A sole three-dimensional digital model construction and intelligent recommendation system, method and electronic device

Abstract

This invention provides a system, method, and electronic device for constructing and intelligently recommending 3D digital models of shoe soles. By introducing motion phase weights and velocity-stiffness mapping into the modeling process, the generated shoe sole model can match the dynamic needs of the foot under different motion phases and speeds, overcoming the limitation of existing customized shoe soles that only adapt to a single motion state. A multi-objective optimization algorithm balances comfort, lightweight design, and athletic performance to generate an optimal shoe sole solution that matches the user's personalized preferences. By establishing a brand database and a multi-dimensional matching algorithm, accurate shoe recommendations based on 3D foot data and athletic needs are achieved, compensating for the shortcomings of traditional size recommendations. This system organically integrates customized modeling and intelligent recommendation to form a complete closed loop. Users can directly obtain 3D printing files for customized shoe soles as well as purchase suggestions for commercially available shoes, significantly improving the availability of personalized footwear products and the user experience.

Description

Technical Field

[0001] This application belongs to the field of digital design and intelligent recommendation technology for footwear products, and in particular relates to a system, method and electronic device for constructing and intelligently recommending three-dimensional digital models of shoe soles. Background Technology

[0002] With consumers' increasing demands for comfort and personalization in footwear, customized sole design based on foot morphology and precise shoe recommendations have become a hot research topic in the industry. Current technologies for sole customization typically involve the following steps: acquiring static point cloud data of the user's foot using 3D scanning equipment, then performing reverse modeling to obtain a foot model, allocating stiffness to different areas of the sole based on plantar pressure test data, and finally manufacturing a personalized sole using 3D printing. However, this customization method has significant shortcomings: firstly, the foot geometry data comes from static scanning, while the pressure data comes from static standing tests, resulting in a mismatch in the time dimension and ignoring the dynamic changes in foot morphology and stress during movement; secondly, the customization process only addresses a single movement state and cannot adapt to the differences in foot biomechanical response at different movement speeds.

[0003] In the field of footwear recommendation, current technologies mainly rely on size charts and simple foot length measurements, resulting in limited recommendation accuracy. Some advanced platforms, such as Volumemental and Aetrex FitGenius, achieve footwear matching through 3D foot scanning, but their matching algorithms are mostly based on static geometric dimensions and lack quantitative assessment of athletic functional requirements. Furthermore, existing customization and recommendation systems operate independently, failing to form a complete closed loop from personalized modeling to product recommendation. Users need to switch between different platforms, leading to a poor user experience.

[0004] Therefore, there is an urgent need for a comprehensive system that can integrate dynamic biomechanical modeling, multi-objective optimization, and intelligent recommendation to solve the problems of disconnect between static data and dynamic motion, and separation between customization and recommendation in existing technologies. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings and deficiencies of the existing technology and to provide a three-dimensional digital model construction and intelligent recommendation system, method and electronic device for shoe soles. It aims to achieve gait adaptation and velocity adaptation modeling by introducing motion phase weight and velocity-stiffness mapping, and to achieve accurate recommendation through multi-objective optimization and multi-dimensional matching algorithms.

[0006] The technical solution adopted in this invention is as follows: The first aspect of this invention provides a system for constructing and intelligently recommending three-dimensional digital models of shoe soles, comprising: The data acquisition module is used to acquire user foot shape data as well as user-inputted exercise scenarios and exercise intensity parameters; The modeling module, connected to the data acquisition module, is used to generate a basic three-dimensional digital model of the user's shoe sole based on the foot morphology data and exercise intensity parameters; The parameter optimization module, connected to the modeling module, is used to perform multi-objective optimization of the regional stiffness distribution of the basic three-dimensional digital model of the sole, with foot pressure uniformity, sole weight and energy feedback efficiency as objective functions, to generate a user sole optimized three-dimensional digital model. The brand database is used to store 3D digital model data, material performance data, and functional labels for multiple shoe models; The intelligent recommendation module connects the data acquisition module and the brand database to calculate the geometric matching degree, functional matching degree, and preference matching degree between the user's feet and the shoes of each brand, and outputs a list of recommended shoes. The modeling module includes: A phase weight allocation unit is used to allocate the design weights of each motion phase in the gait cycle according to the motion scenario. A speed weight allocation unit is used to allocate weights for different motion speeds according to the motion intensity. The velocity-stiffness mapping unit is used to establish a mapping model between the dynamic stiffness of the foot and the velocity of movement under each motion phase based on biomechanical experimental data. The weighted fusion unit, which connects the phase weight allocation unit, the velocity weight allocation unit, and the velocity-stiffness mapping unit, is used to perform two-layer weighted fusion of the morphological parameters of each region of the foot based on the phase weight, velocity weight, and velocity-stiffness mapping model, to generate comprehensive design parameters related to velocity. The model generation unit, connected to the weighted fusion unit, is used to construct a three-dimensional digital model of the shoe sole based on the comprehensive design parameters.

[0007] Preferably, the two-layer weighted fusion performed by the weighted fusion unit specifically includes: in, The comprehensive design parameters for the k-th region at velocity v are: V i The weight of the i-th type of motion speed, W ij The weight of the j-th phase at the i-th velocity. S ijk Let be the morphological parameters of the k-th region at the j-th phase under the i-th velocity. In the first k The rate of change of the region's morphology This is a dynamic adjustment coefficient. K k ( v) represents the velocity-stiffness mapping value of the k-th region, and β is the stiffness contribution coefficient.

[0008] Preferably, the mapping model established by the velocity-stiffness mapping unit is represented as follows: K j ( v )= f j ( v ),in K j ( v ) represents the equivalent dynamic stiffness of the j-th motion phase at velocity v. The mapping model is trained from biomechanical experimental data through curve fitting or machine learning methods.

[0009] Preferably, the parameter optimization module uses a multi-objective evolutionary algorithm for global optimization, generates a Pareto front for the user to choose from, and outputs the corresponding regional stiffness distribution according to the user's preferred scheme; the preferred scheme includes at least one of comfort, lightweight, or sports performance.

[0010] Preferably, the formula for calculating the matching degree by the intelligent recommendation module is: MatchScore= w g ⋅ G match + w f ⋅ F match + w p ⋅ P match in, G match Geometric matching is calculated based on the degree of fit between foot length, foot circumference, and arch height and the corresponding shoe size. F match Functional fit is calculated based on the degree of matching between the user's gait type, sports scenario, and the shoe's stability design and speed adaptability. P match The preference matching score is calculated based on user brand preference, price range, and appearance style preference. w g , w f , w p For the corresponding weights, and w g + w f + w p =1.

[0011] Preferably, the geometric matching degree is obtained by comparing the distance between key points on the user's foot and the corresponding point cloud inside the shoe cavity. The key points on the foot include the foot length endpoint, the metatarsophalangeal circumference point, the arch apex, and the heel point.

[0012] Preferably, the data acquisition module includes a mobile image acquisition unit and a professional-grade scanning unit; the mobile image acquisition unit captures the user's feet using a mobile phone camera, extracts key points of the feet using computer vision algorithms, and generates a foot point cloud model; the professional-grade scanning unit includes a 3D foot scanner, a dynamic plantar pressure plate, and a video gait analysis system.

[0013] Preferably, the system further includes an output module, which connects the parameter optimization module and the intelligent recommendation module, for exporting the optimized sole model as a 3D printing file and generating a recommended shoe report containing a foot fit heatmap.

[0014] A second aspect of the present invention provides a method for constructing and intelligently recommending three-dimensional digital models of shoe soles, applied to the aforementioned system, comprising the following steps: S1. Collect user foot morphology data and motion parameters; S2. Assign design weights to each motion phase in the gait cycle according to the motion scenario, and assign weights to different motion speeds according to the motion intensity; S3. Based on the phase weight and velocity weight, and combined with the pre-established velocity-stiffness mapping model, the morphological parameters of each region of the foot are weighted and fused in a two-layer manner to generate comprehensive design parameters related to velocity. S4. Construct a three-dimensional digital model of the shoe sole based on the comprehensive design parameters; S5. Using foot pressure uniformity, sole weight, and energy feedback efficiency as objective functions, perform multi-objective optimization on the regional stiffness distribution of the three-dimensional digital model of the sole. S6. Calculate the geometric matching degree, functional matching degree, and preference matching degree between the user's feet and each shoe in the brand database, and output a list of recommended shoes.

[0015] A third aspect of the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for constructing and intelligently recommending three-dimensional digital models of shoe soles.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for constructing and intelligently recommending three-dimensional digital models of shoe soles.

[0017] A fifth aspect of the present invention provides a computer program product that, when run on an electronic device, causes the electronic device to execute the above-described method for constructing and intelligently recommending three-dimensional digital models of shoe soles.

[0018] The beneficial effects of this invention are as follows: By introducing motion phase weighting and velocity-stiffness mapping into the modeling process, this application enables the generated sole model to match the dynamic needs of the foot under different motion phases and speeds, solving the deficiency of existing customized soles that only adapt to a single motion state. Through a multi-objective optimization algorithm, it balances comfort, lightweight design, and athletic performance to generate an optimal sole solution that meets the user's personalized preferences. By establishing a brand database and a multi-dimensional matching algorithm, it achieves accurate shoe recommendations based on three-dimensional foot data and athletic needs, compensating for the shortcomings of traditional size recommendations. This system organically integrates customized modeling with intelligent recommendations, forming a complete closed loop. Users can directly obtain 3D printing files for customized soles as well as purchase suggestions for commercially available shoes, significantly improving the availability of personalized footwear products and the user experience. Attached Figure Description

[0019] 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. The drawings described below are only some embodiments of the present invention. For those skilled in the art, obtaining other drawings based on these drawings without creative effort still falls within the scope of the present invention.

[0020] Figure 1 This is a structural block diagram of a three-dimensional digital model construction and intelligent recommendation system for shoe soles provided in an embodiment of this application; Figure 2 A flowchart illustrating a method for constructing a three-dimensional digital model of a shoe sole and providing intelligent recommendations, as provided in an embodiment of this application. Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0021] In the diagram, 100 is the data acquisition module, 110 is the mobile image acquisition unit, 120 is the professional-grade scanning unit, 121 is the 3D foot scanner, 122 is the dynamic plantar pressure plate, and 123 is the video gait analysis system. 200 - Modeling module, 210 - Phase weight allocation unit, 220 - Velocity weight allocation unit, 230 - Velocity-stiffness mapping unit, 240 - Weighted fusion unit, 250 - Model generation unit; 300-Parameter Optimization Module; 400-brand database; 500-Intelligent Recommendation Module; 600-Output Module; 8-Electronic device, 80-Processor, 81-Memory, 82-Computer program. Detailed Implementation

[0022] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail.

[0023] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0024] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0025] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determination" or "if the described condition or event is detected" may be interpreted, depending on the context, as "once determination," "in response to determination," "once the described condition or event is detected," or "in response to the detection of the described condition or event."

[0026] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0027] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0028] To better understand the shoe sole 3D digital model construction and intelligent recommendation system provided in the embodiments of this application, the specific components of the shoe sole 3D digital model construction and intelligent recommendation system provided in the embodiments of this application will be described by way of example below.

[0029] like Figure 1 As shown in the figure, this application provides a 3D digital model construction and intelligent recommendation system for shoe soles, including: a data acquisition module 100, a modeling module 200, a parameter optimization module 300, a brand database 400, an intelligent recommendation module 500, and an output module 600.

[0030] The data acquisition module 100 is used to acquire user foot morphology data, as well as user-inputted exercise scenarios and exercise intensity parameters. The data acquisition module 100 includes a mobile image acquisition unit 110 and a professional-grade scanning unit 120. The mobile image acquisition unit 110 captures the user's feet using a mobile phone camera, extracts key points of the feet using computer vision algorithms, and generates a foot point cloud model, suitable for home or remote scenarios. The professional-grade scanning unit 120 includes a 3D foot scanner 121, a dynamic plantar pressure plate 122, and a video gait analysis system 123, suitable for store or laboratory environments, and can acquire high-precision foot models and gait data. Users input exercise scenarios (such as walking, running, hiking, rehabilitation, etc.) and exercise intensity (such as weekly exercise volume, pace range, etc.) through an interactive interface.

[0031] The reason for this design is that different application scenarios have different requirements for data accuracy. Mobile data collection lowers the barrier to entry for users, while professional-grade data collection provides more accurate data for in-depth customization. By combining the two data collection methods, the system can adapt to multi-level needs ranging from mass consumption to professional sports.

[0032] As an example, the specific method for a mobile phone camera to capture a user's feet can be as follows: Following the on-screen prompts, the user places both feet on a designated background (including a calibration pattern), and images of the left and right feet from three perspectives—front, side, and bottom—are captured. After image acquisition, the mobile image acquisition unit 110 uses a deep learning-based 3D reconstruction algorithm to generate a foot point cloud model, specifically including: First, a pre-trained convolutional neural network (such as HRNet or ResNet-50) is used to detect foot keypoints in multi-view images, extracting the two-dimensional coordinates of 26 anatomical keypoints, including the foot length endpoint, metatarsophalangeal circumference, arch apex, and heel point. The training dataset for the keypoint detection network contains more than 10,000 human foot images annotated with foot keypoints.

[0033] Secondly, the Structure from Motion (SfM) algorithm is used to sparsely reconstruct multi-view images, estimate camera pose, and obtain sparse point clouds of the feet. Specifically, an incremental SfM method is used, which obtains camera parameters and sparse 3D points through steps such as feature extraction (SIFT features), feature matching, geometric verification, and bundle adjustment.

[0034] Then, a multi-view stereo (MVS) algorithm is used to densify the sparse point cloud to generate a dense point cloud for the foot. MVSNet or its variants, based on deep learning, can be used to obtain a dense point cloud containing at least 10 points per millimeter by constructing cost volume, depth map estimation, and depth fusion.

[0035] Finally, the dense point cloud is post-processed, including: removing outliers using statistical filtering, smoothing using moving least squares (MLS), generating a mesh model using the Poisson surface reconstruction algorithm, and aligning the mesh model with the foot anatomy template using a non-rigid registration algorithm to fill in holes caused by occlusion. After the above processing, the generated foot point cloud model can achieve an accuracy of less than 1mm, meeting the basic requirements for shoe sole modeling.

[0036] As an alternative embodiment, those skilled in the art can also use the LiDAR scanning function of the iPhone to directly obtain foot depth data and combine it with the ARKit or ARCore framework for 3D reconstruction. This method can further improve scanning accuracy and speed.

[0037] As an example, the dynamic plantar pressure plate 122 and the video gait analysis system 123 in the professional-grade scanning unit 120 achieve strict time axis alignment through a hardware synchronization trigger. Specifically, the system uses a unified master clock source and sends synchronization trigger signals simultaneously to the pressure plate acquisition card and the high-speed camera via a BNC interface, ensuring that the start time of pressure data sampling and image acquisition is completely consistent. The pressure plate sampling frequency is set to 200Hz, and the high-speed camera sampling frequency is set to 100fps; both are aligned to the same time axis through interpolation. In the data acquisition software, each pressure data frame and image frame is assigned a unified timestamp with a timestamp accuracy down to the microsecond level. After acquisition, the system cross-validates the pressure center trajectory and the heel contact time in the video using a cross-correlation algorithm; if the time deviation exceeds 1 millisecond, automatic correction and alignment are performed.

[0038] The modeling module 200 is connected to the data acquisition module 100 and is used to generate a basic model of the shoe sole based on foot morphology data and motion parameters. The modeling module 200 includes: a phase weight allocation unit 210, a velocity weight allocation unit 220, a velocity-stiffness mapping unit 230, a weighted fusion unit 240, and a model generation unit 250.

[0039] The phase weight allocation unit 210 assigns designed weights to each movement phase in the gait cycle based on the user-selected movement scenario. The gait cycle includes the heel strike phase, full foot support phase, heel lift-off phase, and toe push-off phase. For example, in running scenarios, the heel strike phase and toe push-off phase have higher weights; in walking scenarios, the full foot support phase has higher weights. Users can fine-tune the weights on the interface or select a default template. The weight allocation is based on biomechanical principles and pre-experimental data, and weight templates for various scenarios can be preset.

[0040] The speed weight allocation unit 220 assigns weights to different exercise speeds based on the user-inputted exercise intensity parameters. Exercise intensity parameters may include the user's primary daily exercise speed and speed distribution preferences. The system divides speed into low-speed (3-5 km / h), medium-speed (6-8 km / h), and high-speed (>9 km / h) ranges, and assigns weights to each speed range according to the user's preferences. V i For example, low speeds are given higher weight when daily walking is the primary activity, while medium to high speeds are given higher weights when running training is the primary activity. Users can adjust their speed preferences using a slider.

[0041] The velocity-stiffness mapping unit 230 is used to establish a mapping model between the dynamic stiffness of the foot and the velocity of movement in each motion phase based on biomechanical experimental data. K j ( v )= f j ( v This mapping model reflects the nonlinear response characteristics of foot soft tissue under different loading rates and can be constructed through curve fitting or machine learning methods. For example, for the heel strike phase, plantar pressure and foot deformation data at multiple speeds can be collected, and the power-law function K(v)=a·v can be obtained by least squares fitting. b +c. This mapping model is stored in the system for later use.

[0042] As an example, the velocity-stiffness mapping model K j ( v )= f j ( v This was obtained through the following biomechanical experiments and modeling methods: Thirty healthy participants (15 males and 15 females, aged 20-45 years, foot sizes S, M, and L) were recruited, excluding those with foot diseases or a history of surgery. Participants exercised on a treadmill at 10 different speeds: 3km / h, 4km / h, 5km / h, 6km / h, 7km / h, 8km / h, 9km / h, 10km / h, 11km / h, and 12km / h. Data was collected for 30 seconds after each speed stabilized. A dynamic plantar pressure plate (sampling frequency 200Hz) was used to collect plantar pressure distribution data, and a high-speed motion capture system (sampling frequency 200Hz, 8 infrared cameras) was used to capture the 3D trajectory of 36 marker points on the foot surface, distributed across major skeletal landmarks.

[0043] For each motion velocity v and each motion phase j (heel strike, full foot support, heel lift, toe extension), calculate the equivalent dynamic stiffness K in that phase. j (v). The specific method is as follows: extract the peak value F of the plantar reaction force within this phase time period. max The stiffness calculation formula is K=F, which is related to the change in arch height Δh. max / Δh. The change in arch height is calculated using the vertical displacement of the navicular bone marker. For each subject, the average of 10 gait cycles for each velocity and phase is taken as the stiffness value under that condition. All collected data from subjects are categorized by foot type, and a mapping model is established for each foot type. For each motion phase j, velocity v is used as input, and stiffness K... j (v) is the output, and the following method is used for fitting: (1) Power-law function fitting: K(v) = a·v b +c, using the Levenberg-Marquardt algorithm for nonlinear least-squares fitting, we obtain parameters a, b, and c. For normal foot type, the fitting result for heel strike phase is K = 15.2·v 1.3 +8.5 (R²=0.92).

[0044] As an alternative, a mapping model can be established using Support Vector Regression (SVR) or Gaussian Process Regression, and the optimal kernel function and hyperparameters can be selected through cross-validation. Experiments show that for data with strong nonlinearity, Gaussian Process Regression has higher fitting accuracy (R²>0.95).

[0045] As a preferred solution, this system adopts an ensemble learning method, which combines power-law function fitting with neural networks. First, the baseline curve is obtained by fitting with a power-law function, and then the residuals are learned by a shallow neural network (2 hidden layers, 10 neurons per layer) to improve the model accuracy.

[0046] The above sample size is only used to illustrate the acquisition of the velocity-stiffness mapping model. Increasing the sample size is beneficial to improving the accuracy of the velocity-stiffness mapping model.

[0047] The weighted fusion unit 240 connects the phase weight allocation unit 210, the velocity weight allocation unit 220, and the velocity-stiffness mapping unit 230, and is used for phase weight-based fusion. W ij and speed weight V i Morphological parameters of different areas of the foot S ijk Perform two-layer weighted fusion to generate comprehensive design parameters related to speed. The fusion formula is: in, The comprehensive design parameters for the k-th region (such as the medial forefoot, lateral forefoot, arch, heel, etc.) at velocity v include target stiffness, surface curvature, etc. V i The weight of the i-th type of motion speed, W ij The weight of the j-th phase at the i-th velocity. S ijk The morphological parameters of the k-th region at the j-th phase under the i-th velocity are extracted from the collected point cloud data. In the first k The rate of change of the region's morphology This is the dynamic adjustment coefficient (usually taken as 0.1-0.3). K k ( v ) represents the velocity-stiffness mapping value for the k-th region, and β is the stiffness contribution coefficient (usually taken as 0.5-0.8). This formula organically combines static morphology, dynamic deformation, and velocity-related stiffness, enabling the generated design parameters to comprehensively reflect the complex response of the foot under multi-velocity conditions.

[0048] The reason for this design is that a single velocity or single phase design cannot meet the diverse needs of the foot in actual sports. Through dual-layer weighted fusion, the system can comprehensively consider the importance of different velocities and phases, while introducing dynamic deformation compensation and stiffness mapping, giving the generated sole model true adaptive capability.

[0049] Among them, morphological parameters S ijk Specifically, it refers to the foot surface morphological feature vector in the k-th region during the j-th phase of motion at the i-th motion speed. The morphological parameters of each region include the following dimensions: (1) Regional curvature characteristics: The normal vector distribution of the point cloud in the region is calculated by principal component analysis to obtain the Gaussian curvature and the mean curvature, which reflect the degree of surface undulation in the region. The specific calculation formula is: Gaussian curvature K=κ1·κ2, mean curvature H=(κ1+κ2) / 2, where κ1 and κ2 are the principal curvatures.

[0050] (2) Regional height characteristics: Calculate the average height and height variance of all points in the region relative to the plantar reference plane. The plantar reference plane is determined by fitting three points: the heel point, the first metatarsal head, and the fifth metatarsal head.

[0051] (3) Region contour features: Extract the length, curvature and enclosed area of ​​the region boundary line to describe the size and shape of the region.

[0052] (4) Relative position characteristics of the region: Calculate the spatial coordinate offset of the center point of the region relative to key points of the foot (such as the arch apex and heel point).

[0053] As an example, the method for extracting morphological parameters is as follows: First, the foot point cloud model is segmented according to a pre-defined scheme of 45 regions. The boundaries of each region are determined based on anatomical landmarks, such as the line connecting the 1st to 5th metatarsal heads to divide the forefoot region, and the scaphoid tuberosity to locate the origin of the arch. Then, the point cloud within each region is reconstructed into a mesh, and the aforementioned feature values ​​are calculated. After normalization, all feature values ​​form the morphological parameter vector for that region. S ijk It has 10 dimensions.

[0054] As an example, the dynamic adjustment coefficient The value of the stiffness contribution coefficient β is determined by the following method: First, establish a standard test sample, such as 10 subjects with different foot characteristics (3-4 each of flat feet, normal feet, and high arches), and collect their foot dynamic data at speeds of 3km / h, 5km / h, and 8km / h.

[0055] Secondly, cross-validation is used to optimize the parameters: The value of β is taken in the interval [0, 0.5] with a step size of 0.05, and the value of β is taken in the interval [0, 1] with a step size of 0.1. For each group ( The corresponding shoe sole model is generated by combining (β) and the foot pressure uniformity index is calculated by finite element simulation.

[0056] Then, with the goal of achieving optimal plantar pressure uniformity, the option that minimizes the standard deviation of the pressure distribution is selected. The optimal parameters are (β) combinations. Experimental results show that for normal foot type, the optimal parameters are... =0.2, β=0.7; for flat feet, =0.25, β=0.6; for high arches, =0.15, β=0.75.

[0057] In practical applications, the system first automatically identifies the foot type (flat feet / normal feet / high arches) based on the user's foot scan data, and then loads the corresponding default parameters. Users can also manually adjust these parameters in advanced settings. The system displays the adjusted pressure distribution prediction effect in real time based on the values ​​of β and β.

[0058] The above-mentioned number of subjects is only used as a dynamic adjustment coefficient. The method for determining the stiffness contribution coefficient β is illustrated. Increasing the number of subjects helps to improve the accuracy of the determination.

[0059] Model generation unit 250 is connected to weighted fusion unit 240, used to generate models based on comprehensive design parameters. A three-dimensional digital model of the shoe sole was constructed. First, weighted principal component analysis was used to spatially align the morphological data at various velocities. Then, a quadrilateral mesh reconstruction algorithm was used to generate a continuous and smooth mesh for the sole surface. The mesh model was divided into 45 functional regions, each with a target stiffness value related to velocity. The model generation unit 250 also matched the corresponding internal filling structure from the lattice structure library based on the target stiffness value to generate a complete digital model of the shoe sole.

[0060] The parameter optimization module 300 is connected to the modeling module 200 and is used for multi-objective optimization of the regional stiffness distribution of the shoe sole model. The optimization objective function is: min F ( x )=[ f 1( x ), f 2( x ), f 3( x )] in, f 1( x The plantar pressure uniformity is calculated using finite element analysis (the smaller the standard deviation of the plantar pressure distribution, the more uniform it is). f 2( x The weight of the shoe sole is calculated using the volume fraction of the lattice structure. f 3( xThe energy feedback efficiency is estimated using material resilience and structural deformation recovery time. Design variables x include the lattice structure volume fraction, lattice topology type, and material hardness for 45 regions. Constraints include maximum deformation ≤ 3mm and stiffness gradient between adjacent regions ≤ 20%. The parameter optimization module 300 employs a multi-objective evolutionary algorithm (such as NSGA-II) for global optimization, generating a Pareto front for users to select according to their preferences. Users can choose schemes emphasizing comfort, lightweight design, or motion performance; the system outputs the corresponding regional stiffness distribution based on the selection.

[0061] The reason for this design is that sole design involves multiple conflicting performance metrics, and single-objective optimization cannot meet the personalized needs of users. By generating a Pareto front through multi-objective optimization, users can intuitively see the trade-offs between different performance aspects and make choices based on their own preferences, achieving true personalized customization.

[0062] The stiffness gradient between adjacent regions is defined as follows: for any two adjacent functional regions m and n, their stiffness value G... m and G n It should satisfy |G m -G n | / max(G m G n ≤20%.

[0063] Adjacent regions are determined based on the adjacency relationship of the shoe sole mesh model: in the quadrilateral mesh model, two regions sharing an edge are defined as adjacent regions. The system pre-establishes a region adjacency matrix A, where A... mn =1 indicates that regions m and n are adjacent; otherwise, it is 0.

[0064] During the optimization process, after each iteration generates a new stiffness distribution, the system traverses all adjacent region pairs and calculates the stiffness gradient value. If there are more than 20% of adjacent region pairs, a penalty term is introduced into the objective function. The formula for calculating the penalty term is: Where λ is the penalty coefficient (set to 1000), ensuring that the optimization algorithm prioritizes satisfying the gradient constraint. This design can prevent excessively large abrupt changes in stiffness between adjacent regions, avoiding discomfort when wearing the garment.

[0065] The Brand Database 400 stores 3D model data, material performance data, and functional labels for footwear from multiple brands. The database acquires 3D models of footwear through reverse scanning or brand collaborations, obtains material hardness, resilience, and density parameters through laboratory testing or publicly available data, and assigns functional labels (such as cushioning, support, racing, and stability) to each footwear model through expert annotation. The database supports regular updates to stay in sync with the market.

[0066] The intelligent recommendation module 500 connects to the data acquisition module 100 and the brand database 400 to calculate the matching degree between the user's feet and various shoe brands. The matching degree calculation uses a multi-layer weighted algorithm: MatchScore= w g ⋅ G match + w f ⋅ F match + w p ⋅ P match in, G match The geometric matching degree (40% weight) is calculated based on the fit between foot length, foot circumference, and arch height and the corresponding shoe size. It is obtained by comparing the distance between the user's key foot points (foot length endpoint, metatarsophalangeal circumference point, arch apex, and heel point) and the corresponding point cloud of the shoe's inner cavity. F match Functional matching degree (35% weight, calculated based on the degree of matching between user gait type (pronation / pronation / normal), sports scenario and shoe stability design, speed adaptability, which can be achieved through preset rules or machine learning models); P match The preference matching score (25% weight) is calculated based on user brand preference, price range, and appearance style preference. w g , w f , w p For the corresponding weights, and w g + w f + w p =1, in this embodiment w g Take 40%, w f Take 35%, w p Take 25%. The system outputs a list of recommended shoes sorted by matching degree and generates a foot fit heat map to visually display the fit of each area.

[0067] As an example, the intelligent recommendation module calculates the geometric matching degree 500. G match The specific steps are as follows: (1) Acquisition of point cloud of shoe cavity: The 3D model of the shoe stored in the brand database 400 is either a shoe last model or a scanned model of the inner surface of the shoe cavity. For shoes where the inner cavity data cannot be obtained, a reverse engineering method is used: the actual shoe is placed in a 3D scanner to obtain the point cloud of the outer surface of the shoe, and then the inner cavity point cloud is obtained by offsetting inward according to the material and thickness of the upper (measured by an ultrasonic thickness gauge). The offset amount is adjusted according to the material thickness difference of different parts of the shoe, such as the upper material thickness of 1-2mm and the insole thickness of 3-5mm.

[0068] (2) Determination of Key Point Correspondence: Key points of the user's foot include the foot length endpoint (the foremost point of the toes), the metatarsophalangeal junction (the heads of the first and fifth metatarsal bones), the arch apex (the tuberosity of the navicular bone), and the heel point (the posterior border of the calcaneus). For each key point, the corresponding matching point is determined in the point cloud of the shoe's inner cavity: Foot length endpoint matching: Find the foremost point along the longitudinal central axis of the shoe's inner cavity as the corresponding point; Heel point matching: Find the point at the very end along the central axis as the corresponding point; Metatarsophalangeal circumferential point matching: On a cross-section perpendicular to the length of the foot, find the two points with the largest width, corresponding to the first and fifth metatarsal heads respectively; Foot arch apex matching: Find the center point of the region with the greatest curvature change along the curve of the inner cavity bottom surface.

[0069] (3) Distance calculation: For each key point, calculate the Euclidean distance d between the user's foot point and the corresponding point of the shoe model. i Because different key points have different levels of importance, they are assigned different weights. i For example: the foot length endpoint has a weight of 0.3, the metatarsophalangeal circumference point has a weight of 0.3, the arch apex has a weight of 0.25, and the heel point has a weight of 0.15. The weighted average distance is D=Σ(w i ·d i ).

[0070] (4) Matching score conversion: Convert the weighted average distance D into a matching score between 0 and 1. The conversion formula is as follows: ， Where k is the steepness coefficient (taken as 2), and D0 is the ideal matching distance (taken as 3mm). When D=0mm, G match ≈1; When D=5mm, G match ≈0.5; When D=10mm, G match ≈0.12.

[0071] The output module 600 connects to the parameter optimization module 300 and the intelligent recommendation module 500. It exports the optimized sole model as a 3D printing file (such as STL format) and generates a recommended shoe report containing a foot fit heatmap. Users can choose to customize printing or purchase recommended brand shoes according to their needs.

[0072] As an example, the foot fit heatmap output by output module 600 is generated through the following steps: (1) Fit Calculation: The user's foot model and the shoe's inner cavity model are non-rigidly registered to align them at key anatomical points. After alignment, for each vertex p on the foot model, the shortest distance d(p) to the shoe's inner cavity model is calculated. A positive distance indicates a gap between the point and the cavity, while a negative distance indicates compression (penetration) of the point by the cavity. The fit index is defined as f(p) = 1 - |d(p)| / D max D max The maximum permissible deviation is 5mm. The closer the f(p) value is to 1, the better the fit; the closer it is to 0, the worse the fit.

[0073] (2) Color Mapping: The f(p) value is mapped to the color space: f(p)≥0.8 is mapped to blue (good fit), 0.5≤f(p)<0.8 is mapped to green (moderate fit), 0.2≤f(p)<0.5 is mapped to yellow (loose / tight fit), and f(p)<0.2 is mapped to red (severe mismatch). The system uses a rainbow color scheme as an alternative to meet the visual preferences of different users.

[0074] (3) Heatmap rendering: Assign color values ​​to each vertex of the foot model and render in real time using OpenGL or WebGL to generate a rotatable and scalable 3D heatmap. Users can click on any area of ​​the heatmap with the mouse to view the specific distance value of that point and the suggested adjustment amount (e.g., "3.5mm gap at the arch, it is recommended to choose a support insole").

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules and units is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit module can exist physically separately, or two or more unit modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application.

[0076] Corresponding to the aforementioned three-dimensional digital model construction and intelligent recommendation system for shoe soles, this embodiment also provides a method for constructing three-dimensional digital models of shoe soles and providing intelligent recommendations, such as... Figure 2 As shown, it includes the following steps: Step S1: Collect user foot morphology data and motion parameters.

[0077] Understandably, foot morphology data can be obtained through mobile photography or professional-grade scanning, while motion parameters, including the motion scenario and intensity, are input by the user through the interface. This design is based on the fact that data collection is fundamental to all subsequent analysis, and accurate data ensures the reliability of the model.

[0078] Step S2: Assign design weights to each motion phase in the gait cycle according to the motion scenario, and assign weights to different motion speeds according to the motion intensity.

[0079] It's understandable that the allocation of phase and velocity weights directly affects the subsequent fusion results. The system provides a default template for users to choose from, and also allows users to customize it. This design is because different users have different motion needs, and giving users the ability to adjust can enhance personalization.

[0080] Step S3: Based on the phase weight and velocity weight, and combined with the pre-established velocity-stiffness mapping model, the morphological parameters of each region of the foot are fused in a two-layer weighted manner to generate comprehensive design parameters related to velocity.

[0081] As can be understood, the dual-layer weighted fusion formula, as described above, involves a velocity-stiffness mapping model pre-established and stored in the system through biomechanical experiments. This design is intended to integrate information from multiple dimensions, allowing the comprehensive design parameters to fully reflect the dynamic needs of the foot and provide a scientific basis for subsequent modeling.

[0082] Step S4: Construct a three-dimensional digital model of the shoe sole based on the comprehensive design parameters.

[0083] It is understandable that the construction process includes spatial alignment, mesh generation, and internal infill structure matching. The reason for this design is that only by converting the design parameters into a specific geometric model can they be used for subsequent optimization and manufacturing.

[0084] Step S5: Using foot pressure uniformity, sole weight, and energy feedback efficiency as objective functions, perform multi-objective optimization on the regional stiffness distribution of the three-dimensional digital model of the sole.

[0085] It's understandable that the optimization process employs a multi-objective evolutionary algorithm to generate a Pareto front for users to choose from. This design is based on the fact that multi-objective optimization can balance multiple performance metrics, satisfying different user preferences.

[0086] Step S6: Calculate the geometric matching degree, functional matching degree, and preference matching degree between the user's feet and each shoe in the brand database, and output a list of recommended shoes.

[0087] It's understandable that a weighted algorithm is used to calculate the matching score, and a heatmap is generated for visual visualization. This design is because accurate recommendations require consideration of multiple dimensions, and the heatmap helps users understand the basis for the recommendations, increasing their trust in the recommendations.

[0088] The specific working process of the above method can be referred to the corresponding specific working process of the modules and units in the aforementioned system embodiments, and will not be repeated here.

[0089] This application also provides an electronic device. Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 3 As shown, the electronic device 8 of this embodiment includes: at least one processor 80 ( Figure 3 Only one is shown in the image), at least one memory 81 ( Figure 3 (Only one is shown in the image) and a computer program 82 stored in the at least one memory 81 and executable on the at least one processor 80. When the processor 80 executes the computer program 82, it causes the electronic device 8 to perform the steps in any of the above embodiments of the three-dimensional digital model construction and intelligent recommendation method for shoe soles, or causes the electronic device 8 to perform the functions of each unit in the above embodiments of the device.

[0090] For example, the computer program 82 may be divided into one or more units, which are stored in the memory 81 and executed by the processor 80 to complete this application. The one or more units may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 82 in the construction of a three-dimensional digital model of the shoe sole and intelligent recommendation.

[0091] Electronic device 8 can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. This electronic device may include, but is not limited to, a processor 80 and a memory 81. Those skilled in the art will understand that... Figure 3 This is merely an example of electronic device 8 and does not constitute a limitation on electronic device 8. It may include more or fewer components than shown, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, buses, etc.

[0092] The processor 80 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0093] In some embodiments, the memory 81 may be an internal storage unit of the electronic device 8, such as a hard disk or memory of the electronic device 8. In other embodiments, the memory 81 may be an external storage device of the electronic device 8, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 8. Furthermore, the memory 81 may include both internal and external storage units of the electronic device 8. The memory 81 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 81 can also be used to temporarily store data that has been output or will be output.

[0094] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0095] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0096] This application also provides a computer program product that, when run on an electronic device, causes the electronic device to perform the steps in any of the above method embodiments.

[0097] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A system for constructing and intelligently recommending three-dimensional digital models of shoe soles, characterized in that, include: The data acquisition module is used to acquire user foot shape data as well as user-inputted exercise scenarios and exercise intensity parameters; The modeling module, connected to the data acquisition module, is used to generate a basic three-dimensional digital model of the user's shoe sole based on the foot morphology data and exercise intensity parameters; The parameter optimization module, connected to the modeling module, is used to perform multi-objective optimization of the regional stiffness distribution of the basic three-dimensional digital model of the sole, with foot pressure uniformity, sole weight and energy feedback efficiency as objective functions, to generate a user sole optimized three-dimensional digital model. The brand database is used to store 3D digital model data, material performance data, and functional labels for multiple shoe models; The intelligent recommendation module connects the data acquisition module and the brand database to calculate the geometric matching degree, functional matching degree, and preference matching degree between the user's feet and the shoes of each brand, and outputs a list of recommended shoes. The modeling module includes: A phase weight allocation unit is used to allocate the design weights of each motion phase in the gait cycle according to the motion scenario. A speed weight allocation unit is used to allocate weights for different motion speeds according to the motion intensity. The velocity-stiffness mapping unit is used to establish a mapping model between the dynamic stiffness of the foot and the velocity of movement under each motion phase based on biomechanical experimental data. The weighted fusion unit, which connects the phase weight allocation unit, the velocity weight allocation unit, and the velocity-stiffness mapping unit, is used to perform two-layer weighted fusion of the morphological parameters of each region of the foot based on the phase weight, velocity weight, and velocity-stiffness mapping model, to generate comprehensive design parameters related to velocity. The model generation unit, connected to the weighted fusion unit, is used to construct a three-dimensional digital model of the shoe sole based on the comprehensive design parameters.

2. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 1, characterized in that, The weighted fusion unit performs a two-layer weighted fusion specifically as follows: in, The comprehensive design parameters for the k-th region at velocity v are: V i The weight of the i-th type of motion speed, W ij The weight of the j-th phase at the i-th velocity. S ijk Let be the morphological parameters of the k-th region at the j-th phase under the i-th velocity. In the first k The rate of change of the region's morphology This is a dynamic adjustment coefficient. K k ( v ) represents the velocity-stiffness mapping value of the k-th region, and β is the stiffness contribution coefficient.

3. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 2, characterized in that, The mapping model established by the velocity-stiffness mapping unit is expressed as follows: K j ( v )= f j ( v ),in K j ( v ) represents the equivalent dynamic stiffness of the j-th motion phase at velocity v. The mapping model is trained from biomechanical experimental data through curve fitting or machine learning methods.

4. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 1, characterized in that, The parameter optimization module uses a multi-objective evolutionary algorithm to perform global optimization, generates a Pareto front for the user to choose from, and outputs the corresponding regional stiffness distribution according to the user's preferred scheme; the preferred scheme includes at least one of comfort, lightweight or sports performance.

5. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 1, characterized in that, The formula for calculating the matching degree by the intelligent recommendation module is as follows: MatchScore= w g ⋅ G match + w f ⋅ F match + w p ⋅ P match in, G match Geometric matching is calculated based on the degree of fit between foot length, foot circumference, and arch height and the corresponding shoe size. F match Functional fit is calculated based on the degree of matching between the user's gait type, sports scenario, and the shoe's stability design and speed adaptability. P match The preference matching score is calculated based on user brand preference, price range, and appearance style preference. w g , w f , w p For the corresponding weights, and w g + w f + w p =1.

6. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 5, characterized in that, The geometric matching degree is obtained by comparing the distance between key points on the user's foot and the corresponding point cloud inside the shoe cavity. The key points on the foot include the foot length endpoint, metatarsophalangeal circumference point, arch apex, and heel point.

7. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 1, characterized in that, The data acquisition module includes a mobile image acquisition unit and a professional-grade scanning unit. The mobile image acquisition unit captures images of the user's feet using a mobile phone camera, extracts key points of the feet using computer vision algorithms, and generates a foot point cloud model. The professional-grade scanning unit includes a 3D foot scanner, a dynamic plantar pressure plate, and a video gait analysis system.

8. The three-dimensional digital model construction and intelligent recommendation system for shoe soles according to claim 1, characterized in that, The system also includes an output module, which connects the parameter optimization module and the intelligent recommendation module, for exporting the optimized sole model as a 3D printing file and generating a recommended shoe report containing a foot fit heatmap.

9. A method for constructing and intelligently recommending three-dimensional digital models of shoe soles, applied to the system described in any one of claims 1 to 8, characterized in that, Includes the following steps: S1. Collect user foot morphology data and motion parameters; S2. Assign design weights to each motion phase in the gait cycle according to the motion scenario, and assign weights to different motion speeds according to the motion intensity; S3. Based on the phase weight and velocity weight, and combined with the pre-established velocity-stiffness mapping model, the morphological parameters of each region of the foot are weighted and fused in a two-layer manner to generate comprehensive design parameters related to velocity. S4. Construct a three-dimensional digital model of the shoe sole based on the comprehensive design parameters; S5. Using foot pressure uniformity, sole weight, and energy feedback efficiency as objective functions, perform multi-objective optimization on the regional stiffness distribution of the three-dimensional digital model of the sole. S6. Calculate the geometric matching degree, functional matching degree, and preference matching degree between the user's feet and each shoe in the brand database, and output a list of recommended shoes.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for constructing and intelligently recommending three-dimensional digital models of shoe soles as described in claim 9.

Images