CN111950139A - A clothing grading method and terminal based on objective function optimization - Google Patents

Abstract

The invention provides a garment grading method based on objective function optimization and a terminal, wherein the grading method can be used for finally obtaining grading results of human bodies with different body types by generating initial guess solutions of garments and paper patterns, physical simulation, objective function optimization, triple iteration and the like for given garment models with any types and human body models with any postures and body types. The triple iteration optimization method used by the invention can dynamically adjust the optimization precision, simultaneously reduces the gradient backtracking brought by direct interpolation, increases the optimization stability and improves the operation efficiency.

Description

A clothing grading method and terminal based on objective function optimization

technical field

The invention relates to the fields of computer graphics, cloth simulation, clothing grading and the like, in particular to a clothing grading method and a terminal based on objective function optimization.

Background technique

Grading a given 3D garment onto a 3D human body of any size is of great value, especially in the development of digital technology in the garment industry. The traditional CAD computer-aided design software grading method improves the correctness and reusability of the design level, but this process is very time-consuming and labor-intensive, and it is very dependent on the user's experience.

Most of the existing automatic grading methods are still based on simple 3D-2D coordinate transformation. Specifically, such methods "press" 3D garments into 2D through mapping between 3D coordinate space and 2D coordinate space. coordinate space. However, these methods do not take into account the physical drape effect of clothing.

To solve this problem, Bartle et al. proposed a physics-driven fixed-point optimization method (see Bartle A, Sheffer A, Kim V G, et al. Physics-driven pattern adjustment for direct 3Dgarment editing[J]. ACM Transactions on Graphics (TOG ), 2016, 35(4):50:1-50:11.), however, this method ignores the friction between the clothing and the human body.

-Alpes,Universitéde Grenoble,2016:1-11.)然而该方法使用了逆向弹性设计，存在无法收敛的情况。Subsequently, Casati et al. proposed a physical method that considered the friction between clothing and the human body (see Casati R, Daviet G, Bertails-Descoubes F. Inverse elastic cloth design with contact and friction [D]. Inria Grenoble

-Alpes, Université de Grenoble, 2016: 1-11.) However, this method uses an inverse elastic design, and there is a situation where it cannot converge.

Recently, Wang H (see Rule-free sewing pattern adjustment with precision and efficiency [J]. ACM Transactions on Graphics(TOG), 2018, 37(4): 1-13.) proposed a self-organized high-precision fast This method considers the drape effect of clothing and the simultaneous changes of clothing and pattern in the process of objective function optimization. However, this method has the phenomenon of optimization instability caused by gradient backtracking.

In the future garment factory, there will be thousands of sets of clothing models, the designer will input the 3D human model, select a set of clothing, and the system will output the grading result. This process needs to be fast and fully automatic, otherwise it will affect design efficiency. Therefore, how to quickly and automatically grade a given garment onto mannequins of different body types is an important and valuable problem in the field of garment grading.

SUMMARY OF THE INVENTION

In view of the defects in the prior art, the present invention proposes a clothing grading method and terminal optimized based on an objective function. For given various types of clothing models and clothing patterns, this solution can quickly grading them to different Body shape on a 3D mannequin.

Specifically, the present invention includes the following embodiments:

The embodiment of the present invention proposes a clothing grading method optimized based on an objective function, including:

Step 1. Obtain a three-dimensional standard clothing model, a two-dimensional standard clothing pattern, a three-dimensional standard human body model and a three-dimensional target human body model;

Step 2, determining the initial clothing data of the three-dimensional target clothing based on the three-dimensional standard clothing model, the three-dimensional standard human body model and the three-dimensional target human body model;

Step 3, using the clothing data and the preset objective function to determine the initial pattern data of the two-dimensional target clothing pattern;

Step 4, updating the clothing data of the 3D target clothing by means of quasi-static energy simulation through the current latest pattern data and the 3D target human body model;

Step 5. Update the boundary coordinates of the two-dimensional target clothing pattern through the current latest clothing data, the current latest pattern data, the three-dimensional target human body model and the objective function;

Step 6, update the internal coordinates of the two-dimensional target clothing pattern through the current latest clothing data, the current latest boundary coordinates and the objective function;

Step 7. Update the pattern data of the two-dimensional target garment pattern based on the boundary coordinates and the internal coordinates, so as to perform the iterative operation from step 4 to step 7 based on the latest current pattern data, and select a preset iteration The pattern data after the number of times is used as the garment grading result.

In a specific embodiment,

The three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model are all models composed of meshes.

In a specific embodiment, the grid is a triangular grid or a quadrilateral grid

In a specific embodiment,

The step 2 specifically includes:

Determine the relative distance between each grid on the three-dimensional standard clothing model and a preset point on the three-dimensional standard human body model; the preset point is the point on the three-dimensional standard human body model with the closest distance to the grid;

The relative distances are superimposed on the three-dimensional target human body model to generate initial clothing data of the three-dimensional target clothing.

In a specific embodiment,

The step 3 specifically includes:

The initial clothing data is used as a fixed value in the preset objective function to generate the initial pattern data of the two-dimensional target clothing pattern.

In a specific embodiment, the objective function is used to describe the difference between the target result and the standard result in six aspects: pattern size, pattern shape, garment-to-body distance, garment stretch, garment position, and garment smoothness;

The smaller the value of the objective function, the higher the matching degree between the clothing and the human body;

In the objective function, the data of the three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model are fixed values. The data of the two-dimensional target garment pattern are variables.

In a specific embodiment, the method of quasi-static energy simulation is performed through an energy formula;

The energy formula is: E _all =E _stretching +E _sticking +E _bending +E _collision +E _gravity ;

Among them, E _all is the total energy; E _stretching is the elastic stretching energy, which is used to describe the stretching effect of clothing fabrics; E _sticking is the _sticking energy, which is used to describe the style of clothing accessories; The cloth penetrates each other; E _gravity is the gravitational potential energy, which describes the gravitational effect of the cloth.

In a specific embodiment, the step 6 specifically includes:

The current latest clothing data is used as the quantity in the objective function, and the internal coordinates of the two-dimensional target clothing pattern are updated in combination with the current latest boundary coordinates.

In a specific embodiment, the preset number of iterations is set based on precision requirements.

An embodiment of the present invention further provides a terminal, including: a processor for executing the above method.

With this, the present invention has the following technical effects:

The grading method provided by the present invention can finally obtain a given clothing model of any type, a human body model with any posture and body shape, through the steps of generating the initial guessing solution of clothing and pattern, physical simulation, objective function optimization, triple iteration, etc. The grading results of clothing on different body types. The triple iterative optimization method used in the present invention can dynamically adjust the accuracy of optimization, at the same time reduce the gradient backtracking caused by direct interpolation, increase the stability of the optimization, and improve the operation efficiency.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings used in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore do not It should be regarded as a limitation of the scope, and for those of ordinary skill in the art, other related drawings can also be obtained according to these drawings without any creative effort.

1 is a schematic flowchart of a clothing grading method based on objective function optimization proposed by an embodiment of the present invention;

FIG. 2 is a schematic structural diagram of a terminal according to an embodiment of the present invention.

Detailed ways

Hereinafter, various embodiments of the present disclosure will be described more fully. The present disclosure is capable of various embodiments, and adaptations and changes may be made therein. It should be understood, however, that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but the present disclosure should be construed to cover various embodiments falling within the spirit and scope of the present disclosure. All adjustments, equivalents and/or alternatives.

The terminology used in the various embodiments of the present disclosure is for the purpose of describing particular embodiments only and is not intended to limit the various embodiments of the present disclosure. As used herein, the singular is intended to include the plural as well, unless the context clearly dictates otherwise. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which various embodiments of this disclosure belong. The terms (such as those defined in commonly used dictionaries) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized or overly formal meaning, unless explicitly defined in various embodiments of the present disclosure.

Example 1

Embodiment 1 of the present invention discloses a clothing grading method based on objective function optimization, as shown in FIG. 1 , including the following steps:

Step 1. Obtain a three-dimensional standard clothing model, a two-dimensional standard clothing pattern, a three-dimensional standard human body model and a three-dimensional target human body model;

Specifically, the three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model are all models composed of meshes. Further, the grid may be a triangular grid or a quadrilateral grid. Specifically, clothing and patterns can be generated by professional clothing software such as CLO 3D and Marvelous Designer, and clothing can also be generated by scanning real clothing with a 3D scanner. The 3D target human body model is generated by scanning the real human body with a 3D scanner, or using professional software such as MakeHuman and Maya.

The three-dimensional standard clothing model is a three-dimensional model including at least one piece of clothing. The three-dimensional clothing model corresponds to one piece or a combination of at least two pieces selected from tops, trousers or skirts. The three-dimensional clothing model may correspond to a single-layer piece or a single-layer piece, such as a jacket, trousers or skirt, a jacket and a pair of trousers, and the like. 3D standard mannequins can be in any pose and body shape. Arbitrary posture and body type means having different postures and different height, weight or measurements. The three-dimensional target human body model in step 1 has the same posture as the three-dimensional standard human body model in step 1, but has a different body shape.

Step 2, determining the initial clothing data of the three-dimensional target clothing based on the three-dimensional standard clothing model, the three-dimensional standard human body model and the three-dimensional target human body model;

Specifically, the step 2 specifically includes: determining the relative distance between each grid on the three-dimensional standard clothing model and a preset point on the three-dimensional standard human body model; the preset point is the distance between the three-dimensional standard human body model and the three-dimensional standard human body model. The grid distance is the closest point; the relative distances are superimposed on the three-dimensional target human body model to generate the initial clothing data of the three-dimensional target clothing.

The initial clothing data of the 3D target clothing is generated by superimposing the relative distance between the 3D standard clothing model and the 3D standard human body model input in step 1 onto the 3D target human body model. This approach can make full use of the relative distance between the standard clothing and the standard human body, and the generated initial guess value (that is, the initial clothing data of the three-dimensional target clothing) is used as the initial value of the iterative operation, and the corresponding objective function is small, reducing the convergence operation. amount, reducing operation time.

Step 3, using the clothing data and the preset objective function to determine the initial pattern data of the two-dimensional target clothing pattern;

Step 3 is done by running a simplified objective function optimization. In the complete objective function used in steps 3 and 5, the three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model are fixed values, and the three-dimensional target clothing model and the two-dimensional target clothing pattern are variables. The simplification in step 3 means that the variable three-dimensional target clothing model in the complete objective function is regarded as a fixed value, and the value is the initial guess value generated in step 2, and the generated two-dimensional target clothing pattern is used as a two-dimensional clothing pattern. The guess value of (that is, the initial pattern data).

Specifically, the step 3 specifically includes: taking the initial clothing data as a fixed value in a preset objective function to generate the initial pattern data of the two-dimensional target clothing pattern. Among them, the objective function is used to describe the difference between the target result and the standard result in six aspects: pattern size, pattern shape, clothing-to-body distance, clothing stretching degree, clothing position, and clothing smoothness; small, it means that the matching degree between clothing and human body is higher; in the objective function, the data of the three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model As a fixed value, the data of both the three-dimensional target garment and the two-dimensional target garment pattern are variables.

Specifically, the objective function f has the form:

f=C _change +C _shape +C _dist +C _stretch +C _position +C _smooth

Among them, the formula of each component of the objective function is:

分别为二维目标服装纸样与二维标准服装纸样对应边界点，具体的对应的边界点为二维向量；Among them, σ _change is the parameter distance weight coefficient, i is the number of all boundary points of the pattern, Ψ _i is the same as the

are the corresponding boundary points of the two-dimensional target garment pattern and the two-dimensional standard garment pattern respectively, and the specific corresponding boundary points are two-dimensional vectors;

Among them, σ _shape is the weight coefficient of the boundary shape, e _i,j is all adjacent edges of the pattern boundary, Δθ is the difference between the adjacent edges of the two-dimensional target garment pattern and the two-dimensional standard garment pattern.

为三维标准服装模型网格点与三维标准人体模型表面最近邻距离。Among them, σ _dist is the weight coefficient of clothing distance, n is the number of all grid points of the clothing model, dist(X _n ) is the nearest neighbor distance between the 3D target clothing grid point and the surface of the 3D target human body model,

is the nearest neighbor distance between the grid points of the 3D standard clothing model and the surface of the 3D standard human body model.

与

为三维标准服装网格点坐标，

与

为二维标准纸样网格点坐标。Among them, σ _stretch is the weight coefficient of clothing stretching, i, j are the two endpoints of each edge in the clothing model grid, X _i and X _j are the coordinates of the three-dimensional target clothing grid point, and P _i and P _j are two-dimensional Target pattern grid point coordinates,

and

is the coordinate of the three-dimensional standard clothing grid point,

and

It is the grid point coordinates of the two-dimensional standard pattern.

为步骤(2)中初始猜测值(也即初始的服装数据)的网格点坐标。Among them, σ _position is the weight coefficient of clothing position, i is all grid points of the clothing model, X _i is the coordinates of the three-dimensional target clothing grid point,

is the grid point coordinates of the initial guess value (ie the initial clothing data) in step (2).

分别为三维目标服装与三维标准服装模型的网格点坐标，c为表征三维目标服装与三维标准服装模型两者凹凸属性的值，当两者凹凸属性相同时，c取1，当两者凹凸属性相反时，c取-1。三维目标服装与三维标准服装模型的光滑程度越接近，用来避免放码后服装光滑程度差异过大。Among them, σ _smooth is the weight coefficient of clothing smoothness, e ₀ is all hinge edges of the clothing model, the sum K ₀ is the discrete energy factor based on the Laplace harmonic value, X and

are the grid point coordinates of the 3D target clothing and the 3D standard clothing model respectively, and c is the value representing the concave and convex attributes of the 3D target clothing and the 3D standard clothing model. When the properties are opposite, c takes -1. The closer the smoothness of the 3D target clothing and the 3D standard clothing model is, is used to avoid too much difference in the smoothness of the clothing after grading.

The rationality of the objective function is that the objective function fully describes the difference between the target result and the standard result in six aspects: pattern size, pattern shape, clothing-body distance, clothing stretch, clothing position, and clothing smoothness.

Step 4, updating the clothing data of the 3D target clothing by means of quasi-static energy simulation through the current latest pattern data and the 3D target human body model;

The method of the quasi-static energy simulation is carried out through the energy formula;

The energy formula is: E _all =E _stretching +E _sticking +E _bending +E _collision +E _gravity ;

Among them, E _all is the total energy; E _stretching is the elastic stretching energy, which is used to describe the stretching effect of clothing fabrics; E _sticking is the _sticking energy, which is used to describe the style of clothing accessories; The cloth penetrates each other; E _gravity is the gravitational potential energy, which describes the gravitational effect of the cloth.

Specifically, the formula of each energy term is:

Among them, γ _sewing is the elastic energy coefficient of the sewing spring connecting different patterns of the pattern, X _si and X _sj are the coordinates of the three-dimensional clothing model corresponding to the two ends of the sewing spring, γ _patch is the elastic energy coefficient of the inner spring in the same pattern of the pattern, X _pn and X _pk are the three-dimensional clothing model coordinates corresponding to the two ends of the inner spring, and P _pn and P _pk are the two-dimensional clothing pattern coordinates corresponding to the two ends of the inner spring.

Among them, γ _sticking is the viscosity coefficient of the viscous potential energy, and X _sti and X _stj are the grid coordinates of the corresponding sticky points, respectively.

A₀和A₁分别为共用铰链边的三角形t₀和t₁的面积。K₀的公式为：K₀＝(cotα_ij+cotβ_ij)，K₀是一个四维向量，对于其中任一项，α_ij和β_ij为铰链边ij在共用它的两个三角形中的两个对角。Wherein, γ _bending is the energy coefficient of the bending discrete potential energy, and X=(X ₀ , X ₁ , X ₂ , X ₃ ) ^T , is a vector composed of coordinates of four grid points related to the discrete energy. The formula of Q _i,j is:

A ₀ and A ₁ are the areas of triangles t ₀ and t ₁ , respectively, that share a hinge side. The formula for K ₀ is: K ₀ =(cotα _ij +cotβ _ij ), where K ₀ is a four-dimensional vector, and for either term, α _ij and β _ij are the hinge edge ij in the two triangles that share it Diagonal.

Among them, γ _collision is the energy coefficient of the repulsive potential energy, and dist(X _i ) is the distance between the clothing grid point and the nearest neighbor human body model.

Among them, γ _gravity is the energy coefficient of gravitational potential energy, X _i is the coordinates of the clothing grid point, and _kg is the unit vector of the gravity direction.

Step 5. Update the boundary coordinates of the two-dimensional target clothing pattern through the current latest clothing data, the current latest pattern data, the three-dimensional target human body model and the objective function;

Specifically, the coordinates of the boundary points of the garment pattern are updated by running the complete objective function optimization. For the complete objective function, the 3D target clothing model and the 2D target clothing pattern are changed at the same time. In the actual optimization, the variable scale method, such as the L-BFGS method, can be used (see Liu DC, Nocedal J. On the limited memory BFGSmethod for large scale optimization[J].Mathematical programming,1989,45(1-3):503-528.), to avoid the direct calculation of the objective function Hessian matrix, in the solution, the calculation of the derivative can use the adjoint method.

Step 6, update the internal coordinates of the two-dimensional target clothing pattern through the current latest clothing data, the current latest boundary coordinates and the objective function;

The step 6 specifically includes: taking the current latest clothing data as the quantitative value in the objective function, and updating the internal coordinates of the two-dimensional target clothing pattern in combination with the current latest boundary coordinates. Specifically, the coordinates of the boundary points of the pattern obtained in step 5 are used as constants, and the simplified objective function is run to optimize and update the coordinates of the internal points of the garment pattern. Because after the optimization in step 5, the direct calculation of the coordinates of the internal points of the pattern requires the direct calculation of the higher-order equations, which consumes a lot of time, so when the coordinates of the boundary points of the target pattern are obtained in each iteration, it is regarded as a fixed value. , which is substituted into the complete objective function to obtain a simplified objective function. Compared with direct interpolation, the optimization of it can greatly reduce the gradient backtracking of step 5 in the next run, increase the stability of objective function optimization, and improve computational efficiency. .

In addition, steps 4-6 can be performed multiple iterations, so in the k+1th iteration:

The first baryon replacement, using the current two-dimensional target pattern coordinates P _k and the target human body Φ, through the quasi-static energy simulation process, the coordinates of the updated three-dimensional target clothing are X _k+1 (corresponding to step 4);

The second baryon replacement, using the updated coordinates X _k+1 of the 3D target clothing, the current 2D target pattern coordinates P _k and the 3D target human body model Φ, through the objective function optimization process, the update parameter is Ψ _k+1 , that is, the current Boundary coordinates of the 2D target pattern. In this optimization process, clothing X and pattern P all change with the parameter Ψ, following the principle of derivative transfer;

The third baryon substitution, using the updated coordinates X _k+1 of the 3D target garment, the updated boundary coordinates Ψ _k+1 of the 2D target pattern, and the 3D target body model Φ, this process treats the 3D target garment as a constant and runs a simplification In the optimization process, the internal point coordinates of the pattern are updated.

Subsequent, pattern generation, the pattern boundary and internal coordinates generated by the second baryon substitution and the third baryon substitution are combined as the pattern coordinate value P _k+1 in the next iteration.

Step 7. Update the pattern data of the two-dimensional target garment pattern based on the boundary coordinates and the internal coordinates, so as to perform the iterative operation from step 4 to step 7 based on the latest current pattern data, and select a preset iteration The pattern data after the number of times is used as the garment grading result. Specifically, the preset number of iterations is set based on accuracy requirements.

A clothing grading method based on objective function optimization provided by the present invention has the following technical effects: (1) The grading method provided by the present invention can assign a given garment model of any type, a human body model of any posture and body shape , through the steps of generating the initial guess solution of clothing and pattern, physical simulation, objective function optimization, triple sub-substitution and other steps, the grading results of clothing in different body types are finally obtained. (2) The triple substitution optimization method used in the present invention can dynamically adjust the accuracy of optimization, and at the same time reduce the gradient backtracking caused by direct interpolation, increase the stability of the optimization, and improve the operation efficiency.

Example 2

Embodiment 2 of the present invention further discloses a terminal 200, as shown in FIG. 2, including: a processor 201 for executing the method described in Embodiment 1. Specifically, this Embodiment 2 also discloses other technical features. For the description of the specific technical features, please refer to the description in Embodiment 1, and details are not repeated here.

Those skilled in the art can understand that the accompanying drawing is only a schematic diagram of a preferred implementation scenario, and the modules or processes in the accompanying drawing are not necessarily necessary to implement the present invention.

Those skilled in the art can understand that the modules in the device in the implementation scenario may be distributed in the device in the implementation scenario according to the description of the implementation scenario, or may be located in one or more devices different from the implementation scenario with corresponding changes. The modules of the above implementation scenarios may be combined into one module, or may be further split into multiple sub-modules.

The above serial numbers of the present invention are only for description, and do not represent the pros and cons of the implementation scenarios.

The above disclosures are only a few specific implementation scenarios of the present invention, however, the present invention is not limited thereto, and any changes that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims (10)

1. a clothing grading method optimized based on objective function, is characterized in that, comprises: Step 1. Obtain a three-dimensional standard clothing model, a two-dimensional standard clothing pattern, a three-dimensional standard human body model and a three-dimensional target human body model; Step 2, determining the initial clothing data of the three-dimensional target clothing based on the three-dimensional standard clothing model, the three-dimensional standard human body model and the three-dimensional target human body model; Step 3, using the clothing data and the preset objective function to determine the initial pattern data of the two-dimensional target clothing pattern; Step 4, updating the clothing data of the 3D target clothing by means of quasi-static energy simulation through the current latest pattern data and the 3D target human body model; Step 5. Update the boundary coordinates of the two-dimensional target clothing pattern through the current latest clothing data, the current latest pattern data, the three-dimensional target human body model and the objective function; Step 6, update the internal coordinates of the two-dimensional target clothing pattern through the current latest clothing data, the current latest boundary coordinates and the objective function; Step 7. Update the pattern data of the two-dimensional target garment pattern based on the boundary coordinates and the internal coordinates, so as to perform the iterative operation from step 4 to step 7 based on the latest current pattern data, and select a preset iteration The pattern data after the number of times is used as the garment grading result. 2. The method of claim 1, wherein The three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model are all models composed of meshes. 3. The method of claim 2, wherein the mesh is a triangular mesh or a quadrilateral mesh. 4. The method of claim 1 or 2, wherein The step 2 specifically includes: Determine the relative distance between each grid on the three-dimensional standard clothing model and a preset point on the three-dimensional standard human body model; the preset point is the point on the three-dimensional standard human body model with the closest distance to the grid; The relative distances are superimposed on the three-dimensional target human body model to generate initial clothing data of the three-dimensional target clothing. 5. The method of claim 1, wherein The step 3 specifically includes: The initial clothing data is used as a fixed value in the preset objective function to generate the initial pattern data of the two-dimensional target clothing pattern. 6. method as claimed in claim 1 is characterized in that, described objective function is used to describe target result and standard result in pattern size, pattern shape, garment-human body distance, garment stretch degree, garment position, garment smoothness degree six differences; The smaller the value of the objective function, the higher the matching degree between the clothing and the human body; In the objective function, the data of the three-dimensional standard clothing model, the two-dimensional standard clothing pattern, the three-dimensional standard human body model and the three-dimensional target human body model are fixed values. The data of the two-dimensional target garment pattern are variables. 7. The method of claim 1, wherein the quasi-static energy simulation mode is carried out by an energy formula; The energy formula is: E _all =E _stretching +E _sticking +E _bending +E _collision +E _gravity ; Among them, E _all is the total energy; E _stretching is the elastic stretching energy, which is used to describe the stretching effect of clothing fabrics; E _sticking is the _sticking energy, which is used to describe the style of clothing accessories; The cloth penetrates each other; E _gravity is the gravitational potential energy, which describes the gravitational effect of the cloth. 8. The method of claim 1, wherein the step 6 specifically comprises: The current latest clothing data is used as the quantity in the objective function, and the internal coordinates of the two-dimensional target clothing pattern are updated in combination with the current latest boundary coordinates. 9. The method of claim 1, wherein the preset number of iterations is set based on accuracy requirements. 10. A terminal, comprising: a processor for executing the method according to any one of claims 1-9.

Images