A knee joint stress distribution calculation method, device, equipment and storage medium

Abstract

Embodiments of the present application provide a knee joint stress distribution calculation method, device, equipment and storage medium. The method comprises: establishing a knee joint finite element model; setting different knee joint conditions for the knee joint finite element model, using the knee joint finite element model to perform simulation calculation under different knee joint conditions, and obtaining stress distribution simulation results of the knee joint under different knee joint conditions; taking the knee joint conditions as samples and the stress distribution simulation results corresponding to the knee joint conditions as labels, constructing a data set, training a machine learning model using the data set, and obtaining a knee joint stress distribution calculation model; and using the knee joint stress distribution calculation model to calculate the stress distribution of the knee joint under a specified knee joint condition. In this way, the combination of finite element simulation and machine learning can be used to quickly and accurately calculate the stress distribution of the knee joint.

Description

Technical Field

[0001] This invention relates to the field of medical assessment technology, and in particular to a method, apparatus, device, and storage medium for calculating stress distribution in a knee joint. Background Technology

[0002] The knee joint is a vital part of the human body, and its health is closely related to the normal functioning of the body. Knee injuries are common problems in all population groups, including damage to the knee cartilage, anterior cruciate ligament (ACL) injuries, and meniscus injuries. Related diseases such as knee osteoarthritis severely affect quality of life, causing motor impairments such as walking speed, gait, and balance problems, and in severe cases, can lead to disability.

[0003] Meanwhile, the knee joint is the most complex joint in the human body, exhibiting significant differences between individuals. Therefore, personalized and specific medical plans are needed to address knee injuries. Knee injuries are significantly related to their biomechanical environment and factors. Stress analysis is conducted in clinical practice and research to obtain biomechanical information at the tissue level of the knee joint. The finite element method (FEM) is commonly used to analyze the biomechanical environment of the knee joint structure to understand its health status and risks, and to assess motor function. However, the FEM method for simulating and analyzing the biomechanical behavior of human tissues typically involves complex geometric structures, material nonlinearity, and dynamic loading conditions. It is a nonlinear, high-dimensional problem, making the modeling process cumbersome and computationally time-consuming, which is difficult to meet the needs of personalized and rapid clinical practice. Therefore, how to quickly and accurately calculate the stress distribution of the knee joint has become an urgent technical problem to be solved. Summary of the Invention

[0004] In a first aspect, embodiments of the present invention provide a method for calculating the stress distribution of a knee joint, the method comprising:

[0005] Establish a finite element model of the knee joint;

[0006] Different knee joint conditions were set for the knee joint finite element model, and simulation calculations were performed using the knee joint finite element model under different knee joint conditions to obtain the simulation results of stress distribution of the knee joint under different knee joint conditions;

[0007] Using knee joint conditions as samples and stress distribution simulation results corresponding to knee joint conditions as labels, a dataset is constructed. The dataset is then used to train the machine learning model to obtain a knee joint stress distribution calculation model.

[0008] The stress distribution of the knee joint under specified knee joint conditions is calculated using a knee joint stress distribution calculation model.

[0009] Among the possible implementations of the first aspect, a finite element model of the knee joint is established, including:

[0010] Based on knee joint imaging data, a three-dimensional reconstruction of the knee joint was performed to obtain a three-dimensional geometric model of the knee joint. The three-dimensional geometric model of the knee joint was then smoothed and fitted with surfaces, and meshed. The corresponding material properties of each structure of the three-dimensional geometric model of the knee joint were then assigned, and boundary conditions and loading conditions were set for the three-dimensional geometric model of the knee joint to obtain a finite element model of the knee joint.

[0011] In some possible implementations of the first aspect, knee joint conditions include knee joint geometric parameters, knee joint material properties, and knee joint load states; different knee joint conditions are set for the knee joint finite element model, including:

[0012] Set at least one different knee joint condition among multiple knee joint geometric parameters, knee joint material properties, and knee joint load states for the knee joint finite element model.

[0013] In some possible implementations of the first aspect, a machine learning model is trained using a dataset to obtain a knee joint stress distribution calculation model, including:

[0014] The dataset is preprocessed, including data cleaning, data standardization, data structuring, and data augmentation.

[0015] The preprocessed dataset is divided into training set, validation set and test set;

[0016] The machine learning model is trained using the training set, the hyperparameters of the trained machine learning model are optimized using the validation set, and the performance of the trained machine learning model is evaluated using the test set. If the trained machine learning model passes the performance evaluation, it is used as the knee joint stress distribution calculation model.

[0017] Among some possible implementations of the first aspect, the method also includes:

[0018] Knee injury assessment is performed based on the stress distribution of the knee joint under specified knee joint conditions.

[0019] Among some possible implementations of the first aspect, the method also includes:

[0020] Multiple knee joint conditions were selected as target knee joint conditions from different knee joint conditions;

[0021] The simulation accuracy of the knee joint finite element model is calculated based on the simulation results of stress distribution under various target knee joint conditions and the actual stress distribution under various target knee joint conditions obtained through physical experiments and / or literature searches.

[0022] If the simulation accuracy is greater than or equal to the preset threshold, the simulation results of stress distribution of the knee joint under different knee joint conditions will be corrected.

[0023] If the simulation accuracy is less than the preset threshold, the source of error in the knee joint finite element model is analyzed based on the simulation results of stress distribution under each target knee joint condition and the actual stress distribution results. Based on this, the knee joint finite element model is optimized, and the optimized knee joint finite element model is reused to perform simulation calculations under different knee joint conditions to obtain the stress distribution simulation results of the knee joint under different knee joint conditions.

[0024] In some feasible ways of implementing the first aspect, the simulation results of stress distribution of the knee joint under different knee joint conditions are corrected, including:

[0025] The stress distribution simulation results of the knee joint under different knee joint conditions were corrected using a stress distribution simulation result correction model.

[0026] The stress distribution simulation result correction model is trained through the following steps:

[0027] The stress distribution simulation results are input into the generator of the generative adversarial network to generate false stress distribution results. The false stress distribution results are mixed with the true stress distribution results and input into the discriminator of the generative adversarial network to obtain the authenticity probabilities corresponding to the false stress distribution results and the true stress distribution results respectively. Based on these probabilities, the loss value of the discriminator is calculated, and the discriminator is updated according to the loss value of the discriminator.

[0028] The stress distribution simulation results are input into the generator of the generative adversarial network to generate false stress distribution results. The false stress distribution results are then input into the discriminator of the generative adversarial network to obtain the probability of authenticity of the false stress distribution results. Based on this, the loss value of the generator is calculated, and the generator is updated according to the loss value of the generator.

[0029] The model parameters of the discriminator and the generator are continuously updated iteratively until the preset stopping condition is met, resulting in a well-trained generative adversarial network. The generator in the network is then used as a correction model for the stress distribution simulation results.

[0030] In a second aspect, embodiments of the present invention provide a knee joint stress distribution calculation device, the device comprising:

[0031] A module is created to build a finite element model of the knee joint;

[0032] The simulation module is used to set different knee joint conditions for the knee joint finite element model, and to perform simulation calculations using the knee joint finite element model under different knee joint conditions to obtain the simulation results of stress distribution of the knee joint under different knee joint conditions.

[0033] The training module is used to construct a dataset using knee joint conditions as samples and the stress distribution simulation results corresponding to the knee joint conditions as labels. The dataset is then used to train the machine learning model to obtain a knee joint stress distribution calculation model.

[0034] The calculation module is used to calculate the stress distribution of the knee joint under specified knee joint conditions using a knee joint stress distribution calculation model.

[0035] Thirdly, embodiments of the present invention provide an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; the memory storing instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described above.

[0036] Fourthly, embodiments of the present invention provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method described above.

[0037] In this embodiment of the invention, the combination of finite element simulation and machine learning can be used to quickly and accurately calculate the stress distribution of the knee joint, meeting the needs of personalized and rapid clinical practice, helping to better assess the health status of the knee joint, evaluate the safety and effectiveness of exercise, and guide the formulation of clinical strategies.

[0038] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of the present invention, nor is it intended to restrict the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0039] The above and other features, advantages, and aspects of the various embodiments of the present invention will become more apparent from the accompanying drawings and the following detailed description. The drawings are provided for a better understanding of the invention and are not intended to limit the invention. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0040] Figure 1 A flowchart illustrating a method for calculating knee joint stress distribution provided in an embodiment of the present invention;

[0041] Figure 2 A structural diagram of a knee joint stress distribution calculation device provided in an embodiment of the present invention;

[0042] Figure 3 This is a structural diagram of an exemplary electronic device capable of implementing embodiments of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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.

[0044] Furthermore, the term "and / or" in this invention is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this invention generally indicates that the preceding and following related objects have an "or" relationship.

[0045] To address the technical problems mentioned in the background art, embodiments of the present invention provide a method, apparatus, device, and storage medium for calculating knee joint stress distribution. Specifically, a finite element model of the knee joint is established; different knee joint conditions are set for the finite element model, and simulation calculations are performed using the finite element model under different knee joint conditions to obtain simulation results of stress distribution of the knee joint under different knee joint conditions; using the knee joint conditions as samples and the corresponding stress distribution simulation results as labels, a dataset is constructed, and the dataset is used to train a machine learning model to obtain a knee joint stress distribution calculation model; the knee joint stress distribution calculation model is used to calculate the stress distribution of the knee joint under specified knee joint conditions.

[0046] In this way, the combination of finite element simulation and machine learning can be used to quickly and accurately calculate the stress distribution of the knee joint, meeting the needs of personalized and rapid clinical practice, helping to better assess the health status of the knee joint, evaluate the safety and effectiveness of exercise, and guide the development of clinical strategies.

[0047] The following detailed description, with reference to the accompanying drawings and specific embodiments, illustrates a method, apparatus, device, and storage medium for calculating knee joint stress distribution provided by the present invention.

[0048] Figure 1 A flowchart of a method for calculating knee joint stress distribution provided in an embodiment of the present invention is shown below. Figure 1 As shown, the knee joint stress distribution calculation method 100 may include:

[0049] S1, Establish a finite element model of the knee joint.

[0050] In some embodiments, a three-dimensional reconstruction of the knee joint can be performed based on knee imaging data to obtain a three-dimensional geometric model of the knee joint. The three-dimensional geometric model of the knee joint is then smoothed and fitted with a surface, and a mesh is generated. The corresponding material properties are then assigned to each structure of the three-dimensional geometric model of the knee joint, and boundary conditions and loading conditions are set for the three-dimensional geometric model of the knee joint to obtain a finite element model of the knee joint.

[0051] S2 sets different knee joint conditions for the knee joint finite element model, and uses the knee joint finite element model to perform simulation calculations under different knee joint conditions to obtain the simulation results of stress distribution of the knee joint under different knee joint conditions.

[0052] Different knee joint conditions represent the knee joint characteristics of different populations and different sports.

[0053] In some embodiments, knee joint conditions may include knee joint geometric parameters (e.g., the size, shape, and spatial location of the various structures that make up the knee joint), knee joint material properties, and knee joint load states. Accordingly, setting different knee joint conditions for the knee joint finite element model may include setting at least one different knee joint condition among multiple knee joint geometric parameters, knee joint material properties, and knee joint load states for the knee joint finite element model.

[0054] S3 uses knee joint conditions as samples and the stress distribution simulation results corresponding to the knee joint conditions as labels to construct a dataset. The dataset is then used to train the machine learning model to obtain a knee joint stress distribution calculation model.

[0055] In some embodiments, the above model training may include:

[0056] The dataset is preprocessed (e.g., data cleaning, data standardization, data structuring, data augmentation). The preprocessed dataset is divided into training, validation, and test sets. The training set is used to train a machine learning model (e.g., a deep neural network). The validation set is used to optimize the hyperparameters of the machine learning model during training. The test set is used to evaluate the performance of the trained machine learning model. If the trained machine learning model passes the performance evaluation, it is used as the knee joint stress distribution calculation model.

[0057] S4 uses a knee joint stress distribution calculation model to calculate the stress distribution of the knee joint under specified knee joint conditions.

[0058] In some embodiments, the knee joint stress distribution calculation model can be deployed on a mobile terminal or a computing server. When using the model, the knee joint conditions are input to the input end of the model through the upstream interface. The model calculates the stress distribution of the knee joint under the specified knee joint conditions and outputs the calculation results to the downstream interface through the output end, so that the corresponding device can perform knee joint injury assessment based on the stress distribution of the knee joint under the specified knee joint conditions.

[0059] In summary, according to the embodiments of the present invention, at least the following technical effects are achieved:

[0060] This invention achieves stress analysis of the knee joint from a data-driven perspective by combining traditional numerical simulation. Based on a finite element simulation dataset for various knee joint conditions, it covers a wide range of people and sports, ensuring good generalization of the mapping results. It combines multi-layer networks that describe the characteristics of the finite element model for learning, which greatly reduces the time required for stress analysis, reducing the original finite element analysis solution time from hours to seconds, and provides accurate high-dimensional nonlinear mapping capabilities.

[0061] It should be noted that, in order to improve the realism of the stress distribution simulation results and optimize the accuracy of the knee joint stress distribution calculation model, the knee joint stress distribution calculation method 100 may also include:

[0062] Multiple knee joint conditions are selected as target knee joint conditions from different knee joint conditions. The simulation accuracy of the knee joint finite element model is calculated based on the stress distribution simulation results under each target knee joint condition and the actual stress distribution results under each target knee joint condition obtained through physical experiments and / or literature searches.

[0063] If the simulation accuracy is greater than or equal to a preset threshold, the simulation results of stress distribution of the knee joint under different knee joint conditions are corrected. For example, a stress distribution simulation result correction model is used to correct the simulation results of stress distribution of the knee joint under different knee joint conditions. The stress distribution simulation result correction model can be trained through the following steps:

[0064] (1) Input the stress distribution simulation results into the generator of the generative adversarial network to generate false stress distribution results. Mix the false stress distribution results with the true stress distribution results and input them into the discriminator of the generative adversarial network to obtain the authenticity probabilities corresponding to the false stress distribution results and the true stress distribution results respectively. Calculate the loss value of the discriminator based on this and update the discriminator according to the loss value of the discriminator.

[0065] (2) Input the stress distribution simulation results into the generator of the generative adversarial network to generate false stress distribution results. Input the false stress distribution results into the discriminator of the generative adversarial network to obtain the authenticity probability corresponding to the false stress distribution results. Calculate the loss value of the generator based on this and update the generator according to the loss value of the generator.

[0066] (3) Iteratively update the model parameters of the discriminator and the model parameters of the generator until the preset stopping condition is met, and obtain the trained generative adversarial network, and use the generator in it as the correction model for the stress distribution simulation results.

[0067] If the simulation accuracy is less than the preset threshold, the source of error in the knee joint finite element model (such as modeling error, discretization error, parameter error, etc.) is analyzed based on the simulation results of stress distribution under each target knee joint condition and the actual stress distribution results. Based on this, the knee joint finite element model is optimized, and the optimized knee joint finite element model is reused to perform simulation calculations under different knee joint conditions to obtain the stress distribution simulation results of the knee joint under different knee joint conditions.

[0068] To facilitate further understanding, the above steps will be described in detail below with reference to specific embodiments:

[0069] For example, S1 is subdivided here as follows:

[0070] 1.1 Data Acquisition

[0071] Knee imaging data (such as knee imaging data obtained from CT or MRI scans) is acquired to create a three-dimensional geometric model of the knee joint. The resolution of the acquired data must meet the requirements for accurately reconstructing the anatomical structure of the knee joint to ensure the accuracy of the model.

[0072] 1.2 Geometric Modeling and Mesh Generation

[0073] 1.2.1 Three-dimensional reconstruction

[0074] Knee imaging data is converted into a 3D geometric model of the knee joint using 3D reconstruction software (such as 3D Slicer or Mimics). During the reconstruction process, it is necessary to ensure the integrity of the model's details, especially to maintain the true geometry in complex anatomical structures (such as the cartilage layer around the patella, the cartilage layer at the distal femur and tibial plateau, and the meniscus).

[0075] 1.2.2 Smoothing and Surface Fitting

[0076] The generated 3D geometric model of the knee joint is smoothed to eliminate surface irregularities and sharp edges. Surface fitting algorithms (such as B-spline interpolation) are used to ensure that the model surface is smooth and close to the real structure.

[0077] 1.2.3 Mesh Generation

[0078] The 3D geometric model of the knee joint is imported into relevant preprocessing software (such as HyperMesh), and meshed according to the regional characteristics of the model. A coarser tetrahedral mesh can be used for skeletal structures, while finer meshes are needed for cartilage and ligament areas to ensure accuracy. The meshed model should have a high node density to ensure computational accuracy and simulation stability.

[0079] 1.2.4 Mesh Optimization

[0080] Interpolation algorithms and the mesh optimization tools built into the finite element solver are used to ensure mesh uniformity and quality. Multi-scale optimization techniques (such as adaptive local refinement) are used in densely meshed regions to reduce computational cost, and a combination of tetrahedral and hexahedral meshes is employed to ensure strong stability of the model during stress analysis.

[0081] 1.3 Definition of Material Properties

[0082] 1.31 Material Property Settings

[0083] The various structures in the 3D geometric model of the knee joint are given realistic material properties, including the elastic modulus of bone and the stress-strain relationship of cartilage. Typically, bones are modeled using materials with higher rigidity, while cartilage and ligaments are modeled using different nonlinear material models to accurately simulate their biomechanical properties.

[0084] 1.3.2 Nonlinear Material Model

[0085] For soft tissues (such as articular cartilage and ligaments), Neo-Hookean or Mooney-Rivlin models can be applied, as these models can effectively describe the nonlinear mechanical behavior of soft tissues. Furthermore, anisotropic parameters (such as ligament fiber orientation and tensile strength) can be introduced to enhance the accuracy of the simulation.

[0086] 1.3.3 Boundary Conditions and Loading Conditions

[0087] Set the actual boundary conditions for knee joint movement, such as fixing the specific position of the joint, setting certain rotation or tilt angles, etc. Loading conditions include ground reaction forces during gait and contact forces between the knee joints, ensuring that the model can obtain accurate stress distribution data under different movement states.

[0088] For example, S2 is subdivided here as follows:

[0089] 2.1 Experimental Conditions

[0090] For different human postures (such as standing, walking, running, etc.), different knee joint conditions are set for the finite element model of the knee joint, that is, knee joint mechanical simulation conditions, including knee joint geometric parameters, knee joint material properties, knee joint load state, etc., in order to simulate the mechanical response of the knee joint tissue level under various motion scenarios.

[0091] 2.2 Dynamic Loading Conditions

[0092] Dynamic mechanical loading, including time-varying force and angle changes, is introduced into the finite element model of the knee joint to more closely approximate the stress conditions of the knee joint during actual movement. Gradual loading or dynamic loading is used to simulate the stress response at each stage of movement.

[0093] 2.3 Simulation Solution and Data Acquisition

[0094] Based on the finite element model of the knee joint, mechanical simulation is performed using finite element solvers (such as Abaqus and Ansys). The simulation results of stress distribution in different parts of the knee joint under different knee joint conditions are obtained through calculation, including the contact stress distribution of cartilage, the shear stress of the meniscus, etc., to ensure that stress information of the main parts of the knee joint is covered.

[0095] 2.4 Result Validation

[0096] The accuracy of simulation results should be verified through experiments or existing literature data. For example, the simulation results can be compared with knee joint stress data in the literature or with stress data in physical experiments (such as loading experiments on knee joint models) to ensure that the simulation results have clinical reference value.

[0097] If the verification passes, no further processing is required; if the verification fails, the simulation results of stress distribution of the knee joint under different knee joint conditions are corrected; or, the finite element model of the knee joint is optimized, and the optimized finite element model of the knee joint is reused to perform simulation calculations under different knee joint conditions to obtain the simulation results of stress distribution of the knee joint under different knee joint conditions.

[0098] For example, S3 is subdivided here as follows:

[0099] 3.1 Data Processing

[0100] 3.1.1 Data Cleaning and Standardization

[0101] The stress distribution simulation results generated by the finite element method (FEM) are cleaned to remove invalid and outlier data. The data is then standardized to ensure all values ​​are within a uniform dimension, eliminating the influence of different numerical ranges on model training.

[0102] 3.1.2 Structured Datasets

[0103] Based on the knee joint conditions, the knee joint geometric parameters, knee joint material properties, knee joint load states, and corresponding stress distribution simulation results are integrated into a structured dataset. The data structure is organized in the form of conditional inputs (geometric features, material properties, load features) and outputs (stress distribution values), which facilitates the training of machine learning models.

[0104] 3.1.3 Data Augmentation

[0105] The dataset is augmented by adding small amounts of noise and random transformations to simulate individual differences among different populations and under different movement postures, thereby improving the model's generalization ability. Furthermore, K-fold cross-validation is used to ensure the model can handle diverse input conditions and improve its predictive performance on unknown data.

[0106] 3.1.4 Data Partitioning

[0107] The dataset was divided into training, validation, and test sets in an 8:1:1 ratio. The training set was used for model training, the validation set for hyperparameter optimization, and the test set for evaluating the model's predictive ability on new data.

[0108] 3.2 Network Structure Design

[0109] 3.2.1 Input Layer

[0110] The input layer receives feature information from the finite element simulation model, namely the knee joint conditions, including knee joint geometric parameters, knee joint material properties, and knee joint load state. These input features are vectorized to ensure that the model can process them effectively.

[0111] 3.2.2 Network Architecture

[0112] By combining convolutional neural networks (CNNs) and fully connected networks, CNNs are used to extract input features, and multi-layer fully connected networks are used to process high-dimensional nonlinear data, thereby increasing the model's expressive power. Residual networks or long short-term memory networks are introduced into the hidden layers, utilizing residual learning or time-series characteristics to improve the model's prediction accuracy for the complex nonlinear mechanical behavior of the knee joint.

[0113] 3.2.3 Output Layer

[0114] The output layer is used to output the predicted stress distribution results for different parts of the knee joint. Each output node corresponds to the stress information at a specific location in the knee joint model, achieving accurate prediction of the global stress distribution.

[0115] 3.3 Model Training and Optimization

[0116] 3.3.1 Loss Function

[0117] Mean squared error (MSE) is chosen as the loss function to measure the deviation between the model's predicted stress distribution and the actual stress value. Optimizing the loss function helps improve the model's accuracy in predicting stress distribution.

[0118] 3.3.2 Optimization Algorithm and Hyperparameter Tuning

[0119] Adaptive optimization algorithms such as Adam or RMSprop are employed to gradually reduce the loss function value through multiple rounds of iterative training, ensuring network convergence. Hyperparameters such as the learning rate, regularization parameter, and number of network layers are tuned using methods such as grid search and Bayesian optimization.

[0120] 3.3.3 Model Evaluation and Cross-Validation

[0121] Cross-validation is used to evaluate the model's generalization ability and avoid overfitting. K-fold cross-validation or leave-one-out cross-validation is used to validate the model, ensuring that the model has stable stress prediction capabilities under different motion scenarios and individual differences.

[0122] For example, S4 is subdivided here as follows:

[0123] 4.1 Model Deployment

[0124] The trained stress prediction model is deployed on mobile devices or computing servers, providing users with a real-time computing interface. The model can quickly output stress distribution prediction results after receiving input data.

[0125] 4.2 Input / Output Interface Design

[0126] The input side supports knee joint imaging data, knee joint geometric parameters, knee joint material properties, and knee joint load status, and provides an interface to input structured data into the model. The output side displays the stress distribution of various parts of the knee joint, and the prediction results can be presented through a graphical interface for easy and intuitive viewing by users.

[0127] 4.3 Application of Prediction Results

[0128] The results can be used in clinical applications such as knee injury assessment, rehabilitation program development, and sports safety analysis. For personalized medicine, the model can provide targeted stress distribution data in real time based on the patient's knee joint structure and movement status, assisting medical personnel in developing treatment or training plans.

[0129] 4.4 System Expansion and Iteration

[0130] The model has good scalability and can be updated by inputting more personalized data to make it suitable for the personalized needs of different groups of people.

[0131] In summary, the knee joint stress distribution calculation method provided by the embodiments of the present invention can utilize the combination of finite element simulation and deep learning technology to achieve rapid calculation of the mechanical response of complex structures, and it has strong adaptability and stability.

[0132] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0133] The above is an introduction to the method embodiments. The following describes the solution of the present invention further through device embodiments.

[0134] Figure 2 A structural diagram of a knee joint stress distribution calculation device provided in an embodiment of the present invention is shown below. Figure 2 As shown, the knee joint stress distribution calculation device 200 may include:

[0135] Module 201 is used to create a finite element model of the knee joint.

[0136] Simulation module 202 is used to set different knee joint conditions for the knee joint finite element model, and to perform simulation calculations using the knee joint finite element model under different knee joint conditions to obtain the simulation results of stress distribution of the knee joint under different knee joint conditions.

[0137] Training module 203 is used to construct a dataset using knee joint conditions as samples and stress distribution simulation results corresponding to knee joint conditions as labels. The dataset is then used to train the machine learning model to obtain a knee joint stress distribution calculation model.

[0138] The calculation module 204 is used to calculate the stress distribution of the knee joint under specified knee joint conditions using a knee joint stress distribution calculation model.

[0139] Understandable, Figure 2 Each module / unit in the knee joint stress distribution calculation device 200 shown has the ability to implement... Figure 1 The functions of each step in the knee joint stress distribution calculation method 100 shown are explained, and their corresponding technical effects are achieved. For the sake of brevity, they will not be elaborated here.

[0140] Figure 3This is a structural diagram of an exemplary electronic device capable of implementing embodiments of the present invention. Electronic device 300 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 300 may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown in this invention, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0141] like Figure 3 As shown, the electronic device 300 may include a computing unit 301, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 302 or a computer program loaded from a storage unit 308 into a random access memory (RAM) 303. The RAM 303 may also store various programs and data required for the operation of the electronic device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0142] Multiple components in electronic device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of displays, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows electronic device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0143] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, method 100 may be implemented as a computer program product, including a computer program tangibly contained in a computer-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of method 100 described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform method 100 by any other suitable means (e.g., by means of firmware).

[0144] The various embodiments described above in this invention can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), payload programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0145] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0146] In the context of this invention, a computer-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of computer-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0147] It should be noted that the present invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute method 100 and achieve the corresponding technical effects achieved by executing the method in the embodiments of the present invention. For the sake of brevity, they will not be described in detail here.

[0148] In addition, the present invention also provides a computer program product, which includes a computer program that implements method 100 when executed by a processor.

[0149] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this invention does not impose any limitations on them.

[0150] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for calculating stress distribution in a knee joint, characterized in that, The method includes: Establish a finite element model of the knee joint; Different knee joint conditions are set for the knee joint finite element model, and simulation calculations are performed using the knee joint finite element model under different knee joint conditions to obtain the stress distribution simulation results of the knee joint under different knee joint conditions. The knee joint conditions include knee joint geometric parameters, knee joint material properties, and knee joint load states. Setting different knee joint conditions for the knee joint finite element model includes setting at least one different knee joint condition among multiple knee joint geometric parameters, knee joint material properties, and knee joint load states. Using the knee joint conditions as samples and the stress distribution simulation results corresponding to the knee joint conditions as labels, a dataset is constructed. The dataset is then used to train a machine learning model to obtain a knee joint stress distribution calculation model. The dataset is preprocessed, including data cleaning, data standardization, data structuring, and data augmentation. The preprocessed dataset is divided into a training set, a validation set, and a test set. The training set is used to train the machine learning model, the validation set is used to optimize the hyperparameters of the trained machine learning model, and the test set is used to evaluate the performance of the trained machine learning model. If the trained machine learning model passes the performance evaluation, it is used as the knee joint stress distribution calculation model. The stress distribution of the knee joint under specified knee joint conditions is calculated using the aforementioned knee joint stress distribution calculation model. The method further includes: Multiple knee joint conditions were selected as target knee joint conditions from different knee joint conditions. Based on the simulation results of stress distribution under various target knee joint conditions and the actual stress distribution results under various target knee joint conditions obtained through physical experiments and / or literature searches, the simulation accuracy of the knee joint finite element model is calculated; wherein, the physical experiment method is the knee joint model loading experiment method. If the simulation accuracy is greater than or equal to the preset threshold, the stress distribution simulation result correction model is used to correct the stress distribution simulation results of the knee joint under different knee joint conditions; If the simulation accuracy is less than a preset threshold, the source of error of the knee joint finite element model is analyzed based on the simulation results of stress distribution under each target knee joint condition and the actual stress distribution results. The knee joint finite element model is then optimized based on this, and the optimized knee joint finite element model is reused to perform simulation calculations under different knee joint conditions to obtain the stress distribution simulation results of the knee joint under different knee joint conditions. The stress distribution simulation result correction model is trained through the following steps: The stress distribution simulation results are input into the generator of the generative adversarial network to generate false stress distribution results. The false stress distribution results are mixed with the true stress distribution results and input into the discriminator of the generative adversarial network to obtain the authenticity probabilities corresponding to the false stress distribution results and the true stress distribution results respectively. Based on these probabilities, the loss value of the discriminator is calculated, and the discriminator is updated according to the loss value of the discriminator. The stress distribution simulation results are input into the generator of the generative adversarial network to generate false stress distribution results. The false stress distribution results are then input into the discriminator of the generative adversarial network to obtain the probability of authenticity of the false stress distribution results. Based on this, the loss value of the generator is calculated, and the generator is updated according to the loss value of the generator. The model parameters of the discriminator and the generator are continuously updated iteratively until the preset stopping condition is met, resulting in a well-trained generative adversarial network. The generator in the network is then used as a correction model for the stress distribution simulation results.

2. The method according to claim 1, characterized in that, The establishment of the knee joint finite element model includes: Based on knee joint imaging data, a three-dimensional reconstruction of the knee joint was performed to obtain a three-dimensional geometric model of the knee joint. The three-dimensional geometric model of the knee joint was then smoothed and fitted with surfaces, and meshed. The corresponding material properties of each structure of the three-dimensional geometric model of the knee joint were then assigned, and boundary conditions and loading conditions were set for the three-dimensional geometric model of the knee joint to obtain a finite element model of the knee joint.

3. The method according to claim 1, characterized in that, The method further includes: Knee injury assessment is performed based on the stress distribution of the knee joint under specified knee joint conditions.

4. A knee joint stress distribution calculation device, characterized in that, The device includes: A module is created to build a finite element model of the knee joint; The simulation module is used to set different knee joint conditions for the knee joint finite element model, and to perform simulation calculations using the knee joint finite element model under different knee joint conditions to obtain the stress distribution simulation results of the knee joint under different knee joint conditions. The knee joint conditions include knee joint geometric parameters, knee joint material properties, and knee joint load states. Setting different knee joint conditions for the knee joint finite element model includes setting at least one different knee joint condition among multiple knee joint geometric parameters, knee joint material properties, and knee joint load states. The training module is used to construct a dataset using the knee joint conditions as samples and the stress distribution simulation results corresponding to the knee joint conditions as labels. The dataset is then used to train a machine learning model to obtain a knee joint stress distribution calculation model. The dataset is preprocessed, including data cleaning, data standardization, data structuring, and data augmentation. The preprocessed dataset is divided into a training set, a validation set, and a test set. The training set is used to train the machine learning model, the validation set is used to optimize the hyperparameters of the trained machine learning model, and the test set is used to evaluate the performance of the trained machine learning model. If the trained machine learning model passes the performance evaluation, it is used as the knee joint stress distribution calculation model. The calculation module is used to calculate the stress distribution of the knee joint under specified knee joint conditions using the knee joint stress distribution calculation model. The device further includes: an optimization module, used for: Multiple knee joint conditions were selected as target knee joint conditions from different knee joint conditions. Based on the simulation results of stress distribution under various target knee joint conditions and the actual stress distribution results under various target knee joint conditions obtained through physical experiments and / or literature searches, the simulation accuracy of the knee joint finite element model is calculated; wherein, the physical experiment method is the knee joint model loading experiment method. If the simulation accuracy is greater than or equal to the preset threshold, the stress distribution simulation result correction model is used to correct the stress distribution simulation results of the knee joint under different knee joint conditions; If the simulation accuracy is less than a preset threshold, the source of error of the knee joint finite element model is analyzed based on the simulation results of stress distribution under each target knee joint condition and the actual stress distribution results. The knee joint finite element model is then optimized based on this, and the optimized knee joint finite element model is reused to perform simulation calculations under different knee joint conditions to obtain the stress distribution simulation results of the knee joint under different knee joint conditions. The stress distribution simulation result correction model is trained through the following steps: The stress distribution simulation results are input into the generator of the generative adversarial network to generate false stress distribution results. The false stress distribution results are mixed with the true stress distribution results and input into the discriminator of the generative adversarial network to obtain the authenticity probabilities corresponding to the false stress distribution results and the true stress distribution results respectively. Based on these probabilities, the loss value of the discriminator is calculated, and the discriminator is updated according to the loss value of the discriminator. The stress distribution simulation results are input into the generator of the generative adversarial network to generate false stress distribution results. The false stress distribution results are then input into the discriminator of the generative adversarial network to obtain the probability of authenticity of the false stress distribution results. Based on this, the loss value of the generator is calculated, and the generator is updated according to the loss value of the generator. The model parameters of the discriminator and the generator are continuously updated iteratively until the preset stopping condition is met, resulting in a well-trained generative adversarial network. The generator in the network is then used as a correction model for the stress distribution simulation results.

5. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1-3.

6. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method described in any one of claims 1-3.

Images