Parametric Design Optimization Method and System of Ankle-Foot Orthosis for 3D Printing

Abstract

The present application discloses a parametric design optimization method and system for ankle-foot orthosis for 3D printing, which relates to the field of ankle-foot correction technology. The method and system pre-process medical image data, segment pathological tissue of the deformed foot, and convert it into an initial three-dimensional model of the pathological tissue of the deformed foot; mesh the initial three-dimensional model into a finite element mesh model, calculate the material properties of the pathological tissue of the deformed foot, obtain the stress tensor and strain tensor of the deformed foot through multi-scale analysis, and evaluate the pressure sore risk index; output the orthosis design parameters based on the material properties, stress tensor, strain tensor and pressure sore risk index using the orthosis parameter prediction model; optimize the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the ventilation holes, and build a pressure sensor and a temperature and humidity sensor in the orthosis to adjust the local environment in real time; generate a three-dimensional solid model based on the design parameters of the orthosis, convert it into a 3D printing file, and use a 3D printer to manufacture a solid orthosis.

Description

Technical Field

[0001] This application relates to the technical field of ankle-foot orthosis, and particularly to a parametric design optimization method and system for ankle-foot orthosis for 3D printing. Background Art

[0002] An ankle-foot orthosis is a commonly used orthopedic device for correcting foot and ankle deformities or dysfunctions. The traditional design and manufacturing process of ankle-foot orthosis usually involves an orthotist manually making a plaster model based on the patient's foot and ankle conditions, and then manufacturing the orthosis according to the plaster model. In recent years, 3D printing technology has been widely used in the medical field, providing new ideas for solving the design and manufacturing problems of traditional ankle-foot orthosis. Combining computer-aided design methods such as parametric design and topology optimization with 3D printing manufacturing technology is expected to achieve personalized and precise design and rapid manufacturing of ankle-foot orthosis.

[0003] Chinese patent application with publication number CN110897772A discloses a manufacturing method of an ankle-foot orthosis based on 3D printing, including the following steps: Step 1, using plantar pressure data; Step 2, scanning the affected limb with a three-dimensional scanner; Step 3, obtaining an STL file from the point cloud data of the affected limb obtained in Step 2; Step 4, using ankle-foot biomechanics and the three-point force system and the abnormal plantar pressure situation obtained in Step 1 to digitally design the three-dimensional model of the ankle-foot orthosis structure; Step 5, topology optimization; Step 6, pre-processing before 3D printing to obtain an ankle-foot orthosis with supports; Step 7, performing post-processing, adding accessories such as pads and Velcro, and finally having the affected limb try it on. It reduces the requirements for the experience and technology of operators, and at the same time can achieve lightweight design, reduce the weight of the ankle-foot orthosis, and improve the acceptance of the ankle-foot orthosis by patients.

[0004] Existing methods mainly rely on abnormal plantar pressure conditions when designing ankle-foot orthosis, and the dimension of personalized design is relatively single, and the consideration of patients' individual needs is not comprehensive enough. Summary of the Invention

[0005] This application aims to solve at least one of the technical problems in the related art to some extent. For this reason, an object of this application is to propose a parametric design optimization method and system for ankle-foot orthosis for 3D printing, which improves the comprehensiveness of considering patients' individual needs.

[0006] One aspect of this application provides a parametric design optimization method for ankle-foot orthosis for 3D printing, including:

[0007] Step S100: collecting medical image data of the patient's ankle and foot, preprocessing the medical image data, performing segmentation of the pathological tissue of the clubfoot based on the preprocessed medical image data, obtaining a segmentation result, and converting the segmentation result into an initial three-dimensional model of the pathological tissue of the clubfoot;

[0008] Step S200: meshing the initial three-dimensional model to generate a finite element mesh model of the pathological tissue of the clubfoot, calculating the material properties of the pathological tissue of the clubfoot in the finite element mesh model according to the gray value, performing multi-scale analysis on the finite element mesh model to obtain the stress tensor and strain tensor of the clubfoot, and evaluating the pressure ulcer risk index of the clubfoot;

[0009] Step S300: outputting design parameters of the orthosis using a pre-trained orthosis parameter prediction model according to the material properties, stress tensor, strain tensor and pressure sore risk index of the clubfoot;

[0010] Step S400: Optimizing the thickness of the anti-pressure sore buffer layer and the graduation density and size distribution of the ventilation holes of the orthosis based on the pressure sore risk index, and building a pressure sensor and a temperature and humidity sensor in the orthosis to adjust the local environment in real time;

[0011] Step S500: Generate a three-dimensional solid model based on the design parameters of the orthosis, convert it into a 3D printing file, and use a 3D printer to manufacture a physical orthosis.

[0012] The specific method of collecting medical image data of the patient's ankle and foot, preprocessing the medical image data, segmenting the pathological tissue of the clubfoot based on the preprocessed medical image data, obtaining the segmentation result, and converting the segmentation result into the initial three-dimensional model of the pathological tissue of the clubfoot is:

[0013] Step S110: using medical imaging equipment to collect medical imaging data of the patient's ankle and foot, and preprocessing the medical imaging data to obtain preprocessed medical imaging data;

[0014] Step S120: using a deep learning-based image segmentation algorithm to perform clubfoot pathological tissue segmentation on the preprocessed medical image data to obtain a binary segmentation mask image;

[0015] The specific method of using the deep learning-based image segmentation algorithm to segment the preprocessed medical image data into pathological tissues of clubfoot to obtain a binary segmentation mask image is as follows:

[0016] Step S121: obtaining medical image samples of patients with clubfoot, annotating the pathological tissue of each medical image sample, and obtaining a corresponding binary segmentation mask image;

[0017] Step S122: The deep learning-based image segmentation algorithm selects the U-Net model based on the attention mechanism, and constructs the input layer, encoder, decoder, output layer and loss function of the U-Net model;

[0018] The input layer is used to receive the preprocessed medical image data;

[0019] The encoder includes four downsampling blocks, and each downsampling block consists of two convolutional layers, an attention module and a max pooling layer;

[0020] The decoder includes four upsampling blocks, and each upsampling block consists of a transposed convolutional layer, an attention model, a multi-scale feature fusion module and two convolutional layers;

[0021] The output layer outputs the binary segmentation mask image of the medical image data;

[0022] The loss function uses weighted cross-entropy loss;

[0023] Step S123: Using the medical image samples as inputs and the corresponding binary segmentation mask images as outputs, construct training samples for training the U-Net model based on the attention mechanism. Taking the value of minimizing the loss function between the real binary segmentation mask image and the predicted binary segmentation mask image as the training objective, when the loss function converges, obtain the trained U-Net model based on the attention mechanism;

[0024] Step S124: Use the U-Net model based on the attention mechanism to segment the preprocessed medical image data to obtain the binary segmentation mask image of the medical image data;

[0025] Step S130: Use the binary segmentation mask image corresponding to each medical image data as the segmentation result of the medical image data. In the binary segmentation mask image, the voxels with a value of 1 represent the deformed pathological tissue area, and the voxels with a value of 0 represent the normal tissue area;

[0026] Step S140: Scan the binary segmentation mask image based on the 3D reconstruction algorithm. Calculate the binary states of the vertices of the cube unit of each voxel in the binary segmentation mask image and convert them into index values to generate a triangular facet mesh model;

[0027] Step S150: Post-process the triangular facet mesh model to obtain the initial 3D model of the deformed foot pathological tissue;

[0028] The specific method for meshing the initial three-dimensional model to generate a finite element mesh model of the pathological tissue of clubfoot, calculating the material properties of the pathological tissue of clubfoot in the finite element mesh model according to the gray value, and performing multi-scale analysis on the finite element mesh model to obtain the stress tensor and strain tensor of clubfoot is as follows:

[0029] Step S210: Mesh the initial three-dimensional model to generate a finite element mesh model of the pathological tissue of clubfoot ;

[0030] Step S220: Register the medical image data of the patient with the finite element mesh model of the pathological tissue of clubfoot. For each element of the finite element mesh model, extract the gray value of the corresponding region in the medical image data and calculate the average gray value of the element;

[0031] Step S230: Calculate the bone density of the element according to the average gray value of each element;

[0032] Step S240: Calculate the material properties of the pathological tissue of clubfoot of the element according to the bone density. The material properties include elastic modulus, Poisson's ratio and shear modulus;

[0033] Step S250: Perform multi-scale mechanical analysis on the finite element mesh model of the pathological tissue of clubfoot to obtain the stress tensor and strain tensor of the pathological tissue of clubfoot at the macro scale, meso scale and micro scale;

[0034] The specific method for performing multi-scale mechanical analysis on the finite element mesh model of the pathological tissue of clubfoot to obtain the stress tensor and strain tensor of the pathological tissue of clubfoot at the macro scale, meso scale and micro scale is as follows:

[0035] Step S251: The multi-scale mechanical analysis includes macro-scale analysis, meso-scale analysis and micro-scale analysis;

[0036] Step S252: In the macro-scale analysis, the finite element mesh model is regarded as a whole, and based on the static equilibrium equation, the stress tensor at the macro scale is solved and the strain tensor ;

[0037] Step S253: Select a representative region of the pathological tissue of clubfoot, establish a meso-mechanical model, use the stress tensor obtained from the macro-scale analysis as the boundary condition of the meso-mechanical model, and solve the stress tensor and the strain tensor ;

[0038] Step S254: For the stress and strain concentration regions in the mesoscopic scale, extract their microstructural features, establish a micromechanical model, use the stress tensor and strain tensor in the mesoscopic scale as the boundary conditions of the micromechanical model, and solve the stress tensor at the microscale and the strain tensor ;

[0039] The specific method for evaluating the pressure ulcer risk index of clubfoot is as follows:

[0040] Step S260: Calculate the fused stress tensor according to the stress tensors and representative volumes of the clubfoot pathological tissue at the macroscopic scale, mesoscopic scale, and microscale ;

[0041] Step S270: Calculate the pressure and shear force of different units of the clubfoot pathological tissue according to the fused stress tensor, and obtain its pressure distribution and shear force distribution;

[0042] Step S280: Preset the critical thresholds of pressure and shear force, and calculate the pressure ulcer risk index according to the pressure and shear force of each unit ;

[0043] The specific method for using the pre-trained orthosis parameter prediction model to output the design parameters of the orthosis according to the material properties, stress tensor, strain tensor, and pressure ulcer risk index of the clubfoot is as follows:

[0044] Step S310: Take the stress tensor, strain tensor, and material properties of the clubfoot as the mechanical characteristics of the clubfoot, take the mechanical characteristics of the clubfoot and the pressure ulcer risk index as input data, and take the design parameters of the orthosis as output data. The design parameters include geometric parameters, material parameters, and mechanical parameters, and obtain training samples;

[0045] Step S320: Use a neural network model as the initial model to train the orthosis parameter prediction model between the mechanical characteristics of the clubfoot, the pressure ulcer risk index, and the design parameters of the orthosis. The orthosis parameter prediction model includes an input layer, a hidden layer, an output layer, and a loss function;

[0046] Among them, the input layer inputs the mechanical characteristics of the clubfoot and the pressure ulcer risk index into the input layer nodes of the neural network model. In the hidden layer, the nonlinear representations of the mechanical characteristics of the clubfoot and the pressure ulcer risk index are extracted through n layers of fully connected neural networks. The output layer maps the nonlinear representations to the design parameters of the orthosis; The mean square loss is used as the loss function to calculate the difference between the predicted design parameters and the true design parameters. Minimizing the loss function is used as the training objective to optimize the model parameters. When the loss function converges to the minimum value, the training is completed, and the trained orthosis parameter prediction model is obtained;

[0047] Step S330: Using the mechanical characteristics of the deformed foot and the pressure ulcer risk index of the current patient as input data, predict the design parameters of the orthosis for the deformed foot of the patient;

[0048] The specific method for optimizing the thickness of the pressure ulcer prevention buffer layer, the distribution density and size distribution of the ventilation holes of the orthosis based on the pressure ulcer risk index, and installing pressure sensors and temperature and humidity sensors in the orthosis to adjust the local environment in real time is as follows:

[0049] Step S410: Optimize the thickness t(x,y) of the pressure ulcer prevention buffer layer according to the pressure ulcer risk index of different units of the patient's ankle and foot;

[0050] Step S420: Design ventilation holes on the outer wall of the orthosis and optimize the distribution density and size distribution ;

[0051] Step S430: Install pressure sensors and temperature and humidity sensors in the orthosis to monitor the pressure distribution on the inner wall of the orthosis in real time. When the pressure is greater than the critical threshold of the pressure , trigger the optimization mechanism of the thickness of the pressure ulcer prevention buffer layer of the orthosis;

[0052] Step S440: The temperature and humidity sensors monitor the temperature and humidity inside the orthosis in real time. When the temperature is greater than the critical threshold of the skin surface temperature or the humidity is greater than the critical threshold of the skin surface humidity, trigger the optimization mechanism of the ventilation holes of the orthosis.

[0053] One aspect of the present application provides a parametric design optimization system for ankle-foot orthoses for 3D printing, including:

[0054] A three-dimensional model generation module, which is used to collect medical image data of the patient's ankle and foot parts, preprocess the medical image data, perform segmentation of the pathological tissues of the deformed foot based on the preprocessed medical image data, obtain the segmentation result, and convert the segmentation result into an initial three-dimensional model of the pathological tissues of the deformed foot;

[0055] A pressure ulcer risk assessment module, which is used to perform mesh division on the initial three-dimensional model to generate a finite element mesh model of the pathological tissues of the deformed foot, calculate the material properties of the pathological tissues of the deformed foot in the finite element mesh model according to the gray value, perform multi-scale analysis on the finite element mesh model, obtain the stress tensor and strain tensor of the deformed foot, and evaluate the pressure ulcer risk index of the deformed foot;

[0056] A design parameter output module, which is used to output the design parameters of the orthosis by using a pre-trained orthosis parameter prediction model according to the material properties, stress tensor, strain tensor and pressure ulcer risk index of the deformed foot;

[0057] A local environment adjustment module is used to optimize the thickness of the pressure ulcer prevention buffer layer of the orthosis, the indexing density and size distribution of the ventilation holes based on the pressure ulcer risk index, and pressure sensors and temperature and humidity sensors are built into the orthosis to adjust the local environment in real time;

[0058] A 3D printing conversion module is used to generate a three-dimensional solid model based on the design parameters of the orthosis, convert it into a 3D printing file, and use a 3D printer to manufacture a solid orthosis.

[0059] One aspect of the present application provides a readable storage medium storing a computer program adapted to be loaded by a processor to execute the steps in the parametric design optimization method of an ankle-foot orthosis for 3D printing.

[0060] The parametric design optimization method and system of the ankle-foot orthosis for 3D printing proposed in the present application have the following advantages compared with the prior art:

[0061] The present application comprehensively uses technologies such as medical imaging, deep learning, and finite element analysis to achieve precise three-dimensional reconstruction and mechanical property analysis of the pathological tissues of clubfoot. By collecting medical imaging data of the ankle-foot part of the patient, a U-Net model based on the attention mechanism is used to segment the pathological tissues of clubfoot to obtain an accurate three-dimensional model of the pathological tissues. Further, finite element mesh division and material property calculation are carried out on the model, and multi-scale mechanical analysis is carried out to comprehensively evaluate the stress-strain distribution and pressure ulcer risk of clubfoot.

[0062] The present application constructs a training sample with the mechanical characteristics and pressure ulcer risk index of clubfoot as the input and the orthosis design parameters as the output, and uses a neural network model to train the orthosis parameter prediction model, realizing the end-to-end mapping from the pathological and mechanical information of individual patients to the orthosis design parameters, greatly improving the pertinence and timeliness of orthosis design, and reducing the dependence on professional knowledge and experience in the design process.

[0063] In view of the high incidence of pressure ulcers in clubfoot patients, the present application integrates functional structures such as a pressure ulcer prevention buffer layer and ventilation holes in the orthosis design, and realizes the optimal matching of structural parameters according to the pressure ulcer risk index in different regions of the patient's ankle-foot.

[0064] Pressure sensors and temperature and humidity sensors are built into the orthosis in the present application, realizing the real-time monitoring of the local stress state and skin condition of clubfoot, and establishing a closed-loop feedback adjustment mechanism between the orthosis and the affected foot. When the pressure exceeds the critical threshold, the local stiffness of the inner wall of the orthosis is optimized to dynamically relieve stress concentration; when the temperature and humidity exceed the critical threshold, the opening degree of the ventilation hole is adjusted to accelerate heat and moisture dissipation. This active adaptive design concept further improves the correction effect and comfort of the orthosis and reduces the incidence of secondary pressure ulcers. Description of the Drawings

[0065] Figure 1 It is a flowchart of the parametric design optimization method for an ankle-foot orthosis for 3D printing provided by the present application;

[0066] Figure 2 It is a flowchart of the construction method of the initial three-dimensional model provided by the present application;

[0067] Figure 3 It is a flowchart of the optimization method for the thickness of the anti-pressure sore buffer layer and the ventilation holes inside the orthosis provided by the present application;

[0068] Figure 4 It is the left view of the three-dimensional solid model of the orthosis;

[0069] Figure 5 It is the right view of the three-dimensional solid model of the orthosis;

[0070] Figure 6 It is a schematic structural diagram of the three-dimensional solid model of the orthosis provided by the present application;

[0071] Figure 7 It is a functional module diagram of the parametric design optimization system for an ankle-foot orthosis for 3D printing provided by the present application.

[0072] Reference Signs: 1. Leg fixing part; 2. Ankle joint fixing part; 3. Fixing ring at the upper end of the calf; 4. Fixing device on the fixing ring at the upper end of the calf; 5. Fixing device on the fixing ring at the lower end of the calf; 6. Adjusting device on the fixing ring at the lower end of the calf; 7. Inner wall of the orthosis. Detailed Embodiments

[0073] To better understand the present application, various aspects of the present application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of the exemplary embodiments of the present application and do not limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0074] In the drawings, for the sake of clarity, the sizes, dimensions and shapes of the elements have been slightly adjusted. The drawings are only examples and are not drawn to scale strictly. As used herein, the terms "substantially", "approximately" and similar terms are used as terms indicating approximation and not as terms indicating degree, and are intended to illustrate the inherent deviations in measured or calculated values that would be recognized by those of ordinary skill in the art. Additionally, in the present application, the order of description of the various steps does not necessarily represent the order in which these processes occur in actual operation, unless otherwise clearly defined or derivable from the context.

[0075] It should also be understood that expressions such as "comprising", "including", "having", "containing" and / or "including" are open-ended rather than closed-ended expressions in this specification, which means that the stated features, elements and / or components exist, but do not exclude the existence of one or more other features, elements, components and / or their combinations. In addition, when an expression such as "at least one of..." appears after a list of listed features, it modifies the entire list of features, rather than just an individual element in the list. In addition, when describing the embodiments of the present application, the use of "may" means "one or more embodiments of the present application". And the term "exemplary" is intended to refer to an example or illustration.

[0076] Unless otherwise defined, all terms used herein (including engineering terms and scientific and technical terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. It should also be understood that unless clearly stated in this application, words defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and should not be interpreted in an idealized or overly formal sense.

[0077] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0078] Embodiment 1

[0079] As Figure 1 shown, the parametric design optimization method for an ankle-foot orthosis for 3D printing provided by the present application includes:

[0080] Step S100: Collect medical image data of the patient's ankle-foot part, preprocess the medical image data, perform segmentation of the pathological tissue of the deformed foot based on the preprocessed medical image data to obtain a segmentation result, and convert the segmentation result into an initial three-dimensional model of the pathological tissue of the deformed foot.

[0081] The specific method for collecting medical image data of the patient's ankle-foot part, preprocessing the medical image data, performing segmentation of the pathological tissue of the deformed foot based on the preprocessed medical image data to obtain a segmentation result, and converting the segmentation result into an initial three-dimensional model of the pathological tissue of the deformed foot is as follows:

[0082] Step S110: Use a medical imaging device to collect medical image data of the patient's ankle-foot part, preprocess the medical image data to obtain preprocessed medical image data;

[0083] Step S120: Use an image segmentation algorithm based on deep learning to segment the pathological tissues of clubfoot in the preprocessed medical image data to obtain a binary segmentation mask image;

[0084] The specific method of using an image segmentation algorithm based on deep learning to segment the pathological tissues of clubfoot in the preprocessed medical image data to obtain a binary segmentation mask image is as follows:

[0085] Step S121: Obtain medical image samples of clubfoot patients, label the pathological tissues of each medical image sample to obtain the corresponding binary segmentation mask image;

[0086] The binary segmentation mask image refers to a C-channel image with the same size as the input medical image data, representing the probability that each voxel belongs to C pathological tissue categories; preferably, the number of channels C of the binary segmentation mask image is equal to 1, and the binary segmentation mask image is a single-channel image with the same size as the input medical image data. The value of each voxel is 0 or 1. When the value of the voxel is 0, it means that the voxel does not belong to the pathological tissue, and when the value of the voxel is 1, it means that it belongs to the pathological tissue.

[0087] Step S122: The image segmentation algorithm based on deep learning selects a U-Net model based on the attention mechanism, and constructs an input layer, an encoder, a decoder, an output layer and a loss function of the U-Net model;

[0088] The input layer is used to receive the preprocessed medical image data;

[0089] The encoder includes four downsampling blocks, and each downsampling block is composed of two convolutional layers, an attention module and a max pooling layer;

[0090] The decoder includes four upsampling blocks, and each upsampling block is composed of a transposed convolutional layer, an attention model, a multi-scale feature fusion module and two convolutional layers;

[0091] The output layer outputs the binary segmentation mask image of the medical image data;

[0092] The loss function uses weighted cross-entropy loss;

[0093] The calculation formula of the weighted cross-entropy loss is: , where is the weighted cross-entropy loss of the l-th layer, is the weight coefficient of the l-th layer, Nd is the number of layers of the decoder, and Nd is equal to 4;

[0094] The calculation formula of the weighted cross-entropy loss of the l-th layer is: , where \(N_k\) is the number of medical image samples, \(N_c\) is the number of voxel categories in the binary segmentation mask image, is the class weight, is the true binary segmentation mask image of the \(k\)-th medical image sample, is the binary segmentation mask image predicted by the \(l\)-th layer of the decoder, and \(\log(.)\) is the logarithmic function;

[0095] The U-Net model based on the attention mechanism uses the preprocessed medical image data as the input, with the size of , where , , are the height, width, and depth of the medical image data respectively; the convolutional layers of the encoder perform convolutional operations on the input medical image data to obtain convolutional feature maps. An attention module is added after each convolutional layer to calculate the attention weights for each channel, and the attention weights are applied to the convolutional feature maps. The max pooling layer is used to perform pooling operations on the convolutional feature maps to obtain feature maps;

[0096] Let the \(l\)-th convolutional operation be , the attention module be , and the max pooling operation be , then the feature map output by the encoder can be expressed as: ; where, ;

[0097] The transposed convolutional layers of the decoder perform upsampling on the feature maps, and the feature maps at different levels of the encoder are fused with the feature maps in the decoder through the multi-scale feature fusion module. After upsampling, convolutional operations are performed through the convolutional layers of the decoder for further feature extraction, and the decoder outputs the final feature map;

[0098] Let the \(m\)-th transposed convolutional operation be , and the multi-scale feature fusion module be , then the feature map output by the decoder can be expressed as: = ; where when \(m = 4\), , \(m\) equals 4, 3, 2, 1;

[0099] The multi-scale feature fusion module fuses the feature map of the \(l\)-th layer of the encoder with the feature map of the \(m\)-th layer of the decoder to obtain the fused feature map , where, represents the concatenation operation, is the fusion weight;

[0100] The fusion weights are calculated by 1×1×1 convolution and the softmax function: ;

[0101] The output layer uses a 1×1×1 convolutional layer to map the feature map output by the decoder into a binary segmentation mask image;

[0102] Step S123: Using the medical image sample as the input and the corresponding binary segmentation mask image as the output, construct a training sample to train the U-Net model based on the attention mechanism. Take minimizing the value of the loss function between the real binary segmentation mask image and the predicted binary segmentation mask image as the training objective. When the loss function converges, obtain the trained U-Net model based on the attention mechanism;

[0103] Step S124: Use the U-Net model based on the attention mechanism to segment the preprocessed medical image data to obtain the binary segmentation mask image of the medical image data.

[0104] Step S130: Take the binary segmentation mask image corresponding to each medical image data as the segmentation result of the medical image data. In the binary segmentation mask image, the voxels with a value of 1 represent the deformed pathological tissue area, and the voxels with a value of 0 represent the normal tissue area;

[0105] Step S140: Scan the binary segmentation mask image based on the three-dimensional reconstruction algorithm. Calculate the binary states of the vertices of the cube unit of each voxel in the binary segmentation mask image and convert them into index values to generate a triangular patch mesh model.

[0106] The specific method of scanning the binary segmentation mask image based on the three-dimensional reconstruction algorithm, calculating the binary states of the vertices of the cube unit of each voxel in the binary segmentation mask image, and converting them into index values to generate a triangular patch mesh model is as follows:

[0107] Step S141: Use the Marching Cubes algorithm as the three-dimensional reconstruction algorithm to scan the binary segmentation mask image, traverse each voxel in the binary segmentation mask image, and for each voxel, extract the cube unit composed of its adjacent eight voxels;

[0108] Step S142: Calculate the binary states of the eight vertices of the cube unit, and convert them into index values according to the isosurface threshold ,

[0109] The specific method of calculating the binary states of the eight vertices of the cube unit and converting them into index values according to the isosurface threshold , , compare the grayscale value of each vertex with the threshold to obtain the binary state of the vertex. If the grayscale value of the vertex is greater than or equal to the isosurface threshold, its binary state is 1; otherwise, its binary state is 0. Arrange the binary states of the eight vertices in the order of the index number to form an eight-bit binary number, and convert this binary number into a decimal number to obtain the index value of the cube unit;

[0110] Step S143: Look up the triangular patch lookup table of the Marching Cubes algorithm according to the index value to obtain the topological structure and vertex positions of the triangular patches inside the cube unit;

[0111] Step S144: For the vertices of each triangular patch, calculate their spatial coordinates by linear interpolation , and save the spatial coordinates and topological structure of the vertices of the generated triangular patches into the triangular patch mesh model;

[0112] The calculation formula for the spatial coordinates is: , , , where , , are the coordinates of the current voxel in the binary segmentation mask image, , , represent the actual lengths of the voxel in the x, y, and z directions, , , represent the relative positions of the vertices of the triangular patch inside the cube unit;

[0113] Step S145: Repeat steps S141 - S144 to traverse all voxels until the entire binary segmentation mask image is processed.

[0114] Step S150: Post - process the triangular patch mesh model to obtain the initial three - dimensional model of the clubfoot pathological tissue;

[0115] Figure 2 is the flow chart of the construction method of the initial three - dimensional model provided by this application;

[0116] The post - processing includes smoothing processing, simplification processing, and patching processing, which are used to eliminate sawteeth and spikes to make the surface smoother, reduce the number of triangular patches to improve the calculation efficiency, and fill possible holes to ensure the integrity of the model respectively.

[0117] Step S200: Perform mesh division on the initial three-dimensional model to generate a finite element mesh model of the clubfoot pathological tissue. Calculate the material properties of the clubfoot pathological tissue in the finite element mesh model according to the gray values. Conduct multi-scale analysis on the finite element mesh model to obtain the stress tensor and strain tensor of the clubfoot, and evaluate the pressure ulcer risk index of the clubfoot.

[0118] The specific method for performing mesh division on the initial three-dimensional model to generate a finite element mesh model of the clubfoot pathological tissue, calculating the material properties of the clubfoot pathological tissue in the finite element mesh model according to the gray values, and conducting multi-scale analysis on the finite element mesh model to obtain the stress tensor and strain tensor of the clubfoot is as follows:

[0119] Step S210: Perform mesh division on the initial three-dimensional model to generate a finite element mesh model of the clubfoot pathological tissue ;

[0120] The conditions that the mesh division needs to meet are: the value range of the element size is , where , are the minimum and maximum values of the element size respectively;

[0121] The minimum value of the element size is generally taken as 1 to 2 times the voxel size, and the maximum value is generally taken as of the minimum characteristic size of the pathological tissue;

[0122] The finite element mesh model of the clubfoot pathological tissue is represented by a node set, an element set, a connection relationship matrix, and a boundary condition set. Among them, nodes are discrete points in the finite element mesh model, representing the positions of the clubfoot pathological tissue in space, and the degree of freedom at each node is defined as a displacement vector; elements represent the basic building blocks in the finite element mesh model, generated by nodes according to the topological structure, and their material properties are defined on each element; the connection relationship represents the topological relationship between nodes and elements; the boundary condition represents the interaction between the clubfoot pathological tissue and the external environment.

[0123] Step S220: Register the medical image data of the patient with the finite element mesh model of the clubfoot pathological tissue. For each element of the finite element mesh model, extract the gray value of the corresponding region in the medical image data and calculate the average gray value of the element.

[0124] The calculation formula for the average gray value of the element is: , where is the number of voxels in the corresponding region of the a-th element, is the gray value of the n_a-th voxel in the corresponding region of the a-th element;

[0125] Step S230: Calculate the bone density of each unit according to the average gray value of each unit;

[0126] The calculation formula for the bone density is: , where and are the linear conversion coefficients of the gray value and the bone density;

[0127] The linear conversion coefficients are set by those skilled in the art according to experience.

[0128] Step S240: Calculate the material properties of the pathological tissues of the clubfoot of each unit according to the bone density. The material properties include the elastic modulus , Poisson's ratio and shear modulus ;

[0129] The calculation formula for the elastic modulus is: , where is the maximum elastic modulus of the bone tissue, k1 is the first material constant, is the minimum bone density of the bone tissue;

[0130] The value range of the maximum elastic modulus of the bone tissue is 15 - 20 GPa, and the specific value is set by those skilled in the art according to experience.

[0131] The value range of the first material constant is 2 - 4 cm 3 / g.

[0132] The value of the minimum bone density of the bone tissue is 0.1 g / cm 3 .

[0133] The calculation formula for Poisson's ratio is: ; where , are the maximum and minimum Poisson's ratios of the bone tissue respectively, is the second material parameter;

[0134] The value ranges of the maximum and minimum Poisson's ratios of the bone tissue are 0.2 - 0.4 and 0.1 - 0.2 respectively;

[0135] The value range of the second material parameter is 1 - 2 cm 3 / g.

[0136] The calculation formula for the shear modulus is: ;

[0137] In the pathological tissues of the clubfoot, the elastic modulus represents the hardness of the bone or soft tissue, Poisson's ratio can reflect the degree of deformation of the pathological tissues, and the shear modulus reflects the strength and deformation characteristics of the cartilage, muscle or lesion area.

[0138] Step S250: Perform multi-scale mechanical analysis on the finite element mesh model of the clubfoot pathological tissue to obtain the stress tensor and strain tensor of the clubfoot pathological tissue at the macroscopic scale, mesoscopic scale, and microscopic scale;

[0139] The specific method for performing multi-scale mechanical analysis on the finite element mesh model of the clubfoot pathological tissue to obtain the stress tensor and strain tensor of the clubfoot pathological tissue at the macroscopic scale, mesoscopic scale, and microscopic scale is as follows:

[0140] Step S251: The multi-scale mechanical analysis includes macroscopic scale analysis, mesoscopic scale analysis, and microscopic scale analysis;

[0141] Step S252: In the macroscopic scale analysis, the finite element mesh model is regarded as a whole, and based on the static equilibrium equation, the stress tensor at the macroscopic scale is solved and the strain tensor ;

[0142] and represent the stress tensor and strain tensor in the i-th coordinate direction and the j-th coordinate direction respectively;

[0143] The equation of the static equilibrium equation is: , where is the component of the external force acting on the clubfoot pathological tissue in the i-th coordinate direction, represents the stress tensor in the j-th coordinate direction;

[0144] The calculation formula for the strain tensor is: , where represents the rate of change of the displacement vector in the j-th coordinate direction, represents the rate of change of the displacement vector in the i-th coordinate direction;

[0145] Step S253: Select a representative region of the clubfoot pathological tissue, establish a meso-mechanical model, use the stress tensor obtained from the macroscopic scale analysis as the boundary condition of the meso-mechanical model, and solve the stress tensor and the strain tensor at the mesoscopic scale;

[0146] The formula for the meso-mechanical model is: , where represents the elastic tensor at the mesoscopic scale, represents the shear strain tensor between the k-th coordinate direction and the l-th coordinate direction at the mesoscopic scale;

[0147] The calculation formula of the elastic tensor is as follows: , where represents the elastic tensor of the matrix material, represents the correction terms of bone density in each coordinate direction, represents the correction terms of the trabecular bone orientation in each coordinate direction, is the angular coordinate of the trabecular bone orientation;

[0148] Preferably, the representative regions are selected as the calcaneus and talus regions;

[0149] Step S254: For the stress and strain concentration regions at the mesoscale, extract their microstructural characteristics, establish a micromechanical model, use the stress tensor and strain tensor at the mesoscale as the boundary conditions of the micromechanical model, and solve the stress tensor and strain tensor ;

[0150] The formula of the micromechanical model is: , where is the elastic tensor at the microscale, is the shear strain tensor between the k-th coordinate direction and the l-th coordinate direction at the microscale;

[0151] The calculation formula of the elastic tensor at the microscale is: , where represents the bone density at the microscale and the microstructure parameters in the correction terms in each coordinate direction;

[0152] The specific method for evaluating the pressure ulcer risk index of clubfoot is as follows:

[0153] Step S260: According to the stress tensors and representative volumes of the clubfoot pathological tissue at the macroscale, mesoscale and microscale, calculate the fused stress tensor ;

[0154] The calculation formula of the fused stress tensor is: , where V is the total volume of the clubfoot pathological tissue, , , are the representative volumes of the clubfoot pathological tissue at the macroscale, mesoscale and microscale respectively, and the relationship between the total volume and the representative volume of the clubfoot pathological tissue satisfies ;

[0155] The representative volumes of the clubfoot pathological tissue at the macroscopic scale, mesoscopic scale, and microscopic scale refer to the smallest volume units for analyzing and calculating the tissue mechanical properties at each scale. By analyzing the initial three-dimensional model of the clubfoot pathological tissue at different scales, the corresponding representative volumes are extracted. At the macroscopic scale, a uniform volume unit is selected from the entire three-dimensional model through volume sampling to simulate the overall mechanical properties. At the mesoscopic scale, a part of the local region characteristics, such as trabecular bone, is extracted. At the microscopic scale, smaller units are scanned at high resolution, and representative regions are selected.

[0156] Step S270: Calculate the pressure and shear force of different units of the clubfoot pathological tissue according to the fused stress tensor, and obtain its pressure distribution and shear force distribution.

[0157] The calculation formula for the pressure of different units of the clubfoot pathological tissue is: , where , , are the fused stress tensors in the first, second, and third coordinate directions respectively.

[0158] The , , are the fused stress tensors in three coordinate directions, that is, the fused stress tensor in the orthogonal coordinate system.

[0159] The calculation formula for the shear force of different regions of the clubfoot pathological tissue is: ;

[0160] Step S280: Preset the critical thresholds of pressure and shear force, and calculate the pressure ulcer risk index according to the pressure and shear force of each unit.

[0161] The calculation formula for the pressure ulcer risk index is: , where , , are the weight coefficients of pressure, shear force, and the interaction term of pressure and shear force respectively, , are the critical thresholds of pressure and shear force respectively, , are the sensitivity coefficients of pressure and shear force respectively.

[0162] The weight coefficients of pressure, shear force, and the interaction term of pressure and shear force, the critical thresholds of pressure and shear force, and the sensitivity coefficients of pressure and shear force are set by those skilled in the art according to experience.

[0163] Step S300: Based on the material properties, stress tensor, strain tensor, and pressure ulcer risk index of the clubfoot, use the pre-trained orthosis parameter prediction model to output the design parameters of the orthosis;

[0164] The specific method for outputting the design parameters of the orthosis based on the material properties, stress tensor, strain tensor, and pressure ulcer risk index of the clubfoot by using the pre-trained orthosis parameter prediction model is as follows:

[0165] Step S310: Take the stress tensor, strain tensor, and material properties of the clubfoot as the mechanical characteristics of the clubfoot, take the mechanical characteristics of the clubfoot and the pressure ulcer risk index as input data, and take the design parameters of the orthosis as output data. The design parameters include geometric parameters, material parameters, and mechanical parameters, and obtain training samples;

[0166] Step S320: Use a neural network model as the initial model to train the orthosis parameter prediction model between the mechanical characteristics of the clubfoot, the pressure ulcer risk index, and the design parameters of the orthosis. The orthosis parameter prediction model includes an input layer, a hidden layer, an output layer, and a loss function;

[0167] Among them, the input layer inputs the mechanical characteristics of the clubfoot and the pressure ulcer risk index into the input layer nodes of the neural network model. In the hidden layer, the non-linear representation of the mechanical characteristics of the clubfoot and the pressure ulcer risk index is extracted through n layers of fully connected neural networks. The output layer maps the non-linear representation to the design parameters of the orthosis; the mean square loss is used as the loss function to calculate the difference between the predicted design parameters and the true design parameters. Minimizing the loss function is used as the training objective to optimize the model parameters. When the loss function converges to the minimum value, the training is completed, and the trained orthosis parameter prediction model is obtained;

[0168] The calculation formula for the non-linear representation is: , where, is the output of the q-th hidden layer node, f(.) is the activation function, is the weight from the r-th node in the input layer to the q-th node in the hidden layer, is the bias term of the q-th node in the hidden layer, is the input data;

[0169] The calculation formula for mapping the non-linear representation to the design parameters of the orthosis is: , where, is the output of the s-th node in the output layer, that is, the predicted value of the orthosis design parameter, g(.) is the activation function of the output layer, is the weight from the q-th node in the hidden layer to the s-th node in the output layer, is the bias term of the s-th node in the output layer;

[0170] The calculation formula of the loss function is: ,in, is the predicted value of the orthosis design parameters in the nyth training sample, is the true value of the orthosis design parameter in the nyth training sample, Ny is the number of training samples, ;

[0171] Step S330: using the current patient's clubfoot mechanical characteristics and pressure sore risk index as input data to predict the design parameters of the patient's clubfoot orthosis;

[0172] Step S400: Optimizing the thickness of the anti-pressure sore buffer layer and the graduation density and size distribution of the ventilation holes of the orthosis based on the pressure sore risk index, and building a pressure sensor and a temperature and humidity sensor in the orthosis to adjust the local environment in real time;

[0173] The specific method of optimizing the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the ventilation holes based on the pressure sore risk index, and building a pressure sensor and a temperature and humidity sensor in the orthosis to adjust the local environment in real time is as follows:

[0174] Step S410: Optimizing the thickness t(x, y) of the anti-pressure sore buffer layer according to the pressure sore risk index of different units of the patient's ankle and foot;

[0175] The calculation formula for the thickness of the anti-pressure sore buffer layer is: ,in, is the base thickness, is the proportionality coefficient, is the pressure sore risk index at the coordinate (x, y) on the pathological tissue of the clubfoot at the microscopic scale;

[0176] Preferably, the base thickness is 3 mm and the proportional coefficient is 0.5.

[0177] Step S420: Designing ventilation holes on the outer wall of the orthosis and optimizing the distribution density of the ventilation holes and size distribution ;

[0178] The calculation formula of the distribution density is: ,in, is the basic density, is the adjustment coefficient of the pressure ulcer risk index, Skin surface temperature The adjustment coefficient, is the critical threshold of skin surface temperature;

[0179] The calculation formula for the size distribution is: ,in, is the basic dimension, is the adjustment coefficient of the pressure ulcer risk index, is the skin surface humidity 's adjustment coefficient, is the critical threshold of the skin surface humidity;

[0180] The basic density, basic dimension, and adjustment coefficient of the pressure ulcer risk index and and the adjustment coefficients of the skin surface temperature and skin surface humidity, and the critical thresholds of the skin surface temperature and skin surface humidity are set by those skilled in the art according to experience.

[0181] Preferably, the basic density is taken as 0.8 per cm², and the basic dimension is taken as 1 mm, and the adjustment coefficient of the pressure ulcer risk index is taken as 0.6, is taken as 0.4, the adjustment coefficient of the skin surface temperature is taken as 0.02, and the adjustment coefficient of the skin surface humidity is taken as 0.03; the critical threshold of the skin surface temperature is taken as 32 °C, and the critical threshold of the skin surface humidity is taken as 60%;

[0182] Step S430: Install pressure sensors and temperature and humidity sensors inside the orthosis to monitor the pressure distribution on the inner wall of the orthosis in real time. When the pressure is greater than the critical threshold of the pressure , trigger the optimization mechanism of the thickness of the pressure ulcer prevention buffer layer of the orthosis;

[0183] The triggering of the optimization mechanism of the thickness of the pressure ulcer prevention buffer layer of the orthosis means jumping to step S410 to recalculate the optimal thickness of the pressure ulcer prevention buffer layer.

[0184] Step S440: The temperature and humidity sensors monitor the temperature and humidity inside the orthosis in real time. When the temperature is greater than the critical threshold of the skin surface temperature or the humidity is greater than the critical threshold of the skin surface humidity, trigger the optimization mechanism of the ventilation holes of the orthosis;

[0185] Figure 3 is the flow chart of the optimization method for the thickness of the pressure ulcer prevention buffer layer and the ventilation holes inside the orthosis provided by this application;

[0186] The triggering of the optimization mechanism of the ventilation holes of the orthosis means jumping to step S420 to recalculate the optimal distribution density and size distribution of the ventilation holes.

[0187] Step S500: Generate a three-dimensional solid model based on the design parameters of the orthosis, convert it into a 3D printing file, and use a 3D printer to manufacture a solid orthosis;

[0188] The printing material of the physical orthosis supports real-time adjustment functions. For example, electroactive polymers or shape memory polymers can be selected. When these materials receive specific external stimuli such as temperature, humidity, and pressure, they can achieve dynamic adjustment of shape or mechanical properties.

[0189] According to the orthosis parameter prediction model pre-trained in this application, the mechanical characteristics of the deformed foot and the pressure ulcer risk index of the current patient are used as input data to predict the design parameters of the orthosis for the deformed foot of this patient. The design parameters include geometric parameters, material parameters, and mechanical parameters. The above parameters determine the overall shape, size, and material characteristics of the orthosis. Based on the design parameters, a three-dimensional solid model of the orthosis is constructed in three-dimensional modeling software.

[0190] Figure 4 It is the left view of the three-dimensional solid model of the orthosis. Figure 5 It is the right view of the three-dimensional solid model of the orthosis. The three-dimensional solid model of the orthosis is composed of a leg fixing part 1 and an ankle fixing part 2. Among them, the upper calf fixing ring 3 can fix the orthosis on the patient's calf according to the fixing device 4 on the upper calf fixing ring and the fixing device 5 on the lower calf fixing ring.

[0191] In order to embed pressure sensors and temperature and humidity sensors in the orthosis, it is necessary to reserve installation spaces for the sensors at predetermined positions in the three-dimensional solid model of the orthosis, design a circuit system for sensor signal acquisition, processing, and wireless transmission in the three-dimensional solid model, and integrate the digital design of the sensors and the feedback system with the main model of the orthosis to form a complete intelligent orthosis digital model.

[0192] In Figure 5 a temperature and humidity sensor is provided at the adjusting device 6 on the lower calf fixing ring. When the temperature is greater than the critical threshold of the skin surface temperature or the humidity is greater than the critical threshold of the skin surface humidity, the adjustment mechanism of the adjusting device 6 on the lower calf fixing ring is triggered to achieve the ventilation effect of the orthosis.

[0193] Figure 6 It is the structural schematic diagram of the three-dimensional solid model of the orthosis provided by this application. In Figure 6 a pressure ulcer prevention buffer layer and a pressure sensor are integrated in the inner wall 7 of the orthosis, and the optimization mechanism of the thickness of the pressure ulcer prevention buffer layer of the orthosis is triggered according to the pressure distribution monitored by the pressure sensor.

[0194] Convert the complete digital model of the smart orthosis into an STL file and import it into the slicing software for slicing. Select appropriate slicing parameters in the slicing software to generate digital instructions for 3D printing. Select appropriate 3D printing materials, input the generated digital instructions into the 3D printer, and start printing the orthosis layer by layer. Integrate devices such as pressure sensors, temperature and humidity sensors into the orthosis, and encapsulate and fix them.

[0195] Example 2

[0196] like Figure 7 As shown, the parametric design optimization system for ankle-foot orthosis for 3D printing provided by this application includes:

[0197] A three-dimensional model generation module is used to collect medical image data of the patient's ankle and foot, pre-process the medical image data, segment the pathological tissue of the deformed foot based on the pre-processed medical image data, obtain the segmentation result, and convert the segmentation result into an initial three-dimensional model of the pathological tissue of the deformed foot;

[0198] The pressure sore risk assessment module is used to mesh the initial three-dimensional model, generate a finite element mesh model of the pathological tissue of the clubfoot, calculate the material properties of the pathological tissue of the clubfoot in the finite element mesh model according to the gray value, perform multi-scale analysis on the finite element mesh model, obtain the stress tensor and strain tensor of the clubfoot, and assess the pressure sore risk index of the clubfoot;

[0199] A design parameter output module, which is used to output the design parameters of the orthosis according to the material properties of the clubfoot, the stress tensor, the strain tensor and the pressure sore risk index using a pre-trained orthosis parameter prediction model;

[0200] The local environment adjustment module is used to optimize the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the ventilation holes based on the pressure sore risk index. The pressure sensor and temperature and humidity sensor are built into the orthosis to adjust the local environment in real time.

[0201] The 3D printing conversion module is used to generate a three-dimensional solid model based on the design parameters of the orthosis, and convert it into a 3D printing file, and use a 3D printer to manufacture a solid orthosis.

[0202] Example 3

[0203] According to one embodiment of the present application, a readable storage medium is also provided. Computer-readable instructions are stored on the readable storage medium. When the computer-readable instructions are executed by the processor, the parametric design optimization method for ankle-foot orthosis for 3D printing according to the embodiment of the present application can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache). Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0204] In addition, according to the implementation mode of the present application, the process described above can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: collecting medical image data of the patient's ankle and foot, preprocessing the medical image data, segmenting the pathological tissue of the deformed foot based on the preprocessed medical image data, obtaining the segmentation result, converting the segmentation result into an initial three-dimensional model of the pathological tissue of the deformed foot; meshing the initial three-dimensional model to generate a finite element mesh model of the pathological tissue of the deformed foot, and calculating the gray value of the deformed foot pathology in the finite element mesh model. The material properties of the physical tissue are analyzed at multiple scales on the finite element mesh model to obtain the stress tensor and strain tensor of the deformed foot, and the pressure sore risk index of the deformed foot is evaluated; the design parameters of the orthosis are outputted using the pre-trained orthosis parameter prediction model according to the material properties, stress tensor, strain tensor and pressure sore risk index of the deformed foot; the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the vents are optimized based on the pressure sore risk index, and a pressure sensor and a temperature and humidity sensor are built into the orthosis for real-time adjustment of the local environment; a three-dimensional solid model is generated based on the design parameters of the orthosis, and converted into a 3D printing file, and a physical orthosis is manufactured using a 3D printer. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0205] The methods, systems, and storage media of the present application can be implemented in a variety of ways. For example, the methods, systems, and storage media of the present application can be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order for the method steps is for illustration only, and the method steps of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application can also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the methods according to the present application. Therefore, the present application also covers recording media that store programs for executing methods according to the present application.

[0206] In addition, parts of the above technical solutions provided in the embodiments of the present application that have the same implementation principle as the corresponding technical solutions in the prior art are not described in detail to avoid excessive repetition.

[0207] As described above in the specific embodiments, the purpose, technical solutions, and beneficial effects of the present invention have been further described in detail. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A parametric design optimization method for ankle-foot orthoses for 3D printing, characterized in that, include: Collect medical image data of the patient's ankle and foot, pre-process the medical image data, segment the pathological tissue of the clubfoot based on the pre-processed medical image data, obtain the segmentation result, and convert the segmentation result into an initial three-dimensional model of the pathological tissue of the clubfoot; The initial three-dimensional model was meshed to generate a finite element mesh model of the pathological tissue of the clubfoot. The material properties of the pathological tissue of the clubfoot in the finite element mesh model were calculated according to the grayscale value. The finite element mesh model was subjected to multi-scale analysis to obtain the stress tensor and strain tensor of the clubfoot, and the pressure ulcer risk index of the clubfoot was evaluated. According to the material properties, stress tensor, strain tensor and pressure sore risk index of the clubfoot, the pre-trained orthosis parameter prediction model is used to output the design parameters of the orthosis; Based on the pressure sore risk index, the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the ventilation holes are optimized. Pressure sensors and temperature and humidity sensors are built into the orthosis to adjust the local environment in real time. Generate a three-dimensional solid model based on the design parameters of the orthosis, convert it into a 3D printing file, and use a 3D printer to manufacture the physical orthosis; The specific method of meshing the initial three-dimensional model, generating a finite element mesh model of the clubfoot pathological tissue, calculating the material properties of the clubfoot pathological tissue in the finite element mesh model according to the gray value, performing multi-scale analysis on the finite element mesh model, and obtaining the stress tensor and strain tensor of the clubfoot is: For the initial 3D model Perform meshing to generate a finite element mesh model of the pathological tissue of clubfoot ; The medical imaging data of the patient is registered with the finite element mesh model of the pathological tissue of the clubfoot. For each unit of the finite element mesh model, the gray value of the corresponding area in the medical imaging data is extracted, and the average gray value of the unit is calculated; The bone density of each unit is calculated based on the average gray value of the unit; Calculating material properties of the clubfoot pathological tissue of the unit according to the bone density, wherein the material properties include elastic modulus, Poisson's ratio and shear modulus; Multi-scale mechanical analysis was performed on the finite element mesh model of the clubfoot pathological tissue to obtain the stress tensor and strain tensor of the clubfoot pathological tissue at the macroscale, mesoscale and microscale. The specific method for evaluating the pressure ulcer risk index of clubfoot is: Calculate the fused stress tensor based on the stress tensors and representative volumes of clubfoot pathological tissues at the macroscopic scale, mesoscopic scale, and microscopic scale. ; Calculate the pressures of different units of the pathological tissue of clubfoot based on the fused stress tensor and shear forces to obtain their pressure distributions and shear force distributions; Critical thresholds of preset pressure and shear force, and calculate the pressure ulcer risk index according to the pressure and shear force of each unit .

2. The parameterized design optimization method of an ankle-foot orthosis for 3D printing according to claim 1, characterized in that The specific method of collecting medical image data of the patient's ankle and foot, preprocessing the medical image data, segmenting the pathological tissue of the clubfoot based on the preprocessed medical image data, obtaining the segmentation result, and converting the segmentation result into the initial three-dimensional model of the pathological tissue of the clubfoot is: Using medical imaging equipment to collect medical imaging data of the patient's ankle and foot, preprocessing the medical imaging data to obtain preprocessed medical imaging data; The image segmentation algorithm based on deep learning is used to segment the pathological tissue of the clubfoot in the preprocessed medical image data to obtain a binary segmentation mask image. The binary segmentation mask image corresponding to each medical image data is used as the segmentation result of the medical image data, wherein the voxel with a value of 1 in the binary segmentation mask image represents the deformed pathological tissue area, and the voxel with a value of 0 represents the normal tissue area; Scan the binary segmentation mask image based on the 3D reconstruction algorithm. Calculate the binary states of the vertices of the cube unit of each voxel in the binary segmentation mask image, and convert them into index values to generate a triangular facet mesh model; Perform post-processing on the triangular facet mesh model to obtain the initial 3D model of the clubfoot pathological tissue.

3. The parametric design optimization method of the ankle-foot orthosis for 3D printing according to claim 2, wherein, The specific method for segmenting the clubfoot pathological tissue from the preprocessed medical image data by using the image segmentation algorithm based on deep learning to obtain the binary segmentation mask image is as follows: Obtain the medical image samples of clubfoot patients, label the pathological tissues of each medical image sample to obtain the corresponding binary segmentation mask image; The image segmentation algorithm based on deep learning selects the U-Net model based on the attention mechanism, and constructs the input layer, encoder, decoder, output layer and loss function of the U-Net model; The input layer is used to receive the preprocessed medical image data; The encoder includes four downsampling blocks, and each downsampling block consists of two convolutional layers, an attention module and a max pooling layer; The decoder includes four upsampling blocks, and each upsampling block consists of a transposed convolutional layer, an attention model, a multi-scale feature fusion module and two convolutional layers; The output layer outputs the binary segmentation mask image of the medical image data; The loss function uses weighted cross-entropy loss; Use the medical image sample as the input and the corresponding binary segmentation mask image as the output to construct a training sample for training the U-Net model based on the attention mechanism. Use minimizing the value of the loss function between the real binary segmentation mask image and the predicted binary segmentation mask image as the training goal. When the loss function converges, obtain the trained U-Net model based on the attention mechanism; Use the U-Net model based on the attention mechanism to segment the preprocessed medical image data to obtain the binary segmentation mask image of the medical image data.

4. The parameterized design optimization method of an ankle-foot orthosis for 3D printing according to claim 3, characterized in that, The specific method for outputting the design parameters of the orthosis by using the pre-trained orthosis parameter prediction model according to the material properties, stress tensor, strain tensor and pressure ulcer risk index of the clubfoot is as follows: Take the stress tensor, strain tensor and material properties of the clubfoot as the mechanical characteristics of the clubfoot. Take the mechanical characteristics of the clubfoot and the pressure ulcer risk index as the input data, and the design parameters of the orthosis as the output data. The design parameters include geometric parameters, material parameters and mechanical parameters, and obtain the training sample; Use the neural network model as the initial model to train the orthosis parameter prediction model. The orthosis parameter prediction model includes an input layer, a hidden layer, an output layer and a loss function; Among them, the input layer inputs the mechanical characteristics of the clubfoot and the pressure ulcer risk index into the input layer nodes of the neural network model. In the hidden layer, the non-linear representation of the mechanical characteristics of the clubfoot and the pressure ulcer risk index is extracted through n layers of fully connected neural networks. The output layer maps the non-linear representation to the design parameters of the orthosis; The mean square loss is used as the loss function to calculate the difference between the predicted design parameters and the actual design parameters. The minimization of the loss function is used as the training goal to optimize the model parameters. When the loss function converges to the minimum value, the training is completed and the trained orthosis parameter prediction model is obtained. The mechanical characteristics of the current patient's clubfoot and the pressure sore risk index are used as input data to predict the design parameters of the orthosis for the patient's clubfoot.

5. The parameterized design optimization method of an ankle-foot orthosis for 3D printing according to claim 4, wherein The specific method of optimizing the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the ventilation holes based on the pressure sore risk index, and building a pressure sensor and a temperature and humidity sensor in the orthosis to adjust the local environment in real time is as follows: According to the pressure sore risk index of different units of the patient's ankle and foot, optimize the thickness of the anti-pressure sore buffer layer t(x,y); Design ventilation holes on the outer wall of the orthosis and optimize the distribution density and size distribution ; A pressure sensor and a temperature and humidity sensor are built into the orthosis. The pressure sensor monitors the pressure distribution on the inner wall of the orthosis in real time. When the pressure is greater than the critical threshold of the pressure an optimization mechanism for the thickness of the anti-pressure ulcer buffer layer of the orthosis is triggered; The temperature and humidity sensor monitors the temperature and humidity inside the orthosis in real time. When the temperature is greater than the critical threshold of the skin surface temperature or the humidity is greater than the critical threshold of the skin surface humidity, the optimization mechanism of the orthosis vent is triggered.

6. A parametric design optimization system for an ankle-foot orthosis for 3D printing, which is used to implement the parametric design optimization method for an ankle-foot orthosis for 3D printing described in any one of claims 1-5, characterized in that, include: A three-dimensional model generation module is used to collect medical image data of the patient's ankle and foot, pre-process the medical image data, segment the pathological tissue of the deformed foot based on the pre-processed medical image data, obtain the segmentation result, and convert the segmentation result into an initial three-dimensional model of the pathological tissue of the deformed foot; The pressure sore risk assessment module is used to mesh the initial three-dimensional model, generate a finite element mesh model of the pathological tissue of the clubfoot, calculate the material properties of the pathological tissue of the clubfoot in the finite element mesh model according to the gray value, perform multi-scale analysis on the finite element mesh model, obtain the stress tensor and strain tensor of the clubfoot, and assess the pressure sore risk index of the clubfoot; A design parameter output module, which is used to output the design parameters of the orthosis according to the material properties of the clubfoot, the stress tensor, the strain tensor and the pressure sore risk index using a pre-trained orthosis parameter prediction model; The local environment adjustment module is used to optimize the thickness of the anti-pressure sore buffer layer of the orthosis and the graduation density and size distribution of the ventilation holes based on the pressure sore risk index. The pressure sensor and temperature and humidity sensor are built into the orthosis to adjust the local environment in real time. The 3D printing conversion module is used to generate a three-dimensional solid model based on the design parameters of the orthosis, and convert it into a 3D printing file, and use a 3D printer to manufacture a solid orthosis.

7. A readable storage medium, characterized in that, The readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the steps in the parametric design optimization method for ankle-foot orthosis for 3D printing as described in any one of claims 1 to 5.

Images