CN114662362A - Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system - Google Patents

Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system Download PDF

Info

Publication number
CN114662362A
CN114662362A CN202210296439.8A CN202210296439A CN114662362A CN 114662362 A CN114662362 A CN 114662362A CN 202210296439 A CN202210296439 A CN 202210296439A CN 114662362 A CN114662362 A CN 114662362A
Authority
CN
China
Prior art keywords
vertebral body
segment
model
dimensional
lumbar vertebra
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210296439.8A
Other languages
Chinese (zh)
Inventor
华庆
赵蒙蒙
司海朋
赵俊勇
王晶晶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong Normal University
Original Assignee
Shandong Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong Normal University filed Critical Shandong Normal University
Priority to CN202210296439.8A priority Critical patent/CN114662362A/en
Publication of CN114662362A publication Critical patent/CN114662362A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B17/00—Surgical instruments, devices or methods, e.g. tourniquets
    • A61B17/56—Surgical instruments or methods for treatment of bones or joints; Devices specially adapted therefor
    • A61B17/58—Surgical instruments or methods for treatment of bones or joints; Devices specially adapted therefor for osteosynthesis, e.g. bone plates, screws, setting implements or the like
    • A61B17/68—Internal fixation devices, including fasteners and spinal fixators, even if a part thereof projects from the skin
    • A61B17/70—Spinal positioners or stabilisers ; Bone stabilisers comprising fluid filler in an implant
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20112—Image segmentation details
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30008—Bone
    • G06T2207/30012—Spine; Backbone

Abstract

The invention provides a deep learning-based lumbar vertebra segment internal fixation mode simulation method and system, which comprises the steps of obtaining a CT lumbar vertebra segment scanning image; segmenting the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image; reconstructing a three-dimensional vertebral body model based on the segmented vertebral body structure and preprocessing the three-dimensional vertebral body model; adding and assembling lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional lumbar vertebra segment model; based on the three-dimensional lumbar vertebra segment model, carrying out meshing on the components of the model; carrying out finite element analysis based on the three-dimensional lumbar vertebra segment model after the grid division to obtain a simulation result of a fixing mode in the lumbar vertebra segment; according to the invention, the image is segmented by using the deep learning U-Net network, so that the current situation that image annotation is troublesome and laborious is greatly improved, and a more accurate three-dimensional vertebral body model can be obtained.

Description

Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system
Technical Field
The invention belongs to the technical field of image processing and biomechanics simulation, and particularly relates to a deep learning-based lumbar vertebra segment internal fixation mode simulation method and system.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Spinal disorders include degenerative diseases of the lumbar spine, deformities, fractures, and tumors, and the onset of patients is becoming younger and younger, as is common in the lumbar segments of L4-L5. The clinical manifestations of the patient are mostly aching pain of the lumbosacral part, pain of lower limbs caused by involvement, incapability of standing upright and the like, serious patients may cause incontinence of urine and feces, even paralysis, and the life quality of the patient is reduced to a great extent. At present, aiming at the disease, the clinical treatment mostly adopts the mode of minimally invasive surgery to apply internal fixation for treatment, which generally adopts the internal fixation technology of vertebral pedicle and CBT, and the operation mode is an advanced operation mode for internationally treating spine diseases such as lumbar spondylolisthesis, lumbar vertebra instability and thoracolumbar fracture. The internal fixation technology is that unstable lumbar vertebra segments are connected and fixed by using instruments such as metal screws, rods and the like, and the internal fixation is utilized to keep local reduction of movement and stabilize the lumbar vertebra. Pedicle screws and CBT screws have been used in clinical minimally invasive treatment of the spine in a combined bi-segmental internal fixation.
The inventor finds that the clinical research index result shows that the pedicle screw and CBT screw combined double-section internal fixation mode has better stability, greatly shortens the minimally invasive incision, but cannot search the mechanical property of the minimally invasive incision.
Disclosure of Invention
In order to solve the problems, the invention provides a deep learning-based lumbar vertebra segment internal fixation mode simulation method and system.
According to some embodiments, the invention provides a deep learning-based lumbar vertebra segment internal fixation simulation method, which adopts the following technical scheme:
a deep learning-based lumbar vertebra segment internal fixation mode simulation method comprises the following steps:
acquiring a CT lumbar vertebra segment scanning image;
segmenting the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image;
reconstructing a three-dimensional vertebral body model based on the segmented vertebral body structure and preprocessing the three-dimensional vertebral body model;
adding and assembling lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional vertebral body segment model;
based on the three-dimensional vertebral body segment model, carrying out meshing on the components of the three-dimensional vertebral body segment model;
and carrying out finite element analysis based on the three-dimensional vertebral body segment model after meshing to obtain a simulation result of the internal fixation mode of the lumbar vertebral body segment.
Further, the segmenting the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure includes:
carrying out vertebral body segmentation based on a deep learning 3D U-Net convolution neural network;
inputting the acquired CT lumbar vertebra segment scanning image into a deep learning 3D U-Net convolution neural network;
performing feature extraction on the CT lumbar vertebra segment scanning image through down-sampling to obtain a feature map of the CT lumbar vertebra segment scanning image;
and (4) performing up-sampling on the characteristic image of the CT lumbar vertebra segment scanning image through the deconvolution layer to obtain a segmented vertebral body structure image.
Further, the deep learning 3D U-Net convolutional neural network comprises an encoder and a decoder;
the decoder comprises four network layers with different resolutions, wherein each resolution layer comprises two convolution layers, a ReLU activation layer and a maximum pooling layer;
the decoder comprises four network layers with different resolutions, wherein each resolution layer comprises two deconvolution layers, two convolution layers and a ReLU activation layer.
Further, reconstructing a three-dimensional vertebral body model based on the segmented vertebral body structure and preprocessing the three-dimensional vertebral body model, including:
importing the divided vertebral body structure image into a Mimics, and directly reconstructing a three-dimensional vertebral body model;
based on the complete three-dimensional vertebral body model, using a smoothening tool to carry out local and complete optimization on the surface of the vertebral body;
and obtaining the preprocessed three-dimensional vertebral body model.
Further, reconstructing a three-dimensional vertebral body model based on the segmented vertebral body structure and preprocessing the three-dimensional vertebral body model, further comprising:
based on the preprocessed three-dimensional vertebral body model, performing smooth optimization again;
and fitting the smooth and optimized three-dimensional cone model to obtain a Nurbs solid curved surface through the operations of accurate curved surface, detection curvature, curved surface sheet construction, grating construction and curved surface fitting, thus obtaining the three-dimensional solid curved surface cone model.
Further, the adding and assembling of the lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain the fixed three-dimensional vertebral body segment model includes:
determining an upper endplate, a lower endplate, nucleus pulposus and an annulus fibrosus based on the lower surface of the L4 vertebral body and the upper surface of the L5 vertebral body of the intervertebral disc, and reconstructing and adding a complete intervertebral disc structural model;
simulating an operation position, placing a CBT screw into an L4 vertebral body, and placing a pedicle screw into an L5 vertebral body;
manufacturing a nail rod according to the positions of the upper and lower screws of the L4-L5 centrum, connecting the pedicle screw with the CBT screw, and screwing the nut to ensure that the screw threads of the nut and the screw are mutually matched, thereby completing the assembly of the pedicle screw and the CBT screw internal fixing system;
and obtaining the fixed three-dimensional vertebral body segment model.
Further, the finite element analysis is performed on the three-dimensional vertebral segment model after meshing, so as to obtain a simulation result of the internal fixation mode of the lumbar vertebral segment, and the method comprises the following steps:
setting the analysis step as a static general type, and defining a unit as geometric nonlinearity;
creating field output and course output in the analysis step, and setting indexes of result analysis;
adding interaction and setting surface-surface binding constraint on each part of the model;
adding loads and boundary conditions, and simulating six motion conditions of buckling, stretching, left-right lateral bending and left-right rotation;
solving is carried out based on the six motion condition models, the error indication is adjusted and modified, and a correct solution is solved;
and drawing a cloud picture of the three-dimensional vertebral body segment model based on the indexes of correct solution and result analysis, namely a simulation result of the fixing mode in the lumbar vertebral body segment.
According to some embodiments, the second aspect of the present invention provides a deep learning-based lumbar vertebra segment internal fixation simulation system, which adopts the following technical solutions:
a deep learning-based lumbar vertebra segment internal fixation simulation system comprises:
an image acquisition module configured to acquire a CT lumbar spine segment scan image;
the image segmentation module is configured to segment the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image;
the image reconstruction module is configured to reconstruct a three-dimensional vertebral body model based on the segmented vertebral body structure and perform preprocessing;
the component assembling module is configured to add and assemble lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional vertebral body segment model;
a meshing module configured to mesh its components based on the three-dimensional vertebral segment model;
and the finite element analysis module is configured to perform finite element analysis based on the three-dimensional vertebral body segment model after meshing to obtain a simulation result of the internal fixation mode of the lumbar vertebral segment.
According to some embodiments, a third aspect of the invention provides a computer-readable storage medium.
A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, carries out the steps of a deep learning-based lumbar spine segment internal fixation simulation method according to the first aspect.
According to some embodiments, a fourth aspect of the invention provides a computer apparatus.
A computer apparatus comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor when executing the program implementing the steps in a deep learning based lumbar spine segmental internal fixation simulation method as described in the first aspect above.
Compared with the prior art, the invention has the beneficial effects that:
the invention uses the deep learning U-Net network to segment the image, greatly improves the current situation that the image labeling is troublesome and laborious, and can obtain a more accurate three-dimensional vertebral body model, thereby realizing the accurate simulation of the lumbar vertebra segment, and providing a reference scheme with guiding significance for ensuring the safe operation and the subsequent rehabilitation and the clinical operation treatment.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
Fig. 1 is a schematic flow chart of a deep learning-based lumbar vertebra segment internal fixation simulation method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an image segmentation network architecture based on 3D U-Net according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a model in a finite element analysis process according to an embodiment of the present invention;
FIG. 4 is a cloud of results of the internal fixation model described in the example of the invention under six activity conditions.
Detailed Description
The invention is further described with reference to the following figures and examples.
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiments and features of the embodiments of the present invention may be combined with each other without conflict.
Example one
As shown in fig. 1 to 4, the present embodiment provides a deep learning-based lumbar vertebra segment internal fixation simulation method, including:
acquiring a CT lumbar vertebra segment scanning image;
segmenting the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image;
reconstructing a three-dimensional vertebral body model based on the segmented vertebral body structure and preprocessing the three-dimensional vertebral body model;
adding and assembling lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional vertebral body segment model;
based on the three-dimensional vertebral body segment model, carrying out meshing on the components of the three-dimensional vertebral body segment model;
and carrying out finite element analysis based on the three-dimensional vertebral body segment model after meshing to obtain a simulation result of the internal fixation mode of the lumbar vertebral body segment.
Specifically, the method of this embodiment includes the following steps:
s1: medical image acquisition. First, medical images of a patient are acquired, and the patient is imaged with an electronic Computed Tomography (CT) scan of the lumbar spine at the level L4-L5. The selected patients were normal patients without spinal disease.
The invention selects normal patients without spine diseases, and carries out CT scanning on the lumbar vertebra L4-L5 two segments to obtain medical digital images, and the medical digital images are stored as Dicom format files.
S2: and (4) segmenting the medical image. For the CT medical image of lumbar vertebra L4-L5 stage obtained by scanning, the image is divided to obtain the vertebral body structure. The experiment adopts a deep learning 3D U-Net convolution neural network to carry out vertebral body segmentation.
Predictive pictures are generated by upsampling in combination with downsampling using a standard 3D U-Net network. In this network configuration, feature extraction is performed by down-sampling using an original image as an input. Then, the decoder gradually performs up-sampling on the feature map through the deconvolution layer to obtain a final segmentation image.
The network structure used by the invention comprises two parts of feature coding and feature decoding, wherein the two parts respectively comprise four network layers with different resolutions. Each resolution layer of the encoder part comprises two convolution layers with convolution kernel size of 3 multiplied by 3 and a ReLU, and then a maximum pooling layer with step size of 2 and convolution kernel size of 2 multiplied by 2 is accompanied, so that on one hand, the feature map is reduced, and the complexity of network calculation is simplified; on the other hand, feature compression and extraction are performed, and feature dimensions are reduced. Wherein the formula for maximum pooling is as follows:
Figure BDA0003563527080000081
wherein, YkHWDRepresents the output result of the k-th feature map through the pooling operation, xkmniRepresenting pooled regions
Figure BDA0003563527080000082
The element at the (m, n, i) position. Through this process, the numerical distribution of the C-profile of the layer is obtained.
In the decoding section, each resolution layer contains two deconvolution layers with convolution kernel sizes of 2 × 2 × 2, followed by two convolution layers with convolution kernels of 3 × 3 × 3 and one ReLU. And performing feature fusion on the same resolution layer in the coding path and the feature map of the decoding part through jumping connection to the decoding path so as to provide the original high-resolution features. And finally, using a convolution layer with the convolution kernel size of 1 multiplied by 1 on an output layer, and obtaining a final prediction result through a Sigmoid activation function, wherein the number of the final output results is the number of label categories. The overall network structure is shown in fig. 2.
The network structure used by the invention comprises two parts of feature coding and feature decoding, wherein the two parts respectively comprise four network layers with different resolutions. Each convolutional layer in the encoder section contains two 3 x 3 convolutional kernels and a ReLU active layer, followed by a max pooling layer with convolutional kernel size of 2 x 2, step size of 2. In the decoding section, each resolution layer contains two deconvolution layers of step size 2 and convolution kernel size 2 × 2 × 2, followed by two convolution layers of convolution kernel 3 × 3 × 3 and one ReLU active layer. And performing feature fusion on the same resolution layer in the coding path and the feature map of the decoding part through jumping connection to the decoding path so as to provide the original high-resolution features. And finally, putting the obtained feature graph into a convolution layer with convolution kernel of 1 multiplied by 1 to reduce the number of output channels, wherein the number of the output channels is the number of label categories finally.
S3: and (5) reconstructing a medical image. And importing the divided vertebral body image into the Mimics, directly reconstructing a three-dimensional vertebral body model, and then storing the complete three-dimensional vertebral body model as an stl format file.
S4: the surface of the three-dimensional vertebral body model reconstructed by the Mimics is uneven and poor in flatness, and subsequent analysis cannot be used, so that the stl file is preprocessed by using 3-Matic.
And (3) carrying out local and total optimization on the surface of the vertebral body by using a smoothening tool on the introduced three-dimensional vertebral body model with the surface flaw, finishing the pretreatment, storing and exporting the stl format data image file.
S5: the stl format data image file model after pretreatment is a unit set consisting of triangular patches, only displays a closed surface without entity characteristics, and can be introduced into reverse engineering software Geomagic to convert the surface model into an entity model.
The imported triangular patch model needs to be subjected to polygon processing, the model is subjected to smooth optimization processing again by using a gridding doctor, a simplified and removed nail and a noise-reducing tool, the processed polygonal model is subjected to operations of accurate surface, detection curvature, a constructed surface patch, a constructed grating and a fitted surface, a Nurbs solid surface is fitted, and the solid surface vertebral body model is stored as an igs format data image file.
Through the processing flow, the modeling of the vertebral body model for subsequent analysis is basically completed.
S6: after the modeling of the vertebral body is finished, screws, intervertebral discs and other parts need to be added and assembled in Solidworks.
Firstly, the addition of the intervertebral disc needs to be determined according to the anatomical structure and the relationship, the lower surface of the L4 vertebral body and the upper surface of the L5 vertebral body, the upper and lower end plates, the nucleus pulposus and the annulus fibrosus are further determined, and a complete intervertebral disc structure model is reconstructed and added. Secondly, the pedicle screws, the CBT screws and the nuts are all manufactured in Catia software. According to the operation designated position, using tools such as copying, rotating and moving and the like to place the CBT screw into an L4 vertebral body, placing the pedicle screw into an L5 vertebral body, and after the placement is finished, using a Boolean subtraction operation tool to subtract the overlapped part of the vertebral body model; thirdly, a nail rod is manufactured according to the positions of the upper and lower screws of the L4-L5 vertebral body, the pedicle screw is connected with the CBT screw, the nut is screwed, the screw threads of the nut and the screw are matched with each other, the assembly of the pedicle screw and the CBT screw internal fixing system is completed, and the double-segment internal fixing operation of the pedicle screw and the CBT screw for the lumbar vertebra L4-L5 is simulated. And finally, saving the model parts in the internal fixed system one by one into a step format data image file.
S7: and (4) carrying out meshing on each assembled step format model component by using Hypermesh.
Step format files which are assembled are imported into Hypermesh, firstly, geometric cleaning is needed, and curved surface pieces are optimized to promote smooth division of grids; secondly, automatically dividing 2D unit surface meshes for each model part by using an automesh tool, selecting a triangular type for unit division, setting the size of a centrum mesh to be 2mm, and reducing the size of the meshes of other parts to be 1mm, so as to ensure that the subsequent analysis result is more accurate; thirdly, checking the quality of the grid unit under the operation of a qualityindex tool, and adjusting the unsatisfied grid unit according to the rule so as to achieve the effect of optimizing the grid unit; and finally, automatically generating a 3D volume mesh according to the 2D surface mesh, setting the cell type as a tetrahedral cell (C3D4), deleting the 2D surface mesh and reserving the 3D volume mesh, and saving the 3D volume mesh model as an inp format data image file.
S8: to ensure that the finite element analysis is highly reproducible, material properties are added to the L4-L5 cones. Because the junction of cortical bone and cancellous bone in the vertebral body is not obvious, the bone is not uniform, and anisotropy exists, the vertebral body model is closer to the real vertebral body characteristic by using the vertebral body empirical formula for the assignment of the vertebral body.
Using a Mimics import inp file, calculating a Hutch unit gray value for each unit of a volume grid according to CT scanning image data, then defining corresponding materials of a vertebral body according to different gray ranges, assigning the vertebral body model to ten gradient materials by a lumbar vertebra empirical assignment formula, wherein the elastic modulus value is determined by apparent density, and the Poisson ratio is set to be 0.3.
The empirical formula for the lumbar vertebrae is as follows:
ρ＝47+1.122×Hu
E＝1.92×ρ-170
v＝0.3
hu is the gray value, ρ is the apparent density, E is the elastic modulus, and v is the Poisson's ratio.
S9: importing the inp file into Abaqus for finite element analysis requires the following 8-module setup:
the grid module is used for establishing seven ligaments including an anterior longitudinal ligament, a posterior longitudinal ligament, a ligamentum flavum, a supraspinatus ligament, an interspinous ligament, a transverse interspinous ligament and a joint capsule ligament according to the ligament organization structures and the relationships under the operation of an editing grid unit tool, wherein the ligament types are set as trusses.
And secondly, an attribute module, namely adding material attributes to other parts except the vertebral body in the model, setting the type of the elastic material to be isotropic, and adding data of the elastic modulus and the Poisson ratio.
Assembling the parts into a complete non-independent model by the assembling module, combining and cutting the model, and ensuring that the model has no phase separation and intersection parts.
And the analysis step module is added with a finite element analysis step, the analysis step is set to be a static general type, the unit is defined to be geometric nonlinearity, field output and process output are established in the analysis step, and indexes of result analysis are set.
And the interaction module is used for adding interaction and setting surface-surface binding constraint to each part, binding the two parts of the model together, and preventing relative motion between the two parts.
Sixthly, adding load and boundary conditions, namely completely fixing the lower edge of the L5 to stabilize the model; next, in order to simulate the self weight of the human body, 500N of downward concentration force is applied to the upper surface of the L4 vertebral body, and six motion conditions of flexion, extension, left-right lateral bending and left-right rotation are simulated by applying 10Nm of torque.
And the operation module submits the six motion state models respectively and solves the models, adjusts and modifies error prompts and calculates a correct solution.
And the visualization module checks the required index result and draws a cloud picture. The complete model can be viewed and single or few component clouds can be compared to the values.
S10: as shown in fig. 4, after the finite element analysis is completed to solve the final assembly, the results of the model such as stress, strain, displacement, etc. can be obtained in the visualization module. For activity (ROM) measurements, UG NX software implementation is required. Through observing and analyzing the cloud picture and the numerical value, a reliable result of biomechanical quantitative data based on deep learning and a lumbar vertebra segment two-screw combined internal fixation mode can be obtained, and meanwhile, the biomechanical advantages of the lumbar vertebra segment two-screw combined internal fixation mode can be explored through comparing with other internal fixation modes and combining with clinical data.
Example two
The embodiment provides a lumbar vertebra segment internal fixation mode simulation system based on deep learning, including:
an image acquisition module configured to acquire a CT lumbar spine segment scan image;
the image segmentation module is configured to segment the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image;
the image reconstruction module is configured to reconstruct a three-dimensional vertebral body model based on the segmented vertebral body structure and perform preprocessing;
the component assembling module is configured to add and assemble lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional vertebral body segment model;
a meshing module configured to mesh components thereof based on the three-dimensional vertebral body segment model;
and the finite element analysis module is configured to perform finite element analysis based on the three-dimensional vertebral body segment model after meshing to obtain a simulation result of the internal fixation mode of the lumbar vertebral body segment.
The modules are the same as the corresponding steps in the implementation example and application scenarios, but are not limited to the disclosure of the first embodiment. It should be noted that the modules described above as part of a system may be implemented in a computer system such as a set of computer executable instructions.
In the foregoing embodiments, the descriptions of the embodiments have different emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
The proposed system can be implemented in other ways. For example, the above-described system embodiments are merely illustrative, and for example, the division of the above-described modules is merely a logical division, and in actual implementation, there may be other divisions, for example, multiple modules may be combined or integrated into another system, or some features may be omitted, or not executed.
EXAMPLE III
The present embodiment provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps in a deep learning-based lumbar spine segment internal fixation simulation method as described in the first embodiment above.
Example four
The embodiment provides a computer device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the steps of the deep learning-based lumbar vertebra segment internal fixation simulation method according to the embodiment.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by a computer program, which may be stored in a computer readable storage medium and executed by a computer to implement the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.

Claims (10)

1. A deep learning-based lumbar vertebra segment internal fixation mode simulation method is characterized by comprising the following steps:
acquiring a CT lumbar vertebra segment scanning image;
segmenting the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image;
reconstructing a three-dimensional vertebral body model based on the segmented vertebral body structure and preprocessing the three-dimensional vertebral body model;
adding and assembling lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional vertebral body segment model;
based on the three-dimensional vertebral body segment model, carrying out meshing on the components of the three-dimensional vertebral body segment model;
and carrying out finite element analysis based on the three-dimensional vertebral body segment model after meshing to obtain a simulation result of the internal fixation mode of the lumbar vertebral body segment.
2. The deep learning-based simulation method for the internal fixation of the lumbar vertebra segment as claimed in claim 1, wherein the segmenting the acquired CT lumbar vertebra segment scan image to obtain the vertebral body structure comprises:
carrying out vertebral body segmentation based on a deep learning 3D U-Net convolution neural network;
inputting the acquired CT lumbar vertebra segment scanning image into a deep learning 3D U-Net convolution neural network;
performing feature extraction on the CT lumbar vertebra segment scanning image through downsampling to obtain a feature map of the CT lumbar vertebra segment scanning image;
and (4) performing up-sampling on the characteristic image of the CT lumbar vertebra segment scanning image through the deconvolution layer to obtain a segmented vertebral body structure image.
3. The deep learning-based lumbar spine segment internal fixation simulation method of claim 2, wherein the deep learning 3D U-Net convolutional neural network comprises an encoder and a decoder;
the decoder comprises four network layers with different resolutions, wherein each resolution layer comprises two convolution layers, a ReLU activation layer and a maximum pooling layer;
the decoder comprises four network layers with different resolutions, wherein each resolution layer comprises two deconvolution layers, two convolution layers and a ReLU activation layer.
4. The deep learning-based lumbar spine segmental internal fixation simulation method as claimed in claim 1, wherein the reconstructing and preprocessing of the three-dimensional vertebral body model based on the segmented vertebral body structure comprises:
importing the divided vertebral body structure image into a Mimics, and directly reconstructing a three-dimensional vertebral body model;
based on the complete three-dimensional vertebral body model, using a smoothening tool to carry out local and complete optimization on the surface of the vertebral body;
and obtaining the preprocessed three-dimensional vertebral body model.
5. The deep learning-based lumbar spine segmental internal fixation simulation method as claimed in claim 4, wherein the reconstructing and preprocessing of the three-dimensional vertebral body model based on the segmented vertebral body structure further comprises:
based on the preprocessed three-dimensional vertebral body model, performing smooth optimization again;
and fitting the three-dimensional vertebral body model subjected to smooth optimization to obtain a Nurbs solid curved surface through the operations of accurate curved surface, detection curvature, curved surface sheet construction, grating construction and curved surface fitting, so as to obtain the three-dimensional solid curved surface vertebral body model.
6. The deep learning-based simulation method for the internal fixation of the lumbar vertebra segment according to claim 1, wherein the adding and assembling of the lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain the fixed three-dimensional vertebral body segment model comprises:
determining an upper endplate, a lower endplate, nucleus pulposus and an annulus fibrosus based on the lower surface of the L4 vertebral body and the upper surface of the L5 vertebral body of the intervertebral disc, and reconstructing and adding a complete intervertebral disc structural model;
simulating an operation position, placing a CBT screw into an L4 vertebral body, and placing a pedicle screw into an L5 vertebral body;
manufacturing a nail rod according to the positions of the upper and lower screws of the L4-L5 centrum, connecting the pedicle screw with the CBT screw, and screwing the nut to ensure that the screw threads of the nut and the screw are mutually matched, thereby completing the assembly of the pedicle screw and the CBT screw internal fixing system;
and obtaining the fixed three-dimensional vertebral body segment model.
7. The deep learning-based simulation method for internal fixation of lumbar vertebra segments according to claim 1, wherein the finite element analysis is performed on the three-dimensional vertebral segment model after meshing to obtain a simulation result of internal fixation of lumbar vertebra segments, comprising:
setting the analysis step as a static general type, and defining a unit as geometric nonlinearity;
creating field output and course output in the analysis step, and setting indexes of result analysis;
adding interaction and setting surface-surface binding constraint on each part of the model;
adding loads and boundary conditions, and simulating six motion conditions of buckling, stretching, left-right lateral bending and left-right rotation;
solving is carried out based on the six motion condition models, the error indication is adjusted and modified, and a correct solution is solved;
and drawing a cloud picture of the three-dimensional vertebral body segment model based on the indexes of correct solution and result analysis, namely a simulation result of the fixing mode in the lumbar vertebral body segment.
8. A deep learning-based lumbar vertebra segment internal fixation simulation system is characterized by comprising:
an image acquisition module configured to acquire a CT lumbar spine segment scan image;
the image segmentation module is configured to segment the acquired CT lumbar vertebra segment scanning image to obtain a vertebral body structure image;
the image reconstruction module is configured to reconstruct a three-dimensional vertebral body model based on the segmented vertebral body structure and perform pretreatment;
the component assembling module is configured to add and assemble lumbar vertebra segment components based on the preprocessed three-dimensional vertebral body model to obtain a fixed three-dimensional vertebral body segment model;
a meshing module configured to mesh components thereof based on the three-dimensional vertebral body segment model;
and the finite element analysis module is configured to perform finite element analysis based on the three-dimensional vertebral body segment model after meshing to obtain a simulation result of the internal fixation mode of the lumbar vertebral segment.
9. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, carries out the steps of a deep learning based lumbar spine segment internal fixation simulation method according to any one of claims 1-7.
10. A computer apparatus comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of a deep learning based lumbar spine intra-segmental fixation simulation method as claimed in any one of claims 1-7.
CN202210296439.8A 2022-03-24 2022-03-24 Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system Pending CN114662362A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210296439.8A CN114662362A (en) 2022-03-24 2022-03-24 Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210296439.8A CN114662362A (en) 2022-03-24 2022-03-24 Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system

Publications (1)

Publication Number Publication Date
CN114662362A true CN114662362A (en) 2022-06-24

Family

ID=82031520

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210296439.8A Pending CN114662362A (en) 2022-03-24 2022-03-24 Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system

Country Status (1)

Country Link
CN (1) CN114662362A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115329648A (en) * 2022-10-17 2022-11-11 博志生物科技(深圳)有限公司 Mechanical testing method, device, equipment and storage medium for spinal internal fixation screw
CN115482466A (en) * 2022-09-28 2022-12-16 广西壮族自治区自然资源遥感院 Three-dimensional model vegetation area lightweight processing method based on deep learning

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115482466A (en) * 2022-09-28 2022-12-16 广西壮族自治区自然资源遥感院 Three-dimensional model vegetation area lightweight processing method based on deep learning
CN115329648A (en) * 2022-10-17 2022-11-11 博志生物科技(深圳)有限公司 Mechanical testing method, device, equipment and storage medium for spinal internal fixation screw

Similar Documents

Publication Publication Date Title
Pahr et al. A comparison of enhanced continuum FE with micro FE models of human vertebral bodies
Sigal et al. Mesh-morphing algorithms for specimen-specific finite element modeling
Lavecchia et al. Lumbar model generator: a tool for the automated generation of a parametric scalable model of the lumbar spine
Thong et al. Three-dimensional morphology study of surgical adolescent idiopathic scoliosis patient from encoded geometric models
CN114662362A (en) Deep learning-based lumbar vertebra segment internal fixation mode simulation method and system
CN114998301B (en) Vertebral body sub-region segmentation method and device and storage medium
Podshivalov et al. Multiscale FE method for analysis of bone micro-structures
Campbell et al. An automated method for landmark identification and finite-element modeling of the lumbar spine
Kallemeyn et al. An interactive multiblock approach to meshing the spine
Vergari et al. Evaluation of a patient-specific finite-element model to simulate conservative treatment in adolescent idiopathic scoliosis
CN113379892A (en) Vertebral body mechanical strength evaluation method and device, computer equipment and storage medium
Dahmen et al. An automated workflow for the biomechanical simulation of a tibia with implant using computed tomography and the finite element method
Karimi et al. Evaluation of the efficiency of various force configurations on scoliotic, lordotic and kyphotic curves in the subjects with scoliosis
Wu et al. A graphical guide for constructing a finite element model of the cervical spine with digital orthopedic software
CN114429533A (en) Simulation method and system for lumbar isthmus fissure screw vertebral lamina hook internal fixation system
CN113223724A (en) Modeling method of hexahedral finite element model of lumbar posterior longitudinal ligament
Kallemeyn et al. Advancements in spine FE mesh development: toward patient-specific models
Grif et al. Planning technology for neurosurgical procedures by using a software platform to create an optima configuration of customized titanium implants
Zhang et al. Development and model form assessment of an automatic subject-specific vertebra reconstruction method
Groesel et al. Alternative solution of virtual biomodeling based on CT-scans
Azmi et al. CT-FEA of Inhomogeneous Lumbar with Different Loadings of the Spinal Cage
Hui et al. Biomechanical analysis of degenerative intervertebral disc enlargement of human spine
Schwen Composite finite elements for trabecular bone microstructures
CN113408172B (en) Manufacturing method and manufacturing system of interbody fusion cage
Toth-Tascau et al. Modeling and numerical analysis of a cervical spine unit

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination