Method for rapidly repairing scoliosis spine model based on semantic feature optimization

Abstract

The invention discloses a lateral bending spine model rapid restoration method based on semantic feature optimization, and relates to the technical field of medical image analysis and biomedical engineering, and the method comprises the steps: carrying out the segmentation recognition operation of a pre-collected medical image through the morphological feature and geometric analysis technology, obtaining the three-dimensional semantic information of the spine structure based on the segmentation recognition result; extracting morphological parameters on the basis of the three-dimensional semantic information of the spine structure, and exploring an association relationship among the morphological parameters by using a correlation analysis technology; and generating a spinal semantic parameter combination based on the association relationship and the morphological parameters, and searching an optimal spinal semantic parameter combination by using an optimization algorithm to repair the scoliosis spinal model. According to the method, based on a global optimization strategy of an optimization algorithm, a reasonable evaluation function is designed to quantify the performance of each parameter combination on a scoliosis model repair task, and the optimal semantic parameter combination weight is automatically found, so that effective adjustment of scoliosis spine model parameters is guided.

Description

Technical Field

[0001] The present invention relates to the field of medical image analysis and biomedical engineering technology, and in particular to a method for quickly repairing a scoliotic spine model based on semantic feature optimization. Background Art

[0002] With the advancement of digital technology, it has promoted the three-dimensional model repair of scoliosis. Through the bone cross-sectional contour lines generated by medical images, a detailed three-dimensional model can be constructed, and geometric transformations such as scaling and shearing can be performed to simulate different correction strategies.

[0003] However, traditional shape adjustment-based methods have limitations when dealing with complex lesions, especially for scoliosis problems that require consideration of anatomical semantic information, which is difficult to express accurately through simple geometric deformation. Existing methods can effectively avoid the problem of poor local detail effects when using only mesh deformation, but the semantic parameter set referenced by the deformation in the existing technology is not comprehensive enough, and there is a lack of effective optimization algorithms to determine the optimal parameter combination during the adjustment process. During repair, different parameter combinations can only be repeatedly tested, which undoubtedly increases the complexity and uncertainty of the repair plan, and has also become a bottleneck currently faced by scoliosis repair.

[0004] Constructing a comprehensive set of spinal semantic parameters with clear anatomical meanings and exploring methods to intelligently optimize these parameters are the core issues of current three-dimensional scoliosis model repair. Furthermore, how to find the optimal or near-optimal solution in a high-dimensional parameter space with semantic relevance and constraints remains to be studied.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the problems in the related art, the present invention proposes a rapid repair method for a scoliotic spine model based on semantic feature optimization to overcome the above-mentioned technical problems existing in the existing related art.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] A rapid repair method for a scoliotic spine model based on semantic feature optimization includes:

[0009] Use morphological features and geometric analysis technology to perform segmentation and recognition operations on pre-collected medical images, and obtain three-dimensional semantic information of the spinal structure based on the segmentation and recognition results;

[0010] The morphological parameters were extracted based on the three-dimensional semantic information of the spinal structure, and the correlation analysis technology was used to explore the correlation between the morphological parameters.

[0011] A spinal semantic parameter combination is generated based on the association relationship and morphological parameters, and the optimal spinal semantic parameter combination is found using an optimization algorithm. The scoliotic spinal model is repaired based on the optimal spinal semantic parameter combination.

[0012] Preferably, using morphological features and geometric analysis technology to perform segmentation and recognition operations on pre-collected medical images, and obtaining three-dimensional semantic information of the spinal structure based on the segmentation and recognition results includes:

[0013] Collect tomography images and magnetic resonance images as medical images, apply affine transformation technology to perform registration processing on the medical images, and verify the registration results based on analyzing the matching degree between the medical images;

[0014] The registration process is completed by adjusting the change parameters according to the verification results, and the registered medical images are denoised and contrast enhanced by filtering and equalization to obtain optimized medical images;

[0015] The spinal region and ligament information in medical images are obtained based on morphological operations and probability distribution techniques, and the growth feature lines in medical images are identified using the watershed algorithm to obtain the spinal structure.

[0016] The spinal structure, spinal region and ligament information are used as the three-dimensional semantic information of the spinal structure.

[0017] Preferably, the spinal region and ligament information in the medical image are respectively obtained based on morphological operations and probability distribution technology, and the spinal structure is obtained by identifying the growth feature lines in the medical image using a watershed algorithm, including:

[0018] Morphological operations are used to identify contour feature lines in medical images, and slicing and projection operations are performed on the contour feature lines to obtain a point set of the medical image, and regional information of the spine is obtained based on the point set;

[0019] Generate a graph model based on adaptive sampling strategy and point set, and analyze the graph model with graph theory algorithm to complete hole identification and repair operations, and extract ligament information based on the repair results;

[0020] The watershed algorithm is used to identify the point set to obtain the growth characteristic line, and a multidimensional tree is constructed after optimizing the growth characteristic line. The detailed information of the spine structure is analyzed based on the multidimensional tree.

[0021] Preferably, the contour feature lines in the medical image are identified by using morphological operations, and the contour feature lines are sliced ​​and projected to obtain a point set of the medical image. The regional information of the spine is obtained according to the point set, including:

[0022] Calculate the gradient magnitude and direction of each pixel in the medical image, and define the threshold index based on historical experience to identify the contour of the medical image, and erode the pixels on the contour boundary based on morphology;

[0023] The gaps between objects on the contour boundary are filled by using the expansion technology, and then the adjacent objects are connected to obtain the contour feature line. The position and direction of the slice plane are determined based on the position and direction of the contour feature line.

[0024] The point set intersecting with the slice plane is calculated based on the contour feature line, and the point set is projected onto the slice plane. The regional information of the spine is obtained based on the projection result.

[0025] Preferably, a graph model is generated based on an adaptive sampling strategy and a point set, and a graph model is analyzed using a graph theory algorithm to complete hole identification and repair operations. The information of the ligament is extracted based on the repair results, including:

[0026] The point set is taken as the original spine data set, and a group of data points are randomly selected from the original spine data set as the first sampling points and added to the sampling point set;

[0027] Iteratively selecting sampling points from the original spine data set based on a preset number of sampling points, and the sum of the distances between the sampling points and all data points in the original spine data set is the largest;

[0028] When the number of iterations reaches a preset number of sampling points, the sampling points are added to the sampling point set to form a spine sampling data set, the spine sampling data set is converted into a graph model, and nodes are created for the data points;

[0029] Determine the distance between any two nodes, connect the two nodes to create an edge when the distance value is less than the set threshold, and apply graph theory algorithms to detect the edge creation status and analyze the hole information of the graph model;

[0030] The boundaries of the holes are determined and connected by the minimum spanning tree to fill the holes, and the ligament information is extracted based on the filled graph model.

[0031] Preferably, a watershed algorithm is applied to identify the point set to obtain a growth characteristic line, and a multidimensional tree is constructed after optimizing the growth characteristic line. The detailed information of the spine structure analyzed based on the multidimensional tree includes:

[0032] After preprocessing the data points in the point set, the corresponding gradient values ​​are calculated, and the points whose gradient values ​​exceed the threshold are selected as the starting points of the growth feature lines;

[0033] Use the gradient value to construct a gradient image, select the data point with the largest gradient value as the seed point, and set the seed point as the marker point of the watershed transformation;

[0034] Based on the marking points, the gradient image is segmented into several groups of regions using watershed changes, and the growth feature lines are identified after judging the boundary information between regions and the gradient changes within the regions;

[0035] After performing noise removal and discontinuous point removal operations on the growth feature line, the point set corresponding to the growth feature line is divided into several subsets, and a multidimensional tree is constructed on the subsets;

[0036] Based on the multidimensional tree and K nearest neighbor search technology, similar points on the growth feature line are found to form a growth region, and the detailed information of the spine structure is extracted after merging and segmenting the growth region.

[0037] Preferably, extracting morphological parameters based on the three-dimensional semantic information of the spinal structure and using correlation analysis technology to explore the correlation between the morphological parameters includes:

[0038] The morphological parameters are extracted from the three-dimensional semantic information of the spinal structure, and the morphological parameters are normalized. The correlation coefficients between the two morphological parameters are analyzed using correlation technology.

[0039] The linear relationship between the morphological parameters is explained according to the size of the correlation coefficient, and the relevant parameters are divided into different groups according to the preset classification criteria;

[0040] The variance analysis was performed on each group to determine the ratio of the variance between groups to the variance within groups and the probability value of the ratio value, and the differences in morphological parameters between different groups were explained based on the ratio value and the probability value;

[0041] A clustering operation is performed on the morphological parameters, and the correlation and difference between the morphological parameters are explained according to the clustering results to obtain the parameter correlation relationship, and the spinal morphological characteristics represented by each morphological parameter are shown based on the parameter correlation relationship.

[0042] Preferably, generating a spinal semantic parameter combination based on the association relationship and the morphological parameters, finding the best spinal semantic parameter combination using an optimization algorithm, and repairing the scoliotic spinal model according to the best spinal semantic parameter combination includes:

[0043] Determine the combination of spinal semantic parameters based on spinal morphological features, association relationships and morphological parameters, and generate a parameter space based on the combination of spinal semantic parameters;

[0044] Encoding the parameter combination into individuals in the parameter space, defining the optimization goal according to the application of the spinal semantic parameters in the spinal model repair, and defining the evaluation function according to the optimization goal to calculate the evaluation value of each individual;

[0045] Based on the evaluation value and optimization algorithm, genetic operations are performed to simulate natural selection and genetic processes, and the evaluation of individuals is re-evaluated according to the genetic process. The individual with the highest evaluation is selected as the optimal solution to obtain the best combination of spine semantic parameters.

[0046] A prototype scoliotic spine model is generated based on the spine data, and the optimal spine semantic parameter combination is applied to the prototype scoliotic spine model to repair the spine state.

[0047] Preferably, the calculation formula of the evaluation function is:

[0048] F(x)=α.P accuracy +β.P correlation -γ.P complexity +δ.P robustness ;

[0049] In the formula, F(x) represents the evaluation value of the xth individual, P accuracy represents the segmentation accuracy, P correlation represents parameter correlation, P complexity represents the computational complexity, P robustness represents robustness, α represents the weight coefficient of segmentation accuracy, β represents the weight coefficient of parameter correlation, γ represents the weight coefficient of computational complexity, and δ represents the weight coefficient of robustness.

[0050] Preferably, generating a prototype scoliosis spine model based on the spine data, and applying the optimal spine semantic parameter combination to the prototype scoliosis spine model to repair the spine state includes:

[0051] After preprocessing the medical imaging data of the healthy spine, the spine contour data is extracted, and a prototype scoliotic spine model is constructed in combination with auxiliary design technology;

[0052] Based on the preset repair requirements, the boundary conditions, initial displacement and velocity are defined, and the deformation of the prototype scoliosis spine model is guided by the physical model deformation algorithm, and the new position of each data point is obtained to obtain the scoliosis spine deformation model;

[0053] The optimal combination of spine semantic parameters is applied to the scoliotic spine deformation model to simulate the morphological changes of the spine, and the scoliotic spine deformation model is repaired based on the morphological simulation results.

[0054] The beneficial effects of the present invention are:

[0055] 1. The present invention is based on the global optimization strategy of the optimization algorithm, designs a reasonable evaluation function to quantify the performance of each parameter combination in the scoliosis model repair task, and automatically finds the optimal semantic parameter combination weight, thereby guiding the effective adjustment of the scoliosis spine model parameters, which can not only improve the quality of model repair, but also improve the parameter optimization efficiency and avoid the tedious work caused by manual blind modification of parameters.

[0056] 2. The present invention utilizes defined and accessible intact semantic parameters with clear anatomical significance to describe in detail the degree, morphological characteristics and potential correction needs of scoliosis, and at the same time uses an optimization algorithm to achieve the purpose of finding the most appropriate combination of model parameters to accurately simulate the true state of the scoliotic spine and the best possible correction results.

[0057] 3. The present invention realizes efficient and accurate segmentation of spinal images, which not only improves the quality of spinal repair, but also ensures the integrity and continuity of the spinal structure. At the same time, it guides the initial deformation of the spinal model based on physical models and finite element analysis, providing a basis for subsequent optimization based on genetic algorithms, and ensuring the morphological rationality and accuracy of the spinal model.

[0058] 4. The lateral spine model rapid repair method implemented by the present invention significantly improves the accuracy of spine repair and the authenticity of the three-dimensional model, effectively solves the problems of structural continuity and detail loss in traditional methods, greatly promotes the accuracy and efficiency of medical image analysis, surgical planning and treatment evaluation, and is of great significance to promoting the development of personalized medicine and precision surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 is a flow chart of a method for quickly repairing a scoliotic spine model based on semantic feature optimization according to an embodiment of the present invention;

[0061] Figure 2 is a schematic diagram of a scoliotic spine model rapid repair method based on semantic feature optimization according to an embodiment of the present invention;

[0062] Figure 3 are lateral and frontal images of a scoliosis patient in a scoliosis spine model rapid repair method based on semantic feature optimization according to an embodiment of the present invention;

[0063] Figure 4 yes Figure 3 Schematic representation of the spine centerline in a patient with moderate scoliosis. DETAILED DESCRIPTION

[0064] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. By referring to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0065] According to an embodiment of the present invention, a method for quickly repairing a scoliotic spine model based on semantic feature optimization is provided.

[0066] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figures 1 to 4 As shown, according to the embodiment of the present invention, the scoliosis spine model rapid repair method based on semantic feature optimization comprises:

[0067] Step S1, using feature and geometric analysis technology to perform segmentation and recognition operations on pre-collected medical images, and obtaining three-dimensional semantic information of the spinal structure based on the segmentation and recognition results;

[0068] Step S2, extracting morphological parameters based on the three-dimensional semantic information of the spinal structure, and using correlation analysis technology to explore the correlation between the morphological parameters;

[0069] Step S3, generating a spinal semantic parameter combination based on the association relationship and the morphological parameters, finding the best spinal semantic parameter combination using an optimization algorithm, and repairing the scoliotic spinal model according to the best spinal semantic parameter combination.

[0070] In this embodiment, when morphological features and geometric analysis techniques are used to perform segmentation and recognition operations on pre-collected medical images, and three-dimensional semantic information of the spinal structure is obtained based on the segmentation and recognition results, tomographic images and magnetic resonance images can be collected as medical images, and affine transformation technology is applied to perform registration processing on the medical images, and the registration results are verified based on the analysis of the degree of matching between the medical images; the change parameters are adjusted according to the verification results to complete the registration processing, and the registered medical images are denoised and contrast enhanced by filtering and equalization to obtain optimized medical images; based on morphological operations and probability distribution techniques, the spinal region and ligament information in the medical images are respectively obtained, and the watershed algorithm is used to identify the growth characteristic lines in the medical images to obtain the spinal structure; the spinal structure, spinal region and ligament information are used as the three-dimensional semantic information of the spinal structure.

[0071] Specifically, the medical image is medical image data used for three-dimensional reconstruction of the spine, such as CT (tomography) or MRI (magnetic resonance) images. At the same time, the image is registered to ensure the correct alignment between different slices, in order to facilitate subsequent structural analysis. Image registration is the process of aligning spinal images acquired at different time points or by different imaging devices to ensure the position consistency of the same anatomical structure in different images. The steps of image registration are as follows:

[0072] Extract key feature points from medical images, including corners, edges, etc., select registration algorithms, such as rigid transformation, affine transformation and nonlinear transformation, and define a similarity metric based on the registration algorithm to evaluate the degree of matching between images. Commonly used ones are mutual information (MI) and mean squared error (MSE).

[0073] The calculation formula of mutual information is:

[0074]

[0075] Where MI(X, Y) represents the mutual information between the image X before registration and the image Y after registration, p(x, y) represents the joint probability density function, p(x) and p(y) represent the marginal probability density functions of X and Y respectively, and x and y represent the feature points before registration and the feature points after registration respectively.

[0076] The formula for calculating the mean square error is:

[0077]

[0078] Where I1 and I2 represent the corresponding pixel values ​​of the images before and after registration, respectively, n is the total number of pixels, and i represents the number of observed data.

[0079] At the same time, optimization algorithms such as gradient descent, Newton's method or genetic algorithm can be used to adjust the transformation parameters to maximize the similarity measure.

[0080] Specifically, when performing preprocessing operations such as image denoising, contrast enhancement and standardization to improve image quality, image preprocessing includes denoising, contrast enhancement and standardization, as follows:

[0081] Image denoising: Specifically, mean filtering can be used to reduce random noise in the image:

[0082]

[0083] Where f(x, y) represents the original image, g(x, y) represents the denoised image, m and n represent the number of pixels in the filter window, s represents the index of the filter window in the row direction (vertical direction), t represents the index of the filter window in the column direction (horizontal direction), a and b are and

[0084] Contrast enhancement aims to improve the visual effect of spinal images. It uses the commonly used histogram equalization to enhance the contrast by adjusting the grayscale distribution of the image. The transformation function T is defined as:

[0085]

[0086] Where r represents the grayscale of the input image, s represents the grayscale of the output image, pr(j) represents the probability that the grayscale of the input image is j, and j represents the grayscale of the input image.

[0087] At the same time, the purpose of standardization is to adjust the grayscale range or distribution of the image to meet specific statistical characteristics. The linear standardization formula is:

[0088] g = (f - μ) / σ;

[0089] In the formula, f represents the pixel value of the original image, g represents the standardized pixel value, μ represents the mean of the pixel value, and σ represents the standard deviation.

[0090] Image denoising, contrast enhancement and standardization can effectively improve the quality of spinal medical imaging data and lay a good foundation for subsequent three-dimensional reconstruction of the spine.

[0091] In this embodiment, when the spinal region and ligament information in the medical image are respectively obtained based on morphological operations and probability distribution technology, and the watershed algorithm is used to identify the growth feature lines in the medical image to obtain the spinal structure, the morphological operation can be used to identify the contour feature lines in the medical image, and the slicing and projection operations are performed on the contour feature lines to obtain the point set of the medical image, and the regional information of the spine is obtained according to the point set; a graph model is generated based on the adaptive sampling strategy and the point set, and the graph model is analyzed in combination with the graph theory algorithm to complete the hole identification and repair operations, and the ligament information is extracted based on the repair results; the watershed algorithm is used to identify the point set to obtain the growth feature lines, and the growth feature lines are optimized to construct a multidimensional tree, and the detailed information of the spinal structure is analyzed based on the multidimensional tree.

[0092] In this embodiment, when using morphological operations to identify contour feature lines in medical images, and performing slicing and projection operations on the contour feature lines to obtain a point set of the medical image, and obtaining regional information of the spine based on the point set, the gradient amplitude and direction of each pixel point in the medical image can be calculated, and a threshold indicator can be defined based on historical experience to identify the contour of the medical image, and the pixels on the contour boundary are eroded based on morphology; the gaps between objects on the contour boundary are filled using dilation technology and then adjacent objects are connected to obtain contour feature lines, and the position and direction of the slicing plane are determined based on the position and direction of the contour feature lines; the point set intersecting the slicing plane is calculated based on the contour feature lines, and the point set is projected onto the section plane, and the regional information of the spine is obtained based on the projection result.

[0093] It should be explained that morphological features can be used to identify different areas of the spine when performing contour feature line recognition and slice projection. The human spine is composed of several vertebrae. An adult has 26 vertebrae, including 7 cervical vertebrae, 12 thoracic vertebrae, 5 lumbar vertebrae, 1 sacrum and 1 coccyx.

[0094] The specific implementation steps are as follows:

[0095] Step 1: Extract contour feature lines, using morphological operations (such as edge detection) to identify contour feature lines;

[0096] The Canny edge detector is used to identify the contours in the image, and the Gaussian filter is used to reduce the image noise. The gradient amplitude and direction of each pixel in the image are calculated, and the edges are refined at the same time, and non-edge points are suppressed to 0. Two thresholds, a high threshold and a low threshold, are selected to determine which edges are "determined edges" and "possible edges". Finally, the edges are determined by hysteresis threshold processing;

[0097] Morphological operations, including erosion and dilation, are used to further refine the edges.

[0098] Erosion: Used to eliminate small objects and reduce pixels on object boundaries.

[0099] Dilation: Used to fill the gaps between objects and connect adjacent objects.

[0100] Use findContours function (e.g. in OpenCV) to extract contours after edge detection.

[0101] Step 2: Slice the extracted contour feature lines and project the 3D data onto a 2D plane to enhance spatial positioning accuracy;

[0102] Based on the position and direction of the contour feature line, the position and direction of the slice plane are determined. For each contour feature line Ci, the point set P that intersects with the slice plane is calculated. The slice operation formula can be expressed as:

[0103]

[0104] Project the 3D point set P onto the 2D plane. For example, project (x, y, z) onto the (x, y) plane. The projection formula can be expressed as:

[0105] P'(P)={(x,y)(x,y,z)∈P}

[0106] Where P represents the 2D point set, which represents the part of Ci that passes through the original 3D data set D, and x, y, and z represent the coordinate points of the contour feature line.

[0107] In this embodiment, a graph model is generated based on an adaptive sampling strategy and a point set, and a graph model is analyzed in combination with a graph theory algorithm to complete hole identification and repair operations. When extracting ligament information based on the repair results, the point set can be used as the original spinal data set, and a group of data points are randomly selected from the original spinal data set as the first sampling points and added to the sampling point set; sampling points are iteratively selected from the original spinal data set based on a preset number of sampling points, and the sum of the distances between the sampling points and all data points in the original spinal data set is the largest; when the number of iterations reaches the preset number of sampling points, the sampling points are added to the sampling point set to form a spinal sampling data set, the spinal sampling data set is converted into a graph model, and nodes are created for the data points; the distance between any two nodes is determined, and when the distance value is less than a set threshold, the two nodes are connected to create an edge, and a graph theory algorithm is applied to detect the edge creation status to analyze the hole information of the graph model; the boundaries of the holes are determined and the minimum spanning tree is combined to perform connectivity analysis to fill the holes, and the ligament information is extracted based on the filled graph model.

[0108] It should be explained that in order to improve the representativeness and efficiency of sampling, an adaptive sampling strategy is adopted in this embodiment. Adaptive sampling not only takes randomness into consideration, but also determines the next sampling point based on the currently selected sampling point, thereby better covering the spatial distribution of the entire 3D point cloud data set, so that the key information of the ligament can be extracted based on probability distribution and graph theory methods.

[0109] Through the adaptive mechanism, the new point selected each time is the point with the maximum information gain relative to the existing sample points. Assume that D is the original spine 3D dataset, S i is the point set obtained by the i-th sampling (sampling point set), and k represents the number of sampling points required.

[0110] Then randomly select a point from D as the first sampling point and add it to S1. For each iteration i, i Select the next point p i+1 , so that p i+1 With S i The sum of the distances of all points in is maximized, which can be achieved by calculating the following objective function:

[0111]

[0112] Where d(p, q) represents the Euclidean distance between point p and point q. i+1 Add to S i S i+1 , when |S i Stop sampling when |=k.

[0113] Specifically, connectivity analysis in graph theory is used to identify and repair holes to improve the integrity of ligament information. First, the 3D point cloud data is converted into a graph model, in which each point in the point cloud corresponds to a node in the graph, and the connection between points (usually pairs of points with a Euclidean distance less than a certain threshold) corresponds to the edge in the graph. A node is created for each point in the point cloud, and if the distance between two nodes is less than the set threshold, an edge is created between the nodes.

[0114] In the graph model, holes can be regarded as disconnected parts of the graph. Graph theory algorithms (such as depth-first search DFS or breadth-first search BFS) can be used to detect connected components in the graph. If a connected component is completely surrounded by other connected components, there is a hole. Repairing holes usually involves determining the hole boundary and filling the hole area, identifying the set of nodes that constitute the hole boundary, and using graph theory algorithms (such as minimum spanning tree MST or shortest path algorithm) to determine how to fill the hole.

[0115] In this embodiment, a watershed algorithm is applied to identify a point set to obtain a growth characteristic line, and a multidimensional tree is constructed after an optimization operation is performed on the growth characteristic line. When analyzing the detailed information of the spinal structure based on the multidimensional tree, the data points in the point set can be preprocessed and the corresponding gradient values ​​can be calculated, and the points with gradient values ​​exceeding the threshold are selected as the starting points of the growth characteristic lines; a gradient image is constructed using the gradient values, and the data points with the largest gradient values ​​are selected as seed points, and the seed points are set as marking points of the watershed transformation; the gradient image is divided into several groups of regions based on the marking points using the watershed change, and the growth characteristic lines are identified after the boundary information between the regions and the change of the gradient within the region are determined; after performing noise removal and discontinuous point removal operations on the growth characteristic lines, the point set corresponding to the growth characteristic lines is divided into several groups of subsets, and a multidimensional tree is constructed on the subsets; based on the multidimensional tree and K nearest neighbor search technology, similar points on the growth characteristic lines are found to form growth regions, and the detailed information of the spinal structure is extracted after the growth regions are merged and segmented.

[0116] It should be explained that the growth feature line and parallel K-nearest neighbor region growth are based on the growth algorithm and graph theory to segment the detailed structure such as the intervertebral disc. At the same time, the watershed algorithm is used to identify the growth feature line in the 3D data. In order to enhance the ability of the watershed algorithm to identify feature lines in 3D data, the machine learning model can be combined to improve the gradient calculation, and a more advanced segmentation algorithm can be used to improve the recognition accuracy of the feature line. The specific steps are as follows:

[0117] Step 1: Gradient calculation: Gradient calculation is the first step in identifying feature lines. Traditionally, the Laplacian operator is used to estimate the gradient of each voxel point. This embodiment uses a machine learning model to further improve the accuracy of gradient estimation. First, the 3D data is preprocessed to ensure the quality and consistency of the data. For each voxel point v, the Laplacian operator is used to calculate its gradient. At the same time, a machine learning model (such as a convolutional neural network CNN) is trained. The model accepts a local 3D voxel block as input and outputs the gradient vector of the center point of the voxel block. The specific formula is:

[0118]

[0119] Among them, V block is a local voxel block centered at v, and ML represents the machine learning model.

[0120] At the same time, the possible starting point of the feature line is determined based on the gradient magnitude, and the point whose gradient magnitude exceeds a certain threshold can be selected as the candidate starting point of the feature line.

[0121] Step 2: Use the watershed algorithm to identify feature lines: Use gradient information to construct a gradient image. The intensity value of each voxel point corresponds to the size of its gradient, and the point with the largest gradient amplitude is selected as the seed point. The seed point will be used as the marker point of the watershed transformation. The watershed algorithm is used to divide the gradient image into multiple regions. The goal of the watershed transformation is to find the "valley" in the image, that is, the place where the gradient changes most slowly. These places are regarded as the boundaries of different regions, which are specifically expressed as:

[0122] R=Watershed(Gradient Image, Seeds);

[0123] Among them, R represents the segmented area, Watershed represents the watershed algorithm, Gradient Image represents the gradient image, and Seeds represents the seed points.

[0124] By analyzing the boundaries between regions and the changes in gradients, feature lines are identified. Feature lines are usually the locations where the gradient changes the most, which are the "ridges" in the watershed algorithm. The identified feature lines are optimized to remove noise or discontinuous parts to make the feature lines smoother and more coherent.

[0125] Step 3: Parallel K-nearest neighbor region growing: In order to speed up the K-nearest neighbor search process and ensure the accuracy of the search, parallel processing technology can be used. By introducing parallel processing technology, the speed of K-nearest neighbor search can be significantly accelerated without sacrificing accuracy, thereby improving the efficiency of the overall region growing algorithm. The details are as follows:

[0126] A KD tree (multidimensional tree) is constructed on the point set around the growth feature line to quickly perform K nearest neighbor search. In order to support parallelization, the point set is divided into multiple subsets and a KD tree is constructed independently on each subset: the point set P is divided into n subsets P1, P2, ..., P n , for each subset P i , build the KD tree T in parallel i , the formula is:

[0127] T i =BuildKDTree(P i ),

[0128] All subtrees T i Merge into a complete KD tree T. During the merging process, it is necessary to ensure that the root nodes of all subtrees are correctly connected. For each point on the growth feature line, use the parallelized K nearest neighbor search to find the K nearest points, and set a threshold to judge similar points: For each point p on the growth feature line, parallelize in each subtree T i Perform K nearest neighbor search on the tree to find the K points closest to p in each subtree. The specific formula is:

[0129] N i =KNearestNeighbors(T i , p, K),

[0130] Return the nearest neighbor set N of all subtrees i , and select the K points closest to p as the final neighbor point set N. The formula is expressed as:

[0131]

[0132] A threshold ∈ is set. Only when the distance between points is less than this threshold, they are considered similar points. The similar points found are merged into the growing region. In order to speed up this process, the points on each growth feature line are processed in parallel, a growing region R is initialized, and the points on the growth feature line are added to the region. For each newly added point q, K nearest neighbor search is performed in parallel, and the similar points found are added to the growing region R. The above process is repeated in parallel until no new points can be added to the growing region R.

[0133] Optimize the growth area and remove noise or discontinuous parts. Use parallelized region merging or segmentation technology to optimize the quality of the region. Use parallelized smoothing filters to remove noise points in the region. If two regions are very close or overlap, they can be merged in parallel to reduce unnecessary segmentation.

[0134] Post-processing of the segmentation results, such as filling holes and removing small areas, can also be parallelized. A parallelized filling algorithm can be used to repair holes in the area and detect and remove areas that are too small in parallel.

[0135] In this embodiment, when morphological parameters are extracted based on the three-dimensional semantic information of the spinal structure and the correlation analysis technology is used to explore the correlation between the morphological parameters, the morphological parameters can be extracted from the three-dimensional semantic information of the spinal structure, and the morphological parameters are normalized, and the correlation coefficients between the two morphological parameters are analyzed using the correlation technology; the linear relationship between the morphological parameters is explained according to the size of the correlation coefficient, and the related parameters are divided into different groups according to the preset division criteria; variance analysis is performed on the groups to determine the ratio value of the between-group variance to the within-group variance and the probability value of the ratio value, and the differences in morphological parameters between different groups are explained according to the ratio value and the probability value; clustering operation is performed on the morphological parameters, and the correlation and difference between the morphological parameters are explained according to the clustering results to obtain the parameter correlation relationship, and the spinal morphological characteristics represented by each morphological parameter are shown based on the parameter correlation relationship.

[0136] It needs to be explained that the calculation of spinal morphological parameters and the establishment of association relationships are to extract key morphological parameters from the results of spinal semantic segmentation, such as: cervical lordosis angle, thoracic kyphosis angle, thoracolumbar kyphosis angle, lumbar lordosis angle, sacral curvature, scoliosis angle, rotation angle, rib transverse process angle, vertebral body angle, etc.

[0137] The extracted parameters are normalized, and the formula is as follows:

[0138]

[0139] Among them, Z i represents the normalized parameter; P i represents the original parameter; μ represents the mean of the parameter; σ represents the standard deviation of the parameter.

[0140] Correlation analysis is used to explore the linear relationship between different parameters. Specifically, all extracted spinal morphology parameters can be collected, and correlation analysis methods such as Pearson correlation coefficient or Spearman's rank correlation coefficient can be selected. The correlation coefficient is calculated for each pair of parameters to evaluate the strength of the linear relationship between them, and a significance test (such as t-test) is performed on the correlation of each pair of parameters to determine whether the correlation is statistically significant.

[0141] Based on the size of the correlation coefficient and the results of the significance test, the linear relationship between the parameters was explained, and variance analysis was used to explore the differences between different parameters.

[0142] Specifically, according to the purpose of the study, the parameter data can be divided into different groups, such as grouping by gender, age, etc., and one-way ANOVA or multi-way ANOVA can be selected, mainly depending on the number and type of groups. ANOVA, F value and p value are performed on each parameter to evaluate whether there are significant differences between different groups.

[0143] If the ANOVA shows a significant difference, perform a multiple comparison test (such as Tukey's test or Bonferroni correction) to determine which specific groups are different and explain the variability of the parameters between the different groups based on the F value and p value.

[0144] Standardize or normalize the normalized parameter data to ensure that the data are on the same scale, select a suitable clustering algorithm, such as K-means clustering, hierarchical clustering or DBSCAN, etc., determine the number of clusters (k value) according to the research purpose or use the Elbow method, etc., and use the selected algorithm to cluster the parameter data. Use the Silhouette score or other evaluation indicators to evaluate the quality of clustering, and based on the clustering results, explain the similarities and differences between the parameters, as well as the spinal morphological characteristics they may represent.

[0145] In this embodiment, when generating a spinal semantic parameter combination based on association relationships and morphological parameters, finding the best spinal semantic parameter combination using an optimization algorithm, and repairing the scoliotic spinal model according to the best spinal semantic parameter combination, the spinal semantic parameter combination can be determined based on spinal morphological characteristics, association relationships and morphological parameters, and a parameter space is generated based on the spinal semantic parameter combination; the parameter combination is encoded as an individual in the parameter space, and an optimization target is defined based on the application of the spinal semantic parameters in spinal model repair, and an evaluation function is defined based on the optimization target to calculate the evaluation value of each individual; genetic operations are performed based on the evaluation values ​​and the optimization algorithm to simulate natural selection and genetic processes, and the evaluation of the individuals is re-evaluated based on the genetic process, and the individual with the highest evaluation is selected as the optimal solution to obtain the best spinal semantic parameter combination; a prototype scoliotic spinal model is generated based on the spinal data, and the best spinal semantic parameter combination is applied to the prototype scoliotic spinal model to repair the spinal state.

[0146] It should be explained that a genetic algorithm can be used to find the best combination of spine semantic parameters. The specific steps are as follows:

[0147] Step 1: Population initialization;

[0148] Parameter space definition: Determine all possible combinations of spine semantic parameters, including but not limited to segmentation thresholds, edge detection parameters (such as parameters in Sobel operator and Canny edge detection algorithm), feature extraction parameters (such as parameters of HOG and Haar features), classifier parameters (such as C value and gamma value in SVM) and clustering parameters (such as the number of clusters in K-means).

[0149] Based on the spine semantic parameter combination, randomly generate or use heuristic methods to create multiple candidate solutions (populations): randomly generate a certain number of candidate solutions in the parameter space, select some potential parameter combinations as the initial population based on prior knowledge or experience, and encode each parameter combination as a chromosome, which can be binary coding, real number coding or Gray coding, etc.

[0150] Step 2: Fitness evaluation;

[0151] The fitness function (evaluation function) is defined according to specific goals (such as segmentation accuracy, parameter relevance, computational complexity, and robustness), as follows:

[0152] F(x)=α.P accuracy +β.P correlation -γ.P complexity +δ.P robustness ;

[0153] In the formula, F(x) represents the evaluation value of the xth individual, P accuracy It represents the segmentation accuracy, that is, the ratio of the correctly segmented spine area to the actual spine area, P correlation It represents the correlation of parameters, that is, it measures the correlation within the parameter combination, usually expressed by the correlation coefficient, P complexity represents the computational complexity, which measures the running time and resource consumption of the algorithm, P robustness It represents robustness, which measures the algorithm's resistance to external noise or data changes. α represents the weight coefficient of segmentation accuracy, β represents the weight coefficient of parameter correlation, γ represents the weight coefficient of computational complexity, and δ represents the weight coefficient of robustness.

[0154] The fitness value of each individual is calculated by comparing the segmentation results of the algorithm with the actual annotation results, using correlation coefficients (such as the Pearson correlation coefficient) to measure the correlation between parameters, while recording the running time and resource consumption of the algorithm, and testing the stability of the algorithm by introducing noise or changes in the data set.

[0155] Step 3: Genetic operation;

[0156] Genetic algorithms simulate natural selection and inheritance processes through genetic operations (selection, crossover, mutation): select individuals according to fitness values ​​to produce the next generation, randomly select two individuals, exchange some of their parts to produce new individuals, randomly change part of the individuals to produce new individuals, and through an iterative process, the population gradually converges to the optimal solution.

[0157] Genetic operations are used to generate a new generation of population, and the fitness of each individual in the new population is re-evaluated. When the preset maximum number of iterations is reached or a certain termination condition is met, the iteration is stopped, the results are output, and the individual with the highest fitness in the current population is selected as the optimal solution. The optimal solution, that is, the optimal combination of spine semantic parameters, is output.

[0158] In this embodiment, when a prototype scoliosis spine model is generated based on spinal data and the optimal spinal semantic parameter combination is applied to the prototype scoliosis spine model to repair the spinal state, the medical imaging data of the healthy spine can be preprocessed to extract the spinal contour data, and the prototype scoliosis spine model can be constructed in combination with the auxiliary design technology; the boundary conditions, initial displacement and velocity are defined based on the preset repair requirements, and the deformation of the prototype scoliosis spine model is guided by the physical model deformation algorithm, and the new position of each data point is obtained to obtain the scoliosis spine deformation model; the optimal spinal semantic parameter combination is applied to the scoliosis spine deformation model to simulate the morphological changes of the spine, and the scoliosis spine deformation model is repaired based on the morphological simulation results.

[0159] It needs to be explained that the best semantic feature combination guides the deformation of the spinal model. First, the healthy spinal data is used to guide the deformation of the model to generate a preliminary spinal model. Then, the best spinal semantic parameter combination obtained by the genetic algorithm is applied to the model to simulate the adjustment of the spinal morphology, explore possible correction strategies and predict the ideal state after correction.

[0160] The spinal model is generated as follows: medical imaging data of a healthy spine, such as X-rays, CT or MRI scans, is collected in advance, and the collected imaging data is preprocessed, including denoising, standardization and segmentation, to extract the contour of the spine; based on the preprocessed data, a three-dimensional model of the spine is constructed using computer-aided design (CAD) software or a three-dimensional modeling tool; the accuracy of the model is verified by comparing the model with the anatomical structure of an actual healthy spine; the initial deformation of the spinal model is guided by a deformation algorithm (such as a physics-based deformation model or a machine learning model) using the healthy spine data; the process of guiding the initial deformation of the spinal model based on the physical model deformation algorithm involves modeling and simulation of the physical properties of the spine.

[0161] Among them, the modeling and simulation of the physical properties of the spine requires the establishment of a model that can reflect the physical properties of the spine to describe the response of the spine under different conditions. For example, for elastic bodies, the Navier-Cauchy equation can be used to describe their behavior: Among them, σ1 represents the stress tensor, b represents the body force (such as gravity), ρ represents the material density, and u represents the second-order derivative of the displacement.

[0162] For linear elastic materials, the relationship between the stress tensor and the strain tensor can be expressed by Hooke's law: σ=E.ε, where E represents Young's modulus and ε represents the strain tensor.

[0163] Secondly, specify appropriate boundary conditions to simulate the real physical environment. Boundary conditions can be displacement boundary conditions (fixing the displacement of certain areas) or force boundary conditions (applying external loads), and set initial conditions, such as initial displacement and velocity.

[0164] The equations are solved using numerical methods such as the finite element method. The finite element method discretizes the continuous domain into a finite number of small units, approximates the solution in each unit, and finds the solution to the entire domain by solving a system of linear equations. In the finite element method, the solution can be expressed as:

[0165]

[0166] Where N i represents the shape function, u i Represents the displacement on the node. After substituting it into the elastic mechanics equation, we can get a set of linear equations: KU=F, where K represents the stiffness matrix; U represents the node displacement vector; F represents the external force vector. Once the linear equation is solved, the new position of each node can be obtained, thereby obtaining the deformed spine model.

[0167] Specifically, the steps of adjusting the spinal morphology based on the best parameter combination are as follows: determine the key semantic parameters that affect the spinal morphology, such as vertebral size, intervertebral disc thickness, intervertebral angle, etc., use genetic algorithms to optimize the parameters, and find the best parameter combination. In order to improve the performance of the genetic algorithm, increase its diversity and prevent premature convergence to the local optimal solution, the genetic algorithm is improved. The specific steps are as follows:

[0168] Set the initial population size N, each individual in the population represents a set of parameter configurations, and randomly generate N sets of parameters as the initial population P0, specifically: P0 = {x, x2, ..., x N}; where x i Represents the parameter vector of the i-th individual.

[0169] Calculate the fitness value of each individual. The fitness function can be defined based on the difference between the spine model and the ideal state. The fitness value f(x i ) is higher, the individual x i The closer to the optimal solution, the specific formula is expressed as: f(x i )=evaluate(x i ).

[0170] The roulette wheel selection and the elitism strategy are combined to ensure that the best individuals are retained. The elitism strategy can ensure that at least one best individual in each generation is passed to the next generation. The formula is: P elite ={x best};

[0171] P next =P elite ∪{roulette wheel selection(P current )};

[0172] Where P next represents the population of the next generation, P elite represents the elite individual set, P current represents the current generation population, x best Represents the best individual in the current generation.

[0173] Use a variety of crossover strategies (such as single-point crossover, two-point crossover, uniform crossover, etc.) to increase population diversity. The crossover probability C p It should be moderate. Too high will lead to premature convergence, while too low may lead to slow convergence. The formula is:

[0174]

[0175] The mutation operation can help escape the local optimal solution and increase population diversity. The mutation probability M p It should be moderate. Too high will destroy excellent individuals, and too low may not increase sufficient diversity. The specific formula is: if random()<M p then{x i ′=mutate(x i );

[0176] A new generation of population is generated based on selection, crossover and mutation operations. The new generation of population P t+1 It should include elite individuals and new individuals generated by selection, crossover, and mutation operations. The specific formula is: if t≥Torf(x)converges thenstop.

[0177] In the formula, X1 and X2 represent two individuals in the current generation population, the parent individuals used for crossover operation, X′1 and X′2 represent the offspring individuals generated by the crossover operation, random() is a random number generation function, which usually generates a random number between 0 and 1, and x i ′ represents the new individual generated by the mutation operation, x i represents the current individual, t represents the current generation, T represents the maximum number of iterations, and f(x) is the fitness function.

[0178] Then, by adjusting the parameters in the model, the morphological changes of the spine can be simulated, and virtual reality (VR) technology can be introduced in the process of spinal morphological adjustment to provide doctors, researchers and patients with a more intuitive and interactive learning and treatment environment, which not only increases the functionality of spinal morphological adjustment, but also improves its interest and educational value.

[0179] A virtual environment is created using a VR software development kit (SDK), such as Unity or Unreal Engine, which can display the spinal model and allow users to interact with it through VR devices (such as head-mounted displays (HMDs) and handles). Users wear VR devices to enter the virtual environment, and the spinal parameter model optimized based on the genetic algorithm is imported into the VR environment to ensure that the model can be displayed correctly in VR.

[0180] The user interface (UI) is designed so that users can directly adjust the parameters of the spinal model, such as vertebral size, intervertebral disc thickness, and intervertebral angle, through the VR handle. At the same time, a real-time feedback mechanism is implemented. When the user adjusts the parameters, the spinal model is immediately updated so that the user can observe the changes. After the adjustment is completed, the system saves the final spinal morphology and can export data for actual medical applications or research. At the same time, the data of users adjusting the spinal model in the VR environment is collected for subsequent analysis and optimization of the genetic algorithm, and the difference between the adjusted spinal model and the ideal state is evaluated.

[0181] Among them, the steps of exploring correction strategies and predicting the ideal state after correction are: analyzing the differences between the current spinal model and the ideal state, and determining the parts that need to be corrected; designing possible correction strategies, such as physical therapy, surgical correction, or brace correction; simulating the application of correction strategies on the model and observing the changes in spinal morphology; predicting the ideal spinal state after applying the correction strategies based on the simulation results; verifying the accuracy of the prediction results through actual cases, and adjusting the model and correction strategies based on feedback.

[0182] In summary, with the help of the above technical solution of the present invention, the present invention is based on the global optimization strategy of the optimization algorithm, and designs a reasonable evaluation function to quantify the performance of each parameter combination in the scoliosis model repair task, and automatically finds the optimal semantic parameter combination weight, thereby guiding the effective adjustment of the scoliosis spine model parameters, which can not only improve the quality of model repair, but also improve the parameter optimization efficiency, and avoid the tedious work caused by manual blind modification of parameters. The present invention uses defined and accessible intact semantic parameters with clear anatomical significance to describe in detail the degree, morphological characteristics and potential correction needs of scoliosis, and at the same time uses an optimization algorithm to achieve the purpose of finding the most suitable model parameter combination to accurately simulate the true state of the scoliotic spine and the best possible correction results.

[0183] The present invention achieves efficient and accurate segmentation of spinal images, which not only improves the quality of spinal repair, but also ensures the integrity and continuity of the spinal structure. At the same time, it guides the initial deformation of the spinal model based on the physical model and finite element analysis, providing a basis for subsequent optimization based on genetic algorithms, and ensuring the morphological rationality and accuracy of the spinal model. The lateral spinal model rapid repair method implemented by the present invention significantly improves the accuracy of spinal repair and the authenticity of the three-dimensional model, and effectively solves the problems of structural continuity and loss of details in traditional methods, greatly promoting the accuracy and efficiency of medical image analysis, surgical planning and treatment evaluation, and is of great significance to promoting the development of personalized medicine and precision surgery.

[0184] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A rapid repair method for scoliotic spine model based on semantic feature optimization, characterized in that: The scoliosis spine model rapid repair method includes: Use morphological features and geometric analysis technology to perform segmentation and recognition operations on pre-collected medical images, and obtain three-dimensional semantic information of the spinal structure based on the segmentation and recognition results; The morphological parameters were extracted based on the three-dimensional semantic information of the spinal structure, and the correlation analysis technology was used to explore the correlation between the morphological parameters. A spinal semantic parameter combination is generated based on the association relationship and morphological parameters, and the optimal spinal semantic parameter combination is found using an optimization algorithm. The scoliotic spinal model is repaired based on the optimal spinal semantic parameter combination.

2. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 1 is characterized in that: The method of using morphological features and geometric analysis technology to perform segmentation and recognition operations on pre-collected medical images, and obtaining three-dimensional semantic information of the spinal structure based on the segmentation and recognition results includes: Collect tomography images and magnetic resonance images as medical images, apply affine transformation technology to perform registration processing on the medical images, and verify the registration results based on analyzing the matching degree between the medical images; The registration process is completed by adjusting the change parameters according to the verification results, and the registered medical images are denoised and contrast enhanced by filtering and equalization to obtain optimized medical images; The spinal region and ligament information in medical images are obtained based on morphological operations and probability distribution techniques, and the growth feature lines in medical images are identified using the watershed algorithm to obtain the spinal structure. The spinal structure, spinal region and ligament information are used as the three-dimensional semantic information of the spinal structure.

3. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 2 is characterized in that: The method of obtaining the spinal region and ligament information in the medical image based on morphological operation and probability distribution technology, and using the watershed algorithm to identify the growth feature line in the medical image to obtain the spinal structure includes: Morphological operations are used to identify contour feature lines in medical images, and slicing and projection operations are performed on the contour feature lines to obtain a point set of the medical image, and regional information of the spine is obtained based on the point set; Generate a graph model based on adaptive sampling strategy and point set, and analyze the graph model with graph theory algorithm to complete hole identification and repair operations, and extract ligament information based on the repair results; The watershed algorithm is used to identify the point set to obtain the growth characteristic line, and a multidimensional tree is constructed after optimizing the growth characteristic line. The detailed information of the spine structure is analyzed based on the multidimensional tree.

4. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 3 is characterized in that: The method of using morphological operations to identify contour feature lines in medical images, and performing slicing and projection operations on the contour feature lines to obtain a point set of the medical image, and obtaining regional information of the spine according to the point set includes: Calculate the gradient magnitude and direction of each pixel in the medical image, and define the threshold index based on historical experience to identify the contour of the medical image, and erode the pixels on the contour boundary based on morphology; The gaps between objects on the contour boundary are filled by using the expansion technology, and then the adjacent objects are connected to obtain the contour feature line. The position and direction of the slice plane are determined based on the position and direction of the contour feature line. The point set intersecting with the slice plane is calculated based on the contour feature line, and the point set is projected onto the slice plane. The regional information of the spine is obtained based on the projection result.

5. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 4 is characterized in that: The method of generating a graph model based on the adaptive sampling strategy and point set, and analyzing the graph model in combination with a graph theory algorithm to complete hole identification and repair operations, and extracting ligament information based on the repair results includes: The point set is taken as the original spine data set, and a group of data points are randomly selected from the original spine data set as the first sampling points and added to the sampling point set; Iteratively selecting sampling points from the original spine data set based on a preset number of sampling points, and the sum of the distances between the sampling points and all data points in the original spine data set is the largest; When the number of iterations reaches a preset number of sampling points, the sampling points are added to the sampling point set to form a spine sampling data set, the spine sampling data set is converted into a graph model, and nodes are created for the data points; Determine the distance between any two nodes, connect the two nodes to create an edge when the distance value is less than the set threshold, and apply graph theory algorithms to detect the edge creation status and analyze the hole information of the graph model; The boundaries of the holes are determined and connected by the minimum spanning tree to fill the holes, and the ligament information is extracted based on the filled graph model.

6. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 5 is characterized in that: The watershed algorithm is used to identify the point set to obtain the growth characteristic line, and the growth characteristic line is optimized to construct a multidimensional tree. The detailed information of the spine structure analyzed based on the multidimensional tree includes: After preprocessing the data points in the point set, the corresponding gradient values ​​are calculated, and the points whose gradient values ​​exceed the threshold are selected as the starting points of the growth feature lines; Use the gradient value to construct a gradient image, select the data point with the largest gradient value as the seed point, and set the seed point as the marker point of the watershed transformation; Based on the marking points, the gradient image is segmented into several groups of regions using watershed changes, and the growth feature lines are identified after judging the boundary information between regions and the changes in the gradient within the region; After performing noise removal and discontinuous point removal operations on the growth feature line, the point set corresponding to the growth feature line is divided into several subsets, and a multidimensional tree is constructed on the subsets; Based on the multidimensional tree and K nearest neighbor search technology, similar points on the growth feature line are found to form a growth region, and the detailed information of the spine structure is extracted after merging and segmenting the growth region.

7. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 1 is characterized in that: The method of extracting morphological parameters based on the three-dimensional semantic information of the spine structure and using correlation analysis technology to explore the correlation between the morphological parameters includes: The morphological parameters are extracted from the three-dimensional semantic information of the spinal structure, and the morphological parameters are normalized. The correlation coefficients between the two morphological parameters are analyzed using correlation technology. The linear relationship between the morphological parameters is explained according to the size of the correlation coefficient, and the relevant parameters are divided into different groups according to the preset classification criteria; The variance analysis was performed on each group to determine the ratio of the variance between groups to the variance within groups and the probability value of the ratio value, and the differences in morphological parameters between different groups were explained based on the ratio value and the probability value; A clustering operation is performed on the morphological parameters, and the correlation and difference between the morphological parameters are explained according to the clustering results to obtain the parameter correlation relationship, and the spinal morphological characteristics represented by each morphological parameter are shown based on the parameter correlation relationship.

8. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 1 is characterized in that: The method of generating a spinal semantic parameter combination based on the association relationship and the morphological parameters, finding the best spinal semantic parameter combination using an optimization algorithm, and repairing the scoliotic spinal model according to the best spinal semantic parameter combination includes: Determine the combination of spinal semantic parameters based on spinal morphological features, association relationships and morphological parameters, and generate a parameter space based on the combination of spinal semantic parameters; Encoding the parameter combination into individuals in the parameter space, defining the optimization goal according to the application of the spinal semantic parameters in the spinal model repair, and defining the evaluation function according to the optimization goal to calculate the evaluation value of each individual; Based on the evaluation value and optimization algorithm, genetic operations are performed to simulate natural selection and genetic processes, and the evaluation of individuals is re-evaluated according to the genetic process. The individual with the highest evaluation is selected as the optimal solution to obtain the best combination of spine semantic parameters. A prototype scoliotic spine model is generated based on the spine data, and the optimal spine semantic parameter combination is applied to the prototype scoliotic spine model to repair the spine state.

9. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 8 is characterized in that: The calculation formula of the evaluation function is: F(x)=α.P accuracy +β.P correlation -g.P complexity +d.P robustness ; In the formula, F(x) represents the evaluation value of the xth individual, P accuracy represents the segmentation accuracy, P correlation represents parameter correlation, P complexity represents the computational complexity, P robustness represents robustness, α represents the weight coefficient of segmentation accuracy, β represents the weight coefficient of parameter correlation, γ represents the weight coefficient of computational complexity, and δ represents the weight coefficient of robustness.

10. The method for rapid repair of scoliotic spine model based on semantic feature optimization according to claim 9, characterized in that: The method of generating a prototype scoliosis spine model based on the spine data and applying the optimal spine semantic parameter combination to the prototype scoliosis spine model to repair the spine state includes: After preprocessing the medical imaging data of the healthy spine, the spine contour data is extracted, and a prototype scoliotic spine model is constructed in combination with auxiliary design technology; Based on the preset repair requirements, the boundary conditions, initial displacement and velocity are defined, and the deformation of the prototype scoliosis spine model is guided by the physical model deformation algorithm, and the new position of each data point is obtained to obtain the scoliosis spine deformation model; The optimal combination of spine semantic parameters is applied to the scoliotic spine deformation model to simulate the morphological changes of the spine, and the scoliotic spine deformation model is repaired based on the morphological simulation results.

Images