Clothing pattern customization method and system
By using residual neural networks and automated evaluation algorithms, combined with user preferences and garment structural features, the problem of insufficient fit in garment pattern customization has been solved, achieving efficient and personalized pattern design and production optimization.
Patent Information
- Application Number
- CN202511001918.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-21
- Publication Date
- 2025-10-31
AI Technical Summary
In existing technologies, clothing pattern customization suffers from problems such as insufficient pattern adaptability, poor customization efficiency and economy. In particular, there is a lack of systematic evaluation methods for the matching degree between pattern design and clothing structure, resulting in low design efficiency and insufficient personalization.
Image feature extraction is performed using a ResNet residual neural network model. Combined with user historical preference information and clothing structural features, the structural fit index is calculated using a cosine similarity algorithm. Furthermore, a cost model and a process complexity evaluation algorithm are introduced to optimize the pattern design process.
It improves the matching degree between patterns and garment structure, realizes the automation and personalized design of the customization process, reduces labor costs, improves design accuracy and production efficiency, and controls production costs.
Smart Images
Figure CN120874158A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of clothing pattern customization technology, and in particular to a method and system for customizing clothing patterns. Background Technology
[0002] With the continuous growth of personalized consumption demands, clothing customization has gradually become an important development direction in the apparel industry. Traditional clothing customization methods mainly rely on manual design and experience-based judgment, which suffers from problems such as low design efficiency, insufficient personalization, and poor adaptability. Especially in pattern customization, how to quickly generate patterns that match the garment structure according to the user's personalized needs, while ensuring aesthetics and taking into account cost control and technological feasibility, has become a pressing technical challenge in the current clothing customization field.
[0003] In recent years, with the development of artificial intelligence technology, deep learning models have demonstrated powerful capabilities in image recognition and feature extraction, especially Residual Neural Networks (ResNet), which have achieved remarkable results in image feature extraction tasks. However, current technologies have not fully utilized deep learning models for intelligent analysis and feature extraction of user pattern requirements, and there is a lack of effective methods for integrating multi-dimensional information such as user historical preferences, clothing structural features, and market trends to evaluate pattern adaptation.
[0004] Furthermore, customizing clothing patterns involves not only aesthetics and personalization, but also the cost and complexity of the manufacturing process. In existing technologies, the process evaluation for pattern customization largely relies on manual experience, lacking systematic cost models and complexity assessment algorithms. This leads to significant discrepancies between pattern design and actual production, impacting the production efficiency and economic viability of customized clothing. Summary of the Invention
[0005] The purpose of this invention is to provide a method and system for customizing clothing patterns, which solves the problems of insufficient pattern adaptability, poor customization efficiency and economy in the existing technology.
[0006] To achieve the above objectives, the present invention provides a method for customizing clothing patterns, comprising the following steps:
[0007] S1. Obtain the first feature information of the part of the garment to be customized, and receive the second feature information sent by the user, including the garment pattern customization requirements;
[0008] S2. Use the ResNet residual neural network model to extract image features from the second feature information sent by the user, and obtain the second feature value based on the image feature extraction results and the user's historical preference information;
[0009] S3. Based on the first feature information and the second feature value, the fit degree is calculated to obtain the structural fit index between the second feature value and the part of the garment to be customized. Those with a structural fit index higher than the preset threshold are retained, and those with a structural fit index lower than the preset threshold are selected.
[0010] S4. Based on the cost model and process complexity evaluation algorithm, the preparation process cost and complexity of the second eigenvalue are comprehensively evaluated to obtain the comprehensive evaluation result;
[0011] S5. Customize clothing patterns based on structural adaptation index and comprehensive evaluation results.
[0012] In some embodiments of this application, in S1, the first feature information includes structural feature information, position mapping information, pattern mapping information, and market occurrence rate data;
[0013] The second feature information includes pattern information and position information.
[0014] In some embodiments of this application, in S2, a ResNet residual neural network model is used to extract image features from the second feature information sent by the user. The second feature value is obtained based on the image feature extraction result and the user's historical preference information, including:
[0015] S21. Perform format conversion, resolution normalization, and pixel value normalization preprocessing on the second feature information in sequence to obtain the preprocessed image tensor.
[0016] S22. Input the preprocessed image tensor into the ResNet residual neural network model and extract the convolutional feature map of the last layer or the image feature vector after global average pooling.
[0017] S23. Extract the user's historical selection or collection pattern category tag vector, historical color preference vector, and historical style preference vector from the user's historical profile database. Perform a weighted average and concatenate the above vectors to obtain the user's historical preference vector.
[0018] S24. Perform feature fusion on the image feature vector and the user's historical preference vector to obtain the second feature value.
[0019] In some embodiments of this application, in step S3, the adaptation degree is calculated based on the first feature information and the second feature value to obtain the structural adaptation index between the second feature value and the part of the garment to be customized, including:
[0020] S31. Vectorize the clothing structure feature information, position mapping information, pattern mapping information, and market occurrence rate data respectively, and concatenate the vectors to obtain the first feature vector, expressed as:
[0021] Ff = [S; P; T; M];
[0022] Among them, F f S is the first feature vector, S is the structural feature vector, P is the position mapping vector, T is the pattern mapping vector, and M is the market occurrence rate vector.
[0023] S32. The cosine similarity between the first eigenvector and the second eigenvalue is calculated using the cosine similarity algorithm. The expression is:
[0024]
[0025] Where Similarity is the cosine similarity, F s This is the second feature vector;
[0026] S33. Calculate the structural fit index based on cosine similarity, with the following expression:
[0027]
[0028] SAI stands for Structure Fit Index.
[0029] In some embodiments of this application, selecting structures with a structural fit index lower than a preset threshold includes:
[0030] If the structural fit index is lower than the preset threshold, the corresponding second feature information is fed back to the user for reselection or reselection in the system database.
[0031] In some embodiments of this application, based on a cost model and a process complexity evaluation algorithm, the preparation process cost and complexity of the second eigenvalue are comprehensively evaluated, and the comprehensive evaluation results include:
[0032] S41. Based on the second feature value and the clothing part information, retrieve the appropriate preparation process dataset from the system process database, extract the cost parameters of each process, and obtain the cost index based on the process cost parameter vector.
[0033] The cost parameters include equipment type, number of equipment, number of processes, processing time per piece, and material loss rate.
[0034] S42. Obtain the complexity evaluation parameters for each process, and obtain the complexity index based on the complexity evaluation parameters;
[0035] The complexity assessment parameters include the number of processes, skill level requirements, number of quality control points, and historical defect rate.
[0036] S43. Using weighted fusion, the cost index and complexity index of each process are weighted and fused to obtain a comprehensive evaluation index for each process. The process corresponding to the target comprehensive evaluation index is selected as the comprehensive evaluation result.
[0037] In some embodiments of this application, a system for customizing clothing patterns is also disclosed, comprising:
[0038] The acquisition module is used to acquire the first feature information of the part of the garment to be customized, and to receive the second feature information sent by the user, including the garment pattern customization requirements;
[0039] The feature extraction module is used to extract image features from the second feature information sent by the user using a ResNet residual neural network model, and to obtain the second feature value based on the image feature extraction result and the user's historical preference information.
[0040] The fit evaluation module is used to calculate the fit based on the first feature information and the second feature value, and obtain the structural fit index between the second feature value and the part of the garment to be customized. The structural fit index is retained if it is higher than the preset threshold, and the structural fit index is selected if it is lower than the preset threshold.
[0041] The comprehensive evaluation module is used to comprehensively evaluate the preparation process cost and complexity of the second feature value based on the cost model and process complexity evaluation algorithm, and obtain the comprehensive evaluation result.
[0042] The pattern customization module is used to customize clothing patterns based on the structural fit index and comprehensive evaluation results.
[0043] The advantages and beneficial effects of this invention compared to the prior art are:
[0044] 1. This invention uses a residual neural network to extract deep image features from user-provided pattern requests, and combines this with user historical preference information to generate a second feature value. This second feature value is then compared with multi-dimensional first feature information, including structural features, positional mapping, and pattern mapping of the garment's customizable parts, to calculate the structural fit index. This method effectively improves the matching degree between the pattern and the garment structure, ensuring visual and structural harmony and significantly enhancing the personalization and aesthetics of customized clothing.
[0045] 2. By introducing deep learning models and automated evaluation algorithms, this invention achieves fully automated processing from user requirement input to pattern adaptation evaluation, which greatly improves customization efficiency, reduces labor costs, and improves the accuracy and consistency of pattern design, making it suitable for large-scale personalized customization scenarios.
[0046] 3. This invention introduces a cost model and a process complexity evaluation algorithm into the pattern customization process. It comprehensively evaluates the cost and complexity of the preparation process for each candidate pattern, selecting the process scheme with the optimal comprehensive evaluation index. This method effectively avoids the problem of excessively high production costs or difficult-to-implement processes caused by overly complex pattern designs, improving the economy and manufacturability of pattern customization. It helps enterprises control production costs and improve production efficiency while meeting users' personalized needs.
[0047] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0048] Figure 1 This is a schematic diagram illustrating the steps of a method for customizing clothing patterns according to an embodiment of the present invention;
[0049] Figure 2 This is a structural diagram of a clothing pattern customization system according to an embodiment of the present invention. Detailed Implementation
[0050] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0051] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
[0052] like Figure 1 As shown, the present invention provides a method for customizing clothing patterns, including the following steps:
[0053] S1. Obtain the first feature information of the part of the garment to be customized, and receive the second feature information sent by the user, including the garment pattern customization requirements;
[0054] S2. Use the ResNet residual neural network model to extract image features from the second feature information sent by the user, and obtain the second feature value based on the image feature extraction results and the user's historical preference information;
[0055] S3. Based on the first feature information and the second feature value, the fit degree is calculated to obtain the structural fit index between the second feature value and the part of the garment to be customized. Those with a structural fit index higher than the preset threshold are retained, and those with a structural fit index lower than the preset threshold are selected.
[0056] S4. Based on the cost model and process complexity evaluation algorithm, the preparation process cost and complexity of the second eigenvalue are comprehensively evaluated to obtain the comprehensive evaluation result;
[0057] S5. Customize clothing patterns based on structural adaptation index and comprehensive evaluation results.
[0058] In some embodiments of this application, in S1, the first feature information includes structural feature information, position mapping information, pattern mapping information, and market occurrence rate data;
[0059] The second feature information includes pattern information and position information.
[0060] In some embodiments of this application, in S2, a ResNet residual neural network model is used to extract image features from the second feature information sent by the user. The second feature value is obtained based on the image feature extraction result and the user's historical preference information, including:
[0061] S21. Perform format conversion, resolution normalization, and pixel value normalization preprocessing on the second feature information in sequence to obtain the preprocessed image tensor.
[0062] S22. Input the preprocessed image tensor into the ResNet residual neural network model and extract the convolutional feature map of the last layer or the image feature vector after global average pooling.
[0063] S23. Extract the user's historical selection or collection pattern category tag vector, historical color preference vector, and historical style preference vector from the user's historical profile database. Perform a weighted average and concatenate the above vectors to obtain the user's historical preference vector.
[0064] S24. Perform feature fusion on the image feature vector and the user's historical preference vector to obtain the second feature value.
[0065] In some embodiments of this application, in step S3, the adaptation degree is calculated based on the first feature information and the second feature value to obtain the structural adaptation index between the second feature value and the part of the garment to be customized, including:
[0066] S31. Vectorize the clothing structure feature information, position mapping information, pattern mapping information, and market occurrence rate data respectively, and concatenate the vectors to obtain the first feature vector, expressed as:
[0067] F f = [S; P; T; M];
[0068] Among them, F f S is the first feature vector, S is the structural feature vector, P is the position mapping vector, T is the pattern mapping vector, and M is the market occurrence rate vector.
[0069] S32. The cosine similarity between the first eigenvector and the second eigenvalue is calculated using the cosine similarity algorithm. The expression is:
[0070]
[0071] Where Similarity is the cosine similarity, F s This is the second feature vector;
[0072] S33. Calculate the structural fit index based on cosine similarity, with the following expression:
[0073]
[0074] SAI stands for Structure Fit Index.
[0075] In some embodiments of this application, selecting structures with a structural fit index lower than a preset threshold includes:
[0076] If the structural fit index is lower than the preset threshold, the corresponding second feature information is fed back to the user for reselection or reselection in the system database.
[0077] In some embodiments of this application, based on a cost model and a process complexity evaluation algorithm, the preparation process cost and complexity of the second eigenvalue are comprehensively evaluated, and the comprehensive evaluation results include:
[0078] S41. Based on the second feature value and the clothing part information, retrieve the appropriate preparation process dataset from the system process database, extract the cost parameters of each process, and obtain the cost index based on the process cost parameter vector.
[0079] The cost parameters include equipment type, number of equipment, number of processes, processing time per piece, and material loss rate.
[0080] S42. Obtain the complexity evaluation parameters for each process, and obtain the complexity index based on the complexity evaluation parameters;
[0081] The complexity assessment parameters include the number of processes, skill level requirements, number of quality control points, and historical defect rate.
[0082] S43. Using weighted fusion, the cost index and complexity index of each process are weighted and fused to obtain a comprehensive evaluation index for each process. The process corresponding to the target comprehensive evaluation index is selected as the comprehensive evaluation result.
[0083] Specifically, after obtaining the comprehensive evaluation index for each process, the one with the lowest comprehensive evaluation index is the one with the lowest overall complexity and process cost, that is, the one with the best economy and customization efficiency. The lowest comprehensive evaluation index is selected as the target comprehensive evaluation index.
[0084] In some embodiments of this application, such as Figure 2 As shown, a system for customizing clothing patterns is also disclosed, including:
[0085] The acquisition module is used to acquire the first feature information of the part of the garment to be customized, and to receive the second feature information sent by the user, including the garment pattern customization requirements.
[0086] The feature extraction module is used to extract image features from the second feature information sent by the user using a ResNet residual neural network model, and obtain the second feature value based on the image feature extraction results and the user's historical preference information.
[0087] The fit evaluation module is used to calculate the fit based on the first feature information and the second feature value, and obtain the structural fit index between the second feature value and the part of the garment to be customized. The structural fit index is retained if it is higher than the preset threshold, and the structural fit index is selected if it is lower than the preset threshold.
[0088] The comprehensive evaluation module is used to comprehensively evaluate the preparation process cost and complexity of the second eigenvalue based on the cost model and process complexity evaluation algorithm, and obtain the comprehensive evaluation result.
[0089] The pattern customization module is used to customize clothing patterns based on the structural fit index and comprehensive evaluation results.
[0090] The advantages and beneficial effects of this invention compared to the prior art are:
[0091] 1. This invention uses a residual neural network to extract deep image features from user-provided pattern requests, and combines this with user historical preference information to generate a second feature value. This second feature value is then compared with multi-dimensional first feature information, including structural features, positional mapping, and pattern mapping of the garment's customizable parts, to calculate the structural fit index. This method effectively improves the matching degree between the pattern and the garment structure, ensuring visual and structural harmony and significantly enhancing the personalization and aesthetics of customized clothing.
[0092] 2. By introducing deep learning models and automated evaluation algorithms, this invention achieves fully automated processing from user requirement input to pattern adaptation evaluation, which greatly improves customization efficiency, reduces labor costs, and improves the accuracy and consistency of pattern design, making it suitable for large-scale personalized customization scenarios.
[0093] 3. This invention introduces a cost model and a process complexity evaluation algorithm into the pattern customization process. It comprehensively evaluates the cost and complexity of the preparation process for each candidate pattern, selecting the process scheme with the optimal comprehensive evaluation index. This method effectively avoids the problem of excessively high production costs or difficult-to-implement processes caused by overly complex pattern designs, improving the economy and manufacturability of pattern customization. It helps enterprises control production costs and improve production efficiency while meeting users' personalized needs.
[0094] In this application, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. In case of any inconsistency, the meaning set forth in this specification or derived from the content described herein shall prevail. Furthermore, the terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the scope of this application.
[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for customizing clothing patterns, characterized in that, Includes the following steps: S1. Obtain the first feature information of the part of the garment to be customized, and receive the second feature information sent by the user, including the garment pattern customization requirements; S2. Use the ResNet residual neural network model to extract image features from the second feature information sent by the user, and obtain the second feature value based on the image feature extraction results and the user's historical preference information; S3. Based on the first feature information and the second feature value, the fit degree is calculated to obtain the structural fit index between the second feature value and the part of the garment to be customized. Those with a structural fit index higher than the preset threshold are retained, and those with a structural fit index lower than the preset threshold are selected. S4. Based on the cost model and process complexity evaluation algorithm, the preparation process cost and complexity of the second eigenvalue are comprehensively evaluated to obtain the comprehensive evaluation result; S5. Customize clothing patterns based on structural adaptation index and comprehensive evaluation results.
2. The method for customizing clothing patterns according to claim 1, characterized in that, In S1, the first feature information includes structural feature information, position mapping information, pattern mapping information, and market occurrence rate data; The second feature information includes pattern information and position information.
3. The method for customizing clothing patterns according to claim 2, characterized in that, In step S2, a ResNet residual neural network model is used to extract image features from the second feature information sent by the user. The second feature value is obtained based on the image feature extraction result and the user's historical preference information, including: S21. Perform format conversion, resolution normalization, and pixel value normalization preprocessing on the second feature information in sequence to obtain the preprocessed image tensor. S22. Input the preprocessed image tensor into the ResNet residual neural network model and extract the convolutional feature map of the last layer or the image feature vector after global average pooling. S23. Extract the user's historical selection or collection pattern category tag vector, historical color preference vector, and historical style preference vector from the user's historical profile database. Perform a weighted average and concatenate the above vectors to obtain the user's historical preference vector. S24. Perform feature fusion on the image feature vector and the user's historical preference vector to obtain the second feature value.
4. The method for customizing clothing patterns according to claim 3, characterized in that, In step S3, the fit degree is calculated based on the first feature information and the second feature value, and the structural fit index between the second feature value and the part of the garment to be customized is obtained, including: S31. Vectorize the clothing structure feature information, position mapping information, pattern mapping information, and market occurrence rate data respectively, and concatenate the vectors to obtain the first feature vector, expressed as: F f =[S;P;T;M]; Among them, F f S is the first feature vector, S is the structural feature vector, P is the position mapping vector, T is the pattern mapping vector, and M is the market occurrence rate vector. S32. The cosine similarity between the first eigenvector and the second eigenvalue is calculated using the cosine similarity algorithm. The expression is: Where Similarity is the cosine similarity, F s This is the second feature vector; S33. Calculate the structural fit index based on cosine similarity, with the following expression: SAI stands for Structure Fit Index.
5. The method for customizing clothing patterns according to claim 4, characterized in that, In step S3, re-selecting structures with a structural fit index lower than a preset threshold includes: If the structural fit index is lower than the preset threshold, the corresponding second feature information is fed back to the user for reselection or reselection in the system database.
6. The method for customizing clothing patterns according to claim 5, characterized in that, In step S4, based on the cost model and process complexity evaluation algorithm, the preparation process cost and complexity of the second eigenvalue are comprehensively evaluated, and the comprehensive evaluation results include: S41. Based on the second feature value and the clothing part information, retrieve the appropriate preparation process dataset from the system process database, extract the cost parameters of each process, and obtain the cost index based on the process cost parameter vector. The cost parameters include equipment type, number of equipment, number of processes, processing time per piece, and material loss rate. S42. Obtain the complexity evaluation parameters for each process, and obtain the complexity index based on the complexity evaluation parameters; The complexity assessment parameters include the number of processes, skill level requirements, number of quality control points, and historical defect rate. S43. Using weighted fusion, the cost index and complexity index of each process are weighted and fused to obtain a comprehensive evaluation index for each process. The process corresponding to the target comprehensive evaluation index is selected as the comprehensive evaluation result.
7. A system for customizing clothing patterns, characterized in that, include: The acquisition module is used to acquire the first feature information of the part of the garment to be customized, and to receive the second feature information sent by the user, including the garment pattern customization requirements; The feature extraction module is used to extract image features from the second feature information sent by the user using a ResNet residual neural network model, and to obtain the second feature value based on the image feature extraction result and the user's historical preference information. The fit evaluation module is used to calculate the fit based on the first feature information and the second feature value, and obtain the structural fit index between the second feature value and the part of the garment to be customized. The structural fit index is retained if it is higher than the preset threshold, and the structural fit index is selected if it is lower than the preset threshold. The comprehensive evaluation module is used to comprehensively evaluate the preparation process cost and complexity of the second feature value based on the cost model and process complexity evaluation algorithm, and obtain the comprehensive evaluation result. The pattern customization module is used to customize clothing patterns based on the structural fit index and comprehensive evaluation results.