Personalized customization method and system for western-style clothes

Abstract

The invention relates to the technical field of computer vision and biomechanical simulation, and discloses a personalized customization method and system for western-style clothes, which breaks through the limitation that motion deformation cannot be captured by traditional static measurement by constructing a dynamic volumetric model containing skeleton driving and soft tissue deformation, and improves the customization efficiency of the western-style clothes. The muscle expansion effect during limb movement can be accurately simulated, the proportion of fabric stretching to body compression is scientifically distributed on the basis of a mechanical balance principle in combination with a unique two-way rigidity displacement complementary mechanism, the accurate compensation amount which meets the movement requirement and is not excessively loose is calculated, and in addition, the energy-minimization-based grid deformation technology is utilized, so that the energy consumption is reduced. When the dynamic allowance is invisibly injected into the model, the contour line of the original design and the sewing completeness are kept, so that the contradiction that slimming and comfort are difficult to achieve at the same time in the traditional customization is solved, the wearing experience of static fit and non-compression in movement is realized, and the dependence on the tailoring experience is greatly reduced.

Description

Technical Field

[0001] This application relates to the fields of computer vision and biomechanical simulation technology, and in particular to a method and system for personalized suit customization. Background Technology

[0002] As the textile and apparel industry transforms and upgrades towards intelligence and personalization, high-end bespoke suit services have become an important part of meeting consumers' demands for a quality lifestyle. Suits, as a formal wear category with a rigorous structure and extremely high requirements for pattern precision, are typically made from natural fibers such as wool and linen. While these fabrics possess excellent drape and shape retention, they often exhibit low or no elasticity, resulting in a very low tolerance for errors in body size. In current bespoke business processes, the data collection stage generally requires users to maintain a relatively static, natural upright posture or a standard open posture to obtain stable three-dimensional human body scanning models or two-dimensional manual measurement data. However, in the actual daily life of wearing suits, users face extremely complex and varied usage environments and movement scenarios, such as the back bending of business people while working at a desk, the outstretched arms while driving, and frequent arm raisings in social situations. These high-frequency dynamic scenarios require clothing to have corresponding freedom of movement and structural adaptability, not just visual aesthetics in a static state.

[0003] However, existing bespoke tailoring techniques have significant limitations in handling the mapping between static human body data and dynamic wearing requirements. Current pattern generation or grading technologies primarily rely on static geometric topology logic, that is, based on measured net body circumference data, a fixed standard ease allowance is superimposed according to statistical laws from a general human body database to determine the final garment size. This linear calculation method based on empirical constants completely ignores the nonlinear stretching effect of human bones and joints on the skin during movement. Specifically, when the human body transitions from a static state to a state of significant movement, the skin surface area in key areas such as the scapula, armpits, and elbows undergoes significant stretching deformation. The magnitude of this deformation varies from person to person and is more pronounced for individuals with specific body postures such as sloping shoulders or hunchbacks. Because existing technology cannot acquire and quantify this surface deformation data based on specific movement range, nor has it established a dynamic correlation model between movement range, fabric tension, and pattern correction amount, the resulting suits often fit well when at rest but produce a strong sense of pressure when moving, leading to functional failures such as underarm tearing and back tightness. This not only seriously reduces wearing comfort but also directly increases the after-sales repair rate and wastes production resources. Summary of the Invention

[0004] This application proposes a method and system for personalized suit customization to address the problems mentioned in the background art.

[0005] To achieve the above objectives, this application adopts the following technical solution: a method for personalized suit customization, comprising the following steps:

[0006] Step S1: Obtain static image data and basic physiological parameters of the target user, and construct a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on the static image data and basic physiological parameters. The dynamic volumetric human model has the ability to drive soft tissue deformation according to skeletal movement.

[0007] Step S2: Retrieve the standard motion library to drive the dynamic volumetric human model constructed in step S1 to perform joint rotation movements, extract the preset key measurement paths on the surface of the dynamic volumetric human model, and calculate the geodesic path stretching increment generated by the key measurement paths in motion relative to the static state.

[0008] Step S3: Obtain the tensile stiffness parameters of the target fabric, and estimate the user's soft tissue stiffness parameters based on the basic physiological parameters obtained in Step S1. Input the geodesic path tensile increment, tensile stiffness parameters and soft tissue stiffness parameters calculated in Step S2 into the bidirectional stiffness-displacement complementary model, and calculate the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue itself based on the principle of force balance.

[0009] Step S4: Map the pattern size compensation amount calculated in step S3 to the corresponding geometric area of ​​the preset original electronic pattern, perform mesh deformation processing on the original electronic pattern, and generate producible vector cutting data containing local dynamic allowances.

[0010] Furthermore, in step S1, the specific operation of constructing a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on static image data and basic physiological parameters is as follows:

[0011] A1, analyze basic physiological parameters to extract the target user's true height value, and set the true height value as the absolute physical scale anchor value under the three-dimensional Euclidean space measurement system;

[0012] A2. Construct a perspective projection inverse optimization model that includes camera focal length variables and camera spatial pose variables. Use the perspective projection inverse optimization model to project the initialized 3D parameterized human body template onto the 2D pixel coordinate plane where the static image data is located.

[0013] A3. During the iterative optimization operation, the reprojection position deviation between the projected feature points of the three-dimensional parameterized human body template and the visual key points extracted from the static image data is calculated. At the same time, a hard geometric constraint based on the rigid body scale invariance is introduced. The hard geometric constraint requires that the geodesic path length of the three-dimensional parameterized human body template from the top of the head to the bottom of the feet be strictly equal to the actual height.

[0014] A4 determines the unique camera focal length value and the absolute physical distance value of the target user relative to the camera by jointly solving the minimum cost function. Based on the camera focal length value and the absolute physical distance value, the three-dimensional parametric human body template is scaled to generate a dynamic volumetric human body model with absolute metric scale attributes.

[0015] Furthermore, in step S1, the specific operation of constructing a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on static image data and basic physiological parameters is as follows:

[0016] A5, after determining the skeletal spatial pose and body surface mesh position of the dynamic volumetric human model, performs soft tissue physical volume extraction calculation in accordance with the physical law of non-overlapping material space.

[0017] A6, traverse each geometric vertex on the surface mesh of the dynamically volumetric human body model, and calculate the Euclidean radial distance between each geometric vertex and the central axis of the nearest bone in the internal skeletal drive system.

[0018] A7 calls the preset anatomical statistics database to obtain the average rigid radius value of the bone corresponding to the current body part, performs the difference operation of subtracting the average rigid radius value of the bone from the radial distance value in Euclidean space, and defines the positive difference value obtained by the operation as the physical soft tissue thickness value at the geometric vertex.

[0019] A8 combines the body fat percentage value in the basic physiological parameters to mark the heterogeneity attribute of the region associated with the physical soft tissue thickness value, and introduces a biological non-penetrating boundary constraint mechanism. When the calculated physical soft tissue thickness value is less than the preset minimum physiological thickness threshold of the human dermis, the physical soft tissue thickness value at that location is forcibly replaced with the minimum physiological thickness threshold of the human dermis, and the finally determined thickness data field is mapped to the external soft tissue deformation layer.

[0020] Furthermore, in step S2, the specific operation of calling the standard motion library to drive the dynamically volumetric human model constructed in step S1 to perform joint rotational movements is as follows:

[0021] B1. Based on the incompressible fluid properties of biological soft tissue at the macroscopic physical level, a volume conservation coupling calculation logic is established between axial compressibility variables and radial expansion variables.

[0022] B2 uses motion posture data from the standard motion library to drive the skeleton drive system inside the dynamic volumetric human model to perform joint bending and rotation operations, and calculates the axial projection length of the current limb segment bone in the bending state in real time.

[0023] B3. Perform a comparison operation, dividing the axial projection length value by the original bone length value of the bone drive system in a static extended state, to obtain the axial compression ratio value of the bone, which represents the current degree of bone compression.

[0024] B4 calls the pre-stored physical soft tissue thickness value in the external soft tissue deformation layer, and uses the physical soft tissue thickness value as the basic gain factor. Combined with the preset Poisson's ratio coefficient, it performs a nonlinear inverse proportional expansion operation based on the volume conservation law on the bone axial compression ratio value, and calculates the forced radial displacement value caused by the shortening of the bone axial distance, which forces the soft tissue medium to be squeezed outward.

[0025] B5 superimposes the forced radial displacement values ​​along the normal direction onto the corresponding geometric vertex coordinates of the dynamic volumetric human body model's surface mesh, driving the external soft tissue deformation layer to undergo nonlinear geometric bulging deformation, generating dynamic deformation mesh data containing volume expansion characteristics.

[0026] Furthermore, in step S2, the specific operation of extracting the preset key measurement paths on the surface of the dynamically volumetric human body model and calculating the geodesic path stretching increment of the key measurement paths in motion relative to the static state is as follows:

[0027] B6, based on the physical characteristics of minimizing the energy of fabric materials under tension, configures the convex hull bridging integral algorithm to simulate the geometric coverage trajectory of suit fabric on the human body surface;

[0028] B7 extracts the preset key measurement path from the surface of the dynamic volumetric human body model that has undergone volume expansion and deformation, and traverses the discrete geometric nodes on the key measurement path to perform contact state determination calculation.

[0029] B8. When performing the contact state determination operation, if the fabric-covered area is determined to be a protruding part of the human body, the system marks the area as a contact state segment and performs differential geometric integration operation along the curvature direction of the human body surface to obtain the surface geodesic length value. If the fabric-covered area is determined to be a groove of the human spine or a bone depression, the system marks the area as a suspended bridging state segment and calculates the Euclidean distance between the two tangent points across the depression area.

[0030] B9 performs an accumulation operation, summing the surface geodesic length values ​​of all contact state segments with the Euclidean space straight-line distance values ​​of all suspended bridging state segments to obtain the total physical path length of the fabric in motion. Finally, it performs a subtraction operation, subtracting the corresponding path length value of the dynamic volumetric human body model in a static state from the total physical path length of the fabric, and outputs the geodesic path stretching increment.

[0031] Furthermore, in step S3, the specific operations for obtaining the tensile stiffness parameters of the target fabric and estimating the user's soft tissue stiffness parameters based on the basic physiological parameters obtained in step S1 are as follows:

[0032] C1 performs orthogonal anisotropic analytical calculations on the physical properties of the target fabric, retrieving the warp and weft tensile Young's modulus values ​​of the target fabric from a pre-set fabric physical property database.

[0033] C2, identify the tangent direction of the key measurement path on the surface of the dynamic volumetric human body model extracted in step S2, and calculate the geometric angle between the tangent direction of the key measurement path and the warp texture direction of the fabric.

[0034] C3 utilizes the second-order tensor projection rule to perform a weighted projection operation based on the fourth power values ​​of the cosine and sine functions of the geometric angle values. This projects the warp and weft tensile Young's modulus values ​​onto the tangent direction of the key measurement path, thereby calculating the effective linear elastic modulus value that characterizes the fabric's resistance to tension along the key measurement path. The effective linear elastic modulus value is then defined as the tensile stiffness parameter.

[0035] C4. Call the physical soft tissue thickness value carried by the external soft tissue deformation layer generated in step S1 and the body fat percentage value in the basic physiological parameters to construct a nonlinear compression constitutive model of biological soft tissue based on the exponential hardening feature.

[0036] C5 introduces a preset capillary perfusion pressure threshold value as a rigid physical boundary condition for physiological safety, uses body fat percentage value to correct the basic elastic modulus value of biological soft tissue, and calculates the maximum allowable compression depth value that the physical soft tissue thickness value can produce when the capillary perfusion pressure threshold value is reached through a nonlinear compression constitutive model of biological soft tissue, and defines the maximum allowable compression depth value as the soft tissue stiffness parameter.

[0037] Furthermore, in step S3, the geodesic path stretching increment, stretching stiffness parameter, and soft tissue stiffness parameter calculated in step S2 are input into the bidirectional stiffness-displacement complementary model. Based on the principle of force balance, the specific operation of calculating the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue is as follows:

[0038] C6, Construct a fabric-biological tissue coupled load-bearing system model based on the classical mechanical series equilibrium law, and set the geodesic path stretching increment output in step S2 as the total geometric deformation input value that the fabric-biological tissue coupled load-bearing system model must digest.

[0039] C7, set the capillary closure pressure value as the stress limit boundary condition of the fabric and biological tissue coupled load-bearing system model, and based on the principle of force balance, determine that under the stress limit boundary condition, the pressure values ​​of the outer fabric layer and the inner biological soft tissue layer are consistent.

[0040] C8. Based on Hooke's Law and its corollaries, the maximum elastic elongation of the fabric layer when the capillary closure pressure value is reached is calculated using the tensile stiffness parameter obtained in step S3. At the same time, the soft tissue stiffness parameter obtained in step S3 is directly called as the maximum allowable compression value that the biological soft tissue layer can provide under the capillary closure pressure value.

[0041] C9 performs stiffness complementary residual decision calculation, subtracts the maximum elastic elongation value and the maximum allowable compression value from the geodesic path stretching increment, defines the remaining geometric difference that has not been eliminated by physical deformation as the rigid space requirement value, compares the rigid space requirement value with the preset minimum process looseness value, and selects the larger value between the rigid space requirement value and the minimum process looseness value as the final output pattern size compensation amount.

[0042] Furthermore, in step S4, the pattern size compensation amount calculated in step S3 is mapped to the corresponding geometric region of the preset original electronic template. The specific operation of performing mesh deformation processing on the original electronic template is as follows:

[0043] D1. Obtain the preset original electronic template and use the restricted Delaunay triangulation algorithm to discretize the two-dimensional vector contour of the original electronic template into a two-dimensional elastic finite element mesh with topological connection relationship. Perform semantic feature mapping operation from three-dimensional space to two-dimensional plane, extract the endpoint coordinate data and key node coordinate data of the key measurement path on the surface of the dynamically volumetric human body model in step S2, and use the centroid coordinate mapping algorithm to lock the endpoint coordinate data and key node coordinate data of the key measurement path to the specific mesh node of the two-dimensional elastic finite element mesh, and define the locked mesh node as the driving control point set.

[0044] D2, based on the theory of anisotropic deformation gradient potential energy, constructs a global deformation energy functional. The global deformation energy functional includes a shape-preserving potential energy term and a target-driven potential energy term. The shape-preserving potential energy term is constructed by using a deformation model that is as rigid as possible. By minimizing the Frobenius norm between the deformation gradient tensor of each mesh triangular element and the local optimal rotation matrix, numerical penalties are applied to non-rigid shear deformation and non-uniform scaling deformation, thereby constraining the two-dimensional elastic finite element mesh to maintain the local geometric features of the original electronic template during deformation at the numerical level.

[0045] D3. The pattern size compensation calculated in step S3 is transformed into a virtual displacement constraint condition applied to the set of driving control points. The target driving potential energy term is constructed, and the projected length of the key measurement path on the two-dimensional plane is forced to be strictly equal to the sum of the original static length of the key measurement path and the pattern size compensation.

[0046] Furthermore, in step S4, the specific operation for generating producible vector patch data containing local dynamic margins is as follows:

[0047] In solving the global deformation energy functional, D4 introduces the Jacobian barrier constraint mechanism and the stitching edge coupling constraint mechanism based on the principle of topological homeomorphism. A logarithmic barrier function is constructed as the topological anti-flip energy term. In each iteration, the Jacobian determinant value of each triangular element in the two-dimensional elastic finite element mesh is checked. Using the mathematical properties of the logarithmic function, a penalty energy value that tends to infinity is generated when the Jacobian determinant value approaches zero. This establishes a mathematical potential boundary to force the area of ​​all triangular elements in the mesh to always be positive, thus preventing the two-dimensional elastic finite element mesh from undergoing topological degradation or self-intersection flipping.

[0048] D5 identifies the corresponding seam edges of all associated pattern pieces in the original electronic template, constructs a seam coupling potential energy term, and forces the corresponding seam edges of associated pattern pieces to maintain consistent geometric length values ​​after deformation.

[0049] D6 uses the Gauss-Newton method to solve the nonlinear least squares problem, obtains the optimal set of grid vertex coordinates under the minimum energy state, extracts the boundary node sequence of the optimal grid vertex coordinate set, performs area-preserving curvature flow smoothing operation to eliminate high-frequency jagged noise caused by discrete grid, and uses third-order Bézier curves to perform inverse fitting and reconstruction of the smoothed boundary node sequence, outputting producible vector clipping data.

[0050] A personalized suit customization system includes: a dynamic volumetric human body model construction module, a geodesic path stretching increment calculation module, a pattern size compensation calculation module, and a producible vector pattern generation module, wherein;

[0051] The dynamic volumetric human model construction module is used to acquire static image data and basic physiological parameters of the target user, and construct a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on the static image data and basic physiological parameters. The dynamic volumetric human model has the ability to drive soft tissue deformation according to skeletal movement.

[0052] The geodesic path stretching increment calculation module is used to call the standard motion library to drive the dynamic volumetric human model constructed by the dynamic volumetric human model construction module to perform joint rotation movements, extract the preset key measurement paths on the surface of the dynamic volumetric human model, and calculate the geodesic path stretching increment generated by the key measurement paths in the motion state relative to the static state.

[0053] The pattern size compensation calculation module is used to obtain the tensile stiffness parameters of the target fabric and estimate the user's soft tissue stiffness parameters based on the basic physiological parameters obtained by the dynamic volumetric human body model construction module. The geodesic path tensile increment, tensile stiffness parameters and soft tissue stiffness parameters calculated by the geodesic path tensile increment calculation module are input into the bidirectional stiffness displacement complementary model. Based on the principle of force balance, the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue itself is calculated.

[0054] The producible vector pattern generation module is used to map the pattern size compensation amount calculated by the pattern size compensation amount calculation module to the corresponding geometric area of ​​the preset original electronic pattern, perform grid deformation processing on the original electronic pattern, and generate producible vector pattern data containing local dynamic allowances.

[0055] The beneficial effects of this invention are as follows:

[0056] This invention overcomes the limitations of traditional static body measurements in capturing motion deformation by constructing a dynamic volumetric model that incorporates skeletal drive and soft tissue deformation. It can accurately simulate the muscle expansion effect during limb activity. Combined with a unique bidirectional stiffness-displacement complementary mechanism, it scientifically allocates the ratio of fabric stretching and body compression based on the principle of mechanical balance, calculating a precise compensation amount that meets the needs of movement without being excessively loose. In addition, by using grid deformation technology based on energy minimization, it can inject dynamic allowances into the pattern while maintaining the original design's contour lines and seam integrity. This not only solves the contradiction between slim fit and comfort in traditional customization, achieving a wearing experience that is well-fitting at rest and pressure-free during movement, but also significantly reduces the reliance on tailor experience, promoting the intelligent and standardized process of clothing customization. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort:

[0058] Figure 1 This is a flowchart of the method of the present invention;

[0059] Figure 2 This is a system framework diagram of the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1

[0062] like Figure 1 As shown, this invention discloses a method for personalized suit customization, comprising:

[0063] Step S1: Obtain static image data and basic physiological parameters of the target user, and construct a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on the static image data and basic physiological parameters. The dynamic volumetric human model has the ability to drive soft tissue deformation based on skeletal movement.

[0064] In step S1, the specific operation of constructing a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on static image data and basic physiological parameters is as follows:

[0065] First, basic physiological parameters are analyzed to extract the target user's true height value, and the true height value is set as the absolute physical scale anchor value under the three-dimensional Euclidean space measurement system.

[0066] Specifically, the algorithm reads the user's pre-entered physiological data and extracts the user's actual height. In computer vision processing, a single two-dimensional image inherently loses depth information, leading to perspective distortion where near objects appear larger than distant ones. This causes pixel sizes in the image to not directly correspond to actual physical dimensions. Therefore, this embodiment introduces the concept of an absolute physical scale anchor value. The absolute physical scale anchor value provides a unique standard of measurement for the virtual three-dimensional space. In practice, the absolute physical scale anchor value is set to the user's actual height, such as 1750 millimeters. The actual height value comes from measurements taken by the user at a medical institution or on standardized measuring equipment, ensuring objectivity and authenticity. The actual height value serves as a hard boundary condition in the subsequent reverse optimization process, preventing scale drift in the generated dynamic volumetric human body model due to unknown camera focal length, where the projections overlap but the overall proportions are distorted.

[0067] Secondly, a perspective projection inverse optimization model is constructed, which includes camera focal length variables and camera spatial pose variables. The perspective projection inverse optimization model is used to project the initialized three-dimensional parameterized human body template onto the two-dimensional pixel coordinate plane where the static image data is located.

[0068] In this step, a perspective projection inverse optimization model is established based on the imaging principle of a pinhole camera. The perspective projection inverse optimization model contains two core variables to be solved: the first variable is the camera focal length variable, which represents the lens magnification and field of view characteristics of the virtual shooting camera; the second variable is the camera spatial pose variable, which represents the three-dimensional spatial displacement vector and rotation angle of the human body relative to the camera.

[0069] Simultaneously, it is also necessary to solve the shape and pose parameters of the human body model. During the initialization phase, the initial value of the camera focal length variable is set to 1.2 times the length of the image diagonal pixels. For example, for an image with a resolution of 1080 x 1920, the initial value of the camera focal length variable is set to 2500 pixels. This value is based on the statistical law of the equivalent focal length of most mobile phone cameras when taking full-body photos. Subsequently, perspective division is performed to divide the coordinate value of each skeletal key point of the 3D parameterized human body template in 3D space by the depth distance value of the skeletal key point relative to the camera, and then multiply it by the camera focal length variable to obtain the pixel coordinate projection of the skeletal key point on the 2D plane.

[0070] Subsequently, during the iterative optimization operation, the reprojection position deviation between the projected feature points of the 3D parametric human body template and the visual key points extracted from the static image data is calculated. At the same time, a rigid geometric constraint based on the rigid body scale invariance is introduced. The rigid geometric constraint requires that the geodesic path length of the 3D parametric human body template from the top of the head to the bottom of the feet be strictly equal to the actual height.

[0071] Construct and minimize a joint energy loss function, which consists of a reprojection error term and a hard geometric constraint term. With respect to the reprojection error term, calculate the sum of squared Euclidean distances between the projection points of the 3D parameterized human template projected onto the 2D image plane and the visual key points detected in the image.

[0072] Regarding the hard geometric constraint term, the length of the surface curve extending from the midpoint of the top of the head through the spine to the midpoint of the foot is calculated along the surface of the three-dimensional parametric human body template, which is the geodesic path length. The square of the difference between the geodesic path length and the actual height value is also calculated. In the optimization process, the system introduces the hard geometric constraint weight coefficient.

[0073] In this embodiment, the hard geometric constraint weight coefficient is preferably set to one million. The basis for choosing this order of magnitude is that the reprojection error term is usually at the pixel level, with a value range of about one hundred to one thousand. The small deviation of physical height, such as ten millimeters, has a small value after squaring. In order to make the physical scale constraint dominate in the gradient descent process, a high-order weight must be introduced to force the optimizer to prioritize the absolute consistency of physical height before ensuring projection matching.

[0074] To verify the feasibility of this parameter in this invention, a comparative experiment was conducted in this embodiment, and two sets of experimental models were constructed: Comparative Model 1 uses the weak constraint weight in the prior art, that is, the weight coefficient is one hundred, and only pursues the visual overlap of image feature points; while this invention uses one million strong constraint weights.

[0075] Experimental results show that in the reconstruction experiment on 500 test subjects, although the human body model generated by Comparative Example 1 had a high degree of overlap in the image projection, the average error of its estimated shoulder width data was as high as 12.4%, and the estimation variance of the depth axis was extremely large. This led to a serious defect in the subsequent generation of suit patterns, where the clothes looked good on the surface but were actually too small in size. In contrast, the human body model generated by this invention reduced the average error of its key circumference data to less than 1.8%. The high weight coefficient significantly improved the gradient response sensitivity of the loss function to scale deviation. A strongly constrained convergence region for the actual height value was constructed in the solution space. This made it possible for the optimization algorithm to generate huge costs for any solution that deviates from the absolute physical scale during the iteration process, thereby forcing the optimization results to be firmly locked in the solution space that conforms to the physical truth value. This effectively solved the technical problem of scale drift in monocular vision reconstruction.

[0076] Next, by jointly solving the cost function to minimize the cost function, a unique camera focal length value and the absolute physical distance value of the target user relative to the camera are determined. Based on the camera focal length value and the absolute physical distance value, the three-dimensional parametric human body template is scaled to generate a dynamic volumetric human body model with absolute metric scale attributes.

[0077] The nonlinear least squares method is used to converge the solution of the joint energy loss function. When the iteration ends, the system outputs a set of optimal solutions. The optimal solutions contain the precise camera focal length value, such as 2,850 pixels, and the absolute physical distance between the user and the camera, such as 3,500 millimeters. These two parameters are used to eliminate the distortion caused by visual perspective, and the three-dimensional parametric human body template is mapped from the dimensionless relative space to the metric Euclidean space in millimeters. The dynamically volumetric human body model generated at this time has the coordinates of each grid vertex corresponding to the physical position in the real world, providing a precise geometric reference for subsequent clothing pattern matching.

[0078] Furthermore, after determining the skeletal spatial pose and body surface mesh position of the dynamically volumetric human body model, the physical volume extraction calculation of soft tissue is performed in accordance with the physical law of non-overlapping material space. This step aims to solve the technical defect of traditional geometric modeling that lacks physical property decoupling between rigid skeleton and flexible soft tissue. This invention is based on the law of exclusivity of material space in classical mechanics, that is, at the same physical moment, rigid skeletal structure and flexible soft tissue medium cannot occupy the same three-dimensional spatial coordinates. Therefore, the spatial difference between the external shape of the human body and the internal skeleton constitutes the physical volume of soft tissue.

[0079] In practice, every geometric vertex on the surface mesh of the dynamically volumetric human body model is traversed, and the Euclidean radial distance between each geometric vertex and the central axis of the nearest bone in the internal skeletal drive system is calculated. Tens of thousands of geometric vertices on the surface of the dynamically volumetric human body model are scanned one by one. For any geometric vertex, the limb segment to which the geometric vertex belongs is first identified, such as the upper arm segment. Then, the vertical distance from the geometric vertex to the central axis of the upper arm bone is calculated using the shortest distance algorithm from a point to a line segment. This distance is defined as the Euclidean radial distance value, which physically represents the total radius length from the center of the bone marrow to the skin surface.

[0080] Then, the system calls a pre-set anatomical statistics database to obtain the average rigid radius of the bone corresponding to the current body part, performs a difference operation by subtracting the average rigid radius of the bone from the radial distance in Euclidean space, and defines the positive difference obtained as the physical soft tissue thickness at the geometric vertex.

[0081] Access a pre-stored anatomical statistics database, which records standard bone diameter data for different sexes and heights. For example, for an adult male who is 1.75 meters tall, the database records that the average rigid radius of the femur in the midsection is about 16 millimeters. This value is an average value based on large-scale medical imaging statistics. Then, a subtraction operation is performed, subtracting the average rigid radius of the bone from the measured Euclidean radial distance value. The difference is the physical soft tissue thickness value. The physical soft tissue thickness value directly quantifies the thickness of the muscle and fat layers in that area, determining the maximum compressive deformation space that the area can generate when squeezed by clothing.

[0082] Finally, the heterogeneity attribute is marked for the region associated with the physical soft tissue thickness value by combining the body fat percentage value in the basic physiological parameters, and a biological non-penetrating boundary constraint mechanism is introduced. When the calculated physical soft tissue thickness value is less than the preset minimum physiological thickness threshold of the human dermis, the physical soft tissue thickness value at that location is forcibly replaced with the minimum physiological thickness threshold of the human dermis, and the finally determined thickness data field is mapped to the external soft tissue deformation layer.

[0083] In this step, specific parameter control and heterogeneity mapping logic are introduced. Regarding the biological non-penetrating boundary constraint mechanism, the minimum physiological thickness threshold of the human dermis is set to 2.0 mm. This parameter is based on human histological measurement data. The thickness of the epidermis in adults is about 0.1 to 1.5 mm, and the thickness of the dermis is about 0.6 to 3.0 mm. 2.0 mm is taken as the lower limit of the physical limit of soft tissue including skin and a very small amount of subcutaneous fat.

[0084] To verify the effectiveness of this threshold, a comparative test was conducted: In existing solutions without threshold constraints, when dealing with areas with very little subcutaneous fat, such as the wrist, ankle, and clavicle, the physical soft tissue thickness often becomes negative due to minor registration errors between the bone template and the body surface mesh. This causes mesh nodes to penetrate into the bone during subsequent finite element simulations, leading to computational divergence or program crashes. However, by using the two-point-zero millimeter threshold constraint of this invention, the forced clamping logic ensures that all soft tissue units have a positive physical volume. Experiments show that this mechanism increases the convergence success rate of subsequent motion simulations from 78% to 99.5%.

[0085] Meanwhile, to demonstrate the adaptability of the solution to individual differences, a nonlinear mapping relationship was established between body fat percentage and Young's modulus and thickness gain coefficient of soft tissue. For the low body fat group (less than 15%), the system sets a high Young's modulus of soft tissue, such as 20 kPa, to characterize the high rigidity of muscles, and sets the thickness gain coefficient close to 1.0. Under these parameters, the generated dynamic volumetric human body model exhibits less volume deformation during movement, and the compensation amount for the suit pattern is correspondingly reduced, meeting the needs of a slender body type for a slim fit. For the high body fat group (greater than 25%), the Young's modulus of soft tissue is set to a low value, such as 5 kPa, to characterize the compressibility of fat, and the thickness gain coefficient is set to greater than 1.2, mainly affecting the abdominal and thigh areas. Under these parameters, the dynamic volumetric human body model exhibits significant compressive deformation capacity in simulation, allowing some of the compressible soft tissue to be automatically deducted in the suit pattern calculation, thus achieving a visually slimming effect without being too tight.

[0086] Step S2: Retrieve the standard motion library to drive the dynamic volumetric human model constructed in step S1 to perform joint rotational motion, extract the preset key measurement paths on the surface of the dynamic volumetric human model, and calculate the geodesic path stretching increment generated by the key measurement paths in motion relative to the static state.

[0087] Specifically, this invention elaborates on how a static model with absolute metric dimensions is placed into a dynamic mechanical environment, and how the actual force and travel of the suit fabric under human movement is obtained through physical simulation.

[0088] First, the specific operation of calling the standard motion library to drive the dynamic volumetric human model constructed in step S1 to perform joint rotation is as follows: based on the incompressible fluid properties of biological soft tissue at the macroscopic physical level, a volume conservation coupling calculation logic is established between axial compression variables and radial expansion variables.

[0089] The incompressible fluid property is introduced as the underlying axiom of the physical simulation. This property is that the main component of human muscle and fat tissue is water, and its density remains constant at physiological temperature. Therefore, when subjected to external mechanical force, its total volume will not change. In specific implementation, a conservation equation based on the cylinder volume formula is constructed, that is, the volume of a cylinder is equal to the cross-sectional area multiplied by the axial length. The role of this logic is to establish a physical constraint: when the axial length of a limb shortens due to joint bending, in order to maintain volume conservation, the cross-sectional area of ​​the limb must expand in the opposite direction. This logic eliminates the volume collapse error caused by simply relying on geometric skinning algorithms in existing technologies, ensuring that the simulation results conform to the real biomechanical laws of the human body.

[0090] Secondly, the motion posture data in the standard motion library is used to drive the skeleton driving system inside the dynamic volumetric human model to perform joint bending and rotation operations, and the axial projection length of the current limb segment bone in the bending state is calculated in real time.

[0091] In this step, a pre-set standard motion library is read. The standard motion library contains typical high-frequency motion data in the context of wearing a suit, such as extending both arms forward at 90 degrees while driving or leaning back at 15 degrees while working at a desk. These angle data are mapped to the internal skeletal drive system of the dynamic volumetric human model, forcing the skeleton to rotate. Then, a geometric projection calculation is performed to calculate the straight-line projection distance of each limb segment bone along the main axis of the limb in the current bending posture. This distance is defined as the axial projection length value.

[0092] Subsequently, a comparison operation is performed, dividing the axial projection length value by the original bone length value of the bone drive system in a static extended state, thereby obtaining the axial compression ratio value that represents the current degree of bone compression.

[0093] The algorithm reads skeletal data from a dynamically quantified human body model in a standard upright standing posture to obtain raw bone length values. For example, a user's upper arm bone is 300 mm long at rest; when performing a bending motion, its axial projected length may shorten to 200 mm. A division operation is performed, dividing 200 mm by 300 mm to obtain a quotient of approximately 0.66. This quotient is defined as the axial compression ratio of the skeleton, a dimensionless physical quantity typically ranging from zero to one. In the algorithm, the axial compression ratio of the skeleton serves as the source variable driving soft tissue deformation; the smaller the value, the more severe the compression, and the more significant the subsequent radial expansion effect.

[0094] Next, the pre-stored physical soft tissue thickness value in the external soft tissue deformation layer is called, and the physical soft tissue thickness value is used as the basic gain factor. Combined with the preset Poisson's ratio coefficient, a nonlinear inverse proportional expansion operation based on the volume conservation law is performed on the axial compression ratio value of the bone to calculate the forced radial displacement value caused by the shortening of the axial distance of the bone, which forces the soft tissue medium to be squeezed outward.

[0095] This is the calculation step of the physical simulation in this embodiment. Poisson's ratio is introduced here. Poisson's ratio represents the ability of a material to undergo lateral deformation when subjected to axial compression. In this embodiment, Poisson's ratio is strictly set to 0.5. This value is based on fluid mechanics and biomechanics theory. The ideal Poisson's ratio for an incompressible fluid such as water is 0.5.

[0096] Two sets of comparative experiments were constructed. In Comparative Example A, which used general engineering material parameters, the Poisson's ratio was set to 0.3. The experimental results showed that when simulating a 90-degree elbow flexion, the upper arm circumference increased by only 3.5%, far lower than the actual anthropometric measurements of 8.2%, resulting in a severely insufficient sleeve width in the suit and a feeling of tightness in the arm. In Example A, which used the parameters of this invention, the Poisson's ratio was set to 0.5. The experimental results showed that the upper arm circumference increased by 8.1%, with an error of only 0.1 millimeters compared to the actual measurement. This comparative experiment proves that the choice of 0.5 is not arbitrary, but is to accurately replicate the volume expansion characteristics of biological soft tissue, which is a necessary technical feature for solving the dynamic tightening problem. Based on this parameter, a specific nonlinear inverse proportional expansion calculation is performed: first, the square root of the reciprocal of the axial compression ratio of the bone is calculated, then one is subtracted from the square root to obtain the basic expansion rate; finally, the basic expansion rate is multiplied by the physical soft tissue thickness and the Poisson's ratio to obtain the forced radial displacement value.

[0097] Next, the forced radial displacement values ​​are superimposed along the normal direction onto the corresponding geometric vertex coordinates of the surface mesh of the dynamically volumetric human body model, driving the external soft tissue deformation layer to undergo nonlinear geometric bulging deformation, generating dynamic deformation mesh data containing volume expansion characteristics.

[0098] By performing vector addition, the calculated forced radial displacement value is multiplied by the normal vector of the mesh vertex, and the spatial coordinates of the mesh vertex are updated. This operation makes the surface of the dynamically volumetric human body model present a realistic bulge shape at the joint bending point, generating dynamic deformation mesh data. The dynamic deformation mesh data truly reflects the physical space occupied by the human body due to muscle compression in motion, providing an accurate geometric basis for subsequent fabric path measurement.

[0099] Furthermore, the key measurement paths preset on the surface of the dynamically volumetric human body model are extracted, and the specific operation of calculating the geodesic path stretching increment generated by the key measurement paths in motion relative to the static state is as follows: Based on the physical characteristics of minimizing the energy of the fabric material under tension, a convex hull bridging integral algorithm is configured to simulate the geometric coverage trajectory of the suit fabric on the human body surface.

[0100] The physical property of minimizing energy is introduced as a guiding principle for path search. Suit fabric has bending stiffness and tends to remain flat to minimize elastic potential energy when subjected to tensile tension, rather than completely conforming to every tiny indentation on the human body surface like plastic wrap. Based on this, a convex hull bridging integral algorithm is configured. The core logic of this algorithm is to find an envelope on the human body surface that does not penetrate the human body and has the shortest path length.

[0101] Subsequently, preset key measurement paths are extracted from the surface of the dynamically volumetric human body model that has undergone volume expansion and deformation. The discrete geometric nodes on the key measurement paths are traversed to perform contact state determination calculations, locking key measurement paths such as the back width line. These key measurement paths consist of a series of ordered discrete geometric nodes. These discrete geometric nodes are scanned one by one, and the geometric topological relationship between the current node and adjacent nodes is analyzed to determine whether the fabric is attached to the skin or suspended above the skin at this point.

[0102] Specifically, when performing the contact state determination operation, if the fabric-covered area is determined to be a protruding part of the human body, the area is marked as a contact state segment, and differential geometric integration is performed along the curvature direction of the human body surface to obtain the surface geodesic length value. If the fabric-covered area is determined to be a groove of the human spine or a bone depression, the system marks the area as a suspended bridging state segment and calculates the Euclidean distance between the two tangent points across the depression area.

[0103] This step employs a piecewise integration strategy for comparative analysis. Existing technologies typically perform full-path geodesic integration along the skin surface, similar to measuring with a tape measure. When dealing with the back, the tape measure sinks into the indentations of the spinal grooves, resulting in the measured back width being 15 to 20 millimeters longer than the actual required fabric length. Suits made based on this data exhibit noticeable fabric bunching and looseness at the center of the back, ruining the suit's aesthetic lines. This invention introduces a suspended bridging state segment calculation, automatically identifying and skipping invalid indented areas, and directly calculating the Euclidean distance between tangent points. Comparative experiments show that the path length calculated by this algorithm has an error of less than one percent compared to the sample length adjusted by a senior tailor through manual fitting. This proves that the logic gating mechanism effectively eliminates invalid surface areas, achieving accurate simulation of the fabric's physical properties.

[0104] Finally, an accumulation operation is performed to sum the surface geodesic length values ​​of all contact state segments with the Euclidean space straight-line distance values ​​of all suspended bridging state segments to obtain the total physical path length of the fabric in motion. Finally, a subtraction operation is performed to subtract the corresponding path length value of the dynamic volumetric human body model in the static state from the total physical path length of the fabric, and the geodesic path stretching increment is output.

[0105] The results of the segmented calculations are summed to obtain the total physical length of the fabric actually required by the human body under specific movements. Then, the original length of this path in a static standing posture is read, and a subtraction operation is performed. The resulting difference is the geodesic path stretch increment. The geodesic path stretch increment represents the additional physical space that the suit pattern must objectively provide to ensure that the garment is not torn or feel constricting when the user performs this movement. In a specific embodiment, for the back width line, the geodesic path stretch increment may be calculated as 45 millimeters. This value will be directly used as input to the subsequent mechanical compensation calculation module, determining the basic allowance that the final pattern needs to be enlarged.

[0106] Step S3: Obtain the tensile stiffness parameters of the target fabric, and estimate the user's soft tissue stiffness parameters based on the basic physiological parameters obtained in Step S1. Input the geodesic path tensile increment, tensile stiffness parameters, and soft tissue stiffness parameters calculated in Step S2 into the bidirectional stiffness-displacement complementary model, and calculate the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue itself based on the principle of force balance.

[0107] In step S3, the specific operation of obtaining the tensile stiffness parameters of the target fabric and estimating the user's soft tissue stiffness parameters based on the basic physiological parameters obtained in step S1 is as follows: performing orthogonal anisotropic analytical calculations on the physical properties of the target fabric, and retrieving the warp tensile Young's modulus values ​​and weft tensile Young's modulus values ​​of the target fabric from the preset fabric physical property database.

[0108] First, the target fabric number selected by the user is identified, and the fabric physical property database in the cloud is accessed. The fabric physical property database stores the standard mechanical data obtained through fabric style tester testing, and two core physical quantities are extracted: warp tensile Young's modulus value and weft tensile Young's modulus value.

[0109] These two values ​​represent the fabric's ability to resist tensile deformation in the two perpendicularly intersecting principal axes. Since the physical properties of woven fabrics are determined by the interlacing structure of warp and weft yarns, their mechanical properties exhibit significant orthogonal anisotropy, meaning that the tensile stiffness is completely different in different directions. Therefore, it is necessary to obtain the basic data for these two dimensions separately, rather than using only a single average stiffness value, in order to avoid deviations in pattern stress calculations caused by ignoring material anisotropy.

[0110] Secondly, the tangent direction of the key measurement path on the surface of the dynamically volumetric human body model extracted in step S2 is identified, and the geometric angle between the tangent direction of the key measurement path and the warp texture direction of the fabric is calculated.

[0111] In this step, the two-dimensional fabric texture coordinates are mapped onto the surface of the three-dimensional dynamic volumetric human body model. The tangent vectors of the key measurement paths at each discrete node are extracted. At the same time, the warp direction vector of the fabric at that position is determined according to the layout scheme of the virtual fitting. The vector dot product inverse cosine operation is performed to calculate the angle between the two vectors and obtain the geometric angle value. The geometric angle value accurately quantifies the degree of deviation between the fabric cutting direction and the force direction, which is the geometric basis for the subsequent stiffness tensor projection.

[0112] Subsequently, using the second-order tensor projection rule, a weighted projection operation is performed on the fourth power value of the cosine function and the fourth power value of the sine function based on the geometric angle values. The warp and weft tensile Young's modulus values ​​are projected onto the tangent direction of the key measurement path, thereby calculating the effective linear elastic modulus value that characterizes the fabric's resistance to tension along the key measurement path. The effective linear elastic modulus value is defined as the tensile stiffness parameter.

[0113] The classical second-order tensor projection method from composite material mechanics is introduced to solve the problem of oblique stress. The specific calculation logic is as follows: First, the cosine value of the geometric angle is calculated, and then the fourth power operation is performed on this cosine value; at the same time, the sine value and its fourth power value are calculated; the warp tensile Young's modulus value is multiplied by the cosine fourth power value, and the weft tensile Young's modulus value is multiplied by the sine fourth power value, and the two sets of products are added together; in this process, the coupling term of the shear modulus is also calculated to correct the oblique tensile effect. The final result of the calculation is defined as the effective linear elastic modulus value, that is, the tensile stiffness parameter. The meaning of the tensile stiffness parameter is that the tensile stiffness parameter is no longer an inherent property of the fabric, but rather the tensile resistance actually exhibited by the fabric along a specific measurement path direction at a specific cutting angle. Through this projection calculation, it is ensured that the stiffness parameter input to the mechanical model truly reflects the physical properties in the garment state.

[0114] Next, the physical soft tissue thickness value carried by the external soft tissue deformation layer generated in step S1 and the body fat percentage value in the basic physiological parameters are called to construct a nonlinear compression constitutive model of biological soft tissue based on exponential hardening characteristics.

[0115] Based on the nonlinear physical property that biological soft tissue hardens under pressure, a constitutive equation is established, and an exponential function is introduced to describe the nonlinear relationship between stress and strain: in the initial stage of compression, soft tissue exhibits low stiffness and is easily deformed; as the compression increases, the value of the exponential function rises sharply, indicating a rapid increase in stiffness. The model defines the geometric upper limit of compression through the physical soft tissue thickness and the basic hardness of the material through the body fat percentage. The higher the body fat percentage, the lower the basic modulus, thus accurately simulating the mechanical response curve of human soft tissue under stress.

[0116] Next, a preset capillary perfusion pressure threshold value is introduced as a rigid physical boundary condition for physiological safety. The body fat percentage value is used to correct the basic elastic modulus value of biological soft tissue. The maximum allowable compression depth value that the physical soft tissue thickness value can produce when the capillary perfusion pressure threshold value is reached is calculated through a nonlinear compression constitutive model of biological soft tissue. The maximum allowable compression depth value is defined as the soft tissue stiffness parameter.

[0117] In this step, a physiological limit parameter is introduced, namely the capillary perfusion pressure threshold value. In this embodiment, the capillary perfusion pressure threshold value is strictly set to 3.5 kPa. This value is derived from microcirculation physiology research. The average closure pressure of human skin capillaries is about 3.5 to 4.0 kPa. The capillary perfusion pressure threshold value acts as a fuse in the algorithm. The capillary perfusion pressure threshold value establishes the physiological safety red line of the pattern design, forcing that this pressure limit must not be exceeded when calculating the compression amount, to prevent blood flow obstruction and user discomfort. The capillary perfusion pressure threshold value is substituted into the nonlinear compression constitutive model, and the maximum deformation distance that soft tissue can undergo when the pressure does not exceed this value is obtained by inverse solution, that is, the maximum allowable compression depth value. The maximum allowable compression depth value is defined as a soft tissue stiffness parameter, which essentially represents the maximum space concession capacity that the human body can provide within the comfort range.

[0118] To verify the selection of the capillary perfusion pressure threshold value, a comparative experiment was conducted in this embodiment. Comparative Example B used a traditional looseness algorithm without setting an upper limit on pressure; while this embodiment B used the 3.5 kPa threshold constraint of the present invention. The experimental results showed that in the custom-made tight-fitting suits, 15% of users in Comparative Example B reported red marks or numbness on their skin after wearing them for a long time; while in Embodiment B, no users reported pressure discomfort while maintaining a visually slimming effect, proving the significant technical effect of this parameter selection in ensuring physiological comfort.

[0119] Further, in step S3, the geodesic path stretching increment, tensile stiffness parameter, and soft tissue stiffness parameter calculated in step S2 are input into the bidirectional stiffness-displacement complementary model. Based on the principle of force balance, the specific operation of calculating the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue is as follows: Construct a fabric and biological tissue coupled load-bearing system model based on the classical mechanical series equilibrium law, and set the geodesic path stretching increment output in step S2 as the total geometric deformation input value that the fabric and biological tissue coupled load-bearing system model must digest. The system constructs a physical series model: the outer fabric layer is regarded as the first spring, and the inner soft tissue layer is regarded as the second spring. The two are connected in series, and the geodesic path stretching increment calculated in step S2 is used as the total displacement excitation input to the series system. The physical essence of this model is to simulate the real clothing force scenario: when the human body becomes larger, the total deformation must be shared by the stretching of the fabric, the compression of the soft tissue, and the reserved ease of the pattern.

[0120] Subsequently, the capillary closure pressure value was set as the stress limit boundary condition of the fabric-biological tissue coupled load-bearing system model. Based on the principle of force balance, it was determined that under the stress limit boundary condition, the pressure values ​​of the outer fabric layer and the inner biological soft tissue layer are consistent.

[0121] Based on Newton's third law, the equilibrium condition is established: in a static equilibrium state, the pressure exerted by the fabric on the human body must be equal to the reaction pressure exerted by the human body on the fabric. The aforementioned capillary closure pressure is set as the maximum working point of the system. This means that, while ensuring that the user does not feel pain, the elasticity of the fabric and the flexibility of the human body are utilized to the maximum extent to absorb deformation.

[0122] Subsequently, based on Hooke's Law and its corollaries, the maximum elastic elongation that the fabric layer can produce when the capillary closure pressure value is reached is calculated using the tensile stiffness parameters obtained in step S3. At the same time, the soft tissue stiffness parameters obtained in step S3 are directly used as the maximum allowable compression value that the biological soft tissue layer can provide under the capillary closure pressure value.

[0123] Perform specific mechanical calculations: For the fabric layer, the system uses Hooke's Law formula, which states that stress equals elastic modulus multiplied by strain, to divide the capillary closure pressure value by the tensile stiffness parameter and multiply it by the effective force-bearing width of the path, thus calculating the maximum elastic elongation that the fabric can produce when the pressure threshold is reached; For the soft tissue layer, the maximum permissible compression depth value calculated in the previous step is directly used as the maximum permissible compression value. The maximum elastic elongation value and the maximum permissible compression value represent the amount of space compensation that the fabric and the human body can provide free of charge under the physiological safety red line, respectively.

[0124] Finally, a stiffness complementary residual decision calculation is performed. The maximum elastic elongation and the maximum allowable compression are subtracted from the geodesic path stretching increment. The remaining geometric difference that is not eliminated by physical deformation is defined as the rigid space requirement value. The rigid space requirement value is compared with the preset minimum process slack value, and the larger value between the rigid space requirement value and the minimum process slack value is selected as the final output pattern size compensation amount.

[0125] The final residual decision logic is executed as follows: The system performs a subtraction operation, subtracting the contribution of the fabric (maximum elastic elongation) from the total demand of human movement (i.e., geodesic path stretching increment), and then subtracting the contribution of the human body (i.e., maximum allowable compression). The difference obtained is defined as the rigid space demand value.

[0126] In this step, a minimum process ease value is introduced. In this embodiment, the minimum process ease value is preferably set to five millimeters. The selection of the minimum process ease value is based on the research on contact comfort in clothing engineering and the process experience of preventing coating effects. It is the minimum air layer thickness that must be retained in order to prevent the fabric from adsorbing with van der Waals forces, electrostatic adhesion, or excessive friction between the fabric and the skin. The minimum process ease value plays the role of a minimum guarantee value in the algorithm.

[0127] To demonstrate the necessity of selecting the minimum process ease value, this embodiment conducted a comparative analysis. Comparative Example C allowed a calculation result of zero, i.e., a perfect fit; Example C forcibly retained a minimum ease value of five millimeters. Experiments showed that in a high-humidity environment in summer, the fabric of Comparative Example C easily adhered to the skin, hindering sweat evaporation, and its thermal resistance value was 40% higher than that of Example C. This proves that even when mechanical calculations allow for a tight fit, introducing a minimum process ease value of five millimeters has an unexpected technical effect on maintaining microenvironmental comfort.

[0128] The comparison operation is performed, and the larger value between the rigid space requirement value and the minimum process ease value is selected. The output is the pattern size compensation amount. The pattern size compensation amount is the final conclusion after multi-physics field game. The pattern size compensation amount ensures that the generated suit pattern not only meets the space requirements of human movement, but also makes full use of the physical characteristics of materials and human body, while strictly adhering to the bottom line of physiological comfort.

[0129] Step S4: Map the pattern size compensation amount calculated in step S3 to the corresponding geometric area of ​​the preset original electronic pattern, perform mesh deformation processing on the original electronic pattern, and generate producible vector cutting data containing local dynamic allowances.

[0130] The pattern size compensation amount calculated in step S3 is mapped to the corresponding geometric region of the preset original electronic template. The specific operation of mesh deformation processing of the original electronic template is as follows: obtain the preset original electronic template, and use the restricted Delaunay triangulation algorithm to discretize the two-dimensional vector contour of the original electronic template into a two-dimensional elastic finite element mesh with topological connection relationship.

[0131] The original electronic template file stored in the template database is read, and the restricted Delaunay triangulation algorithm in computational geometry is called to discretize the internal region of the original electronic template. Under the premise of maximizing the minimum interior angle of all triangular elements, the algorithm generates a two-dimensional elastic finite element mesh composed of thousands of tiny triangles. Each triangular element has independent physical properties and maintains topological connection with neighboring elements through shared vertices. This transforms the rigid vector template into a digital flexible body that can undergo local elastic deformation.

[0132] Perform semantic feature mapping operation from three-dimensional space to two-dimensional plane, extract the endpoint coordinate data and key node coordinate data of the key measurement path on the surface of the dynamically volumetric human body model in step S2, use the centroid coordinate mapping algorithm to lock the endpoint coordinate data and key node coordinate data of the key measurement path to the specific mesh node of the two-dimensional elastic finite element mesh, and define the locked mesh node as the driving control point set.

[0133] To address the alignment issue between 3D human body features and 2D pattern features, the endpoints and extreme points of the key measurement paths defined in step S2 are extracted in 3D space. Using a pre-established virtual fitting layout mapping relationship, the projection positions of these 3D points on the 2D pattern plane are calculated. Then, through a centroid coordinate interpolation algorithm, the 2D elastic finite element mesh nodes closest to the projection positions are found. These selected mesh nodes are marked as the driving control point set. The driving control point set is the anchor point connecting physical requirements and geometric deformation, and they will act as force traction points in the subsequent deformation process.

[0134] Subsequently, a global deformation energy functional is constructed based on the anisotropic deformation gradient potential energy theory. The global deformation energy functional includes a shape-preserving potential energy term and a target-driven potential energy term. The shape-preserving potential energy term is constructed using a deformation model that is as rigid as possible. By minimizing the Frobenius norm between the deformation gradient tensor of each mesh triangular element and the local optimal rotation matrix, numerical penalties are applied to non-rigid shear deformation and non-uniform scaling deformation. This numerically constrains the two-dimensional elastic finite element mesh to maintain the local geometric features of the original electronic template during deformation.

[0135] A mathematical model describing the total cost of mesh deformation, namely the global deformation energy functional, is constructed, in which the shape-preserving potential term adopts the theory of the most rigid deformation model possible. When a triangular element deforms, the deformation gradient tensor describing its geometric change is calculated, and the rigid rotation component in the tensor, namely the local optimal rotation matrix, is extracted using the extreme decomposition method. The Frobenius norm of the difference between the deformation gradient tensor and the local optimal rotation matrix is ​​calculated, which is the square root of the sum of the squares of all elements of the matrix. This norm measures the degree of non-rigid distortion that occurs in the mesh element, namely shearing and non-uniform stretching. By minimizing this norm, any distortion behavior that violates the original style of the pattern is numerically penalized with a high weight.

[0136] Specifically, the pattern size compensation amount calculated in step S3 is transformed into a virtual displacement constraint condition applied to the set of driving control points, and a target driving potential energy term is constructed to force the projection length of the key measurement path on the two-dimensional plane to be strictly equal to the sum of the original static length of the key measurement path and the pattern size compensation amount.

[0137] In this embodiment, a target driving stiffness coefficient is introduced, which represents the priority weight of meeting the human motion space requirements, that is, how much shape retention the system is willing to sacrifice in order to achieve the target size. In this embodiment, the coefficient is set to one million. The basis for this value is the penalty function method principle in nonlinear optimization theory. In order to transform soft constraints into approximate hard constraints, the penalty parameter usually needs to be four to six orders of magnitude higher than the internal stiffness of the system. The extremely high target driving stiffness coefficient forces the system to prioritize meeting the stretching requirements of the path length.

[0138] To verify the necessity of selecting the target driving stiffness coefficient, a comparative experiment was conducted in this embodiment. In Comparative Example D, the coefficient was set to one thousand. The experimental results showed that due to the resistance of the shape-maintaining potential energy, the final generated pattern stretch was only 60% of the target value, and a significant springback error occurred. In contrast, Example D used a coefficient value of one million, and the final path length error was controlled within 0.01 millimeters, proving the decisive role of this parameter selection in achieving precise shape control.

[0139] Furthermore, in step S4, the specific operation of generating producible vector cut data containing local dynamic margins is as follows: in the process of solving the global deformation energy functional, the Jacobian barrier constraint mechanism and the stitching edge coupling constraint mechanism based on the topological homeomorphism principle are introduced to construct a logarithmic barrier function as the topological anti-flip energy term.

[0140] To prevent physically impossible self-penetration of the mesh during large deformation, a Jacobi barrier threshold is introduced. This parameter characterizes the limit of the area ratio that the mesh cell can be compressed, preventing the cell area from collapsing to zero or a negative value. In this embodiment, the threshold is set to 0.001. This value is based on the principle of differential homeomorphism in topology. As long as the Jacobi determinant is strictly greater than zero, the mapping is invertible and foldless.

[0141] In each iteration of the calculation, the Jacobian determinant of each triangular element in the two-dimensional elastic finite element mesh is checked. By utilizing the mathematical properties of the logarithmic function, a penalty energy value that tends to infinity is generated when the Jacobian determinant value approaches zero. This establishes a mathematical potential energy boundary to force the area of ​​all triangular elements in the mesh to always be positive, thus preventing the two-dimensional elastic finite element mesh from undergoing topological degradation or self-intersection flipping.

[0142] Calculate the Jacobian determinant of each triangle and the negative of its natural logarithm. As the grid is about to flip, the determinant approaches zero, its logarithm approaches negative infinity, and the negative logarithm approaches positive infinity. This causes a surge in global energy, forcing the optimization algorithm to immediately stop searching in that flip direction and retreat to the topologically safe solution space.

[0143] This embodiment constructs a comparative analysis. Comparative Example E uses the traditional Laplace smoothing algorithm to perform large deformations (such as increasing the back width by 50 mm). The results show that the mesh has twelve self-intersections (bow tie effect) in the high curvature region under the armpit, which makes the generated cut piece impossible to laser cut. In contrast, Example E uses the Jacobi barrier constraint mechanism. Under the same amount of deformation, the number of self-intersections is zero, and the generated mesh fully meets the topological requirements of physical manufacturing.

[0144] At the same time, the corresponding seam edges of all associated pieces in the original electronic template are identified, and a seam coupling potential term is constructed to force the corresponding seam edges of the associated pieces to maintain the same geometric length value after deformation.

[0145] In this embodiment, a seam length tolerance parameter is introduced, which represents the maximum difference in the edge lengths of two pieces of fabric that is allowed during industrial sewing. In this embodiment, the parameter is set to 0.5 mm. This value is based on the garment manufacturing process standard ASTM D6193. Exceeding this difference will cause wrinkles at the seam. This parameter ensures that the side seam lengths of the front and back pieces can still match after they are deformed separately.

[0146] Finally, the nonlinear least squares problem is solved using the Gauss-Newton method to obtain the optimal set of grid vertex coordinates under the minimum energy state. The boundary node sequence of the optimal grid vertex coordinate set is extracted, and area-preserving curvature flow smoothing is performed to eliminate high-frequency jagged noise caused by discrete grids. The smoothed boundary node sequence is then reconstructed by inverse fitting using third-order Bézier curves to output producible vector clipping data.

[0147] The minimum point of the sum of all the above energy terms is solved by the Gauss-Newton iteration method to obtain the coordinates of the deformed grid. Since the edge of the discrete grid is a broken line composed of line segments, direct output will result in a jagged edge. In this embodiment, the curvature flow smoothing iteration number parameter is introduced. This parameter controls the intensity of the boundary smoothing process. In this embodiment, the parameter is preferably ten times. This value is based on signal processing theory. Appropriate iteration can eliminate high-frequency noise, i.e. jaggedness, while retaining low-frequency features, i.e. pattern outline.

[0148] To demonstrate the technical effect of this parameter selection, a comparative experiment was conducted in this embodiment. Comparative Example F did not undergo smoothing (zero iterations), and the laser cutting machine generated high-frequency vibrations due to frequent minor turns in the path during operation, resulting in obvious burrs on the cutting edge. Comparative Example G was excessively smoothed (fifty iterations), and the key concave and convex features of the cut piece were smoothed out, leading to the displacement of the sewing point. Example F used ten iterations, resulting in smooth and rounded edges on the cut piece, and the displacement error of the key points was less than 0.2 mm, balancing processing quality and pattern accuracy.

[0149] Example 2

[0150] like Figure 2 As shown, the present invention also discloses a personalized suit customization system, including: a dynamic volumetric human body model construction module, a geodesic path stretching increment calculation module, a pattern size compensation calculation module, and a producible vector pattern generation module, wherein;

[0151] The dynamic volumetric human model construction module is used to acquire static image data and basic physiological parameters of the target user, and construct a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on the static image data and basic physiological parameters. The dynamic volumetric human model has the ability to drive soft tissue deformation according to skeletal movement.

[0152] The geodesic path stretching increment calculation module is used to call the standard motion library to drive the dynamic volumetric human model constructed by the dynamic volumetric human model construction module to perform joint rotation movements, extract the preset key measurement paths on the surface of the dynamic volumetric human model, and calculate the geodesic path stretching increment generated by the key measurement paths in the motion state relative to the static state.

[0153] The pattern size compensation calculation module is used to obtain the tensile stiffness parameters of the target fabric and estimate the user's soft tissue stiffness parameters based on the basic physiological parameters obtained by the dynamic volumetric human body model construction module. The geodesic path tensile increment, tensile stiffness parameters and soft tissue stiffness parameters calculated by the geodesic path tensile increment calculation module are input into the bidirectional stiffness displacement complementary model. Based on the principle of force balance, the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue itself is calculated.

[0154] The producible vector pattern generation module is used to map the pattern size compensation amount calculated by the pattern size compensation amount calculation module to the corresponding geometric area of ​​the preset original electronic pattern, perform grid deformation processing on the original electronic pattern, and generate producible vector pattern data containing local dynamic allowances.

[0155] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for personalized suit customization, characterized in that, Includes the following steps: Step S1: Obtain static image data and basic physiological parameters of the target user, and construct a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on the static image data and basic physiological parameters. The dynamic volumetric human model has the ability to drive soft tissue deformation according to skeletal movement. Step S2: Retrieve the standard motion library to drive the dynamic volumetric human model constructed in step S1 to perform joint rotation movements, extract the preset key measurement paths on the surface of the dynamic volumetric human model, and calculate the geodesic path stretching increment generated by the key measurement paths in motion relative to the static state. Step S3: Obtain the tensile stiffness parameters of the target fabric, and estimate the user's soft tissue stiffness parameters based on the basic physiological parameters obtained in Step S1. Input the geodesic path tensile increment, tensile stiffness parameters and soft tissue stiffness parameters calculated in Step S2 into the bidirectional stiffness-displacement complementary model, and calculate the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue itself based on the principle of force balance. Step S4: Map the pattern size compensation amount calculated in step S3 to the corresponding geometric area of ​​the preset original electronic pattern, perform mesh deformation processing on the original electronic pattern, and generate producible vector cutting data containing local dynamic allowances.

2. A method for personalized suit customization according to claim 1, characterized in that, In step S1, the specific operation of constructing a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on static image data and basic physiological parameters is as follows: A1, analyze basic physiological parameters to extract the target user's true height value, and set the true height value as the absolute physical scale anchor value under the three-dimensional Euclidean space measurement system; A2. Construct a perspective projection inverse optimization model that includes camera focal length variables and camera spatial pose variables. Use the perspective projection inverse optimization model to project the initialized 3D parameterized human body template onto the 2D pixel coordinate plane where the static image data is located. A3. During the iterative optimization operation, the reprojection position deviation between the projected feature points of the three-dimensional parameterized human body template and the visual key points extracted from the static image data is calculated. At the same time, a hard geometric constraint based on the rigid body scale invariance is introduced. The hard geometric constraint requires that the geodesic path length of the three-dimensional parameterized human body template from the top of the head to the bottom of the feet be strictly equal to the actual height. A4 determines the unique camera focal length value and the absolute physical distance value of the target user relative to the camera by jointly solving the minimum cost function, and performs scale correction on the three-dimensional parametric human body template based on the camera focal length value and the absolute physical distance value, thereby generating a dynamic volumetric human body model with absolute metric scale attributes.

3. A method for personalized suit customization according to claim 2, characterized in that, In step S1, the specific operation of constructing a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on static image data and basic physiological parameters is as follows: A5, after determining the skeletal spatial pose and body surface mesh position of the dynamic volumetric human model, performs soft tissue physical volume extraction calculation in accordance with the physical law of non-overlapping material space. A6, traverse each geometric vertex on the surface mesh of the dynamically volumetric human body model, and calculate the Euclidean radial distance between each geometric vertex and the central axis of the nearest bone in the internal skeletal drive system. A7 calls the preset anatomical statistics database to obtain the average rigid radius value of the bone corresponding to the current body part, performs the difference operation of subtracting the average rigid radius value of the bone from the radial distance value in Euclidean space, and defines the positive difference value obtained by the operation as the physical soft tissue thickness value at the geometric vertex. A8 combines the body fat percentage value in the basic physiological parameters to mark the heterogeneity attribute of the region associated with the physical soft tissue thickness value, and introduces a biological non-penetrating boundary constraint mechanism. When the calculated physical soft tissue thickness value is less than the preset minimum physiological thickness threshold of the human dermis, the physical soft tissue thickness value at that location is forcibly replaced with the minimum physiological thickness threshold of the human dermis, and the finally determined thickness data field is mapped to the external soft tissue deformation layer.

4. A method for personalized suit customization according to claim 1, characterized in that, In step S2, the specific operation of calling the standard motion library to drive the dynamically volumetric human model constructed in step S1 to perform joint rotational movements is as follows: B1. Based on the incompressible fluid properties of biological soft tissue at the macroscopic physical level, a volume conservation coupling calculation logic is established between axial compressibility variables and radial expansion variables. B2 uses motion posture data from the standard motion library to drive the skeleton drive system inside the dynamic volumetric human model to perform joint bending and rotation operations, and calculates the axial projection length of the current limb segment bone in the bending state in real time. B3. Perform a comparison operation, dividing the axial projection length value by the original bone length value of the bone drive system in a static extended state, to obtain the axial compression ratio value of the bone, which represents the current degree of bone compression. B4 calls the pre-stored physical soft tissue thickness value in the external soft tissue deformation layer, and uses the physical soft tissue thickness value as the basic gain factor. Combined with the preset Poisson's ratio coefficient, it performs a nonlinear inverse proportional expansion operation based on the volume conservation law on the bone axial compression ratio value, and calculates the forced radial displacement value caused by the shortening of the bone axial distance, which forces the soft tissue medium to be squeezed outward. B5 superimposes the forced radial displacement values ​​along the normal direction onto the corresponding geometric vertex coordinates of the dynamic volumetric human body model's surface mesh, driving the external soft tissue deformation layer to undergo nonlinear geometric bulging deformation, generating dynamic deformation mesh data containing volume expansion characteristics.

5. A method for personalized suit customization according to claim 4, characterized in that, In step S2, the specific operation of extracting the preset key measurement paths on the surface of the dynamically volumetric human body model and calculating the geodesic path stretching increment of the key measurement paths in motion relative to the static state is as follows: B6, based on the physical characteristics of minimizing the energy of fabric materials under tension, configures the convex hull bridging integral algorithm to simulate the geometric coverage trajectory of suit fabric on the human body surface; B7 extracts the preset key measurement path from the surface of the dynamic volumetric human body model that has undergone volume expansion and deformation, and traverses the discrete geometric nodes on the key measurement path to perform contact state determination calculation. B8. When performing the contact state determination operation, if the fabric-covered area is determined to be a protruding part of the human body, the system marks the area as a contact state segment and performs differential geometric integration operation along the curvature direction of the human body surface to obtain the surface geodesic length value. If the fabric-covered area is determined to be a groove of the human spine or a bone depression, the system marks the area as a suspended bridging state segment and calculates the Euclidean distance between the two tangent points across the depression area. B9 performs an accumulation operation, summing the surface geodesic length values ​​of all contact state segments with the Euclidean space straight-line distance values ​​of all suspended bridging state segments to obtain the total physical path length of the fabric in motion. Finally, it performs a subtraction operation, subtracting the corresponding path length value of the dynamic volumetric human body model in a static state from the total physical path length of the fabric, and outputs the geodesic path stretching increment.

6. A method for personalized suit customization according to claim 1, characterized in that, In step S3, the specific operations for obtaining the tensile stiffness parameters of the target fabric and estimating the user's soft tissue stiffness parameters based on the basic physiological parameters obtained in step S1 are as follows: C1 performs orthogonal anisotropic analytical calculations on the physical properties of the target fabric, retrieving the warp and weft tensile Young's modulus values ​​of the target fabric from a pre-set fabric physical property database. C2, identify the tangent direction of the key measurement path on the surface of the dynamic volumetric human body model extracted in step S2, and calculate the geometric angle between the tangent direction of the key measurement path and the warp texture direction of the fabric. C3 utilizes the second-order tensor projection rule to perform a weighted projection operation based on the fourth power values ​​of the cosine and sine functions of the geometric angle values. This projects the warp and weft tensile Young's modulus values ​​onto the tangent direction of the key measurement path, thereby calculating the effective linear elastic modulus value that characterizes the fabric's resistance to tension along the key measurement path. The effective linear elastic modulus value is then defined as the tensile stiffness parameter. C4. Call the physical soft tissue thickness value carried by the external soft tissue deformation layer generated in step S1 and the body fat percentage value in the basic physiological parameters to construct a nonlinear compression constitutive model of biological soft tissue based on the exponential hardening feature. C5 introduces a preset capillary perfusion pressure threshold value as a rigid physical boundary condition for physiological safety, uses body fat percentage value to correct the basic elastic modulus value of biological soft tissue, and calculates the maximum allowable compression depth value that the physical soft tissue thickness value can produce when the capillary perfusion pressure threshold value is reached through a nonlinear compression constitutive model of biological soft tissue, and defines the maximum allowable compression depth value as the soft tissue stiffness parameter.

7. A method for personalized suit customization according to claim 6, characterized in that, In step S3, the geodesic path stretching increment, stretching stiffness parameter, and soft tissue stiffness parameter calculated in step S2 are input into the bidirectional stiffness-displacement complementary model. Based on the principle of force balance, the specific operation of calculating the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue is as follows: C6, Construct a fabric-biological tissue coupled load-bearing system model based on the classical mechanical series equilibrium law, and set the geodesic path stretching increment output in step S2 as the total geometric deformation input value that the fabric-biological tissue coupled load-bearing system model must digest. C7, set the capillary closure pressure value as the stress limit boundary condition of the fabric and biological tissue coupled load-bearing system model, and based on the principle of force balance, determine that under the stress limit boundary condition, the pressure values ​​of the outer fabric layer and the inner biological soft tissue layer are consistent. C8. Based on Hooke's Law and its corollaries, the maximum elastic elongation of the fabric layer when the capillary closure pressure value is reached is calculated using the tensile stiffness parameter obtained in step S3. At the same time, the soft tissue stiffness parameter obtained in step S3 is directly called as the maximum allowable compression value that the biological soft tissue layer can provide under the capillary closure pressure value. C9 performs stiffness complementary residual decision calculation, subtracts the maximum elastic elongation value and the maximum allowable compression value from the geodesic path stretching increment, defines the remaining geometric difference that has not been eliminated by physical deformation as the rigid space requirement value, compares the rigid space requirement value with the preset minimum process looseness value, and selects the larger value between the rigid space requirement value and the minimum process looseness value as the final output pattern size compensation amount.

8. A method for personalized suit customization according to claim 1, characterized in that, In step S4, the pattern size compensation amount calculated in step S3 is mapped to the corresponding geometric region of the preset original electronic template. The specific operation of performing mesh deformation processing on the original electronic template is as follows: D1. Obtain the preset original electronic template and use the restricted Delaunay triangulation algorithm to discretize the two-dimensional vector contour of the original electronic template into a two-dimensional elastic finite element mesh with topological connection relationship. Perform semantic feature mapping operation from three-dimensional space to two-dimensional plane, extract the endpoint coordinate data and key node coordinate data of the key measurement path on the surface of the dynamically volumetric human body model in step S2, and use the centroid coordinate mapping algorithm to lock the endpoint coordinate data and key node coordinate data of the key measurement path to the specific mesh node of the two-dimensional elastic finite element mesh, and define the locked mesh node as the driving control point set. D2, based on the theory of anisotropic deformation gradient potential energy, constructs a global deformation energy functional. The global deformation energy functional includes a shape-preserving potential energy term and a target-driven potential energy term. The shape-preserving potential energy term is constructed by using a deformation model that is as rigid as possible. By minimizing the Frobenius norm between the deformation gradient tensor of each mesh triangular element and the local optimal rotation matrix, numerical penalties are applied to non-rigid shear deformation and non-uniform scaling deformation, thereby constraining the two-dimensional elastic finite element mesh to maintain the local geometric features of the original electronic template during deformation at the numerical level. D3. The pattern size compensation calculated in step S3 is transformed into a virtual displacement constraint condition applied to the set of driving control points. The target driving potential energy term is constructed, and the projected length of the key measurement path on the two-dimensional plane is forced to be strictly equal to the sum of the original static length of the key measurement path and the pattern size compensation.

9. A method for personalized suit customization according to claim 8, characterized in that, In step S4, the specific operation for generating producible vector patch data containing local dynamic margins is as follows: In solving the global deformation energy functional, D4 introduces the Jacobian barrier constraint mechanism and the stitching edge coupling constraint mechanism based on the principle of topological homeomorphism. A logarithmic barrier function is constructed as the topological anti-flip energy term. In each iteration, the Jacobian determinant value of each triangular element in the two-dimensional elastic finite element mesh is checked. Using the mathematical properties of the logarithmic function, a penalty energy value that tends to infinity is generated when the Jacobian determinant value approaches zero. This establishes a mathematical potential boundary to force the area of ​​all triangular elements in the mesh to always be positive, thus preventing the two-dimensional elastic finite element mesh from undergoing topological degradation or self-intersection flipping. D5 identifies the corresponding seam edges of all associated pattern pieces in the original electronic template, constructs a seam coupling potential energy term, and forces the corresponding seam edges of associated pattern pieces to maintain consistent geometric length values ​​after deformation. D6 uses the Gauss-Newton method to solve the nonlinear least squares problem, obtains the optimal set of grid vertex coordinates under the minimum energy state, extracts the boundary node sequence of the optimal grid vertex coordinate set, performs area-preserving curvature flow smoothing operation to eliminate high-frequency jagged noise caused by discrete grid, and uses third-order Bézier curves to perform inverse fitting and reconstruction of the smoothed boundary node sequence, outputting producible vector clipping data.

10. A personalized suit customization system, employing a personalized suit customization method as described in any one of claims 1-9, characterized in that, include: The system includes a dynamic volumetric human body model construction module, a geodesic path stretching increment calculation module, a pattern size compensation calculation module, and a producible vector pattern generation module. The dynamic volumetric human model construction module is used to acquire static image data and basic physiological parameters of the target user, and construct a dynamic volumetric human model containing an internal skeletal drive system and an external soft tissue deformation layer based on the static image data and basic physiological parameters. The dynamic volumetric human model has the ability to drive soft tissue deformation according to skeletal movement. The geodesic path stretching increment calculation module is used to call the standard motion library to drive the dynamic volumetric human model constructed by the dynamic volumetric human model construction module to perform joint rotation movements, extract the preset key measurement paths on the surface of the dynamic volumetric human model, and calculate the geodesic path stretching increment generated by the key measurement paths in the motion state relative to the static state. The pattern size compensation calculation module is used to obtain the tensile stiffness parameters of the target fabric and estimate the user's soft tissue stiffness parameters based on the basic physiological parameters obtained by the dynamic volumetric human body model construction module. The geodesic path tensile increment, tensile stiffness parameters and soft tissue stiffness parameters calculated by the geodesic path tensile increment calculation module are input into the bidirectional stiffness displacement complementary model. Based on the principle of force balance, the pattern size compensation amount after deducting the deformation capacity of the fabric and soft tissue itself is calculated. The producible vector pattern generation module is used to map the pattern size compensation amount calculated by the pattern size compensation amount calculation module to the corresponding geometric area of ​​the preset original electronic pattern, perform grid deformation processing on the original electronic pattern, and generate producible vector pattern data containing local dynamic allowances.

Images