area, preserving the luminance channel and replacing the chrominance.

[0111] After introducing the method of the exemplary embodiment of the present invention, next, reference is made to Figure 2 An artificial intelligence-based clothing design assistance system according to an exemplary embodiment of the present invention is described. The system includes:

[0112] The feature extraction module is used to parse the design requirements input by the user through a natural language processing model. It uses a convolutional neural network to extract multi-dimensional style feature information from a pre-built clothing image dataset and fashion trend database, and constructs a style vector space model.

[0113] The matching generation module is used to perform vector similarity matching between the design vector and the style vector space model, and generate clothing design sketches using a graph neural network;

[0114] The auxiliary generation module is used to call the reinforcement learning algorithm to automatically deduce the garment pattern structure, sewing sequence, material specifications and processing route based on the design sketch, and generate a structured process data package for proofing;

[0115] The fitting simulation module is used to input design sketches and structural data into the 3D human body modeling and simulation module. Based on human motion capture and physical fabric simulation technology, it can simulate the wearing effect of virtual clothing.

[0116] The interactive correction module is used to build a human-computer interaction interface, allowing users to make modification suggestions for local styles, colors or structures through voice, and quickly respond to modification requests based on the generative adversarial network model.

[0117] After introducing the method and system of the exemplary embodiment of the present invention, the following Figure 3 For a description of a computer-readable storage medium according to an exemplary embodiment of the present invention, please refer to Figure 3 , the computer-readable storage medium shown is a CD 30, on which a computer program (i.e., a program product) is stored. When the computer program is executed by the processor, it will implement the various steps described in the above method implementation method, for example, parsing the design requirements input by the user through a natural language processing model, extracting multi-dimensional style feature information from a pre-constructed clothing image data set and fashion trend database using a convolutional neural network, and constructing a style vector space model; performing vector similarity matching between the design vector and the style vector space model, and generating a clothing design sketch using a graph neural network; based on the design sketch, calling a reinforcement learning algorithm to automatically deduce the clothing pattern structure, sewing sequence, material specifications and processing route, and generating a structured process data packet for proofing; inputting the design sketch and structure data into a three-dimensional human body modeling and simulation module, and realizing the wearing effect simulation of virtual clothing based on human motion capture and physical fabric simulation technology; constructing a human-computer interaction interface to support users to make modification suggestions for local styles, colors or structures through voice, and quickly responding to modification requests based on a generative adversarial network model; the specific implementation methods of each step are not repeated here.

[0118] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0119] After introducing the method, apparatus and medium of the exemplary embodiment of the present invention, the following is a reference to Figure 4 A computing device for assisting clothing design based on artificial intelligence according to an exemplary embodiment of the present invention.

[0120] Figure 4 A block diagram is shown of an exemplary computing device 40 , which may be a computer system or server, suitable for implementing embodiments of the present invention. Figure 4 The computing device 40 shown is only an example and should not limit the functionality and scope of use of the embodiments of the present invention.

[0121] like Figure 4 As shown, the components of computing device 40 may include, but are not limited to, one or more processors or processing units 401 , a system memory 402 , and a bus 403 connecting various system components (including system memory 402 and processing unit 401 ).

[0122] The computing device 40 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computing device 40, including volatile and non-volatile media, removable and non-removable media.

[0123] System memory 402 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 4021 and / or cache memory 4022. Computing device 40 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, ROM 4023 may be used to read and write non-removable, non-volatile magnetic media ( Figure 4 is not shown in the , usually referred to as "hard drive"). Although not in Figure 4As shown in FIG403 , a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) can be provided. In these cases, each drive can be connected to bus 403 via one or more data media interfaces. System memory 402 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0124] A program / utility 4025 having a set (at least one) of program modules 4024 may be stored, for example, in system memory 402. Such program modules 4024 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. Program modules 4024 generally implement the functions and / or methods of the embodiments described herein.

[0125] The computing device 40 may also communicate with one or more external devices 404 (e.g., a keyboard, a pointing device, a display, etc.). Such communication may be performed via an input / output (I / O) interface 405. Furthermore, the computing device 40 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network such as the Internet) via a network adapter 406. Figure 4 As shown, the network adapter 406 communicates with other modules (such as the processing unit 401, etc.) of the computing device 40 via the bus 403. Figure 4 Not shown, other hardware and / or software modules may be used in conjunction with computing device 40 .

[0126] The processing unit 401 executes various functional applications and data processing by running the programs stored in the system memory 402. For example, the design requirements input by the user are parsed through a natural language processing model, and multi-dimensional style feature information is extracted from a pre-built clothing image data set and fashion trend database using a convolutional neural network, and a style vector space model is constructed; the design vector is matched with the style vector space model for vector similarity, and a clothing design sketch is generated using a graph neural network; based on the design sketch, a reinforcement learning algorithm is called to automatically deduce the clothing pattern structure, sewing sequence, material specifications and processing route, and a structured process data packet is generated for proofing; the design sketch and structure data are input into the three-dimensional human body modeling and simulation module, and based on human motion capture and physical fabric simulation technology, the wearing effect simulation of virtual clothing is realized; a human-computer interaction interface is constructed to support users to make modification suggestions on local style, color or structure through voice, and quickly respond to modification requests based on a generative adversarial network model.

[0127] The specific implementation of each step will not be repeated here. It should be noted that although several units / modules or sub-units / sub-modules of the clothing design auxiliary device based on artificial intelligence are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more units / modules described above can be concretized in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules for concretization.

[0128] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0129] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0130] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0131] 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.

[0132] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0133] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

[0134] Furthermore, although the operations of the method of the present invention are described in a particular order in the accompanying drawings, this does not require or imply that these operations must be performed in this particular order, or that all illustrated operations must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

Claims

1. A clothing design assistance method based on artificial intelligence, characterized in that: include: The natural language processing model is used to parse the design requirements input by the user. A convolutional neural network is used to extract multi-dimensional style feature information from a pre-built clothing image dataset and fashion trend database, and a style vector space model is constructed. Perform vector similarity matching between the design vector and the style vector space model, and use graph neural network to generate clothing design sketches; Based on the design sketch, the reinforcement learning algorithm is used to automatically deduce the garment pattern structure, sewing sequence, material specifications and processing route, and generate a structured process data package for proofing; Input the design sketch and structure data into the 3D human body modeling and simulation module, and simulate the wearing effect of virtual clothing based on human motion capture and physical cloth simulation technology; Build a human-computer interaction interface to support users to make modification suggestions for local styles, colors or structures through voice, and quickly respond to modification requests based on the generative adversarial network model.

2. The artificial intelligence-based clothing design assistance method according to claim 1, characterized in that: The parsing of the design requirements input by the user through the natural language processing model includes: The system collects user demand information about clothing design through a graphical interface, voice recognition module or text input box, uses a pre-trained language model to encode text semantics, uses a deep learning model for intent classification and slot filling, and converts design parameters into a unified design vector representation.

3. The artificial intelligence-based clothing design assistance method according to claim 1, characterized in that: The method uses a convolutional neural network to extract multi-dimensional style feature information from a pre-built clothing image dataset and fashion trend database, and constructs a style vector space model, including: The image's style vectors form a high-dimensional style space. Principal component analysis is used for visualization and dimensionality reduction. A clustering algorithm is used to divide the style space into several style clusters. Metric learning methods are used to enhance the distinguishability between vectors. An indexing mechanism is then constructed to enable subsequent matching. The constructed style vector space is connected to the user intention vector to support subsequent semantic query and design generation.

4. The artificial intelligence-based clothing design assistance method according to claim 1, characterized in that: The method of performing vector similarity matching between the design vector and the style vector space model and using a graph neural network to generate a clothing design sketch includes: For the selected k image samples, a graph structure G = (V, E) is constructed, where the node set V = {v T ,h i1 ,...,h ik }, the edge set E constructs edge weights based on the similarity between style vectors; Use graph attention network or graph convolution network to aggregate the information in the graph into a generated vector z T : z T =GNN(G) Among them, GNN (G) is the forward propagation of the graph neural network, z T It is the global feature that is ultimately used to generate the sketch; Based on z T Input, use the generative adversarial network to generate sketch images, and add sketch structure guidance in the generation process.

5. The artificial intelligence-based clothing design assisting method according to claim 1, wherein: Based on the design sketch, the reinforcement learning algorithm is called to automatically deduce the garment pattern structure, sewing sequence, material specifications and processing route, and generate a structured process data package for proofing, including: Use the image structure analysis network to perform image semantic segmentation on the sketch, extract the clothing feature area, and use the policy gradient method in reinforcement learning to optimize the design strategy: Among them, θ is the policy network parameter, γ is the reward discount factor, r t A score based on the comparison between the generated solution and the target; Design reward function: r t ＝α1r 结构相符 +α2r 缝制合理 +α3r 材料利用 +α4r 工艺可行性 Among them, α i is the weight coefficient of the corresponding sub-reward, r 结构相符 is the structural consistency score, r 缝制合理 is the sewing rationality score, r 材料利用 is the material utilization score, r 工艺可行性 Score the process feasibility.

6. The artificial intelligence-based clothing design assistance method according to claim 1, characterized in that: The design sketch and structure data are input into the 3D human body modeling and simulation module, and the wearing effect simulation of the virtual clothing is realized based on human motion capture and physical cloth simulation technology, including: Use 3D human body modeling tools to build an interactive high-precision human body mesh model H(x,y,z). Use human motion capture equipment to collect user static and dynamic posture data, and drive the 3D human body model to perform posture animation reconstruction: H t =f pose (H0,M t ) Among them, H0 is the initial human body model, M t is the posture data of time step t, H t is the 3D human body model after the action, f pose is the posture transformation function; The 2D pattern structure data is converted into a corresponding 3D mesh of patches, where each patch represents a pattern unit. Sewing operations are performed on each mesh boundary point according to the sewing sequence defined in the structure data to construct an initial 3D garment model. Based on the mechanical properties of the fabric, physical parameters are assigned to each garment mesh, initializing the physics solver in the fabric simulation engine. The clothing model is bound to the 3D human body model, and the dynamic deformation of the clothing as the human body moves is simulated through collision detection and constraint solving algorithms; the simulation results are rendered in real time as a 3D animation try-on effect, and the user can choose different body shapes and postures to observe the wearing fit and wrinkle changes of the clothing.

7. The artificial intelligence-based clothing design assisting method according to claim 1, wherein: The human-computer interaction interface is constructed to support users to make modification suggestions for local styles, colors, or structures through voice, and quickly respond to modification requests based on the generative adversarial network model, including: A conditional generative adversarial network is constructed for clothing design. The input is the current design sketch and a semantic condition vector. The condition vector is encoded as a latent vector by the semantic parsing result. The original design drawing and the semantic modification vector are input into the generator. The generator performs target style transfer or image redrawing on local areas of the image according to different types of modification instructions. For color matching modifications, the style coloring module is used to adjust the overall or local area color matrix. The generated clothing sketch is rendered back to the human-computer interface in real time. The user can zoom in, rotate or overlay historical comparison images on the interface to confirm whether the modification is as expected. If the user is not satisfied, he or she can initiate a voice command again to enter a new round of modification cycle.

8. An artificial intelligence-based clothing design assistance system, characterized in that: include: The feature extraction module is used to parse the design requirements input by the user through a natural language processing model. It uses a convolutional neural network to extract multi-dimensional style feature information from a pre-built clothing image dataset and fashion trend database, and constructs a style vector space model. The matching generation module is used to perform vector similarity matching between the design vector and the style vector space model, and generate clothing design sketches using a graph neural network; The auxiliary generation module is used to call the reinforcement learning algorithm to automatically deduce the garment pattern structure, sewing sequence, material specifications and processing route based on the design sketch, and generate a structured process data package for proofing; The fitting simulation module is used to input design sketches and structural data into the 3D human body modeling and simulation module. Based on human motion capture and physical fabric simulation technology, it can simulate the wearing effect of virtual clothing. The interactive correction module is used to build a human-computer interaction interface, allowing users to make modification suggestions for local styles, colors or structures through voice, and quickly respond to modification requests based on the generative adversarial network model.

9. A computing device, comprising: at least one processor, memory, and input-output unit; The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the steps of the artificial intelligence-based clothing design assistance method as described in any one of claims 1 to 7.

10. A computer-readable storage medium comprising instructions, which, when executed on a computer, enables the computer to execute the steps of the artificial intelligence-based clothing design assistance method according to any one of claims 1 to 7.

Images