An AI-based adaptive clothing design animation modeling method and system

Abstract

The embodiment of the present invention relates to the technical field of clothing design, and specifically discloses an AI-based adaptive clothing design animation modeling method and system. The embodiment of the present invention performs type matching, preferential screening and arrangement from multiple publishing channels; performs style selection and iterative design processing; customizes and optimizes the basic design style in terms of material, texture and color; selects scene model models based on AI technology to match clothing; performs dynamic display simulation, eliminates model models, and generates and displays clothing flexible simulation videos. It is capable of publishing collection, performing style selection and iterative design processing on design collection data, and performing customized optimization, and then based on AI technology, performs model model selection and clothing matching, performs dynamic display simulation and model model elimination, generates and displays clothing flexible simulation videos, and realizes clothing design, matching and display that is more in line with the scene, and can intuitively display the designed clothing.

Description

Technical Field

[0001] The present invention belongs to the technical field of clothing design, and in particular relates to an AI-based adaptive clothing design animation modeling method and system. Background Art

[0002] Fashion design is the sum of a series of creative activities such as conception, drawing, production, decoration, display and production guidance, which takes the human body as the object, uses certain thinking forms, aesthetic laws and design procedures, combines style, fabric, color, structure, craftsmanship and other means to meet people's material and spiritual needs. With the rapid development of the fashion industry and the continuous changes in consumer demand, fashion design also needs to be constantly innovated and updated to adapt to market changes.

[0003] In the existing technology, AI-based clothing design can only perform simple style design and rendering, cannot perform iterative design in combination with existing released styles, and cannot design, match and display according to the required scenarios, and cannot intuitively display the designed clothing. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide an AI-based adaptive clothing design animation modeling method and system, aiming to solve the problems raised in the background technology.

[0005] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0006] An AI-based adaptive clothing design animation modeling method, the method specifically comprises the following steps:

[0007] Obtaining the required design type, and performing type matching, optimal screening, and arrangement and sorting from multiple preset release channels according to the required design type, to generate design collection data;

[0008] Receiving design requirement information, and performing style selection and iterative design processing on the design collection data according to the design requirement information to generate a basic design style;

[0009] According to the design requirement information, the basic design style is customized and optimized in terms of material, texture and color to generate an optimized design style;

[0010] Determine the required wearing scene, select a scene model model based on AI technology, match the optimized design style with clothing based on the scene model model, and generate a scene wearing model;

[0011] Automatically perform dynamic rendering processing, perform dynamic display simulation on the scene wearing model, eliminate the model model, generate and display the clothing flexibility simulation video.

[0012] As a further limitation of the technical solution of the embodiment of the present invention, the obtaining of the required design type, performing type matching, preferential screening and arrangement from a plurality of preset release channels according to the required design type, and generating the design collection data specifically comprises the following steps:

[0013] Get the required design type;

[0014] According to the demand design type, type matching is performed from multiple preset publishing channels to obtain matching collection data;

[0015] Extracting evaluation feedback data from the matching collected data;

[0016] According to the evaluation feedback data, the matching collected data is preferentially screened to obtain preferential collected data;

[0017] The preferentially collected data are arranged and sorted to generate design collected data.

[0018] As a further limitation of the technical solution of the embodiment of the present invention, the receiving of design requirement information, performing style selection and iterative design processing on the design collection data according to the design requirement information, and generating a basic design style specifically include the following steps:

[0019] Receive design requirement information;

[0020] Extracting style requirement information from the design requirement information;

[0021] According to the style requirement information, selecting a style for the design collection data to obtain a plurality of selected style data;

[0022] Based on the style requirement information and the plurality of evaluation feedback data, iterative design processing is performed on the plurality of selected style data to generate a basic design pattern.

[0023] As a further limitation of the technical solution of the embodiment of the present invention, the customizing and optimizing the material, texture and color of the basic design style according to the design requirement information to generate the optimized design style specifically includes the following steps:

[0024] extracting detailed requirement information from the design requirement information;

[0025] According to the detailed requirement information, customizing and selecting materials for the basic design style to obtain material selection data;

[0026] According to the detailed requirement information, a texture is customized and selected for the basic design style to obtain texture selection data;

[0027] According to the detailed requirement information, customizing and selecting the color of the basic design style to obtain color selection data;

[0028] The basic design style is customized and optimized according to the material selection data, the texture selection data and the color selection data to generate an optimized design style.

[0029] As a further limitation of the technical solution of the embodiment of the present invention, the determining of the required wearing scene, selecting a scene model model based on AI technology, matching clothing for the optimized design style based on the scene model model, and generating a scene wearing model specifically include the following steps:

[0030] Identify the design requirement information and determine the required wearing scenario;

[0031] Based on AI technology, according to the required wearing scene, a scene model is selected, and model model data of the scene model model is obtained;

[0032] According to the required wearing scene, the model data and the optimized design style, clothing matching selection is performed to obtain matching clothing data;

[0033] Based on the optimized design style and the matching clothing data, clothing matching is performed on the scene model model to generate a scene wearing model.

[0034] As a further limitation of the technical solution of the embodiment of the present invention, the automatic dynamic rendering processing, the dynamic display simulation of the scene wearing model, the model model elimination, the generation and display of the clothing flexibility simulation video specifically include the following steps:

[0035] According to the requirements, the dressing scene is planned and the scene display actions are planned;

[0036] According to the scene display action, dynamically render the model wearing the scene to generate a dynamic video of the model;

[0037] Processing the model dynamic video frame by frame to generate multiple frame-by-frame display pictures;

[0038] Eliminate the model from multiple frame-by-frame display images to generate multiple clothing display images;

[0039] Performing video processing on the plurality of clothing display pictures to generate clothing flexibility simulation videos;

[0040] The clothing flexibility simulation video is displayed.

[0041] An AI-based adaptive clothing design animation modeling system, the system includes a data acquisition and processing unit, a style iteration design unit, a detail customization optimization unit, a model clothing matching unit and a clothing flexible simulation unit, wherein:

[0042] The data collection and processing unit is used to obtain the required design type, and perform type matching, preferential screening and arrangement from a plurality of preset release channels according to the required design type to generate design collection data;

[0043] A style iteration design unit, used for receiving design requirement information, and performing style selection and iterative design processing on the design collection data according to the design requirement information to generate a basic design style;

[0044] A detail customization and optimization unit, used to customize and optimize the material, texture and color of the basic design style according to the design requirement information, and generate an optimized design style;

[0045] A model clothing matching unit is used to determine the required wearing scene, select a scene model model based on AI technology, match the optimized design style with clothing based on the scene model model, and generate a scene wearing model;

[0046] The clothing flexibility simulation unit is used to automatically perform dynamic rendering processing, dynamically display and simulate the scene wearing model, eliminate the model model, and generate and display the clothing flexibility simulation video.

[0047] As a further limitation of the technical solution of the embodiment of the present invention, the data acquisition and processing unit specifically includes:

[0048] Type acquisition module, used to obtain the required design type;

[0049] A type matching module is used to design types according to the requirements, perform type matching from multiple preset publishing channels, and obtain matching collection data;

[0050] An evaluation extraction module, used to extract evaluation feedback data from the matching collected data;

[0051] A selection and screening module is used to select and screen the matching collected data based on the evaluation feedback data to obtain the selected collected data;

[0052] The arrangement and sorting module is used to arrange and sort the preferentially collected data to generate design collected data.

[0053] As a further limitation of the technical solution of the embodiment of the present invention, the style iteration design unit specifically includes:

[0054] A requirement receiving module, used for receiving design requirement information;

[0055] An information extraction module, used to extract style requirement information from the design requirement information;

[0056] A style selection module, used to select a style for the design collection data according to the style requirement information, and obtain a plurality of selected style data;

[0057] The iterative design module is used to perform iterative design processing on the multiple selected style data based on the style requirement information and the multiple evaluation feedback data to generate a basic design style.

[0058] As a further limitation of the technical solution of the embodiment of the present invention, the clothing flexibility simulation unit specifically includes:

[0059] An action planning module is used to plan the scene display action according to the wearing scene required;

[0060] A dynamic rendering module, used to dynamically render the model wearing the scene according to the scene display action, and generate a dynamic video of the model;

[0061] A frame-by-frame processing module, used to process the model dynamic video frame by frame to generate multiple frame-by-frame display pictures;

[0062] A model elimination module is used to eliminate the model in multiple frame-by-frame display pictures to generate multiple clothing display pictures;

[0063] A video processing module, used for performing video processing on the plurality of clothing display pictures to generate clothing flexibility simulation videos;

[0064] The simulation video display module is used to display the clothing flexibility simulation video.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] The embodiment of the present invention performs type matching, preferential screening and arrangement from multiple publishing channels; performs style selection and iterative design processing; customizes and optimizes the basic design style in terms of material, texture and color; selects scene model models based on AI technology to match clothing; performs dynamic display simulation, eliminates model models, and generates and displays clothing flexible simulation videos. It is capable of publishing collection, performing style selection and iterative design processing on the design collection data, and performing customized optimization, and then based on AI technology, performs model model selection and clothing matching, performs dynamic display simulation and model model elimination, generates and displays clothing flexible simulation videos, realizes clothing design, matching and display that is more in line with the scene, and can intuitively display the designed clothing. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0068] Figure 1 A flow chart of a method provided by an embodiment of the present invention is shown.

[0069] Figure 2 A flow chart of type matching, optimal screening and arrangement in the method provided by an embodiment of the present invention is shown.

[0070] Figure 3 A flow chart of style selection and iterative design in the method provided by an embodiment of the present invention is shown.

[0071] Figure 4 A flow chart of customized optimization of materials, textures and colors in the method provided by an embodiment of the present invention is shown.

[0072] Figure 5 A flow chart of generating a scene clothing model in the method provided in an embodiment of the present invention is shown.

[0073] Figure 6 A flow chart of generating and displaying a clothing flexibility simulation video in a method provided in an embodiment of the present invention is shown.

[0074] Figure 7 The application architecture diagram of the system provided by the embodiment of the present invention is shown.

[0075] Figure 8 The structure block diagram of the data acquisition and processing unit in the system provided by the embodiment of the present invention is shown.

[0076] Fig. 9 The structure block diagram of the style iteration design unit in the system provided by the embodiment of the present invention is shown.

[0077] Fig.10 The structure block diagram of the clothing flexibility simulation unit in the system provided by the embodiment of the present invention is shown. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0079] It is understandable that in the existing technology, AI-based clothing design can only perform simple style design and rendering, cannot perform iterative design in combination with existing released styles, and cannot design, match and display according to the demand scenarios, and cannot intuitively display the designed clothing.

[0080] To solve the above problems, the embodiment of the present invention obtains the required design type, and according to the required design type, performs type matching, preferential screening and arrangement from multiple preset release channels to generate design collection data; receives design requirement information, and according to the design requirement information, performs style selection and iterative design processing on the design collection data to generate a basic design style; according to the design requirement information, performs customized optimization of the basic design style in terms of material, texture and color to generate an optimized design style; determines the required wearing scene, selects the scene model model based on AI technology, and performs clothing matching on the optimized design style based on the scene model model to generate a scene wearing model; automatically performs dynamic rendering processing, performs dynamic display simulation on the scene wearing model, and performs model model elimination to generate and display clothing flexible simulation video. It is possible to perform release collection, perform style selection and iterative design processing on the design collection data, and perform customized optimization, and then based on AI technology, perform model model selection and clothing matching, perform dynamic display simulation and model model elimination, generate and display clothing flexible simulation video, realize clothing design, matching and display that is more suitable for the scene, and can intuitively display the designed clothing.

[0081] Figure 1 A flow chart of a method provided by an embodiment of the present invention is shown.

[0082] Specifically, an AI-based adaptive clothing design animation modeling method comprises the following steps:

[0083] Step S101, obtaining the required design type, and performing type matching, preferential screening and arrangement from a plurality of preset release channels according to the required design type to generate design collection data.

[0084] In an embodiment of the present invention, interaction is performed with a designer to obtain a demand design type selected by the designer, and type matching is performed from a plurality of preset publishing channels according to the demand design type to obtain matching collection data, and then evaluation feedback data is extracted from the matching collection data, and the evaluation feedback data is analyzed and multiple clothing products having a ratio greater than a ratio threshold are selected according to the positive review ratio, and corresponding data is preferentially selected from the matching collection data to obtain preferential collection data, and then the data of different clothing products in the preferential collection data are arranged and sorted according to the positive review ratios corresponding to the multiple clothing products to generate design collection data.

[0085] It is understandable that multiple publishing channels can be sales channels for clothing, such as Taobao, JD.com, Vipshop, etc.; matching collected data includes sales display data, sales record data and sales evaluation data of the goods.

[0086] Specifically, Figure 2 A flow chart of type matching, optimal screening and arrangement in the method provided by an embodiment of the present invention is shown.

[0087] Among them, in the preferred implementation manner provided by the present invention, the acquisition of the required design type, according to the required design type, from a plurality of preset release channels, type matching, preferential screening and arrangement are performed to generate the design acquisition data, specifically including the following steps:

[0088] Step S1011, obtaining the required design type;

[0089] Step S1012, performing type matching from a plurality of preset publishing channels according to the required design type, and obtaining matching collection data;

[0090] Step S1013, extracting evaluation feedback data from the matching collected data;

[0091] Step S1014, based on the evaluation feedback data, the matching collected data is preferentially screened to obtain preferential collected data;

[0092] Step S1015, arranging and sorting the preferentially collected data to generate design collected data.

[0093] Among them, in the preferred implementation manner provided by the present invention, the extracting the evaluation feedback data from the matching collected data specifically comprises the following steps:

[0094] Acquire matching collected data, and filter the matching collected data to obtain complete matching data; wherein the complete matching data includes a rating value, feedback content, and a timestamp;

[0095] Extract keywords from the feedback content to obtain keywords;

[0096] Given a keyword list, calculate the matching degree between the keyword and the given keyword list to obtain the keyword matching score;

[0097] Perform exponential decay calculation based on the interval between the timestamp and the current time to get the freshness score;

[0098] The score value, keyword matching score and freshness score are weighted and summed to obtain a comprehensive score;

[0099] The complete matching data are sorted in descending order according to the comprehensive scores, and the first several are selected from the sorting as evaluation feedback data.

[0100] In the embodiment of the present invention, natural language keyword extraction is combined with quantitative analysis. The time impact adopts an exponential decay model to improve the sensitivity to recent data. Comprehensive multi-dimensional scoring is performed to conduct a comprehensive analysis of the feedback data, improve the quality of the selected evaluation feedback data, and provide clear and reliable data support for subsequent processes.

[0101] Furthermore, the AI-based adaptive clothing design animation modeling method further includes the following steps:

[0102] Step S102, receiving design requirement information, and performing style selection and iterative design processing on the design collection data according to the design requirement information to generate a basic design style.

[0103] In an embodiment of the present invention, interaction is performed with a designer to receive design requirement information input by the designer, and style requirement information is extracted from the design requirement information by performing content recognition on the design requirement information. Then, based on the style requirement information, style selection is performed on the design collection data to obtain multiple selected style data that match the style requirement information. Then, based on the style requirement information and multiple evaluation feedback data, iterative design processing is performed on the multiple selected style data to generate a basic design style.

[0104] It can be understood that the iterative design process is a process of gradually designing different style parts, such as buttons, zippers, and pockets, according to the style parts with the highest to lowest praise ratio.

[0105] Specifically, Figure 3 A flow chart of style selection and iterative design in the method provided by an embodiment of the present invention is shown.

[0106] Among them, in the preferred embodiment provided by the present invention, the receiving of design requirement information, performing style selection and iterative design processing on the design collection data according to the design requirement information, and generating a basic design style specifically include the following steps:

[0107] Step S1021, receiving design requirement information;

[0108] Step S1022, extracting style requirement information from the design requirement information;

[0109] Step S1023, selecting a style for the design collection data according to the style requirement information, and obtaining a plurality of selected style data;

[0110] Step S1024, based on the style requirement information and the plurality of evaluation feedback data, iterative design processing is performed on the plurality of selected style data to generate a basic design pattern.

[0111] Among them, in the preferred embodiment provided by the present invention, the iterative design processing of the multiple selected style data based on the style requirement information and the multiple evaluation feedback data to generate the basic design style specifically includes the following steps:

[0112] The selected style data includes style picture information and style description information;

[0113] Perform image recognition on style image information to obtain key design features;

[0114] Perform semantic analysis on style description information to obtain style keywords;

[0115] The key design features and style keywords are combined into style feature vectors, and clustering operations are performed on the style feature vectors using clustering algorithms to obtain multiple style category results;

[0116] The sentiment analysis model is used to extract the sentiment inclination strength of the user's keywords for a certain style in the evaluation feedback data to obtain the keyword sentiment score;

[0117] According to the time interval from the time point of the evaluation feedback data to the current time point, the keyword sentiment score is adjusted to obtain the user feedback score;

[0118] Generate a user demand vector based on the style requirement information, and calculate the matching score between the user demand vector and the style feature vector;

[0119] Obtain the user's negative keyword sentiment intensity for a certain style, and extract the corresponding feedback keywords based on the negative keyword sentiment intensity;

[0120] Search whether the style feature vector contains feedback keywords. If so, it is considered as an improvement target.

[0121] All feedback keywords contained in the style feature vector are combined with the corresponding negative keyword sentiment intensity to obtain an improvement potential score;

[0122] Taking into account the matching score, user feedback score and improvement potential score, each style is quantitatively scored to obtain a comprehensive score of the style;

[0123] Using style category results as the environment, style comprehensive scores as the state, selection or elimination of a style as the action, and the distribution of the filtered style scores close to user needs as the reward, a reinforcement learning algorithm is used for multiple rounds of screening;

[0124] After the screening is completed, the style categories with the top comprehensive scores are output as the basic design styles.

[0125] In the embodiment of the present invention, a comprehensive evaluation is performed from multiple dimensions, wherein the matching score ensures that the style meets the current needs of the user. The user feedback score provides a satisfaction reference based on actual evaluation. The improvement potential determines the space for style optimization to avoid wasting resources on non-optimizable designs. Finally, multiple rounds of screening and adjustment are performed through the reinforcement learning algorithm, gradually converging to the optimal style while retaining a certain amount of innovation space, which can provide a clear direction for subsequent optimization and design innovation while meeting the current needs of users.

[0126] Furthermore, the AI-based adaptive clothing design animation modeling method further includes the following steps:

[0127] Step S103, according to the design requirement information, the basic design style is customized and optimized in terms of material, texture and color to generate an optimized design style.

[0128] In an embodiment of the present invention, by performing content recognition on the design requirement information, detail requirement information is extracted from the design requirement information, and then according to the detail requirement information, the material is customized and selected for the basic design style to obtain material selection data, and according to the detail requirement information, the texture is customized and selected for the basic design style to obtain texture selection data, and according to the detail requirement information, the color is customized and selected for the basic design style to obtain color selection data, thereby achieving customized selection of material, texture and color, and then according to the material selection data, texture selection data and color selection data, the basic design style is customized and optimized to generate an optimized design style.

[0129] Specifically, Figure 4 A flow chart of customized optimization of materials, textures and colors in the method provided by an embodiment of the present invention is shown.

[0130] Among them, in the preferred embodiment provided by the present invention, the customization and optimization of the material, texture and color of the basic design style according to the design requirement information to generate the optimized design style specifically includes the following steps:

[0131] Step S1031, extracting detailed requirement information from the design requirement information;

[0132] Step S1032, according to the detailed requirement information, customize and select the material for the basic design style, and obtain material selection data;

[0133] Step S1033, according to the detailed requirement information, customize and select the texture of the basic design style to obtain texture selection data;

[0134] Step S1034, according to the detailed requirement information, customize the color of the basic design style and obtain color selection data;

[0135] Step S1035 , customizing and optimizing the basic design pattern according to the material selection data, the texture selection data, and the color selection data to generate an optimized design pattern.

[0136] Among them, in the preferred embodiment provided by the present invention, the basic design style is customized and optimized according to the material selection data, the texture selection data and the color selection data, and the generation of the optimized design style specifically includes the following steps:

[0137] Calculate the matching degree between material selection data and user requirements to obtain material scores;

[0138] Calculate the matching degree between texture selection data and user requirements to obtain texture score;

[0139] Calculate the difference between the color selection data and the color required by the user to obtain the color score;

[0140] A multi-objective optimization function is constructed with the optimization objectives of maximizing the sum of material score, texture score and color score and minimizing the total cost of material, texture and color;

[0141] By performing multi-objective optimization on the multi-objective optimization function, an optimized design style is obtained.

[0142] In the embodiment of the present invention, judgment is made through multiple dimensions, not only focusing on a single factor, but also comprehensively considering the full range of user needs, thereby ensuring the integrity of the design style. In addition, the matching score of each dimension is based on a specific similarity calculation, avoiding the influence of subjective judgment. By maximizing the total score and minimizing the total cost, the selected optimized design style can meet customer needs while ensuring the lowest cost.

[0143] Furthermore, the AI-based adaptive clothing design animation modeling method further includes the following steps:

[0144] Step S104, determining the required wearing scene, selecting a scene model model based on AI technology, matching clothing for the optimized design style based on the scene model model, and generating a scene wearing model.

[0145] In an embodiment of the present invention, by performing content recognition on the design requirement information, the required wearing scene is determined from the design requirement information, and then based on the AI ​​technology, a scene model model is selected according to the required wearing scene, and the model model data of the scene model model is obtained. According to the required wearing scene, the model model data and the optimized design style, clothing matching selection is performed, and matching clothing data that is coordinated with the required wearing scene, the model model data and the optimized design style is obtained. Then, based on the optimized design style and the matching clothing data, clothing matching is performed on the scene model model to generate a scene wearing model.

[0146] Specifically, Figure 5 A flow chart of generating a scene clothing model in the method provided in an embodiment of the present invention is shown.

[0147] Among them, in the preferred embodiment provided by the present invention, the determining of the required wearing scene, selecting a scene model model based on AI technology, matching clothing for the optimized design style based on the scene model model, and generating a scene wearing model specifically include the following steps:

[0148] Step S1041, identifying the design requirement information and determining the required wearing scene;

[0149] Step S1042, based on AI technology, select a scene model model according to the required wearing scene, and obtain model model data of the scene model model;

[0150] Step S1043, selecting clothing combinations according to the required wearing scene, the model data and the optimized design style, and obtaining matching clothing data;

[0151] Step S1044, based on the optimized design style and the matching clothing data, clothing matching is performed on the scene model model to generate a scene wearing model.

[0152] Furthermore, the AI-based adaptive clothing design animation modeling method further includes the following steps:

[0153] Step S105, automatically performing dynamic rendering processing, performing dynamic display simulation on the scene wearing model, eliminating the model model, generating and displaying the clothing flexibility simulation video.

[0154] In an embodiment of the present invention, according to the required wearing scene, the scene display action is planned, and then according to the scene display action, the scene wearing model is dynamically rendered to generate a model dynamic video, and the model dynamic video is processed frame by frame to generate a plurality of frame by frame display pictures, and the scene model model is identified in the plurality of frame by frame display pictures, and the scene model model identified in the plurality of frame by frame display pictures is eliminated to generate a plurality of clothing display pictures, and the plurality of clothing display pictures are arranged in an order of their frame rates, and the plurality of clothing display pictures are video processed to generate a clothing flexible simulation video, and the clothing flexible simulation video is displayed when there is a display demand.

[0155] Specifically, Figure 6 A flow chart of generating and displaying a clothing flexibility simulation video in a method provided in an embodiment of the present invention is shown.

[0156] Among them, in the preferred embodiment provided by the present invention, the automatic dynamic rendering processing, the dynamic display simulation of the scene wearing model, the model model elimination, the generation and display of the clothing flexibility simulation video specifically includes the following steps:

[0157] Step S1051, planning scene display actions according to the required wearing scene;

[0158] Step S1052, according to the scene display action, dynamically render the scene wearing model to generate a model dynamic video;

[0159] Step S1053, processing the model dynamic video frame by frame to generate a plurality of frame-by-frame display pictures;

[0160] Step S1054, performing model elimination in a plurality of frame-by-frame display pictures to generate a plurality of clothing display pictures;

[0161] Step S1055, performing video processing on the plurality of clothing display pictures to generate clothing flexibility simulation videos;

[0162] Step S1056, displaying the clothing flexibility simulation video.

[0163] Among them, in the preferred embodiment provided by the present invention, the step of eliminating the model in the multiple frame-by-frame display pictures to generate multiple clothing display pictures specifically includes the following steps:

[0164] The spatiotemporal Transformer model is used to globally model the frame-by-frame display images and generate dynamic features of models and clothing.

[0165] Use a multi-layer segmentation network to perform semantic segmentation on the displayed images frame by frame to generate model areas and clothing areas;

[0166] Apply a fine-grained feature extraction network to map the material, color, and texture information of clothing into a high-dimensional semantic space to extract fine-grained features;

[0167] Based on the dynamic features of the model, the overlapping area of ​​the model's action on the clothing boundary is inferred, and according to the fine-grained features, the appearance of the occluded part is predicted and compensated, so as to realize dynamic compensation of the overlapping area and obtain the compensated clothing area;

[0168] Extract the background features of the occluded area of ​​the model from the compensated image to obtain the clothing area including the blank area;

[0169] Based on the fine-grained features, a generative adversarial network is used to infer the appearance of the blank area in the clothing area containing the blank area to generate a complete clothing area. The generated background texture is optimized for temporal and spatial consistency to ensure the continuity of the background frame by frame, and a complete clothing area is obtained.

[0170] Apply a local generative adversarial network to enhance the details of the complete clothing area, ensuring that the texture and shape of the clothing remain intact after the model is removed, and obtain an enhanced clothing area;

[0171] According to the dynamic characteristics of clothing, the spatiotemporal Transformer is used to globally model the enhanced clothing area frame by frame to align the fine-grained features between adjacent frame images and obtain multiple clothing display pictures.

[0172] In the embodiment of the present invention, the dynamic consistency of clothing shape, material, texture, light and shadow is ensured by combining spatiotemporal features with multimodal information, avoiding the visual incoherence problem caused by the removal of the model. After the model is removed, the contour of the clothing is restored by dynamic compensation to ensure the integrity of the clothing, solving the distortion problem caused by the overlap of the model and clothing boundaries in traditional methods.

[0173] Furthermore, Figure 7 The application architecture diagram of the system provided by the embodiment of the present invention is shown.

[0174] Among them, in another preferred embodiment provided by the present invention, an AI-based adaptive clothing design animation modeling system includes:

[0175] The data collection and processing unit 101 is used to obtain the required design type, and perform type matching, preferential screening and arrangement from a plurality of preset publishing channels according to the required design type to generate design collection data.

[0176] In an embodiment of the present invention, the data acquisition processing unit 101 interacts with the designer to obtain the demand design type selected by the designer, and performs type matching from multiple preset publishing channels according to the demand design type to obtain matching collection data, and then extracts evaluation feedback data from the matching collection data, analyzes the evaluation feedback data, selects multiple clothing products with a ratio greater than a ratio threshold according to the positive review ratio, selects corresponding data from the matching collection data, thereby obtaining preferred collection data, and then arranges and organizes the data of different clothing products in the preferred collection data according to the positive review ratios corresponding to the multiple clothing products to generate design collection data.

[0177] Specifically, Figure 8 It shows a structural block diagram of the data acquisition and processing unit 101 in the system provided by the embodiment of the present invention.

[0178] Among them, in the preferred embodiment provided by the present invention, the data acquisition and processing unit 101 specifically includes:

[0179] The type acquisition module 1011 is used to acquire the required design type;

[0180] The type matching module 1012 is used to design the type according to the requirement, perform type matching from a plurality of preset publishing channels, and obtain matching collection data;

[0181] Evaluation extraction module 1013, used to extract evaluation feedback data from the matching collected data;

[0182] The optimal screening module 1014 is used to perform optimal screening on the matching collected data according to the evaluation feedback data to obtain the optimal collected data;

[0183] The arrangement and sorting module 1015 is used to arrange and sort the preferentially collected data to generate design collected data.

[0184] Furthermore, the AI-based adaptive clothing design animation modeling system also includes:

[0185] The style iteration design unit 102 is used to receive design requirement information, and perform style selection and iterative design processing on the design collection data according to the design requirement information to generate a basic design style.

[0186] In an embodiment of the present invention, the style iteration design unit 102 interacts with the designer, receives the design requirement information input by the designer, extracts the style requirement information from the design requirement information by performing content recognition on the design requirement information, and then performs style selection on the design collection data based on the style requirement information to obtain multiple selected style data that match the style requirement information, and then performs iterative design processing on the multiple selected style data based on the style requirement information and multiple evaluation feedback data to generate a basic design style.

[0187] Specifically, Fig. 9 It shows a structural block diagram of the style iteration design unit 102 in the system provided by an embodiment of the present invention.

[0188] In the preferred embodiment of the present invention, the style iteration design unit 102 specifically includes:

[0189] The requirement receiving module 1021 is used to receive design requirement information;

[0190] An information extraction module 1022, used to extract style requirement information from the design requirement information;

[0191] The style selection module 1023 is used to select a style for the design collection data according to the style requirement information, and obtain a plurality of selected style data;

[0192] The iterative design module 1024 is used to perform iterative design processing on the multiple selected style data based on the style requirement information and the multiple evaluation feedback data to generate a basic design style.

[0193] Furthermore, the AI-based adaptive clothing design animation modeling system also includes:

[0194] The detail customization and optimization unit 103 is used to customize and optimize the material, texture and color of the basic design style according to the design requirement information to generate an optimized design style.

[0195] In an embodiment of the present invention, the detail customization and optimization unit 103 performs content recognition on the design requirement information, extracts the detail requirement information from the design requirement information, and then customizes and selects the material for the basic design style according to the detail requirement information to obtain material selection data, and customizes and selects the texture for the basic design style according to the detail requirement information to obtain texture selection data, and customizes and selects the color for the basic design style according to the detail requirement information to obtain color selection data, thereby achieving customized selection of material, texture and color, and then customizes and optimizes the basic design style according to the material selection data, texture selection data and color selection data to generate an optimized design style.

[0196] The model clothing matching unit 104 is used to determine the required wearing scene, select a scene model model based on AI technology, match the optimized design style with clothing based on the scene model model, and generate a scene wearing model.

[0197] In an embodiment of the present invention, the model clothing matching unit 104 determines the required wearing scene from the design requirement information by performing content recognition on the design requirement information, and then selects a scene model model based on the required wearing scene based on the AI ​​technology, and obtains the model model data of the scene model model. According to the required wearing scene, the model model data and the optimized design style, clothing matching selection is performed, and matching clothing data that is coordinated with the required wearing scene, the model model data and the optimized design style is obtained. Then, based on the optimized design style and the matching clothing data, clothing matching is performed on the scene model model to generate a scene wearing model.

[0198] The clothing flexibility simulation unit 105 is used to automatically perform dynamic rendering processing, perform dynamic display simulation on the scene wearing model, eliminate the model model, and generate and display the clothing flexibility simulation video.

[0199] In an embodiment of the present invention, the clothing flexibility simulation unit 105 plans the scene display action according to the required wearing scene, and then dynamically renders the scene wearing model according to the scene display action to generate a model dynamic video, and generates a plurality of frame-by-frame display pictures by processing the model dynamic video frame by frame, and identifies the scene model model in the plurality of frame-by-frame display pictures, and eliminates the scene model model identified in the plurality of frame-by-frame display pictures to generate a plurality of clothing display pictures, and arranges the plurality of clothing display pictures in an order of their frame rates, performs video processing on the plurality of clothing display pictures to generate a clothing flexibility simulation video, and displays the clothing flexibility simulation video when there is a display demand.

[0200] Specifically, Fig.10 It shows a structural block diagram of the clothing flexibility simulation unit 105 in the system provided by the embodiment of the present invention.

[0201] Among them, in the preferred embodiment provided by the present invention, the clothing flexibility simulation unit 105 specifically includes:

[0202] Action planning module 1051, used for planning scene display actions according to the required wearing scene;

[0203] A dynamic rendering module 1052 is used to dynamically render the scene wearing model according to the scene display action to generate a model dynamic video;

[0204] The frame-by-frame processing module 1053 is used to perform frame-by-frame processing on the model dynamic video to generate multiple frame-by-frame display pictures;

[0205] A model elimination module 1054 is used to eliminate the model of the model in the multiple frame-by-frame display pictures to generate multiple clothing display pictures;

[0206] The video processing module 1055 is used to perform video processing on the plurality of clothing display pictures to generate clothing flexibility simulation videos;

[0207] The simulation video display module 1056 is used to display the clothing flexibility simulation video.

[0208] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0209] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0210] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0211] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0212] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An AI-based adaptive clothing design animation modeling method, characterized in that: The method specifically comprises the following steps: Obtaining the required design type, and performing type matching, optimal screening, and arrangement and sorting from multiple preset release channels according to the required design type, to generate design collection data; Receiving design requirement information, and performing style selection and iterative design processing on the design collection data according to the design requirement information to generate a basic design style; According to the design requirement information, the basic design style is customized and optimized in terms of material, texture and color to generate an optimized design style; Determine the required wearing scene, select a scene model model based on AI technology, match the optimized design style with clothing based on the scene model model, and generate a scene wearing model; Automatically perform dynamic rendering processing, dynamically display and simulate the scene wearing model, eliminate the model model, generate and display the clothing flexibility simulation video; The step of obtaining the required design type, performing type matching, preferential screening and arranging from a plurality of preset release channels according to the required design type, and generating design collection data specifically includes the following steps: Get the required design type; According to the demand design type, type matching is performed from multiple preset publishing channels to obtain matching collection data; Extracting evaluation feedback data from the matching collected data; According to the evaluation feedback data, the matching collected data is preferentially screened to obtain preferential collected data; Arranging and sorting the preferentially collected data to generate design collected data; The receiving of design requirement information, performing style selection and iterative design processing on the design collection data according to the design requirement information, and generating a basic design style specifically includes the following steps: Receive design requirement information; Extracting style requirement information from the design requirement information; According to the style requirement information, selecting a style for the design collection data to obtain a plurality of selected style data; Based on the style requirement information and the plurality of evaluation feedback data, iteratively design the plurality of selected style data to generate a basic design style; The iterative design processing of the plurality of selected style data based on the style requirement information and the plurality of evaluation feedback data to generate a basic design style specifically comprises the following steps: The selected style data includes style picture information and style description information; Perform image recognition on style image information to obtain key design features; Perform semantic analysis on style description information to obtain style keywords; The key design features and style keywords are combined into style feature vectors, and clustering operations are performed on the style feature vectors using clustering algorithms to obtain multiple style category results; The sentiment analysis model is used to extract the sentiment inclination strength of the user's keywords for a certain style in the evaluation feedback data to obtain the keyword sentiment score; According to the time interval from the time point of the evaluation feedback data to the current time point, the keyword sentiment score is adjusted to obtain the user feedback score; Generate a user demand vector based on the style requirement information, and calculate the matching score between the user demand vector and the style feature vector; Obtain the user's negative keyword sentiment intensity for a certain style, and extract the corresponding feedback keywords based on the negative keyword sentiment intensity; Search whether the style feature vector contains feedback keywords. If so, it is considered as an improvement target. All feedback keywords contained in the style feature vector are combined with the corresponding negative keyword sentiment intensity to obtain an improvement potential score; Taking into account the matching score, user feedback score and improvement potential score, each style is quantitatively scored to obtain a comprehensive score of the style; Using style category results as the environment, style comprehensive scores as the state, selection or elimination of a style as the action, and the distribution of the filtered style scores close to user needs as the reward, a reinforcement learning algorithm is used for multiple rounds of screening; After the screening is completed, the style categories with the top comprehensive scores are output as the basic design styles.

2. The AI-based adaptive clothing design animation modeling method according to claim 1, characterized in that: The extracting of the evaluation feedback data from the matching collected data specifically comprises the following steps: Acquire matching collected data, and filter the matching collected data to obtain complete matching data; wherein the complete matching data includes a rating value, feedback content, and a timestamp; Extract keywords from the feedback content to obtain keywords; Given a keyword list, calculate the matching degree between the keyword and the given keyword list to obtain the keyword matching score; Perform exponential decay calculation based on the interval between the timestamp and the current time to get the freshness score; The score value, keyword matching score and freshness score are weighted and summed to obtain a comprehensive score; The complete matching data are sorted in descending order according to the comprehensive scores, and the first several are selected from the sorting as evaluation feedback data.

3. The AI-based adaptive clothing design animation modeling method according to claim 1, characterized in that: The step of customizing and optimizing the material, texture and color of the basic design style according to the design requirement information to generate the optimized design style specifically comprises the following steps: extracting detailed requirement information from the design requirement information; According to the detailed requirement information, customizing and selecting materials for the basic design style to obtain material selection data; According to the detailed requirement information, a texture is customized and selected for the basic design style to obtain texture selection data; According to the detailed requirement information, customizing and selecting the color of the basic design style to obtain color selection data; The basic design style is customized and optimized according to the material selection data, the texture selection data and the color selection data to generate an optimized design style.

4. The AI-based adaptive clothing design animation modeling method according to claim 3 is characterized in that: The step of customizing and optimizing the basic design style according to the material selection data, the texture selection data, and the color selection data to generate an optimized design style specifically comprises the following steps: Calculate the matching degree between material selection data and user requirements to obtain material scores; Calculate the matching degree between texture selection data and user requirements to obtain texture score; Calculate the difference between the color selection data and the color required by the user to obtain the color score; A multi-objective optimization function is constructed with the optimization objectives of maximizing the sum of material score, texture score and color score and minimizing the total cost of material, texture and color; By performing multi-objective optimization on the multi-objective optimization function, an optimized design style is obtained.

5. The AI-based adaptive clothing design animation modeling method according to claim 4 is characterized in that: The step of determining the required wearing scene, selecting a scene model model based on AI technology, matching clothing for the optimized design style based on the scene model model, and generating a scene wearing model specifically comprises the following steps: Identify the design requirement information and determine the required wearing scenario; Based on AI technology, according to the required wearing scene, a scene model is selected, and model model data of the scene model model is obtained; According to the required wearing scene, the model data and the optimized design style, clothing matching selection is performed to obtain matching clothing data; Based on the optimized design style and the matching clothing data, clothing matching is performed on the scene model model to generate a scene wearing model.

6. The AI-based adaptive clothing design animation modeling method according to claim 5 is characterized in that: The automatic dynamic rendering process, the dynamic display simulation of the scene wearing model, the model model elimination, and the generation and display of the clothing flexible simulation video specifically include the following steps: According to the requirements, the dressing scene is planned and the scene display actions are planned; According to the scene display action, dynamically render the model wearing the scene to generate a dynamic video of the model; Processing the model dynamic video frame by frame to generate multiple frame-by-frame display pictures; Eliminate the model from multiple frame-by-frame display images to generate multiple clothing display images; Performing video processing on the plurality of clothing display pictures to generate clothing flexibility simulation videos; The clothing flexibility simulation video is displayed.

7. The AI-based adaptive clothing design animation modeling method according to claim 6 is characterized in that: The method of eliminating the model in the multiple frame-by-frame display pictures to generate multiple clothing display pictures specifically includes the following steps: The spatiotemporal Transformer model is used to globally model the frame-by-frame display images and generate dynamic features of models and clothing. Use a multi-layer segmentation network to perform semantic segmentation on the displayed images frame by frame to generate model areas and clothing areas; Apply a fine-grained feature extraction network to map the material, color, and texture information of clothing into a high-dimensional semantic space to extract fine-grained features; Based on the dynamic features of the model, the overlapping area of ​​the model's action on the clothing boundary is inferred, and according to the fine-grained features, the appearance of the occluded part is predicted and compensated, so as to realize dynamic compensation of the overlapping area and obtain the compensated clothing area; Extract the background features of the occluded area of ​​the model from the compensated image to obtain the clothing area including the blank area; Based on the fine-grained features, a generative adversarial network is used to infer the appearance of the blank area in the clothing area containing the blank area to generate a complete clothing area. The generated background texture is optimized for temporal and spatial consistency to ensure the continuity of the background frame by frame, and a complete clothing area is obtained. Apply a local generative adversarial network to enhance the details of the complete clothing area, ensuring that the texture and shape of the clothing remain intact after the model is removed, and obtain an enhanced clothing area; According to the dynamic characteristics of clothing, the spatiotemporal Transformer is used to globally model the enhanced clothing area frame by frame to align the fine-grained features between adjacent frame images and obtain multiple clothing display pictures.

8. An AI-based adaptive clothing design animation modeling system, the system is applied to the AI-based adaptive clothing design animation modeling method according to any one of claims 1 to 7, characterized in that: The system includes a data acquisition and processing unit, a style iteration design unit, a detail customization optimization unit, a model clothing matching unit and a clothing flexibility simulation unit, wherein: The data collection and processing unit is used to obtain the required design type, and perform type matching, preferential screening and arrangement from a plurality of preset release channels according to the required design type to generate design collection data; A style iteration design unit, used for receiving design requirement information, and performing style selection and iterative design processing on the design collection data according to the design requirement information to generate a basic design style; A detail customization and optimization unit, used to customize and optimize the material, texture and color of the basic design style according to the design requirement information, and generate an optimized design style; A model clothing matching unit is used to determine the required wearing scene, select a scene model model based on AI technology, match the optimized design style with clothing based on the scene model model, and generate a scene wearing model; The clothing flexibility simulation unit is used to automatically perform dynamic rendering processing, dynamically display and simulate the scene wearing model, eliminate the model model, and generate and display the clothing flexibility simulation video.

Images