Cortical bone screw automatic planning method and device based on statistical morphological model

Abstract

The invention provides a cortical bone screw automatic planning method and device based on a statistical morphology model. The method comprises the following steps: acquiring CT image data of any object; based on the CT image data, segmenting point cloud information of a vertebral body to be subjected to screw placement; acquiring a statistical shape model and cortical bone screw planning information corresponding to the serial number; and registering the statistical shape model to the point cloud information, and determining a cortical bone screw plan of the vertebral body to be subjected to screw placement. In the application, the preset screw placement planning of the cortical bone screw is performed through the statistical shape model, and the preset screw placement planning is mapped to the current spinal column three-dimensional model through registration, so that the preset screw placement planning only needs to be performed on the statistical shape model, and a screw placement model is not needed, thereby solving the problem that a large number of cases need to be collected to realize the screw placement model; therefore, the problem of nail placement of the cortical bone can be solved.

Description

Technical Field

[0001] This application relates to the field of medical image recognition technology, and more specifically, to an automatic planning method and apparatus for cortical bone screws based on statistical morphological models. Background Technology

[0002] Cortical bone (CBT) screws have extremely high value in the revision of spinal surgery. Although pedicle screws are the most widely used internal fixation devices for the spine, with the increase in revision cases and osteoporosis cases, more and more patients need to use CBT screws for replacement, revision, or to add spinal anchor points to maintain spinal stability.

[0003] However, current spinal screw placement mainly involves planning the placement of pedicle screws, with fewer examples of cortical bone screw placement, making it impossible to collect a large number of cases to train a screw placement model. Summary of the Invention

[0004] The problem addressed by this application is that current cortical bone screws cannot be planned for placement using a screw placement model.

[0005] To address the aforementioned problems, the first aspect of this application provides an automatic planning method for cortical bone screws based on a statistical morphological model, comprising:

[0006] Acquire CT image data of any object;

[0007] Based on the CT image data, the point cloud information of the vertebral body to be implanted with the screw is segmented;

[0008] Obtain the statistical shape model and cortical bone screw planning information corresponding to the sequence number;

[0009] The statistical shape model is registered to the point cloud information, and the cortical bone screw planning of the vertebral body to be screwed is determined.

[0010] A second aspect of this application provides an automatic cortical screw planning device based on a statistical morphological model, comprising:

[0011] The data acquisition module is used to acquire CT image data of any object.

[0012] The data segmentation module is used to segment the point cloud information of the vertebral body to be implanted based on the CT image data;

[0013] The planning acquisition module is used to acquire the statistical shape model and cortical bone screw planning information corresponding to the sequence number;

[0014] The pin placement planning module is used to register the statistical shape model to the point cloud information and determine the cortical bone screw planning of the vertebral body to be pinned.

[0015] A third aspect of this application provides an electronic device comprising: a memory and a processor;

[0016] The memory is used to store programs;

[0017] The processor, coupled to the memory, is used to execute the program for:

[0018] Acquire CT image data of any object;

[0019] Based on the CT image data, the point cloud information of the vertebral body to be implanted with the screw is segmented;

[0020] Obtain the statistical shape model and cortical bone screw planning information corresponding to the sequence number;

[0021] The statistical shape model is registered to the point cloud information, and the cortical bone screw planning of the vertebral body to be screwed is determined.

[0022] The fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the above-described automatic planning method for cortical bone screws based on a statistical morphological model.

[0023] In this application, a statistical shape model is used to pre-plan the placement of cortical bone screws, and the pre-planned placement is mapped to the current three-dimensional model of the spine through registration. In this way, pre-planning of the placement only needs to be performed on the statistical shape model, without the need for a placement model. This solves the problem that a large number of cases need to be collected to realize a placement model before cortical bone screw placement can be performed. Attached Figure Description

[0024] Figure 1 This is a flowchart of an automatic planning method for cortical bone screws based on a statistical morphological model according to an embodiment of this application;

[0025] Figure 2 This is a schematic diagram of the statistical morphology model of the automatic planning method for cortical bone screws based on a statistical morphology model according to an embodiment of this application;

[0026] Figure 3 This is a diagram showing the results of cortical bone screw planning according to the automatic planning method for cortical bone screws based on statistical morphological models in accordance with embodiments of this application.

[0027] Figure 4 This is a structural block diagram of an automatic cortical bone screw planning device based on a statistical morphological model according to an embodiment of this application;

[0028] Figure 5 This is a structural block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0029] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the following description, in conjunction with the accompanying drawings, will illustrate the specific features of this application.

[0030] The specific embodiments are described in detail below. Although exemplary embodiments of this application are shown in the accompanying drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0031] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0032] CBT (cortical bone screw) has the following advantages:

[0033] I. Mechanical Properties: ① Enhanced Screw Holding Force: CBT screws, through full contact with the cortical bone, increase the contact area and friction between the screw and bone, thereby improving the screw's holding force and pull-out resistance. ② Good Biomechanical Stability: The CBT screw's trajectory contacts multiple cortical bone points, such as the cortex at the dorsal entry point, the medial cortex of the posterior wall of the pedicle, the lateral cortex of the anterior wall of the pedicle, and the vertebral cortex, forming a "four-point fixation," which enhances the screw's stability and load-bearing capacity.

[0034] II. Surgical Aspects: ① Less Surgical Trauma: CBT screw technology involves smaller incisions, closer to the midline, reducing dissection of fascia and muscle tissue, lowering intraoperative bleeding and postoperative infection rates, aligning with modern minimally invasive and rapid recovery principles. ② Shorter Surgical Time: Due to the more direct placement path of CBT screw technology and the elimination of excessive soft tissue dissection, the surgical procedure is relatively simple and the surgical time is shorter.

[0035] III. Safety Aspects: ① Reduced Risk of Nerve Injury: The CBT screw is placed away from the spinal canal and nerves, avoiding the areas of nerve roots and dural sacs, greatly reducing the chance of nerve injury. ② Reduced Adjacent Segment Degeneration: The CBT screw is inserted 1 mm below the lower edge of the transverse process, away from the facet joints of the superior segment, reducing the impact on the facet joints of the superior segment, thereby reducing the incidence of adjacent segment degeneration.

[0036] IV. Applicability: ① Suitable for various patient types: CBT screw technology has special advantages for patients with osteoporosis, obesity, and those undergoing revision surgery, providing them with more reliable fixation. ② Combination of multiple screw placement methods: In spinal revision surgery, CBT screw technology can be used in combination with other screw placement methods without removing the original internal fixation, providing more options for complex spinal surgeries.

[0037] However, current spinal screw placement mainly involves planning the placement of pedicle screws, with fewer examples of cortical bone screw placement, making it impossible to collect a large number of cases to train a screw placement model.

[0038] To address the aforementioned issues, this application provides a novel automatic planning scheme for cortical bone screws based on a statistical morphological model. This scheme enables pre-planning of cortical bone screw placement using a statistical shape model, eliminating the need to collect a large number of cases to develop a placement model before cortical bone screw placement can be performed.

[0039] This application provides an automatic planning method for cortical bone screws based on a statistical morphological model. The specific scheme of this method is as follows: Figures 1-3 As shown, this method can be executed by an automatic cortical screw planning device based on a statistical morphological model, which can be integrated into electronic devices such as computers, servers, computer clusters, and data centers. Combined with... Figure 1 , Figure 2 The diagram shows a flowchart of an automatic cortical screw planning method based on a statistical morphological model according to an embodiment of this application; wherein the automatic cortical screw planning method based on a statistical morphological model includes:

[0040] S101, acquire CT image data of any object;

[0041] In this application, the CT image data includes at least three-dimensional CT images of the lumbar spine of the subject. The three-dimensional CT images of the lumbar spine are in DICOM format, preferably with a slice thickness of ≤1mm.

[0042] S102, Based on the CT image data, segment the point cloud information of the vertebral body to be fitted with the screw;

[0043] S103, obtain the statistical shape model and cortical bone screw planning information corresponding to the sequence number;

[0044] In this application, the vertebral body to be screwed is one or more vertebrae of the thoracic or lumbar spine, and the vertebrae have serial numbers; the vertebrae have multiple corresponding statistical shape models, and each statistical shape model is provided with cortical bone screw planning information.

[0045] S104, Register the statistical shape model to the point cloud information and determine the cortical bone screw plan for the vertebral body to be screwed.

[0046] In this application, a statistical shape model is used to pre-plan the placement of cortical bone screws, and the pre-planned placement is mapped to the current three-dimensional model of the spine through registration. In this way, pre-planning of the placement only needs to be performed on the statistical shape model, without the need for a placement model. This solves the problem that a large number of cases need to be collected to realize a placement model before cortical bone screw placement can be performed.

[0047] In one specific embodiment, step S102, based on the CT image data, segments the point cloud information of the vertebral body to be fitted with the screw, including:

[0048] The CT image data is segmented to obtain voxel data of the vertebral body to be screwed;

[0049] The voxel data of the vertebrae to be placed are converted into a triangular patch network;

[0050] The triangular patch network is converted into point cloud information.

[0051] In this application, Totalsegmentator or medical image processing software Mimics20.0 (Materialise, Leuven, Belgium) is used to segment the vertebral body on CT image data and export the segmentation results; the voxel data of the segmentation results are converted into triangular patch networks and then into point clouds and the position information is saved.

[0052] In this application, the segmentation results can be visualized using ITK-SNAP to check whether they contain complete vertebral bodies and pedicle structures; if there are segmentation errors (such as adhesion between adjacent vertebral bodies), they can be manually corrected or re-segmented.

[0053] Preferably, a pre-trained 3DU-Net or nnUNet model is used, with the entire lumbar spine CT sequence as input and a binary mask for each vertebra as output (taking the segmentation of only the L3 vertebra as an example, the label value of the L3 vertebra is 1, and the rest are 0).

[0054] In this application, after obtaining the voxel data of the vertebrae to be placed, the following steps are also included: morphological closing operation: eliminating segmentation cavities (kernel size 3×3×3 voxels). Connected component analysis: retaining the largest connected component and removing small noise regions.

[0055] Preferably, the method further includes a cortical bone extraction step, in which the region (cortical bone) with a CT value > 500 HU is extracted within the vertebral body mask for threshold segmentation, and the Sobel operator is used to enhance the cortical bone boundary.

[0056] In this application, by extracting the cortical bone, it is possible to visually determine whether the placement of the cortical bone screws meets the requirements.

[0057] In this application, a surface reconstruction algorithm is used to convert a binary voxel mask into a surface mesh model. Specifically, this includes:

[0058] Surface reconstruction:

[0059] Marching Cubes: Input: Binary voxel data (1 = cone, 0 = background). Isosurface threshold: 0.5 (intermediate threshold for voxel values). Output: Initial triangular mesh (typically containing 100,000-500,000 faces).

[0060] Mesh optimization:

[0061] Patch simplification: The Quadric Edge Collapse algorithm is used to reduce the number of patches to 5,000-10,000 levels. Feature preservation: Curvature protection weights are set (e.g., pedicle region weight ×2). Smoothing: Laplacian smoothing is iterated 3 times (smoothing coefficient λ = 0.5) to eliminate step artifacts. Normal calculation: The normal direction is unified (outward) based on the coordinates of neighboring vertices (radius = 2mm).

[0062] In this application, point cloud data suitable for registration is generated from the optimized mesh. Specifically, this includes:

[0063] Point cloud sampling:

[0064] Uniform sampling: 17,000 vertices are randomly generated on the grid surface (density approximately 1 point / mm). 2 Poisson disk sampling is used to avoid local clustering. In this application, an additional 20% of sampling points are added in areas of high curvature (pedicles, articular processes) to enhance the corresponding features.

[0065] Point cloud attributes:

[0066] Coordinates: Stores the (x, y, z) position of the vertex in the original CT coordinate system (unit: mm). Normal Vector: Records the surface normal at each vertex (used for point-to-surface ICP registration). Curvature: Calculates and stores the local curvature (used for subsequent registration weight allocation).

[0067] In one specific implementation, step S103, obtaining the statistical shape model corresponding to the sequence number, includes:

[0068] Acquire training CT data;

[0069] Based on the training CT data, voxel data of the corresponding sequenced vertebral bodies were segmented;

[0070] Convert voxel data into point cloud data of the cone body;

[0071] A statistical shape model of the vertebra is established based on the point cloud data of the vertebra.

[0072] In this application, the training CT data must include lumbar spine CT scans from at least 100 different individuals (covering variables such as age, sex, and ethnicity). The vertebral body voxels will be converted into standardized point clouds.

[0073] In this application, the preferred scanning parameters are: layer thickness ≤ 1mm, resolution ≥ 512×512 pixels, and DICOM format.

[0074] In this application, after acquiring training CT data, cases with severe deformities (such as scoliosis with a Cobb angle >30°) or interference from metal implants are excluded; a uniform HU value range (-1000 to 2000 HU) is established, and isotropic resampling (e.g., 0.5 mm) is performed. 3 (Volumetrics).

[0075] In this application, a pre-trained 3DnnUNet model is used to accurately segment the target vertebral body from the training CT scan. The entire training CT data is input, and the binary mask of each vertebral body is output. Preferably, the target vertebral body voxel data is extracted by the vertebral body number identifier (e.g., L3 label value = 3).

[0076] In this application, the cone voxels are converted into a standardized point cloud, specifically including: generating a triangular mesh using the Marching Cubes algorithm, applying Laplacian smoothing (3 iterations, λ=0.5) to remove staircase artifacts; performing Poisson disk sampling on the mesh surface to generate 10,000-20,000 uniformly distributed vertices; and recording the following for each point: three-dimensional coordinates (original CT coordinate system, unit mm), normal vector (calculated based on local surface), and curvature value (used for subsequent registration weighting).

[0077] In this application, a statistical shape model of the vertebra is established based on the point cloud data of the vertebra using the PCA algorithm.

[0078] In this application, preferably, a statistical shape model of the vertebra is established based on the point cloud data of the vertebra, including:

[0079] Point cloud alignment (spatial normalization): Translate all cone point clouds to coincide with the centroid to eliminate positional differences; calculate the optimal rotation matrix through singular value decomposition (SVD) to initially align each point cloud with the first sample; calculate the average shape of all point clouds after the current alignment, and realign each sample to the average shape, repeating until the average shape change is <0.1mm.

[0080] Establish point correspondence: Perform template deformation on the registered point cloud: Use the first sample as a template and deform the remaining samples to the template point distribution using a non-rigid ICP algorithm, or use anatomical landmarks as guidance: Manually label key points (such as pedicle apex) and establish full surface point correspondence through thin plate spline interpolation (TPS).

[0081] Constructing the shape matrix: Flatten the 3D point cloud coordinates of each sample into a one-dimensional vector, calculate the average shape vector, and construct a centered shape matrix.

[0082] Principal Component Analysis: Calculate the average shape vector across all training samples; then calculate the covariance matrix and solve the characteristic equation; sort by eigenvalues ​​in descending order and retain the top K principal components (typically covering 95% of the cumulative variance).

[0083] Generate statistical shape models: Construct a formula for expressing statistical shape models, and derive the mean model representing the average cone morphology of the population and the variation pattern model representing the extreme shapes of each principal component.

[0084] In this application, the sum of all eigenvalues ​​in the PCA algorithm gives the total variance of the training data. Eigenvectors with larger eigenvalues ​​can describe more significant shape changes, while eigenvectors with smaller eigenvalues ​​express smaller or more local changes.

[0085] In one specific embodiment, there are multiple statistical shape models; step S104, registering the statistical shape models to the point cloud information, includes:

[0086] Calculate the registration matrix between each statistical shape model and the point cloud information of the cone;

[0087] Obtain quantitative evaluation indicators;

[0088] Based on quantitative evaluation indicators, the closest statistical shape model is determined;

[0089] Based on the registration matrix, the closest statistical shape model is mapped to the point cloud information of the cone.

[0090] In this application, there are multiple statistical shape models; the registration matrix between each statistical shape model and the point cloud information of the cone is calculated separately. The specific process can be as follows:

[0091] Pre-alignment is performed using the centroid and principal axis, followed by initial registration using point-to-point ICP, and finally fine registration using point-to-surface ICP.

[0092] In this application, for any statistical shape model and the point cloud information of the cone, rigid registration is performed using the Iterative Closest Point (ICP) algorithm to calculate the corresponding registration matrix (coarse registration).

[0093] The ICP iterative algorithm is an efficient method for handling rigid registration between two point sets. It achieves rigid registration by iteratively registering the data shape to the model shape with rigid transformation. It has two basic steps: finding the correspondence between the point sets, and calculating a new rotation matrix and translation transformation based on the currently calculated correspondence; alternating between the two steps until convergence.

[0094] Fine-grained registration is performed using point-to-plane ICP, which specifically includes: for each point, calculating its local surface normal vector (using k-nearest neighbors analyzed by PCA, usually k=30, taking the eigenvector corresponding to the smallest eigenvalue as the normal vector); and minimizing the distance from the point to the plane through iterative optimization.

[0095] In this application, the quantitative evaluation index can be Hausdorff distance, surface coverage, root mean square error, etc. Through this index, the difference between different statistical shape models and the point cloud information of the cone is evaluated, so that the statistical shape model with the smallest difference is the closest.

[0096] In one specific embodiment, S104, determining the cortical bone screw plan for the vertebral body to be fitted, includes:

[0097] Obtain the cortical bone screw planning information corresponding to the serial number;

[0098] Based on the registration matrix, the cortical bone screw planning information corresponding to the closest statistical shape model is mapped to the point cloud information of the vertebral body;

[0099] The mapped cortical bone screw planning information is converted into voxel data and fused with CT image data.

[0100] In this application, there are multiple statistical shape models corresponding to the serial numbers of the vertebral bodies, and each statistical shape model is set with cortical bone screw planning information.

[0101] It should be noted that cortical bone screw planning information can be added manually after the statistical shape model is known. This eliminates the need for prior cortical bone screw planning cases and allows for direct manual annotation, which is simple and convenient. Furthermore, there are only 5-6 statistical shape models, resulting in minimal workload and minimal human resource consumption.

[0102] Thus, by first obtaining the registration matrix and the cortical bone screw planning information corresponding to the closest statistical shape model, and then mapping the cortical bone screw planning information, the cortical bone screw planning information of the vertebral body can be obtained.

[0103] In this application, the pin track parameters are converted into a binary mask in CT voxel space. Specifically, this may include: converting the pin track mesh to the original CT voxel coordinate system (applying the DICOM physical coordinate transformation matrix), and marking the pin track region within the CT voxel mesh using a scan line algorithm.

[0104] It should be noted that since both the statistical shape model and the pin track model contain the positioning points retained by the hospital, the maximum connectivity is first performed. Connectivity Filter is used to retain the largest connected component (the main body of the vertebra) and remove small volume positioning points (removing the positioning points in the upper right and lower left of the statistical shape model to avoid image registration and quantification calculation). Then, the same sampling process is performed on the STL file of each SSM to obtain the corresponding point cloud.

[0105] In one specific implementation, after acquiring the CT image data of any object, the method further includes:

[0106] The CT image data is adaptively adjusted.

[0107] In one specific implementation, the adaptive adjustment of the CT image data includes:

[0108] The CT image data is divided into blocks to obtain independent blocks;

[0109] For each independent block, obtain the first and second neighboring blocks with different spacings;

[0110] A first feature block is generated based on the independent block and the first neighboring block;

[0111] A second feature block is generated based on the independent block and the second neighboring block.

[0112] The first and second feature blocks are compressed to obtain a compressed block.

[0113] Iterate through all the individual blocks and generate adjusted CT image data based on the resulting compressed blocks.

[0114] In this application, CT image data is divided into blocks, that is, the CT image data is divided into corresponding image blocks using a checkerboard pattern; wherein, the image block can be at the pixel level (that is, each pixel is an image block) or other levels, the specific division depending on the actual processing situation.

[0115] It should be noted that CT image data is point cloud data, which can be regarded as a three-dimensional image.

[0116] In this application, a sliding window or a fixed step size is used to divide the image into blocks of the same size.

[0117] It should be noted that since CT image data is a three-dimensional image, a plane (sagittal plane) is selected for chessboard-like division. Each cell is a strip with a depth (the depth of the three-dimensional image), and this strip is an image block.

[0118] Preferably, in this application, each image block consists of 100-1000 pixels, thereby enabling more feature calculations between local regions while ensuring generation accuracy and reducing computational load.

[0119] In this application, an image block is selected as an independent block. The image blocks above, below, to the left, and to the right of this independent block are the first neighboring blocks; the image blocks one grid away from the top, bottom, left, and right of this independent block are the second neighboring blocks. The spacing between the first and second neighboring blocks and the independent block is different.

[0120] In this application, neighborhood information is extracted for each independent block to capture local structure.

[0121] In this application, generating the first feature block is to generate a local feature representation using an independent block and its first neighboring block. Specifically, this can be done by processing the independent block and the first neighboring block with convolutional layers and attention layers to obtain the first feature block.

[0122] In this application, the specific structure and parameters of the convolutional layer and attention layer can be obtained from the training data or determined according to the actual situation.

[0123] It should be noted that in this application, there are four first neighboring blocks and multiple first feature blocks.

[0124] In this application, the independent block and the first neighboring block are processed by convolutional layers and attention layers to obtain the first feature block. The specific process is as follows: the independent block and four neighboring blocks are concatenated together to form a multi-channel input, and the convolutional layer is used to extract features from the concatenated block; an important feature is enhanced by using a self-attention mechanism or a channel attention mechanism, the attention weight is calculated, and the output of the convolutional layer is weighted to enhance the important feature; the output of the attention layer is split into multiple feature blocks, and each feature block corresponds to the processing result of the independent block and at least one neighboring block.

[0125] In this application, a second feature block is generated to generate a broader local feature representation using the independent block and its second neighboring block. The specific generation process is the same as that of the first feature block, except that the parameters of the convolutional layer and the attention layer are different.

[0126] In this application, the generated feature blocks are compressed into a more compact representation to reduce computational cost while retaining key information. Feature compression is performed using pooling operations (such as max pooling or average pooling) or fully connected layers.

[0127] In this way, multiple first and second feature blocks are compressed into a single compressed block, which corresponds to the size and position of the individual blocks and is used to replace them. All image blocks are replaced by the compressed block, resulting in adjusted CT image data.

[0128] In this application, each image block of the CT image data is traversed to obtain the corresponding compressed block.

[0129] In this application, for image blocks / independent blocks near the edge, their first and second neighboring blocks are incomplete. In this case, the incomplete blocks are completed by copying the first and second neighboring blocks in their relative positions. For example, if the first neighboring block above an independent block does not exist, the first neighboring block below it is copied and used as the block above it.

[0130] In this application, by completing the image blocks, the processing accuracy of adjacent image blocks is greatly improved.

[0131] In this application, the similarity relationship between local regions is captured by an adaptive adjustment module, thereby enhancing the feature representation. For images with rich textures or complex structures, high-quality image data can be generated.

[0132] In this application, the statistical morphological model can analyze a large amount of lumbar CT data to obtain the morphological characteristics and patterns of the lumbar spine in different individuals, thereby providing personalized entry point and path planning for each patient. This automated planning system can automatically plan the path for any CBCT image during testing, eliminating the need for manual intervention in the path planning process.

[0133] In this application, since the placement direction of the CBT screw is far away from the spinal canal and nerve, the statistical morphological model can more accurately determine the entry point and screw path, avoiding the screw from accidentally entering the spinal canal or damaging the nerve root, thereby reducing the risk of nerve injury.

[0134] In this application, the automatic planning system can quickly complete the planning of screw channels, reducing the time spent on repeated intraoperative fluoroscopy and manual measurement, thereby shortening the operation time.

[0135] In this application, the automatic planning system can provide clear screw channels and entry points, making the surgical operation simpler, reducing difficulties in identifying anatomical structures and determining the screw placement path during surgery, and reducing the difficulty of the operation.

[0136] In this application, the statistical morphological model can provide personalized screw channel planning for each patient, making screw placement more precise, reducing interference with surrounding tissues, and thus reducing the incidence of postoperative adjacent segment degeneration.

[0137] In this application, precise screw channel planning makes the contact between the screw and bone tissue closer, improves the stability of the screw, and reduces the risk of screw loosening and breakage.

[0138] In this application, the statistical morphological model can generate personalized screw channel planning schemes based on the specific lumbar spine morphology of each patient, meeting the needs of different patients and improving the adaptability and effectiveness of surgery.

[0139] In this application, for patients with osteoporosis, the statistical morphological model can more accurately determine the screw entry point and screw path, enabling the screw to be better anchored in the cortical bone, improving the screw's stability and pull-out resistance, thereby improving the treatment effect.

[0140] This application provides an automatic planning device for cortical bone screws based on a statistical morphological model, used to execute the automatic planning method for cortical bone screws based on a statistical morphological model described above. The automatic planning device for cortical bone screws based on a statistical morphological model will be described in detail below.

[0141] like Figure 4 As shown, the automatic planning device for cortical bone screws based on statistical morphological models includes:

[0142] Data acquisition module 101 is used to acquire CT image data of any object;

[0143] Data segmentation module 102 is used to segment point cloud information of the vertebral body to be implanted based on the CT image data;

[0144] The planning acquisition module 103 is used to acquire the statistical shape model and cortical bone screw planning information corresponding to the sequence number;

[0145] The pin placement planning module 104 is used to register the statistical shape model to the point cloud information and determine the cortical bone screw planning of the vertebral body to be pinned.

[0146] In one implementation, the data segmentation module 102 is further configured to:

[0147] The CT image data is segmented to obtain voxel data of the vertebral body to be screwed; the voxel data of the vertebral body to be screwed is converted into a triangular patch network; the triangular patch network is converted into point cloud information.

[0148] In one implementation, the planning acquisition module 103 is further configured to:

[0149] Acquire training CT data; based on the training CT data, segment the voxel data of the corresponding vertebral body; convert the voxel data into point cloud data of the vertebral body; based on the point cloud data of the vertebral body, establish a statistical shape model of the vertebral body.

[0150] In one implementation, there are multiple statistical shape models; the pin planning module 104 is further used for:

[0151] Calculate the registration matrix between each statistical shape model and the point cloud information of the vertebra; obtain quantitative evaluation indicators; determine the closest statistical shape model based on the quantitative evaluation indicators; and map the closest statistical shape model to the point cloud information of the vertebra based on the registration matrix.

[0152] In one implementation, the pin placement planning module 104 is further configured to:

[0153] Obtain the cortical bone screw planning information corresponding to the serial number; based on the registration matrix, map the cortical bone screw planning information corresponding to the closest statistical shape model to the point cloud information of the vertebral body; convert the mapped cortical bone screw planning information into voxel data and fuse it with CT image data.

[0154] In one implementation, the data segmentation module 102 is further configured to:

[0155] The CT image data is adaptively adjusted.

[0156] In one implementation, the data segmentation module 102 is further configured to:

[0157] The CT image data is divided into blocks to obtain independent blocks; for each independent block, a first neighboring block and a second neighboring block with different spacing are obtained; a first feature block is generated based on the independent blocks and the first neighboring blocks; a second feature block is generated based on the independent blocks and the second neighboring blocks; the first feature block and the second feature block are compressed to obtain compressed blocks; all independent blocks are traversed, and adjusted CT image data is generated based on the obtained compressed blocks.

[0158] The automatic planning device for cortical bone screws based on statistical morphological models provided in the above embodiments of this application and the automatic planning method for cortical bone screws based on statistical morphological models provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by their stored applications.

[0159] The above describes the internal function and structure of an automatic cortical bone screw planning device based on a statistical morphological model, such as... Figure 5 As shown, in practice, this automatic planning device for cortical bone screws based on statistical morphological models can be implemented as an electronic device, including: a memory 301 and a processor 303.

[0160] Memory 301 can be configured to store a program.

[0161] Additionally, memory 301 can also be configured to store various other data to support operation on the electronic device. Examples of such data include instructions for any application or method used to operate on the electronic device, contact data, phonebook data, messages, pictures, videos, etc. Memory 301 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0162] Processor 303, coupled to memory 301, is used to execute programs in memory 301 for:

[0163] Acquire CT image data of any object;

[0164] Based on the CT image data, the point cloud information of the vertebral body to be implanted with the screw is segmented;

[0165] Obtain the statistical shape model and cortical bone screw planning information corresponding to the sequence number;

[0166] The statistical shape model is registered to the point cloud information, and the cortical bone screw planning of the vertebral body to be screwed is determined.

[0167] In one implementation, the processor 303 is further configured to:

[0168] The CT image data is segmented to obtain voxel data of the vertebral body to be screwed; the voxel data of the vertebral body to be screwed is converted into a triangular patch network; the triangular patch network is converted into point cloud information.

[0169] In one implementation, the processor 303 is further configured to:

[0170] Acquire training CT data; based on the training CT data, segment the voxel data of the corresponding vertebral body; convert the voxel data into point cloud data of the vertebral body; based on the point cloud data of the vertebral body, establish a statistical shape model of the vertebral body.

[0171] In one embodiment, the statistical shape model has multiple components; the processor 303 is further configured to:

[0172] Calculate the registration matrix between each statistical shape model and the point cloud information of the vertebra; obtain quantitative evaluation indicators; determine the closest statistical shape model based on the quantitative evaluation indicators; and map the closest statistical shape model to the point cloud information of the vertebra based on the registration matrix.

[0173] In one implementation, the processor 303 is further configured to:

[0174] Obtain the cortical bone screw planning information corresponding to the serial number; based on the registration matrix, map the cortical bone screw planning information corresponding to the closest statistical shape model to the point cloud information of the vertebral body; convert the mapped cortical bone screw planning information into voxel data and fuse it with CT image data.

[0175] In one implementation, the processor 303 is further configured to:

[0176] The CT image data is adaptively adjusted.

[0177] In one implementation, the processor 303 is further configured to:

[0178] The CT image data is divided into blocks to obtain independent blocks; for each independent block, a first neighboring block and a second neighboring block with different spacing are obtained; a first feature block is generated based on the independent blocks and the first neighboring blocks; a second feature block is generated based on the independent blocks and the second neighboring blocks; the first feature block and the second feature block are compressed to obtain compressed blocks; all independent blocks are traversed, and adjusted CT image data is generated based on the obtained compressed blocks.

[0179] In this application, Figure 5 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 5 The components shown.

[0180] The electronic device provided in this embodiment is based on the same inventive concept as the automatic planning method for cortical bone screws based on statistical morphological models provided in this application embodiment, and has the same beneficial effects as the methods adopted, run or implemented by the application stored therein.

[0181] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0182] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can 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, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0183] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0184] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0185] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory. Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0186] This application also provides a computer-readable storage medium corresponding to the automatic planning method for cortical bone screws based on statistical morphology models provided in the foregoing embodiments, wherein a computer program (i.e., a program product) is stored thereon, and the computer program, when run by a processor, executes the automatic planning method for cortical bone screws based on statistical morphology models provided in any of the foregoing embodiments.

[0187] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0188] As defined in this article, computer-readable media do not include transient computer-readable media, such as modulated data signals and carrier waves.

[0189] The computer-readable storage medium provided in the above embodiments of this application and the automatic planning method for cortical bone screws based on statistical morphological models provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application stored therein.

[0190] It should be noted that numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known structures and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0191] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0192] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. An automatic planning method for cortical bone screws based on statistical morphological models, characterized in that, include: Acquire CT image data of any object; Based on the CT image data, the point cloud information of the vertebral body to be implanted with the screw is segmented; Obtain the statistical shape model and cortical bone screw planning information corresponding to the sequence number; The statistical shape model is registered to the point cloud information, and the cortical bone screw planning of the vertebral body to be screwed is determined.

2. The automatic planning method for cortical bone screws based on statistical morphological models according to claim 1, characterized in that, The step of segmenting the point cloud information of the vertebral body to be fitted with the screw based on the CT image data includes: The CT image data is segmented to obtain voxel data of the vertebral body to be screwed; The voxel data of the vertebrae to be placed are converted into a triangular patch network; The triangular patch network is converted into point cloud information.

3. The automatic planning method for cortical bone screws based on statistical morphological models according to claim 1, characterized in that, The step of obtaining the statistical shape model corresponding to the sequence number includes: Acquire training CT data; Based on the training CT data, voxel data of the corresponding sequenced vertebral bodies were segmented; Convert voxel data into point cloud data of the cone body; A statistical shape model of the vertebra is established based on the point cloud data of the vertebra.

4. The automatic planning method for cortical bone screws based on statistical morphological models according to any one of claims 1-3, characterized in that, There are multiple statistical shape models; the registration of the statistical shape models to the point cloud information includes: Calculate the registration matrix between each statistical shape model and the point cloud information of the cone; Obtain quantitative evaluation indicators; Based on quantitative evaluation indicators, the closest statistical shape model is determined; Based on the registration matrix, the closest statistical shape model is mapped to the point cloud information of the cone.

5. The automatic planning method for cortical bone screws based on statistical morphological models according to claim 4, characterized in that, The planning of the cortical bone screw for determining the vertebral body to be screwed includes: Obtain the cortical bone screw planning information corresponding to the serial number; Based on the registration matrix, the cortical bone screw planning information corresponding to the closest statistical shape model is mapped to the point cloud information of the vertebral body; The mapped cortical bone screw planning information is converted into voxel data and fused with CT image data.

6. The automatic planning method for cortical bone screws based on statistical morphological models according to any one of claims 1-3, characterized in that, After acquiring the CT image data of any object, the process further includes: The CT image data is adaptively adjusted.

7. The automatic planning method for cortical bone screws based on statistical morphological models according to claim 6, characterized in that, The adaptive adjustment of the CT image data includes: The CT image data is divided into blocks to obtain independent blocks; For each independent block, obtain the first and second neighboring blocks with different spacings; A first feature block is generated based on the independent block and the first neighboring block; A second feature block is generated based on the independent block and the second neighboring block. The first and second feature blocks are compressed to obtain a compressed block. Iterate through all the individual blocks and generate adjusted CT image data based on the resulting compressed blocks.

8. An automatic planning device for cortical bone screws based on a statistical morphological model, characterized in that, include: The data acquisition module is used to acquire CT image data of any object. The data segmentation module is used to segment the point cloud information of the vertebral body to be implanted based on the CT image data; The planning acquisition module is used to acquire the statistical shape model and cortical bone screw planning information corresponding to the sequence number; The pin placement planning module is used to register the statistical shape model to the point cloud information and determine the cortical bone screw planning of the vertebral body to be pinned.

9. An electronic device, characterized in that, include: Memory and processor; The memory is used to store programs; The processor, coupled to the memory, is used to execute the program for: Acquire CT image data of any object; Based on the CT image data, the point cloud information of the vertebral body to be implanted with the screw is segmented; Obtain the statistical shape model and cortical bone screw planning information corresponding to the sequence number; The statistical shape model is registered to the point cloud information, and the cortical bone screw planning of the vertebral body to be screwed is determined.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by a processor to implement the automatic planning method for cortical bone screws based on a statistical morphological model as described in any one of claims 1-7.

Images