A clothing design method and system

Abstract

This invention relates to the field of clothing design technology and discloses a clothing design method and system. The method includes establishing a database of standard clothing components; acquiring user-preferred keywords and images, and using a machine learning algorithm to learn the keywords to obtain user preference features; extracting clothing component features of the clothing worn by the person in the image; using a matching algorithm to calculate the user preference features and clothing component features to obtain a matching degree value, and selecting clothing components from the clothing database based on the matching degree value to obtain the optimal matching scheme; applying the optimal matching scheme to a model to obtain a model of the clothing and returning it to the user. This method has the following advantages: This method learns user preferences to obtain user preference features, and selects suitable clothing from the database as the optimal matching scheme based on the preference features, thereby reducing reliance on designers and improving design efficiency.

Description

Technical Field

[0001] This invention relates to the field of clothing design technology, and in particular to a clothing design method and system. Background Technology

[0002] Traditional clothing design relies on the personal experience and professional knowledge of fashion designers. However, in today's rapidly developing information society, traditional clothing design can no longer meet people's needs. The automation and intelligentization of clothing design, and breaking away from traditional design methods, are the development trends of the clothing design industry. A garment product typically consists of different garment components, such as collars, sleeves, and the main body. These components combine to form the garment as a whole, expressing its style and determining its category.

[0003] Computer-aided design is now widely used in clothing design. Existing design systems or methods typically involve creating two-dimensional and three-dimensional garment drawings based on garment component patterns. Different sizes of garment components are then produced from these patterns, and these components are assembled to form a garment pattern. Paper patterns are then created based on the specifications of this pattern. Alternatively, human anatomy measurements can be used to determine the dimensions of various body parts and perform technical calculations. The results are then used to derive three-dimensional pattern pieces, which are then used to create paper patterns. Whether deriving specifications after pattern making or after measuring human body parameters, both methods involve complex and tedious technical specifications and procedures.

[0004] Currently, each garment design process involves complex drafting, cutting, and sewing procedures. This process heavily relies on manual operation and the experience and judgment of fashion designers, resulting in low production efficiency and difficulty in responding flexibly to market changes. This production process is not only complex and time-consuming, but also overly dependent on the experience and knowledge of fashion designers. Furthermore, different garment styles and designs require the creation of new patterns, making it difficult to adapt to the changing clothing market and significantly limiting design and production efficiency. Summary of the Invention

[0005] This invention provides a clothing design method and system to solve the problems of over-reliance on the designer's experience and knowledge and low design efficiency in the clothing design process.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a clothing design method, comprising:

[0007] Establish a clothing database containing standard clothing components;

[0008] Obtain keywords and images that users prefer, and use machine learning algorithms to learn from the keywords to obtain user preference features;

[0009] Based on the image, extract the clothing component features of the person's outfit in the image;

[0010] The matching algorithm is used to calculate the user preference features and the clothing component features to obtain a matching degree value. Based on the matching degree value, standard clothing components are selected from the clothing database to obtain the best matching scheme.

[0011] The optimal matching scheme is applied to a preset model to obtain a dressed model, and the dressed model is returned to the user.

[0012] In one alternative implementation, establishing a clothing database containing standard clothing components includes:

[0013] Obtain clothing images and clothing feature information;

[0014] The clothing is classified according to the clothing images and clothing feature information to obtain clothing category information;

[0015] Each clothing category information is encoded and recorded to obtain the number of clothing categories and a clothing database containing standard clothing components.

[0016] In one optional implementation, the step of using a machine learning algorithm to learn the keywords and obtain user preference features includes:

[0017] The keywords are used as feature vectors, and machine learning algorithms are used to learn the feature vectors to obtain learning results; wherein, the keywords include preferred pattern elements and preferred texture elements;

[0018] Based on the learning results, a preference histogram is constructed for each of the keywords, and the user preference features are obtained based on the preference histogram.

[0019] In one optional implementation, the step of building a preference histogram for each keyword based on the learning results, and obtaining the user preference features based on the preference histogram, includes:

[0020] The learning results are analyzed to obtain the probability of each keyword appearing, and a preference histogram is constructed.

[0021] Based on the preference histogram, the degree of correlation between the user and each of the keywords is recorded to obtain the user preference features;

[0022] The preference histogram is constructed according to the following formula.

[0023]

[0024] in, This represents the preference histogram. Indicates the number of keywords. They represent the first The probability of each keyword appearing.

[0025] In one optional implementation, extracting clothing component features of a person's clothing in the image, based on the image, includes:

[0026] Using a semantic segmentation algorithm, key points of clothing and human body are extracted from the image, and vectors between the key points are calculated to obtain semantic vectors;

[0027] The semantic vector is transformed to generate a feature vector in a high-dimensional space, and the feature vector is integrated to obtain the clothing component features.

[0028] In one optional implementation, the step of using a semantic segmentation algorithm to extract key points of clothing and the human body from the image, and calculating the vectors between the key points to obtain semantic vectors, includes:

[0029] Using a semantic segmentation algorithm, key points of clothing and human body in the image are extracted; wherein the key points include corner coordinates and contour line coordinates.

[0030] The semantic vector is obtained by calculating the vector between the keypoints using the Euclidean distance metric.

[0031] The semantic vector is calculated according to the following formula.

[0032]

[0033] in, This represents the semantic vector. Represents the x-coordinate of the first key point. Indicates the x-coordinate of the second key point. The ordinate of the first key point is represented by its vertical coordinate. The vertical coordinate represents the second key point.

[0034] In one optional implementation, the transformation of the semantic vector to generate a feature vector in a high-dimensional space includes:

[0035] Combine all the semantic vectors to form a feature matrix;

[0036] The covariance of the feature matrix is ​​calculated to obtain the covariance matrix;

[0037] The covariance matrix is ​​decomposed into eigenvalues ​​to obtain the eigenvectors.

[0038] The covariance matrix is ​​calculated according to the following formula:

[0039]

[0040] in, Let the covariance matrix be represented. This indicates the number of semantic vectors. Represents the feature matrix, The mean matrix represents the characteristic matrix.

[0041] In one optional implementation, the step of using a matching algorithm to calculate the user preference features and the clothing component features to obtain a matching degree value, and selecting standard clothing components from a clothing database based on the matching degree value to obtain the optimal matching scheme, includes:

[0042] The Euclidean distance metric is used to calculate the distance between the user preference features and the clothing component features to obtain a matching degree value.

[0043] The matching degree value is multiplied and matched with the preference histogram, and standard clothing parts are selected from the clothing database to obtain the best matching scheme;

[0044] The matching degree value is calculated according to the following formula:

[0045]

[0046] in, This indicates the degree of matching. This indicates the user's preference characteristics. This indicates the features of the clothing component. Indicates the number of clothing categories. Indicates the current number A type of clothing.

[0047] In one optional embodiment, the clothing component features include clothing shape features and human body shape features; wherein, the clothing shape features include the pattern, texture, and color scheme of the clothing; and the human body shape features include the body shape, skin color, face shape, and proportions of limbs.

[0048] Secondly, the present invention provides a clothing design system, comprising:

[0049] The database creation module is used to create a clothing database containing standard clothing parts;

[0050] The user preference feature learning module is used to acquire keywords and images that users prefer, and to use machine learning algorithms to learn the keywords to obtain user preference features.

[0051] The clothing component feature extraction module is used to extract the clothing component features of the person in the image based on the image.

[0052] The optimal matching scheme calculation module is used to calculate the user's preference features and the features of the clothing parts using a matching algorithm, obtain a matching degree value, and select standard clothing parts from the clothing database according to the matching degree value to obtain the optimal matching scheme.

[0053] The data display module is used to apply the best matching scheme to a preset model to obtain a dressed model, and then return the dressed model to the user.

[0054] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the clothing design method described in any one of the above.

[0055] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the clothing design method described in any one of the above.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] This invention provides a clothing design method, comprising: establishing a clothing database containing standard clothing components; acquiring user-preferred keywords and images, and using a machine learning algorithm to learn the keywords to obtain user preference features; extracting clothing component features of the clothing worn by the person in the image; using a matching algorithm to calculate the user preference features and the clothing component features to obtain a matching degree value, and selecting standard clothing components from the clothing database based on the matching degree value to obtain an optimal matching scheme; applying the optimal matching scheme to a preset model to obtain a model wearing the clothing, and returning the model wearing the clothing to the user.

[0058] The method learns user preferences to obtain user preference characteristics, and selects suitable clothing from a clothing database containing standard clothing parts as the user's best matching solution based on these preference characteristics. This reduces the reliance on the experience and knowledge of clothing designers and improves design efficiency. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the clothing design method provided in the first embodiment of the present invention;

[0060] Figure 2 This is a schematic diagram of the structure of the clothing design system provided in the second embodiment of the present invention. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] Reference Figure 1 The first embodiment of the present invention provides a clothing design method, including the following steps:

[0063] S1, Establish a clothing database containing standard clothing components;

[0064] S2, Obtain keywords and images that users prefer, and use machine learning algorithms to learn the keywords to obtain user preference features;

[0065] S3, Based on the image, extract the clothing component features of the person wearing the clothing in the image;

[0066] S4, use a matching algorithm to calculate the user preference features and the clothing component features to obtain a matching degree value, and select standard clothing components from the clothing database according to the matching degree value to obtain the best matching scheme;

[0067] S5, apply the best matching scheme to the preset model to obtain the dressed model, and return the dressed model to the user.

[0068] It's important to note that traditional clothing design schemes typically rely on the designer's experience and knowledge, requiring extensive practice and training. Simultaneously, designers sometimes need to communicate frequently with multiple clients. This one-to-many, high-frequency communication inevitably reduces the efficiency of the design process and limits its potential for improvement. Adding to the complexity, some clients seeking customized designs only have a conceptual idea of ​​their ideal clothing combination and may not be able to accurately express their desired outfits, which often contributes to increased design time.

[0069] In this invention, the use of a clothing design system within a computer program significantly accelerates the efficiency of the clothing design process. Leveraging the efficient machine learning algorithms within the system, users only need to input sufficient preference features, and the system can accurately analyze and understand their personalized preferences, including but not limited to color preferences, style inclinations, and material choices. This allows the system to automatically generate a series of optimal matching schemes that meet the user's expectations and present them to the user. Furthermore, the system offers high flexibility and interactivity during the design process, allowing users to modify their preferences at any time based on changes in their needs. The system can also respond instantly, adjusting the optimal matching schemes accordingly. This customized, real-time feedback mechanism not only greatly accelerates the efficiency of clothing design but also effectively reduces over-reliance on designers.

[0070] In step S1, the manufacturer needs to pre-establish a clothing database containing standard garment components. It is important to note that the clothing database should not only be able to store text data but also allow the storage of image data.

[0071] In one embodiment, establishing a clothing database containing standard clothing components includes: acquiring clothing images and clothing feature information; classifying the clothing according to the clothing images and clothing feature information to obtain clothing category information; and encoding and recording each clothing category information to obtain the number of clothing categories and a clothing database containing standard clothing components.

[0072] Specifically, when classifying clothing based on the clothing images and clothing feature information, an image extraction and template segmentation strategy can be employed. First, the clothing is extracted from the images using image extraction methods, and then split into multiple template components using template segmentation methods, which are recorded separately. These components are then combined with the clothing feature information to form the clothing category information. The template components include jackets, shirts, trousers, dresses, bottoms, tops, trousers, accessories, shoes, and hats; the clothing feature information includes the brand, fabric, price, model, and manufacturer of the current clothing. Finally, after obtaining the clothing category information, each clothing category is assigned a standardized number and then stored in the clothing database in an orderly manner according to the numbering.

[0073] It should be noted that the number is a unique identifier for the garment and plays a crucial role in multiple stages of garment design, production, sales, and inventory management. Therefore, before assigning a standardized number to each garment category, it is essential to ensure that the number used is unique and easily identifiable. Simultaneously, the standard garment components included in the garment database should be those that current manufacturers can produce. Furthermore, the garment database must possess good scalability to meet the growing demand for personalization.

[0074] In step S2, keywords and images favored by users are obtained, and a machine learning algorithm is used to learn the keywords to obtain user preference features, including the following steps:

[0075] S21, the system obtains the user's preferred keywords and images.

[0076] Specifically, the system acquires keywords and image data input by the user. The input data may include the user's body information (e.g., height, weight), preferred clothing information (e.g., clothing style, material), and other additional information (e.g., acceptable price, home address). The image data consists of sample photos of the user's preferred clothing.

[0077] It is important to note that when the system obtains the above information, it needs to perform a simple judgment to remove keywords and images unrelated to clothing design, so as not to affect subsequent results. In addition, the system should also save the above user personal information and establish corresponding security mechanisms.

[0078] S22, the keywords are used as feature vectors, and a machine learning algorithm is used to learn the feature vectors to obtain the learning results; wherein, the keywords include preferred pattern elements and preferred texture elements.

[0079] Specifically, the system first uses a text quantization algorithm to construct feature vectors from the keywords, and then uses machine learning algorithms to learn from these feature vectors, thereby obtaining the learning results. The machine learning algorithms include decision tree algorithms and support vector machine algorithms. The decision tree algorithm recursively partitions the feature space, guided by information gain and Gini impurity standards, to group similar feature vectors into the same category. The support vector machine algorithm, on the other hand, searches for an optimal separating hyperplane in the feature space formed by the feature vectors to maximize the margin between different categories. Both machine learning algorithms can classify the feature vectors into different categories based on the correlation between them, and statistically analyze the frequency and probability of their occurrence as the learning results.

[0080] It's important to note that decision tree and support vector machine algorithms have certain requirements regarding the number of keywords. In particular, if the number of keywords is too small (e.g., less than 24 keywords), the learning results may overfit, meaning they too closely match the specific cases of the samples, resulting in a significant decrease in generalization ability when facing new data. Therefore, to ensure the generalization ability of the algorithm during machine learning, it is necessary to ensure that the number of input keywords is as large as possible. At the same time, since decision tree and support vector machine algorithms are relatively traditional, their computational efficiency and accuracy may slightly decrease when handling large-scale datasets or complex pattern recognition tasks. Therefore, when hardware resources are sufficient, the machine learning algorithm used here can be replaced with a more efficient and accurate deep learning algorithm, such as a convolutional neural network.

[0081] S23, Based on the learning results, a preference histogram is built for each of the keywords, and the user preference features are obtained based on the preference histogram.

[0082] In one implementation, the step of building a preference histogram for each keyword based on the learning results and obtaining the user preference features based on the preference histogram includes: parsing the learning results to obtain the probability of occurrence of each keyword and building a preference histogram; recording the correlation between the user and each keyword based on the preference histogram to obtain the user preference features; wherein the preference histogram is built according to the following formula.

[0083]

[0084] in, This represents the preference histogram. Indicates the number of keywords. They represent the first The probability of each keyword appearing.

[0085] Specifically, the system uses data processing algorithms to analyze and statistically analyze the learning results, and obtains the probability of each keyword appearing based on the learning results. Then, it creates a preference histogram for each keyword, recording the correlation between the keyword and its probability of occurrence. Finally, it summarizes the user preference features based on the preference histogram and stores them in the corresponding file. It is important to note that the preference histogram records the quantity and probability of each keyword, but it is stored in text format rather than image format in the program algorithm. It is only output as an image when the user needs to view it. The user preference features record each user's clothing preferences, and the system also adjusts these user preference features accordingly after the user modifies the input keywords.

[0086] Furthermore, after calculating the preference histogram using the formula, the system records the degree of correlation between the user and each of the keywords based on the preference histogram, that is, it learns and records the user's preferred pattern elements (such as clothing style and model) and preferred texture elements (such as clothing material and pattern), thereby accurately extracting the user's preference features.

[0087] In step S3, the program algorithm extracts the clothing component features of the person in the image. These clothing component features include clothing shape features and human body shape features; the clothing shape features include the garment's cut, texture, and color scheme; the human body shape features include body shape, skin tone, face shape, and limb proportions.

[0088] In one implementation, extracting clothing component features of a person's attire from the image includes the following steps:

[0089] S31, Using a semantic segmentation algorithm, extract key points of clothing and human body from the image, and calculate the vector between the key points to obtain a semantic vector;

[0090] S32, transform the semantic vector to generate a feature vector in a high-dimensional space, and integrate the feature vector to obtain the clothing component features.

[0091] In step S31, the step of using a semantic segmentation algorithm to extract key points of clothing and the human body from the image and calculating the vectors between the key points to obtain a semantic vector includes: using a semantic segmentation algorithm to extract key points of clothing and the human body in the image; wherein the key points include corner coordinates and contour line coordinates; calculating the vectors between the key points using a Euclidean distance metric to obtain the semantic vector; wherein the semantic vector is calculated according to the following formula,

[0092]

[0093] in, This represents the semantic vector. Represents the x-coordinate of the first key point. Indicates the x-coordinate of the second key point. The ordinate of the first key point is represented by its vertical coordinate. The vertical coordinate represents the second key point.

[0094] Specifically, the system uses the semantic segmentation algorithm to perform semantic segmentation on the image, thereby extracting key points of clothing and the human body in the image. These key points are then used to calculate vectors between them, which serve as semantic vectors for subsequent calculations. The semantic segmentation algorithm is a pixel-level object detection algorithm that can effectively distinguish between the human body, clothing, and the environmental background in the image, ensuring accurate extraction of key points for clothing and the human body. After extracting the key points, vectors between each pair of key points are calculated to obtain semantic vectors, which are then temporarily recorded. The semantic vectors obtained from the key points not only contain shape information of the clothing and the human body but also implicitly contain the relative positions and movement trends between objects, providing a computational basis for subsequent recognition tasks such as pose estimation and behavior recognition.

[0095] Furthermore, each key point has a certain degree of correlation with the others, and this correlation degree records the degree of influence between the clothing and the human body. For key points with low correlation, the semantic vectors formed by these key points have little impact on subsequent calculation results. Therefore, the system can choose to prioritize calculating the correlation degree between all key points when calculating the vectors between them, and skip key points with correlation degrees less than a certain threshold from which semantic vectors are built, thereby reducing unnecessary computational burden and resource consumption.

[0096] In step S32, transforming the semantic vectors to generate feature vectors in a high-dimensional space includes: combining all the semantic vectors to form a feature matrix; calculating the covariance of the feature matrix to obtain a covariance matrix; and performing eigenvalue decomposition on the covariance matrix to obtain the feature vectors; wherein the covariance matrix is ​​calculated according to the following formula.

[0097]

[0098] in, Let the covariance matrix be represented. This indicates the number of semantic vectors. Represents the feature matrix, The mean matrix represents the characteristic matrix.

[0099] Specifically, the system first combines all the semantic vectors by substituting the coordinate values ​​of the semantic vectors into the feature matrix. Non-zero elements in the feature matrix represent the semantic vector for the current coordinate, while zero elements indicate that there is no semantic vector for the current coordinate. Then, the system calculates the mean matrix of the feature matrix and substitutes the feature matrix and the mean matrix into a formula to calculate the covariance matrix. Finally, the system performs eigenvalue decomposition on the covariance matrix to obtain the eigenvectors. The system then integrates these eigenvectors to obtain the features of the clothing component and records them in the corresponding file.

[0100] It is important to note that the feature matrix should be a square matrix, meaning that the number of rows and columns in the feature matrix must be equal to perform covariance calculation. Therefore, if the feature matrix obtained after combining the semantic vectors is not a square matrix, an appropriate number of zero elements can be added to the feature matrix to make it a square matrix.

[0101] In another implementation, to eliminate the standardization bias in the covariance matrix and prevent the generation of spurious solutions with the smallest eigenvector, the bias in the covariance matrix can be corrected by introducing a normalization factor. The covariance matrix is ​​corrected according to the following formula:

[0102]

[0103]

[0104] in, This represents the corrected covariance matrix. This represents the normalization factor. After the correction operation, the covariance matrix can be directly replaced with the corrected covariance matrix, and subsequent calculations can continue.

[0105] In step S4, the step of using a matching algorithm to calculate the user preference features and the clothing component features to obtain a matching degree value, and selecting standard clothing components from the clothing database based on the matching degree value to obtain the optimal matching scheme, includes: calculating the distance between the user preference features and the clothing component features using a Euclidean distance metric to obtain a matching degree value; multiplying the matching degree value with the preference histogram for matching; and selecting standard clothing components from the clothing database to obtain the optimal matching scheme; wherein, the matching degree value is calculated according to the following formula.

[0106]

[0107] in, This indicates the degree of matching. This indicates the user's preference characteristics. This indicates the features of the clothing component. Indicates the number of clothing categories. Indicates the current number A type of clothing.

[0108] Specifically, the system first reads the user preference features and clothing component features, previously stored as coordinates, from the corresponding file. Then, it directly calculates the distance between them using Euclidean metric, and this distance is considered the matching degree value, which can be recorded in the file for frequent use. Next, the matching degree value is multiplied and matched with the preference histogram to obtain the matching degree value and the number and probability of the keywords appearing in the preference histogram. Finally, the system uses the Hungarian algorithm to filter from the clothing database, obtaining standard clothing components that match the user preference features, and records them as the best matching scheme in the file. The best matching scheme records the user preference features, the standard clothing components, and the matching degree value in the form of triples. Meanwhile, the Hungarian algorithm is a combinatorial optimization algorithm that solves the task allocation problem in polynomial time; other more suitable task allocation algorithms can also be used instead.

[0109] In another implementation, the matching algorithm can be replaced by the following matching algorithm: First, calculate the Gaussian kernel function corresponding to the semantic vector, and perform an inner product calculation between the Gaussian kernel function and the preference histogram formed by the user preference features to obtain a first matching set; then, calculate the Euclidean distance from the semantic vectors of different categories to all the user preference features, and sort the first matching set according to the distance to obtain a second matching set; finally, fuse the first matching set and the second matching set to obtain the optimal matching scheme. This dual screening mechanism not only enhances the accuracy of matching, but also gives the algorithm the ability to flexibly adapt to diverse user preferences, thus enabling the construction of a more comprehensive and personalized optimal matching scheme.

[0110] In step S5, a standard three-dimensional model needs to be pre-built for the model.

[0111] Specifically, the system extracts the user's human body shape features from the clothing component features, and builds a standard 3D model identical to the current user's shape based on these features. These human body shape features include body type, skin color, face shape, and limb proportions.

[0112] Next, the RGB color values ​​of the corresponding component features are introduced into the standard 3D model.

[0113] Specifically, the system first uses the vertex position of the model as the center point of the Gaussian kernel function, then calculates the mean color value near the vertex position, and finally sets the RGB value of the vertex position as the mean of the Gaussian curve.

[0114] It is important to note that the Gaussian kernel function will produce different results depending on the coordinates in the three-dimensional space. Therefore, a gradient color will appear on the standard 3D model, which can more clearly distinguish different parts of the standard 3D model. This process not only ensures a smooth color transition, but also allows different parts of the model to present a more delicate and layered color effect.

[0115] Finally, the clothing corresponding to the best matching scheme is placed on the standard 3D model to obtain the model of the dressed person, and the final effect is returned to the user.

[0116] Specifically, the system first obtains the corresponding clothing from the clothing database based on the optimal matching scheme; then, it places the clothing on the corresponding part of the standard 3D model to obtain a model of the dressed man; finally, it renders the model of the dressed man to generate a rendering and returns the rendering to the user for viewing.

[0117] In another implementation, the user can set the appearance of the rendering, including adding textures, changing materials, colors, maps, wireframes, mesh display, filling details, lighting effects, reflection effects, fog effects, background blur, partial refresh, scaling, transformation copying, perspective, and movement. The user can also select the body shape of the dressed model, hide hair and accessories, and the viewing angle during preview.

[0118] To facilitate understanding of the present invention, the working process of the present invention is described below using a common scenario as an example:

[0119] In this embodiment, the garment factory plans to use a garment design system to design personalized garments for each user and carry out customized production. This system aims to reduce reliance on the knowledge and experience of garment designers and improve design efficiency. The specific implementation steps are as follows:

[0120] Step 1: The garment factory compiles statistics on the garments it can produce in advance and inputs this garment information data into the garment database for storage.

[0121] Step 2: Users send their personal clothing preferences, such as gender, body type, and preferred clothing style, to the garment factory in the form of text or pictures. The garment factory then inputs the user's preference data into the system.

[0122] Step 3: The system uses machine learning algorithms to learn users' clothing preferences and summarizes user preference characteristics.

[0123] Step 4: The system extracts clothing component features from the user's image, including information about the clothing and the human body.

[0124] Step 5: The system calculates the matching degree between the user and the clothes based on the characteristics of the clothing parts and the user's preferences, and searches for suitable clothing in the clothing database based on the matching degree to form the best matching scheme.

[0125] Step Six: The system applies the generated optimal matching scheme to the preset model, obtains a model of the dressed person, renders it, and finally displays the resulting image to the user. The user can modify the image at any time and regenerate a new image.

[0126] Through the above steps, the system can efficiently design personalized clothing for users without relying too much on the designer's experience and knowledge.

[0127] In summary, this invention provides a clothing design method, comprising: establishing a clothing database containing standard clothing components; acquiring user-preferred keywords and images, and using a machine learning algorithm to learn the keywords to obtain user preference features; extracting clothing component features of the clothing worn by a person in the image; calculating the matching degree value using a matching algorithm on the user preference features and the clothing component features, and selecting standard clothing components from the clothing database based on the matching degree value to obtain the optimal matching scheme; applying the optimal matching scheme to a preset model to obtain a model wearing the clothing, and returning the model wearing the clothing to the user. This method learns user preferences to obtain user preference features, and selects suitable clothing from a clothing database containing standard clothing components as the user's optimal matching scheme based on these preferences, thereby reducing reliance on the designer's experience and knowledge and improving design efficiency.

[0128] Reference Figure 2 The second embodiment of the present invention provides a clothing design system, comprising:

[0129] The database creation module is used to create a clothing database containing standard clothing parts;

[0130] The user preference feature learning module is used to acquire keywords and images that users prefer, and to use machine learning algorithms to learn the keywords to obtain user preference features.

[0131] The user preference feature learning module includes a user information input module, which is used to obtain the user's input preference keywords and images and judge the input; a preference feature extraction module, which is used to learn the keywords using machine learning algorithms and record the learning results; and a user preference feature extraction module, which is used to extract the current user's preference features from the learning results, summarize them into user preference features, and store them.

[0132] The clothing component feature extraction module is used to extract the clothing component features of the person in the image based on the image.

[0133] The clothing component feature extraction module includes a semantic segmentation module for extracting key points of people and clothing from the image; a semantic vector calculation module for calculating semantic vectors formed between the key points; a feature vector calculation module for calculating feature vectors of the matrix obtained through the semantic vectors; and a clothing component feature integration module for summarizing and storing the features of the clothing component based on the feature vectors.

[0134] The optimal matching scheme calculation module is used to calculate the user's preference features and the features of the clothing parts using a matching algorithm, obtain a matching degree value, and select standard clothing parts from the clothing database according to the matching degree value to obtain the optimal matching scheme.

[0135] The optimal matching scheme calculation module includes a matching degree calculation module, which uses a formula to calculate the matching degree value based on the user preference features and the clothing component features; and a clothing component selection module, which selects standard clothing components from the clothing database based on the matching degree value to obtain the optimal matching scheme and stores it.

[0136] The data display module is used to apply the best matching scheme to a preset model to obtain a dressed model, and then return the dressed model to the user.

[0137] The data display module includes a model generation module, which is used to construct a standard 3D model similar to the user's body shape based on the features of the clothing parts; a model coloring module, which is used to perform coloring operations on the standard 3D model using a Gaussian kernel function; and a model rendering module, which is used to apply the best matching scheme to the standard 3D model to obtain a dressed model and render it to obtain the final effect image, and then display the final effect image to the user.

[0138] It should be noted that the clothing design system provided in this embodiment of the invention is used to execute all the process steps of the clothing design method in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0139] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a clothing design program. When the processor executes the computer program, it implements the steps described in the various clothing design method embodiments above, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above system embodiments, such as the user preference feature learning module.

[0140] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0141] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0142] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0143] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0144] Wherein, if the modules / units integrated in the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0145] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0146] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A clothing design method, characterized in that, include: Establish a clothing database containing standard clothing components; Obtain keywords and images that users prefer, and use machine learning algorithms to learn from the keywords to obtain user preference features; Based on the image, extract the clothing component features of the person wearing the clothing in the image; wherein, the clothing component features include clothing shape features and human body shape features, the clothing shape features include the clothing pattern, texture and color scheme; the human body shape features include the body shape, skin color, face shape and body part proportions; The matching algorithm is used to calculate the user preference features and the clothing component features to obtain a matching degree value. Based on the matching degree value, standard clothing components are selected from the clothing database to obtain the best matching scheme. The optimal matching scheme is applied to a preset model to obtain a dressed model, and the dressed model is returned to the user. The process of establishing a clothing database containing standard clothing components includes: acquiring clothing images and clothing feature information; extracting clothing from the images and splitting it into multiple template components using a template segmentation method, recording them separately; then combining these components with the clothing feature information to form clothing category information; wherein the template components include jackets, shirts, trousers, dresses, bottoms, tops, trousers, accessories, shoes, and hats; and the clothing feature information includes the brand, fabric, price, model, and manufacturer of the current clothing; after obtaining the clothing category information, each clothing category is assigned a standardized number, and then stored in the clothing database in an orderly manner according to the number, resulting in the number of clothing categories and a clothing database containing standard clothing components. The step of using machine learning algorithms to learn the keywords and obtain user preference features includes: The keywords are used as feature vectors, and machine learning algorithms are used to learn the feature vectors to obtain learning results; wherein, the keywords include preferred pattern elements and preferred texture elements; a preference histogram is built for each keyword based on the learning results, and the user preference features are obtained based on the preference histograms; Specifically, the user preference features and clothing component features, which were originally stored in coordinate form, are read from the corresponding file, and then the distance between the two is calculated directly using Euclidean metric. The resulting distance is regarded as the matching degree value. Specifically, the matching degree value is multiplied and matched with the preference histogram to obtain the matching degree value and the number and probability of the keyword appearing stored in the preference histogram; the Hungarian algorithm is used to filter from the clothing database to obtain standard clothing parts that meet the user's preference characteristics, and these are recorded as the best matching scheme and stored in a file; wherein, the best matching scheme records the user's preference characteristics, the standard clothing parts, and the matching degree value in the form of a triple.

2. The clothing design method according to claim 1, characterized in that, The step of building a preference histogram for each keyword based on the learning results, and obtaining the user preference features based on the preference histogram, includes: The learning results are analyzed to obtain the probability of each keyword appearing, and a preference histogram is constructed. Based on the preference histogram, the degree of correlation between the user and each of the keywords is recorded to obtain the user preference features; The preference histogram is constructed according to the following formula. in, This represents the preference histogram. Indicates the number of keywords. They represent the first The probability of each keyword appearing.

3. The clothing design method according to claim 1, characterized in that, The step of extracting clothing component features of the person's clothing in the image includes: Using a semantic segmentation algorithm, key points of clothing and human body are extracted from the image, and vectors between the key points are calculated to obtain semantic vectors; The semantic vector is transformed to generate a feature vector in a high-dimensional space, and the feature vector is integrated to obtain the clothing component features.

4. The clothing design method according to claim 3, characterized in that, The semantic segmentation algorithm is used to extract key points of clothing and the human body from the image, and the vectors between the key points are calculated to obtain semantic vectors, including: Using a semantic segmentation algorithm, key points of clothing and human body in the image are extracted; wherein the key points include corner coordinates and contour line coordinates. The semantic vector is obtained by calculating the vector between the keypoints using the Euclidean distance metric. The semantic vector is calculated according to the following formula. in, This represents the semantic vector. Represents the x-coordinate of the first key point. Indicates the x-coordinate of the second key point. The ordinate of the first key point is represented by its vertical coordinate. The vertical coordinate represents the second key point.

5. The clothing design method according to claim 3, characterized in that, The transformation of the semantic vector to generate a feature vector in a high-dimensional space includes: Combine all the semantic vectors to form a feature matrix; The covariance of the feature matrix is ​​calculated to obtain the covariance matrix; The covariance matrix is ​​decomposed into eigenvalues ​​to obtain the eigenvectors. The covariance matrix is ​​calculated according to the following formula. in, Let the covariance matrix be represented. This indicates the number of semantic vectors. Represents the feature matrix, The mean matrix represents the feature matrix.

6. The clothing design method according to claim 1, characterized in that, The matching degree value is calculated according to the following formula. in, This indicates the degree of matching. This indicates the user's preference characteristics. This indicates the features of the clothing component. Indicates the number of clothing categories. Indicates the current number A type of clothing.

7. A clothing design system, characterized in that, For implementing the method as described in any one of claims 1-6, comprising: The database creation module is used to create a clothing database containing standard clothing parts; The user preference feature learning module is used to acquire keywords and images that users prefer, and to learn the keywords using machine learning algorithms to obtain user preference features. The clothing component feature extraction module is used to extract the clothing component features of the person in the image based on the image. The optimal matching scheme calculation module is used to calculate the user's preference features and the features of the clothing parts using a matching algorithm to obtain a matching degree value, and select standard clothing parts from the clothing database according to the matching degree value to obtain the optimal matching scheme. The data display module is used to apply the best matching scheme to a preset model to obtain a dressed model, and then return the dressed model to the user.

Images