A method and system for intelligent automatic typesetting and cutting optimization of clothing

Abstract

This invention relates to the field of intelligent clothing design technology, and in particular to an intelligent automatic layout and cutting optimization method and system for clothing. The method includes: acquiring and judging clothing design requirements; if it is a general production requirement, acquiring an individual information set; classifying the individual information set by size using a first preset algorithm model; determining the clothing layout parameters corresponding to each size based on the individual information set and a historical reference database; and automatically laying out the clothing. If it is a personalized customization requirement, acquiring customized human body shape information; fitting the customized human body shape information using a second preset algorithm model to obtain a customized three-dimensional model; generating a three-dimensional clothing model based on the customized clothing style characteristics and the customized three-dimensional model; and obtaining a two-dimensional clothing drawing after cutting optimization; and splitting the two-dimensional clothing drawing and automatically laying it out. This invention can adapt to the needs of different orders, supports the conversion of multiple sizes and styles, and improves the production efficiency of layout drawings.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent clothing design, and in particular to an intelligent automatic typesetting and cutting optimization method and system for clothing. Background Art

[0002] In the traditional clothing production process, layout and cutting usually rely on manual experience for fabric layout and cutting planning. This method is not only time-consuming and labor-intensive, but also easily leads to fabric waste, inaccurate cutting and low production efficiency. Specifically, the traditional layout and cutting methods have the following main defects:

[0003] (1) Traditional methods rely on manual experience for fabric layout and cutting planning, which is time-consuming and difficult to meet the rapid response requirements of high-frequency and diversified orders;

[0004] (2) Manual typesetting is difficult to achieve optimal fabric utilization, and usually produces a lot of scraps, resulting in fabric waste and increased costs;

[0005] (3) Manual operation is difficult to achieve high precision requirements, especially in the cutting of complex patterns or multi-specification orders, which may cause dimensional errors or pattern deviations.

[0006] Therefore, it is urgent to transform and promote clothing companies towards digitalization and intelligence. In the process of intelligent advancement of clothing companies, the optimization of typesetting and cutting processes is particularly important. Summary of the invention

[0007] In order to solve one of the above-mentioned problems in the prior art, the present invention provides a method and system for optimizing intelligent automatic layout and cutting of clothing.

[0008] In order to achieve the above-mentioned purpose, the present invention provides a method for intelligent automatic typesetting and cutting optimization of clothing, comprising: step S101, obtaining clothing design requirements; step S102, judging whether the clothing design requirements are general production requirements or personal customization requirements according to preset rules, if the clothing design requirements are the general production requirements, executing step S103, if the clothing design requirements are the personal customization requirements, executing step S105; step S103, obtaining an individual information set according to the general production requirements, and classifying the individual information set by size using a first preset algorithm model; step S104, determining the size corresponding to each size according to the individual information set and a historical reference database clothing layout parameters, and automatically layout using the clothing layout parameters; step S105, obtaining customized human body shape information according to the personal customization needs, fitting the customized human body shape information using a second preset algorithm model, and obtaining a customized body three-dimensional model; step S106, obtaining customized clothing style features, and generating a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model; step S107, cutting and optimizing the three-dimensional clothing model through an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting optimization; step S108, projecting the three-dimensional clothing model after cutting optimization to obtain a plane clothing image, splitting the plane clothing image and automatically layouting it.

[0009] On the other hand, the present invention also provides a clothing intelligent automatic typesetting and cutting optimization system, comprising: a startup module, a general production module and a personal customization module; the startup module is used to obtain clothing design requirements, and judge whether the clothing design requirements are general production requirements or personal customization requirements according to preset rules. If the clothing design requirements are the general production requirements, the general production module is started, and if the clothing design requirements are the personal customization requirements, the personal customization module is started; the general production module includes an automatic classification unit and an automatic typesetting unit; the automatic classification unit is used to obtain an individual information set according to the general production requirements, and classify the individual information set into sizes using a first preset algorithm model; the automatic typesetting unit is used to determine the clothing typesetting parameters corresponding to each size according to the individual information set and a historical reference database. number, and automatically typeset using the clothing typesetting parameters; the personal customization module includes: a customer body fitting unit, a clothing generation unit, a cutting optimization unit and a typesetting unit; the customer body fitting unit is used to obtain customized human body information according to the personal customization needs, and use a second preset algorithm model to fit the customized human body information to obtain a customized body three-dimensional model; the clothing generation unit is used to obtain customized clothing style features, and generate a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model; the cutting optimization unit is used to perform cutting optimization on the three-dimensional clothing model through an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting optimization; the typesetting unit is used to project the three-dimensional clothing model after cutting optimization to obtain a plane clothing image, split the plane clothing image and automatically typeset.

[0010] The beneficial effects of the present invention are reflected in that the intelligent automatic layout and cutting optimization method and system for clothing provided by the present invention realizes intelligent automatic layout and cutting optimization by integrating advanced algorithms and automation equipment, and can optimally arrange and design fabrics according to order requirements, minimize waste of scraps, and improve fabric utilization and cutting accuracy. Through the system of the present invention, it is possible to dynamically adapt to the complex requirements of different clothing orders, support the conversion of multiple sizes and styles, and optimize layout through real-time data analysis and feedback, shortening the production cycle, reducing labor costs, and improving the production efficiency of layout drawings. The method of cutting optimization based on the customer's body model further avoids the possibility of fabric waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A flow chart of the intelligent automatic typesetting and cutting optimization method for clothing provided in Example 1 of the present invention;

[0012] Figure 2This is a schematic diagram of the structure of the clothing intelligent automatic typesetting and cutting optimization system provided in Example 1 of the present invention;

[0013] Figure 3 This is a schematic diagram of the structure of a universal production module provided in Example 1 of the present invention;

[0014] Figure 4 This is a schematic diagram of the structure of the personal customization module provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] Example 1

[0017] This embodiment provides a method for optimizing clothing intelligent automatic layout and cutting. Figure 1 ,include:

[0018] Step S101, obtaining clothing design requirements; specifically, the clothing design requirements can be obtained by obtaining orders, and the orders can include specific clothing types, design styles, order quantities, and customer information (height, weight, body shape, size, age, etc.).

[0019] Step S102, according to the preset rules, determine whether the clothing design demand is a general production demand or a personal customization demand. If the clothing design demand is a general production demand, step S103 is executed; if the clothing design demand is a personal customization demand, step S105 is executed; specifically, it can be directly input or selected through the system whether it is a general production demand or a personal customization demand; it can also be judged by detecting the preset fields, for example, by detecting the order quantity, customer information, etc. in the clothing design demand. The general production demand generally includes orders with a large number and low requirements for size matching accuracy, such as company team building clothes, school uniforms, corporate uniforms, etc. The personal customization demand includes orders for customized production lines with a small quantity, and customers with orders for personal customization requirements may have special requirements for tailoring (such as: slim fit, loose fit, etc.). In the specific implementation process, it can include a general production section and a personal customization section. The general production section is responsible for the automatic classification and layout of the size of the fixed size production line, and the personal customization section is responsible for the layout and tailoring optimization of some small customized production lines. After determining whether it is a general production demand or a personal customization demand, the two sections can perform the corresponding typesetting and cutting operations respectively.

[0020] Step S103, obtain an individual information set according to general production requirements, and use the first preset algorithm model to classify the individual information set by size; specifically, the individual information set contains customer information. The preset sizes S, M, and L can be set according to the general clothing sizes (if necessary, the size category can be increased, depending on the order requirements), and the customer information can be classified by size. Different clothing types can be classified in different ways. Since the clothing parameters specified by each size are not a fixed value, but an interval, when typesetting, production should be carried out according to the clothing parameters with the largest number of people in each size in the order as much as possible, so that fabric can be saved while improving the wearing experience of most people. For the design of the size chart, size classification parameters similar to those in Tables 1, 2, and 3 can be used.

[0021] Table 1 Top size chart (unit: cm)

[0022]

[0023] Table 2 Pants size chart (unit: cm)

[0024]

[0025] Table 3 Dress size chart (unit: cm)

[0026]

[0027] In an optional embodiment, the individual information set includes multiple groups of individual characteristic information, each group of individual characteristic information includes at least the age, weight and height of the individual; the individual information set is categorized by size using the first preset algorithm model, including: using the regression algorithm model to calculate the individual characteristic information, obtain the size category corresponding to each individual characteristic information, and categorize the individual characteristic information by size. Specifically, since the standard body parameters (sleeve length, shoulder width, hip circumference, etc.) are inconvenient to measure, and the number of customers is inconvenient to count, it may not be possible to obtain accurate body parameters. Therefore, this embodiment can also only collect age, weight, and height information, and determine the size category corresponding to the customer through age, weight, and height information. Of course, this embodiment does not exclude the implementation method in which the individual characteristic information directly includes standard body parameters. The purpose of introducing age is to take into account that the waist circumference of some male customers may increase with age, but the weight will not change too much. If only height and weight are used, it may lead to inaccurate waist estimation. The regression algorithm model can be used to analyze and predict the customer's body shape using the individual characteristic information obtained. By using individual characteristic information (such as height, weight, and age, etc.) as variables to predict the results (body parameters), more accurate body shape predictions can be achieved.

[0028] In a specific optional implementation manner, the classification in step S103 may be implemented by the following steps:

[0029] (1) Preset the preset sizes S, M, and L.

[0030] (2) Obtain individual characteristic information of each customer in the order, such as age, weight, and height.

[0031] (3) Use the XGBoost (Extreme Gradient Boosting) regression model to implement size classification based on age, weight, and height information. The specific process can be expressed as follows:

[0032]

[0033] in, Indicates the size category after classification. Represents the weighted prediction output of all trees in the XGBoost model, where the prediction of each tree is based on the input features (age, weight, height, etc.) calculated, round Indicates that the continuous value of the regression prediction is rounded to a discrete integer value. Represents the reverse mapping function of the numeric label encoding, converting the numeric label encoding back to the corresponding size category (e.g. 0 for S, 1 for M, 2 for L).

[0034] Step S104, determining the clothing layout parameters corresponding to each size according to the individual information set and the historical reference database, and automatically performing layout using the clothing layout parameters; specifically, after determining the size of each customer, it is necessary to select the corresponding clothing layout parameters for the size, as shown in Tables 1, 2, and 3. An existing size table can be used as a historical reference database, and the size corresponding to each size is selected from the size range of the existing size table as the clothing layout parameter. A specific body parameter historical reference database can also be established, which stores body parameters corresponding to weight and height collected during the past production of clothing.

[0035] In an optional embodiment, the clothing layout parameters corresponding to each size are determined according to the individual information set and the historical reference database, including: counting the occurrence frequency of each individual feature information, determining the individual feature information with the highest frequency in each size; searching and matching from the historical reference database according to the individual feature information with the highest frequency to obtain the general body parameters corresponding to the size; and obtaining the clothing layout parameters corresponding to the size according to the general body parameters. Specifically, after the size category of each person is known, the body parameters of the most people in the order can be estimated using the body parameter historical reference database, specifically counting the most weight and height in each size category, and then calculating the average body parameter corresponding to the height and weight matched in the historical reference database, and using the body parameter as the clothing layout parameter. In the actual process, each individual feature information in each size may appear only once, or there may be multiple individual feature information with the highest frequency, then multiple general body parameters will be obtained, and the average value of multiple general body parameters can be calculated at this time. By counting the individual feature information with the highest frequency to find and match the corresponding body parameters, the produced clothing products can be made to fit the body shape of most customers as much as possible. Alternatively, in the historical reference database, a human body feature information may correspond to multiple universal body parameters, in which case it is also necessary to calculate the average value of the universal body parameters. By counting the individual feature information with the highest frequency to find the matching corresponding body parameters, the produced clothing products can be made to fit the body shapes of most customers as much as possible.

[0036] Step S105, according to the individual customization requirements, the customized human body shape information is obtained, and the customized human body shape information is fitted using the second preset algorithm model to obtain a customized body stereoscopic model; specifically, the second preset algorithm model may include multiple algorithm operation steps. In an optional embodiment, the customized human body shape information may at least include: a customized human body image and a customized body measurement parameter. The customized human body image may include a multi-directional or multi-angle photographic image of the customer's body, for example, it may include a front image, a left side image, a right side image, a back image and other views of the customer in a T-pose (T-Pose), and the customized body measurement parameters may be manually measured body parameters (which may be measured according to the corresponding parameter format of Table 1, Table 2 and Table 3 above according to the target clothing type). The customized body stereoscopic model may be a customized body 3D model. Through the views of the human body in multiple directions, a stereoscopic 3D full view of the customer's customized body can be obtained.

[0037] In an optional embodiment, the customized human body shape information is fitted using a second preset algorithm model to obtain a customized body three-dimensional model, including: using a pre-trained human body segmentation model to segment the customized human body image to obtain a human body mask image; using a human body key point inspection pre-trained model to identify and calculate the human body joint nodes in the human body mask image to obtain a customized body initial model; using customized body measurement parameters to correct the customized body initial model to obtain a customized body three-dimensional model. By first segmenting the customized human body image to obtain a human body mask, the object in the image can be accurately separated from the background, which facilitates the subsequent identification and processing of the key points of the image. After obtaining the initial model, the initial model is corrected using the measurement parameters to further reduce the error, so that the customized body three-dimensional model finally obtained is closer to the customer's body.

[0038] In a specific optional implementation, the customized human body image data may be processed through the following detailed steps to obtain a customized body stereoscopic model:

[0039] (1) Use the pre-trained human segmentation model to perform human segmentation on the human image data to obtain human mask images of four photographic angles (i.e., the four-angle photographic images of the customer's body mentioned above). In this embodiment, the CE2P model (CE2P is an end-to-end human body parsing framework) can be used.

[0040] (2) Use the human key point detection pre-trained model to detect the 3D coordinates of the human joint nodes in the image data And 2D coordinates For recognition, the posenet model (a human posture detection model) can be used in this embodiment.

[0041] (3) and Concatenate into coordinate vectors , and build a The picture ,in = ( is the coordinate vector A collection of To transform the coordinate vector The node set after conversion to graph nodes), Specifically, it is a collection of human joint nodes. Indicates Coordinate vectors of individual body joint nodes; Represented as an adjacency matrix, specifically The connection between joints is defined according to the human skeleton and symmetry relationship, such as: and Same or connected, then ,otherwise .

[0042] (4) Paired images Perform normalized Laplace calculation:

[0043]

[0044] in, For The result of normalized Laplace calculation is: is the identity matrix, is the diagonal matrix of the connectivity of each human joint, .

[0045] (5) Yes Perform spectral domain graph convolution:

[0046]

[0047] in, is the result of the normalized Laplace calculation in step (4) above The result after scaling is: , is the scaling factor; The order is Chebysev polynomials; represents the Chebysev coefficient matrix, is a learnable parameter. The dimensions of the elements in are Therefore, to avoid the Chebysev polynomial order being greater than 5, it can be set to 3 in this embodiment. After completing the above operations, It replaced In .

[0048] (6) By gradually upsampling the image Convert to Grid-Based The picture ,in Represents a collection of human body mesh vertices; The same is expressed as an adjacency matrix. The conversion operation is firstly to pass through the fully connected layer The number of joint nodes in is expanded to the target number ( middle The number of elements in is J , the number of elements is mapped to V , thus obtaining ).

[0049] (7) Paired images Perform spectral domain graph convolution similar to step (5):

[0050]

[0051] in, express Normalize the Laplace result. Use U replace middle of , you can get the customer's customized initial model.

[0052] (8) Use the manually measured body parameters to correct the customized body initial model. The correction method is proportional conversion. For example, if the ratio of chest circumference to waist circumference in the customized body initial model is 1.2:1, and the actual manual measurement result is 1.3:1, the proportion correction is performed according to the measurement result. Generally speaking, the unit of the body model is pixel value, and the unit of the body parameter is millimeter. After the correction is completed, the customer's customized body three-dimensional model is obtained.

[0053] Step S106, obtaining the customized clothing style features, and generating a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model; specifically, a 2D design drawing of a corresponding style can be selected according to the clothing style features selected by the customer, and the 2D design drawing can be combined with the customized body three-dimensional model to produce a three-dimensional clothing model.

[0054] In an optional embodiment, obtaining customized clothing style features, and generating a three-dimensional clothing model based on the customized clothing style features and the customized body three-dimensional model includes: receiving input customized clothing style features, obtaining a corresponding plane design drawing based on the customized clothing style features; generating a clothing skin based on the grid of the customized body three-dimensional model; projecting the plane design drawing onto the customized body three-dimensional model using the coordinates of the human body joint nodes as the alignment standard, and merging the projected layer with the clothing skin to obtain a three-dimensional clothing model. The three-dimensional clothing model can be a three-dimensional 3D model. By means of a 3D clothing model, compared with a 2D plane design drawing, clothing can be designed more comprehensively and accurately according to the customer's body characteristics, so that the designed customized clothing is more suitable for the customer's body shape.

[0055] Specifically, the customer's customized body 3D model has been obtained through the customer body fitting operation. Each intersection of the grid in the model has a 3D coordinate. At this time, a new grid is generated according to a fixed threshold distance from each intersection. This grid will be used as the clothing skin (the clothing skin is a simulated cloth). At this time, input the 2D design drawing corresponding to the clothing style selected by the customer. The 2D design drawing contains the front view and back view of the clothing. There is no special requirement for the size unit of the 2D design drawing. A specific process of generating a 3D clothing model can be as follows:

[0056] (1) Obtain the preset camera parameters given by the posenet model and scale the pixel size of the 2D design image to be consistent with the size of the human body part in the customer's human body image (which can be determined by the size of the rectangular bounding box).

[0057] (2) According to the preset camera model, the 2D design drawing is projected onto the customer's customized body 3D model using the coordinates of the human body's joint nodes as the alignment standard, and the projected layer is fused with the point on the clothing skin that is closest to it (for example, there is a point A on the layer and a point B on the clothing skin, and the distance between the two points is the closest. Point A is fused to point B. Fusion means replacing the color of B with the color of A, and then removing A).

[0058] (3) After the fusion is completed, the unfused points in the clothing skin are removed to obtain the three-dimensional clothing model to be produced.

[0059] Step S107, cutting and optimizing the three-dimensional clothing model through an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting and optimization; specifically, after obtaining the three-dimensional clothing model, the three-dimensional clothing model needs to be further cut and optimized, and the cutting and optimization can usually be performed by optimizing the seams.

[0060] In an optional implementation, the intelligent automatic layout and cutting optimization method of clothing of this embodiment also includes: generating fixed seams on the clothing skin according to the style characteristics of the customized clothing; cutting optimization of the three-dimensional clothing model through an iterative optimization algorithm includes: performing a fitting operation on the seams of the three-dimensional clothing model and the fixed seams through an iterative optimization algorithm to perform cutting optimization, wherein the fitting operation is used to minimize the position difference between the seams of the three-dimensional clothing model and the fixed seams. By optimizing the seams of the three-dimensional clothing model, the designed customized clothing can be more fitted and comfortable to wear.

[0061] In the specific implementation process, fixed seams at fixed positions can be generated in advance according to the specific types of clothing in the customer's customization requirements. For example, the fixed seams of a dress include shoulder seams (located at the top of the shoulder, extending from the neckline to the cuffs), side seams (along the side of the body, extending from the armpit to the bottom of the skirt), waist seams (usually located at the waistline, connecting the top part and the skirt part), etc. The position of the seams plays a vital role in determining the style and fit of the clothing, and the fixed position seams are determined based on the customer's body model.

[0062] The specific implementation steps of the iterative optimization operation can be as follows: optimize the seams according to the clothing model generated by the clothing generation module, specifically, fit the seams of the three-dimensional clothing model with the fixed seams generated by the customer body model through iterative optimization, so as to minimize the difference in the three-dimensional positions of the two:

[0063]

[0064] in, is the loss function, which specifically represents the distance between the two suture points; To fix the position of the suture point; is the weight; Indicates the position of the sewing points of the three-dimensional clothing model; Represents the set of fixed seam points. The final three-dimensional clothing model is obtained through iterative optimization through the above operations.

[0065] Step S108, projecting the cropped and optimized 3D clothing model to obtain a 2D clothing image, splitting the 2D clothing image and automatically typeset it. Specifically, the 3D clothing model and the human joints can be projected back to the 2D plane image (front and back) through the posenet preset camera model. Finally, the 2D plane image is split according to the joints on the 2D plane image and automatically typeset to obtain a layout image.

[0066] The intelligent automatic layout and cutting optimization method for clothing provided in this embodiment can realize intelligent automatic layout and cutting optimization, and can optimally arrange and design fabrics according to order requirements, minimize the waste of scraps, and improve the utilization rate and cutting accuracy of fabrics. Through the method of this embodiment, it is also possible to dynamically adapt to the complex requirements of different clothing orders, support rapid conversion of multiple sizes and styles, and optimize layout through real-time data analysis and feedback, shortening the production cycle, reducing labor costs, and improving the production efficiency of layout drawings. The method of cutting optimization based on the customer's body model further avoids the possibility of fabric waste.

[0067] This embodiment also provides a clothing intelligent automatic typesetting and cutting optimization system, which is used to implement the above-mentioned clothing intelligent automatic typesetting and cutting optimization method. The content that has been detailed in the method will not be repeated in this system. The system is mainly composed of a general production section and a personal customization section. The general production section is responsible for the automatic classification of the sizes of the fixed-size production line and generates the optimal production size based on historical data. The personal customization section generates a body model through customer body data, and produces a 3D three-dimensional clothing model based on the body model and combined with the 2D plane diagram of the clothing style, which is disassembled into a layout diagram after optimization of the suture line.

[0068] like Figure 2 As shown, the clothing intelligent automatic typesetting and cutting optimization system 20 of this embodiment specifically includes: a starting module 201, a general production module 202 and a personal customization module 203.

[0069] The startup module 201 is used to obtain clothing design requirements, and judge whether the clothing design requirements are general production requirements or personal customization requirements according to preset rules. If the clothing design requirements are general production requirements, the general production module 202 is started, and if the clothing design requirements are personal customization requirements, the personal customization module 203 is started; specifically, the startup module 201 can be implemented by combining the software and hardware of the system. On the one hand, it can accept the user's input and feedback the input results to the user, and on the other hand, it can control the startup of the general production module 202 and the personal customization module 203. For the acquisition of clothing design requirements, the clothing design requirements can be obtained by obtaining orders, and the orders can include specific clothing types, design styles, order quantities, and customer information (height, weight, body shape, size, age, etc.). For how the startup module 201 determines the order type of clothing design requirements, the startup module 201 can directly input or select whether it is a general production requirement or a personal customization requirement through the system; it can also be judged by detecting preset fields, for example, by detecting the order quantity and customer information in the clothing design requirements. General production needs generally include orders with large quantities and low requirements for size matching accuracy, such as company team building uniforms, school uniforms, corporate uniforms, etc. Personal customization needs include orders for customized production lines with small quantities. Customers of orders with personal customization needs may have special requirements for tailoring (such as: slim fit, loose fit, etc.).

[0070] like Figure 3 As shown, the general production module 202 is responsible for the automatic classification and layout of the sizes of the fixed-size production line. The general production module 202 may include an automatic classification unit 2021 and an automatic layout unit 2022.

[0071] The automatic classification unit 2021 is used to obtain an individual information set according to general production requirements, and classify the individual information set by size using a first preset algorithm model; specifically, the individual information set contains customer information. The automatic classification unit 2021 can set the preset sizes S, M, and L according to the general clothing sizes (if necessary, the size categories can be increased, depending on the order requirements), and classify the customer information by size. Different clothing types can be classified in different ways. Since the clothing parameters specified by each size are not a fixed value, but an interval, when typesetting, production should be carried out as much as possible according to the clothing parameters of the largest number of people in each size in the order, so that fabric can be saved while improving the wearing experience of most people. For the design of the size chart, size classification parameters similar to those in Tables 1, 2, and 3 can be used.

[0072] In an optional embodiment, the individual information set includes multiple groups of individual characteristic information, each group of individual characteristic information includes at least the age, weight and height of the individual; the individual information set is categorized by size using the first preset algorithm model, including: using the regression algorithm model to calculate the individual characteristic information, obtain the size category corresponding to each individual characteristic information, and categorize the individual characteristic information by size. Specifically, since the standard body parameters (sleeve length, shoulder width, hip circumference, etc.) are inconvenient to measure, and the number of customers is inconvenient to count, it may not be possible to obtain accurate body parameters. Therefore, this embodiment can also only collect age, weight, and height information, and determine the size category corresponding to the customer through age, weight, and height information. Of course, this embodiment does not exclude the implementation method in which the individual characteristic information directly includes standard body parameters. The purpose of introducing age is to take into account that the waist circumference of some male customers may increase with age, but the weight will not change too much. If only height and weight are used, it may lead to inaccurate waist estimation. The regression algorithm model can be used to analyze and predict the customer's body shape using the individual characteristic information obtained. By using individual characteristic information (such as height, weight, and age, etc.) as variables to predict the results (body parameters), more accurate body shape predictions can be achieved.

[0073] In a specific optional implementation, the automatic classification unit 2021 can perform classification through the corresponding steps detailed in the aforementioned clothing intelligent automatic typesetting and cutting optimization method, which will not be repeated here.

[0074] The automatic typesetting unit 2022 is used to determine the clothing typesetting parameters corresponding to each size according to the individual information set and the historical reference database, and perform automatic typesetting using the clothing typesetting parameters; specifically, after determining the size of each customer, it is necessary to select the corresponding clothing typesetting parameters for the size, as shown in Tables 1, 2 and 3. An existing size table can be used as a historical reference database, and the size corresponding to each size is selected from the size range of the existing size table as the clothing typesetting parameter. A specific body parameter historical reference database can also be established, which stores body parameters corresponding to weight and height collected during past clothing production.

[0075] In an optional embodiment, the clothing layout parameters corresponding to each size are determined according to the individual information set and the historical reference database, including: counting the occurrence frequency of each individual feature information, determining the individual feature information with the highest frequency in each size; searching and matching from the historical reference database according to the individual feature information with the highest frequency to obtain the general body parameters corresponding to the size; and obtaining the clothing layout parameters corresponding to the size according to the general body parameters. Specifically, after the size category of each person is known, the body parameters of the most people in the order can be estimated using the body parameter historical reference database, specifically counting the most weight and height in each size category, and then calculating the average body parameter corresponding to the height and weight matched in the historical reference database, and using the body parameter as the clothing layout parameter. In the actual process, each individual feature information in each size may appear only once, or there may be multiple individual feature information with the highest frequency, then multiple general body parameters will be obtained, and the average value of multiple general body parameters can be calculated at this time. By counting the individual feature information with the highest frequency to find and match the corresponding body parameters, the produced clothing products can be made to fit the body shape of most customers as much as possible. Alternatively, in the historical reference database, a human body feature information may correspond to multiple universal body parameters, in which case it is also necessary to calculate the average value of the universal body parameters. By counting the individual feature information with the highest frequency to find the matching corresponding body parameters, the produced clothing products can be made to fit the body shapes of most customers as much as possible.

[0076] like Figure 4 As shown, the personal customization module 203 includes: a customer body fitting unit 2031, a clothing generation unit 2032, a cutting optimization unit 2033 and a layout unit 2034. Since the customer's body parameters will be constantly adjusted during the personal customization design process, and the customer may have special requirements for tailoring (such as: slim fit, loose fit, etc.), the traditional layout process will become complicated and lengthy, and the use of the personal customization module of this embodiment can reduce labor consumption and improve efficiency.

[0077] The customer body fitting unit 2031 is used to obtain customized human body information according to personal customization needs, and use the second preset algorithm model to fit the customized human body information to obtain a customized body stereoscopic model; specifically, the second preset algorithm model may include multiple algorithm operation steps. In an optional embodiment, the customized human body information may at least include: a customized human body image and customized body measurement parameters. The customer body fitting unit 2031 can obtain a customized human body image including a multi-directional or multi-angle photographic image of the customer's body. For example, it can obtain a front image, a left side image, a right side image, a back image and other views of the customer in a T-pose (T-Pose), and the customized body measurement parameters can be manually measured body parameters (which can be measured according to the corresponding parameter format of Table 1, Table 2 and Table 3 above according to the target clothing type). The customized body stereoscopic model can be a customized body 3D model. Through the views of the human body in multiple directions, a stereoscopic 3D full view of the customer's customized body can be obtained.

[0078] In an optional embodiment, the customized human body shape information is fitted using a second preset algorithm model to obtain a customized body stereo model, including: using a pre-trained human body segmentation model to segment the customized human body image to obtain a human body mask image; using a human body key point inspection pre-trained model to identify and calculate the human body joint nodes in the human body mask image to obtain a customized body initial model; using customized body measurement parameters to correct the customized body initial model to obtain a customized body stereo model. The customer body fitting unit 2031 first segments the customized human body image to obtain a human body mask, which can accurately separate the object in the image from the background, facilitating the subsequent identification and processing of the key points of the image. After obtaining the initial model, the initial model is corrected using the measurement parameters, which can further reduce the error, so that the customized body stereo model finally obtained is closer to the customer's body.

[0079] In a specific optional implementation, the customer body fitting unit 2031 can process the customized human body image data through the steps detailed in the aforementioned clothing intelligent automatic typesetting and cutting optimization method to obtain a customized body three-dimensional model, which will not be repeated here.

[0080] The clothing generation unit 2032 is used to obtain customized clothing style features and generate a three-dimensional clothing model based on the customized clothing style features and the customized body three-dimensional model; specifically, the clothing generation unit 2032 can select a 2D design drawing of a corresponding style based on the clothing style features selected by the customer, and combine the 2D design drawing with the customized body three-dimensional model to produce a three-dimensional clothing model.

[0081] In an optional embodiment, obtaining customized clothing style features, and generating a three-dimensional clothing model based on the customized clothing style features and the customized body three-dimensional model includes: receiving input customized clothing style features, obtaining a corresponding plane design drawing based on the customized clothing style features; generating a clothing skin based on the grid of the customized body three-dimensional model; projecting the plane design drawing onto the customized body three-dimensional model using the coordinates of the human body joint nodes as the alignment standard, and merging the projected layer with the clothing skin to obtain a three-dimensional clothing model. The three-dimensional clothing model can be a three-dimensional 3D model. By means of a 3D clothing model, compared with a 2D plane design drawing, clothing can be designed more comprehensively and accurately according to the customer's body characteristics, so that the designed customized clothing is more suitable for the customer's body shape.

[0082] Specifically, the customer's customized body stereoscopic model has been obtained through the operation of the customer body fitting unit 2031. Each intersection of the grid in the model has a 3D coordinate. At this time, a new grid is generated according to a fixed threshold value from each intersection. This grid will be used as a clothing skin (the clothing skin is a simulated cloth). At this time, the clothing generation unit 2032 inputs the 2D design drawing corresponding to the clothing style selected by the customer. The 2D design drawing includes the front view and the back view of the clothing. There is no special requirement for the size unit of the 2D design drawing. A specific process of generating a three-dimensional clothing model has been detailed in the aforementioned method and will not be repeated here.

[0083] The cutting optimization unit 2033 is used to perform cutting optimization on the three-dimensional clothing model through an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting optimization; specifically, after the clothing generation unit 2032 obtains the three-dimensional clothing model, the three-dimensional clothing model needs to be further cut and optimized. The cutting optimization unit 2033 can usually perform cutting optimization by optimizing the seams.

[0084] In an optional implementation, the intelligent automatic layout and cutting optimization method of clothing of this embodiment also includes: generating fixed seams on the clothing skin according to the style characteristics of the customized clothing; cutting optimization of the three-dimensional clothing model through an iterative optimization algorithm includes: performing a fitting operation on the seams of the three-dimensional clothing model and the fixed seams through an iterative optimization algorithm to perform cutting optimization, wherein the fitting operation is used to minimize the position difference between the seams of the three-dimensional clothing model and the fixed seams. By optimizing the seams of the three-dimensional clothing model, the designed customized clothing can be more fitted and comfortable to wear.

[0085] In the specific implementation process, fixed seams of inherent positions can be generated in advance according to the specific type of clothing in the customer's customization requirements. For example, the fixed seam positions of dresses include shoulder seams (located at the top of the shoulder, extending from the neckline to the cuffs), side seams (along the side of the body, extending from the armpit to the bottom of the skirt), waist seams (usually located at the waistline, connecting the top part and the skirt part), etc. The position of the seams plays a vital role in determining the style and fit of the clothing. The inherent position seams are determined based on the customer's body model. The specific implementation steps of the iterative optimization operation have been detailed in the aforementioned method and will not be repeated here.

[0086] The layout unit 2034 is used to project the cropped and optimized three-dimensional clothing model to obtain a two-dimensional clothing image, split the two-dimensional clothing image and automatically layout it. Specifically, the three-dimensional clothing model and the human body joints can be projected back to the 2D plane image (front and back) through the posenet preset camera model. Finally, the 2D plane image is split according to the joints on the 2D plane image and automatically laid out to obtain a layout image.

[0087] The intelligent automatic layout and cutting optimization system for clothing provided in this embodiment realizes intelligent automatic layout and cutting optimization by integrating advanced algorithms and automation equipment. It can optimally arrange and design fabrics according to order requirements, minimize waste of scraps, and improve fabric utilization and cutting accuracy. Through the system of this embodiment, it is also possible to dynamically adapt to the complex requirements of different orders, support the conversion of multiple sizes and styles, and optimize layout through real-time data analysis and feedback, shortening the production cycle, reducing labor costs, and improving the production efficiency of layout drawings. The cutting optimization method based on the customer's body model further avoids the possibility of fabric waste.

[0088] In the description of the embodiments of the present invention, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "center", "top", "bottom", "top", "bottom", "inside", "outside", "inner side", "outer side" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. Among them, "inside" refers to an internal or enclosed area or space. "Periphery" refers to the area surrounding a specific component or a specific area.

[0089] In the description of the embodiments of the present invention, the terms "first", "second", "third", and "fourth" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", "third", and "fourth" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0090] In the description of the embodiments of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "install", "connect", "connect", and "assemble" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0091] In the description of the embodiments of the present invention, specific features, structures, materials or characteristics may be combined in a suitable manner in any one or more embodiments or examples.

[0092] In the description of the embodiments of the present invention, it should be understood that "-" and "~" represent a range between two values, and the range includes the endpoints. For example: "AB" represents a range greater than or equal to A and less than or equal to B. "A~B" represents a range greater than or equal to A and less than or equal to B.

[0093] In the description of the embodiments of the present invention, the term "and / or" herein is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " herein generally indicates that the associated objects before and after are in an "or" relationship.

[0094] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent automatic typesetting and cutting optimization of clothing, characterized in that: include: Step S101, obtaining clothing design requirements; Step S102, judging whether the clothing design requirement is a general production requirement or a personal customization requirement according to a preset rule, if the clothing design requirement is the general production requirement, executing step S103, if the clothing design requirement is the personal customization requirement, executing step S105; Step S103, obtaining an individual information set according to the general production requirements, calculating the individual characteristic information using a regression algorithm model, obtaining a size category corresponding to each individual characteristic information, and classifying the individual characteristic information according to the size, wherein the individual information set includes multiple groups of individual characteristic information, and each group of individual characteristic information includes at least the age, weight and height of the individual; Step S104, determining the clothing layout parameters corresponding to each of the sizes according to the individual information set and the historical reference database, and automatically performing layout using the clothing layout parameters; Step S105, obtaining customized human body shape information according to the personal customization requirements, and fitting the customized human body shape information using a second preset algorithm model to obtain a customized body three-dimensional model; Step S106, obtaining the customized clothing style features, and generating a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model; Step S107, performing cutting optimization on the three-dimensional clothing model through an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting optimization; Step S108, projecting the three-dimensional clothing model after cutting and optimization to obtain a two-dimensional clothing image, splitting the two-dimensional clothing image and automatically typeset it.

2. The intelligent automatic typesetting and cutting optimization method for clothing according to claim 1 is characterized in that: The determining of the clothing layout parameters corresponding to each size according to the individual information set and the historical reference database includes: Counting the occurrence frequency of each of the individual characteristic information, and determining the individual characteristic information with the highest frequency in each of the sizes; According to the individual characteristic information with the highest frequency, searching for matches in the historical reference database to obtain universal body parameters corresponding to the size; The clothing layout parameters corresponding to the size are obtained according to the general body parameters.

3. The intelligent automatic typesetting and cutting optimization method for clothing according to claim 1 is characterized in that: The customized human body information at least includes: a customized human body image and customized body measurement parameters; The method of fitting the customized human body shape information using the second preset algorithm model to obtain a customized body three-dimensional model includes: Using a pre-trained human body segmentation model, segmenting the customized human body image to obtain a human body mask image; Using a human key point inspection pre-trained model to identify and calculate human joint nodes in the human mask image to obtain a customized body initial model; The customized body measurement parameters are used to modify the customized body initial model to obtain the customized body three-dimensional model.

4. The intelligent automatic typesetting and cutting optimization method for clothing according to claim 3 is characterized in that: The step of obtaining the customized clothing style features and generating a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model includes: Receiving the input of the customized clothing style features, and acquiring a corresponding plane design drawing according to the customized clothing style features; Generating a clothing skin according to the mesh of the customized body stereo model; The plane design drawing is projected onto the customized body three-dimensional model using the coordinates of the human body joint nodes as an alignment standard, and the projected layer is fused with the clothing skin to obtain the three-dimensional clothing model.

5. The intelligent automatic typesetting and cutting optimization method for clothing according to claim 4 is characterized in that: The method further includes: generating fixed stitches on the garment skin according to the custom garment style features; The step of performing cutting optimization on the three-dimensional clothing model by an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting optimization comprises: performing a fitting operation on the seams of the three-dimensional clothing model and the fixed seams by an iterative optimization algorithm to perform cutting optimization, wherein the fitting operation is used to minimize the position difference between the seams of the three-dimensional clothing model and the fixed seams.

6. An intelligent automatic typesetting and cutting optimization system for clothing, characterized in that: include: Start-up module, general production module and personal customization module; The startup module is used to obtain clothing design requirements, determine whether the clothing design requirements are general production requirements or personal customization requirements according to preset rules, and if the clothing design requirements are general production requirements, start the general production module; if the clothing design requirements are personal customization requirements, start the personal customization module; The general production module includes an automatic classification unit and an automatic typesetting unit; The automatic classification unit is used to obtain an individual information set according to the general production requirements, calculate the individual characteristic information using a regression algorithm model, obtain a size category corresponding to each individual characteristic information, and classify the individual characteristic information according to the size, wherein the individual information set includes multiple groups of individual characteristic information, and each group of individual characteristic information includes at least the age, weight and height of the individual; The automatic typesetting unit is used to determine the clothing typesetting parameters corresponding to each of the sizes according to the individual information set and the historical reference database, and perform automatic typesetting using the clothing typesetting parameters; The personal customization module includes: a customer body fitting unit, a clothing generation unit, a cutting optimization unit and a typesetting unit; The customer body fitting unit is used to obtain customized human body information according to the personal customization requirements, and use a second preset algorithm model to fit the customized human body information to obtain a customized body three-dimensional model; The clothing generation unit is used to obtain the customized clothing style features and generate a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model; The cutting optimization unit is used to perform cutting optimization on the three-dimensional clothing model through an iterative optimization algorithm to obtain the three-dimensional clothing model after cutting optimization; The layout unit is used to project the three-dimensional clothing model after cutting optimization to obtain a two-dimensional clothing image, split the two-dimensional clothing image and automatically layout it.

7. The intelligent automatic typesetting and cutting optimization system for clothing according to claim 6 is characterized in that: The customized human body information at least includes: a customized human body image and customized body measurement parameters; The method of fitting the customized human body shape information using the second preset algorithm model to obtain a customized body three-dimensional model includes: Using a pre-trained human body segmentation model, segmenting the customized human body image to obtain a human body mask image; Using a human key point inspection pre-trained model to identify and calculate human joint nodes in the human mask image to obtain a customized body initial model; The customized body measurement parameters are used to modify the customized body initial model to obtain the customized body three-dimensional model.

8. The intelligent automatic typesetting and cutting optimization system for clothing according to claim 7 is characterized in that: The step of obtaining the customized clothing style features and generating a three-dimensional clothing model according to the customized clothing style features and the customized body three-dimensional model includes: Receiving the input of the customized clothing style features, and acquiring a corresponding plane design drawing according to the customized clothing style features; Generating a clothing skin according to the mesh of the customized body stereo model; The plane design drawing is projected onto the customized body three-dimensional model using the coordinates of the human body joint nodes as an alignment standard, and the projected layer is fused with the clothing skin to obtain the three-dimensional clothing model.

Images