A method for digital design and fabrication of autotransplantation of teeth

Abstract

This application relates to a digital design and fabrication method for autologous tooth transplantation, belonging to the field of digital oral medical technology. The method performs threshold screening and region growth on oral tomographic scan data to construct a three-dimensional model of the donor tooth, the implantation site, and the opposing dentition. A unified coordinate system is established using covariance matrix decomposition, and the geometric matching position is determined through iterative nearest-point calculation with normal vector constraints. The method constructs a multi-objective function to adjust implantation parameters and establishes a heterogeneous finite element model based on voxel grayscale mapping. If the interface stress exceeds the bone resorption threshold, a rotation correction is generated based on the stress-weighted centroid for closed-loop optimization. During the design phase, features of the implantation site are extracted, and a surgical guide containing shrinkable cooling channels is generated based on signed distance field fusion. This guide is then manufactured after printing and scaling compensation. This invention improves the initial stability of the transplanted tooth and the cooling and adaptation performance of the surgical guide by optimizing the implantation pose through biomechanical simulation and closed-loop feedback.

Description

Technical Field

[0001] This application relates to the field of digital oral healthcare technology, and in particular to a method for digital design and fabrication of autologous tooth transplants. Background Technology

[0002] Autologous tooth transplantation, which involves transplanting impacted teeth or teeth with no functional value to another edentulous site, is an effective biological method for restoring missing teeth. The clinical success rate of this technique highly depends on maintaining the periodontal ligament activity of the donor tooth and ensuring good initial stability after transplantation. With the development of digital technology, three-dimensional reconstruction and surgical simulation based on cone-beam computed tomography (CBCT) data have become important aids in preoperative planning.

[0003] Current digital design processes for autologous tooth transplantation primarily focus on static geometric matching, using computational algorithms to determine the spatial geometric fit between the donor tooth root and the recipient bone cavity, thus defining the implantation location and guiding bone cavity preparation. However, simple geometric matching fails to adequately consider the biomechanical response of the transplanted tooth during mastication. If the implantation position only satisfies geometric fit but the interfacial stress distribution is concentrated, under actual occlusal loads, the local bone wall in the recipient area is prone to compressive stress exceeding physiological limits, potentially inducing root resorption or bone wall necrosis, thus affecting the long-term survival rate of the graft. Furthermore, existing auxiliary analysis methods often simplify complex jawbone tissue into a homogeneous continuous material, neglecting the actual impact of the non-uniform spatial distribution of bone density on mechanical transmission, thus reducing the accuracy of preoperative biomechanical assessment.

[0004] In surgical procedures, digital surgical guides are used to guide the drill bit in preparing the bone cavity for the recipient area. Due to limitations in the guide's structural design, the main body often obstructs the surgical field of view and hinders the flow of external cooling water into the cutting depth, leading to heat accumulation during preparation and increasing the risk of thermal damage to bone tissue. Furthermore, the additive manufacturing process of surgical guides involves material shrinkage, and conventional design processes lack pre-deformation compensation mechanisms for specific printing directions and material properties. This results in deviations in the final placement of the manufactured guide within the patient's mouth, affecting the precise execution of the surgical plan. Summary of the Invention

[0005] The purpose of this application is to provide a digital design and fabrication method for autologous tooth transplantation, which aims to improve the existing technology where digital design of autologous tooth transplantation mainly relies on geometric static matching, lacks biomechanical stress assessment that incorporates the characteristics of heterogeneous bone, and suffers from problems such as obstructed cutting cooling and manufacturing dimensional deviations in surgical guide design, which affect the long-term survival rate of transplantation and the accuracy of surgical execution.

[0006] By adopting the above technical solution, the present invention provides a digital design and fabrication method for autologous tooth transplantation, including receiving oral cavity tomographic scan data of the patient, performing image segmentation processing to construct three-dimensional models of the donor tooth, the alveolar bone of the implantation area and the opposing dentition, establishing a unified spatial coordinate system and performing rigid body transformation and iterative nearest point calculation to determine the geometric matching position of the three-dimensional model of the donor tooth in the three-dimensional model of the alveolar bone of the implantation area.

[0007] Based on this, the interference volume between the donor tooth and the alveolar bone of the implantation area is calculated. The axial angle and implantation depth of the donor tooth are adjusted with reference to the three-dimensional model of the opposing dentition. A simulated occlusal load is applied to the proposed implantation position for finite element analysis. The stress distribution at the interface is calculated and the implantation pose parameters are finely adjusted accordingly. The surface features of the alveolar bone and adjacent teeth in the implantation area are extracted as the base adaptation surface. A hollow guide sleeve structure, a connecting support structure, and a cooling channel structure are generated along the long axis of the three-dimensional model of the donor tooth. The three-dimensional model of the surgical guide is generated by Boolean operation and then manufactured by 3D printing.

[0008] When constructing the 3D model of the donor tooth, the voxels in the region of interest are subjected to grayscale thresholding to obtain the initial hard tissue mask, and the Euclidean distance transformation is performed to generate a distance transformation map. The local maxima with the largest distance value in the distance transformation map are selected as seed points. Based on the connectivity criterion and grayscale difference constraint, the region growth algorithm is executed to separate the independent 3D model of the donor tooth.

[0009] When establishing a unified spatial coordinate system, the covariance matrix of the vertices of the donor tooth 3D model mesh is calculated and eigenvalue decomposition is performed. The eigenvectors corresponding to the maximum and minimum eigenvalues ​​are defined as the Z-axis and Y-axis, respectively. The donor tooth 3D model is divided into multiple layers along the Z-axis and the area is calculated. If the average cross-sectional area at the positive direction end of the Z-axis is less than that at the negative direction end, the Z-axis eigenvector is flipped to point towards the crown direction.

[0010] During the iterative nearest point calculation, the surface of the donor tooth root and the inner surface of the bone cavity of the implanted area are extracted as the source point set and the target point set, respectively. The angle between the normal vector of the point in the source point set and the normal vector of the corresponding point in the target point set is calculated. Matching point pairs with an angle greater than a preset angle threshold are removed. The transformation matrix is ​​calculated based on the retained matching point pairs until the mean square error converges.

[0011] When adjusting the axial angle and implantation depth of the donor tooth 3D model, the geometric position data of each 3D model are extracted, and a multi-objective pose optimization function including occlusal interference depth, adjacency distance and root deviation is calculated. The function is solved by nonlinear least squares method to obtain rigid body transformation increment, and the donor tooth 3D model is adjusted accordingly.

[0012] When performing finite element mesh generation for the proposed implantation location, the model is transformed into tetrahedral mesh elements, and simulated periodontal ligament layer elements are generated in the gap between the tooth root and the bone wall. A mapping relationship between gray values ​​and elastic modulus is established for mesh elements in the bone tissue region, and the elastic modulus of the element is calculated based on the voxel gray values ​​using a power function formula.

[0013] When fine-tuning the implantation pose parameters based on stress distribution, closed-loop optimization is performed. If the maximum von Mises stress value exceeds the preset bone resorption threshold, the stress-weighted centroid of the nodes in the high-stress region and the reverse vector pointing to the geometric center of the donor tooth's three-dimensional model are calculated. The cross product of the reverse vector and the current long axis is calculated to determine the rotation axis. Based on the rotation axis and the stress over-limit ratio, a rotation correction amount is generated to adjust the pose.

[0014] When generating the cooling channel structure, a cubic Bézier curve path is constructed connecting the center of the water inlet and the center of the cooling outlet, and the inner diameter is set to gradually shrink from the inlet end to the outlet end along this path; the axial direction of the cooling outlet center is set to form a preset angle with the central axis of the hollow guide sleeve, so that the axis of the cooling outlet center points to the preset cutting area located in the extension direction of the central axis.

[0015] Before generating and manufacturing the 3D model of the surgical guide, the base adapter surface, hollow guide sleeve structure, connecting support structure, and cooling channel structure are converted into a signed distance field. A union operation is performed in voxel space, and a smooth minimum function is applied to the connection parts to generate transition chamfers. The polyhedron extraction algorithm is used to reconstruct a triangular mesh. The axial scaling factor is obtained by measuring the size ratio of the standard calibration block and a diagonal matrix is ​​constructed. The vertex coordinates of the 3D model of the surgical guide are multiplied by the diagonal matrix to generate pre-deformation compensation manufacturing data.

[0016] In summary, this application includes at least one of the following beneficial technical effects:

[0017] 1. This application establishes a unified coordinate system based on the eigenvalue decomposition of the covariance matrix and combines iterative nearest point calculation with the constraint of the angle between the normal vectors. This can eliminate invalid matching point pairs and achieve geometric registration between the donor tooth and the alveolar bone of the implantation area. Combined with multi-objective pose optimization including occlusal interference and adjacency distance, it reduces the risk of occlusal trauma caused by relying solely on geometric shape matching, reduces the amount of alveolar bone removed during surgery, and improves the initial stability of the donor tooth.

[0018] 2. This application utilizes the mapping relationship between grayscale values ​​and elastic modulus to construct a heterogeneous bone tissue material model, reflecting the influence of bone density distribution in the implantation area on mechanical transmission; through a closed-loop stress optimization mechanism based on bone resorption threshold, high-stress areas are identified and rotation corrections are calculated, so that the implantation posture avoids stress concentration areas, and the interface stress distribution is maintained within the physiological tolerance range, reducing the risk of postoperative root resorption or osteonecrosis caused by excessive local compressive stress.

[0019] 3. The cooling channel integrated into the surgical guide of this application adopts a cubic Bézier curve path with inner diameter contraction, which increases the coolant flow rate, and the outlet axis points to the cutting area, improving the cooling efficiency during the preparation process and reducing bone thermal damage. In addition, Boolean operation based on signed distance field and smooth minimum value function processing ensures smooth transition at the connection of the guide structure. Combined with pre-deformation compensation manufacturing data based on calibration block, the dimensional error caused by the shrinkage of 3D printing material is corrected, ensuring the positioning accuracy of the surgical guide during clinical wear. Attached Figure Description

[0020] Figure 1 This is a system architecture diagram of a digital design and fabrication method for autologous tooth transplantation according to an embodiment of this application;

[0021] Figure 2 This is a flowchart illustrating the main process of a digital design and fabrication method for autologous tooth transplantation according to an embodiment of this application.

[0022] Figure 3 This is a comparison of stress distribution before and after biomechanical optimization of a digital design and fabrication method for autologous tooth transplantation according to an embodiment of this application;

[0023] Figure 4 This is a histogram showing the root bone fit error distribution of a digital design and fabrication method for autologous tooth transplantation according to an embodiment of this application.

[0024] Explanation of reference numerals in the attached figures:

[0025] 100. Computer processing terminal; 10. Data acquisition and reconstruction module; 20. Virtual registration module; 30. Transplantation scheme optimization module; 40. Guide plate generation module; 200. CBCT imaging equipment; 300. 3D printing equipment. Detailed Implementation

[0026] The following is in conjunction with the appendix Figure 1 -Appendix Figure 4 This application will be described in further detail below.

[0027] Reference Figure 1 This invention provides a method for digital design and fabrication of autologous tooth transplants. This method is executed using a digital design and fabrication system for autologous tooth transplants. The system includes a data acquisition and reconstruction module 10, a virtual registration module 20, a transplantation plan optimization module 30, and a guide plate generation module 40, and is connected to a CBCT imaging device 200 and a 3D printing device 300. The method includes the following steps:

[0028] S100 receives patient oral cavity tomographic scan data from CBCT imaging equipment 200 through data acquisition and reconstruction module 10, reads DICOM format raw data, performs image segmentation processing according to preset grayscale threshold range, converts voxel data into triangular mesh data using moving cube algorithm, constructs three-dimensional model of donor tooth, three-dimensional model of alveolar bone in implantation area and three-dimensional model of opposing dentition respectively, performs mesh smoothing and noise reduction operation, and generates digital surface model;

[0029] S200: A unified spatial coordinate system is established through the virtual registration module 20. An initial rigid body transformation is performed based on anatomical feature points to move the donor tooth 3D model to the preset range of the recipient alveolar bone 3D model. The iterative nearest point algorithm is run to calculate the Euclidean distance between the root surface point set of the donor tooth 3D model and the bone cavity surface point set of the recipient alveolar bone 3D model. The mean square error is minimized through iterative calculation to determine the geometric matching position of the donor tooth 3D model within the recipient alveolar bone 3D model.

[0030] S300 calculates the interference volume between the three-dimensional model of the donor tooth and the three-dimensional model of the alveolar bone in the implantation area through the transplantation scheme optimization module 30, quantifies the area of ​​insufficient bone or root obstruction, adjusts the axial angle and implantation depth of the three-dimensional model of the donor tooth based on the occlusal surface data of the three-dimensional model of the opposing dentition, performs finite element mesh generation on the proposed implantation position, applies simulated occlusal load, calculates the von Mises stress distribution at the periodontal ligament and alveolar bone interface, and adjusts the implantation parameters based on the stress distribution results;

[0031] S400 extracts the three-dimensional model of the alveolar bone in the recipient area and the surface features of adjacent teeth as the base adaptation surface through the guide plate generation module 40. A hollow guide sleeve structure is generated along the long axis of the donor tooth three-dimensional model. The base structure, guide sleeve structure and connecting support structure are fused through Boolean operation. The preset buffer gap is deducted to generate a three-dimensional model of the surgical guide. The three-dimensional model of the surgical guide is converted into standard triangular language format data and transmitted to the 3D printing equipment 300 for solid manufacturing.

[0032] Reference Figure 2 The computer processing terminal 100 is used for a processor and a memory, the memory storing computer-executable instructions, and the processor executing the instructions to implement the functions of the following modules.

[0033] The data acquisition and reconstruction module 10 is used to receive patient oral cavity tomographic scan data from the CBCT imaging device 200. The data acquisition and reconstruction module 10 reads the raw data in DICOM format and performs image segmentation processing according to the preset grayscale threshold range. The data acquisition and reconstruction module 10 uses the moving cube algorithm to convert voxel data into triangular mesh data and constructs a three-dimensional model of the donor tooth, a three-dimensional model of the alveolar bone in the implant area, and a three-dimensional model of the opposing dentition, respectively. The data acquisition and reconstruction module 10 also performs mesh smoothing and denoising operations to generate a digital surface model for subsequent processing.

[0034] The virtual registration module 20 is used to establish a unified spatial coordinate system and perform spatial alignment between multiple models. The virtual registration module 20 first performs an initial rigid body transformation based on the anatomical feature points selected by the user, moving the donor tooth 3D model to the preset range of the recipient alveolar bone 3D model. Subsequently, the virtual registration module 20 runs an iterative nearest point algorithm to calculate the Euclidean distance between the root surface point set of the donor tooth 3D model and the bone cavity surface point set of the recipient alveolar bone 3D model. By iteratively calculating and minimizing the mean square error, the optimal geometric matching position of the donor tooth 3D model within the recipient alveolar bone 3D model is determined.

[0035] The transplantation plan optimization module 30 is used to perform biomechanical analysis and morphological adjustments on the registered model. It calculates the interference volume between the donor tooth 3D model and the recipient alveolar bone 3D model, quantifying areas of insufficient bone volume or root obstruction. Based on the occlusal surface data of the opposing dentition 3D model, the module adjusts the axial angle and implantation depth of the donor tooth 3D model. Furthermore, it performs finite element mesh generation at the proposed implantation location, applies simulated occlusal loads, calculates the von Mises stress distribution at the periodontal ligament and alveolar bone interface, and fine-tunes the implantation parameters based on the stress distribution results.

[0036] The surgical guide generation module 40 is used to construct the geometric entity of the surgical guide based on the optimized transplantation plan. The module extracts features from the three-dimensional model of the alveolar bone in the recipient area and the surfaces of adjacent teeth as the base adaptation surface, and generates a hollow guide sleeve structure along the long axis of the donor tooth's three-dimensional model. The module then merges the base structure, guide sleeve structure, and connecting support structure through Boolean operations, and subtracts a preset buffer gap to generate the three-dimensional model of the surgical guide. The module converts the three-dimensional model of the surgical guide into standard trigonometric language format data and transmits it to the 3D printing equipment 300 for solid fabrication.

[0037] The overall workflow of this embodiment includes: first, acquiring high-precision oral anatomy data and reconstructing a three-dimensional model using the data acquisition and reconstruction module 10; second, digitally aligning the donor tooth and the recipient area using the virtual registration module 20; then, performing morphological and mechanical simulation analysis using the transplantation scheme optimization module 30 to establish the optimal implantation scheme; and finally, designing and manufacturing a personalized surgical guide on the 3D printing device 300 using the guide generation module 40 to assist clinicians in completing the extraction and implantation of the donor tooth.

[0038] Image data acquisition parameters and accuracy control:

[0039] In the digital design of autologous tooth transplantation, the geometric accuracy of the digital model directly determines the degree of matching between the donor tooth and the bone cavity of the recipient area.

[0040] S111, Perform cone-beam CT scan of the oral and maxillofacial region:

[0041] The operator fixes the patient's head to the scanning equipment stand and adjusts the field of view (FOV) to fully cover the quadrant where the donor tooth is located, the alveolar bone area of ​​the implantation site, and the opposing dentition. During the scanning process, the patient must maintain a centric, static occlusion.

[0042] S112, Set high-resolution scan parameters:

[0043] To clearly distinguish the fine morphology of the tooth root surface and the trabecular bone structure within the alveolar bone in the reconstructed model, this embodiment strictly sets the scanning parameters and precisely controls the isotropic voxel dimensions. The range of 0.125 mm to 0.2 mm was chosen because excessively large voxels can lead to partial volumetric effects and blurred boundaries, while excessively small voxels can increase radiation dose. This preferred range achieves a balance between image signal-to-noise ratio and geometric resolution that meets clinical needs.

[0044] S113, Accuracy verification based on geometric error theory:

[0045] The theoretical maximum geometric error of a 3D reconstruction model is limited by the spatial resolution of the original data. In this embodiment, the system controls the scanning resolution within the average physiological thickness of the periodontal ligament (approximately 0.25 mm), thereby ensuring that the deviation between the digital model and the actual anatomical structure is less than the thickness of the periodontal ligament. This ensures from the outset that the surgical guide designed subsequently will not cause compression damage to periodontal tissues during implantation due to model errors.

[0046] S114, Output and verify standard image data:

[0047] After scanning, the device outputs medical digital image communication data conforming to the DICOM 3.0 standard. The system automatically parses the data tags. If key parameters (such as pixel spacing and slice thickness) exceed the preset threshold for minimally invasive transplantation design, an alarm will be automatically triggered, prompting a rescan or automatic super-resolution interpolation processing. Simultaneously, to address the strip-like artifacts generated by metal restorations (such as crowns, bridges, and fillings) in the reconstructed images, the system integrates a normalized metal artifact correction algorithm. This algorithm first identifies high-density metal areas through strong threshold segmentation, marks the corresponding metal projection data as missing in the sinusoidal domain, repairs the missing sinusoidal data using linear interpolation, and then performs filtered back-projection reconstruction, fusing high-frequency edge information from the original image. This restores the clarity of the tooth edges while eliminating artifact noise projected onto the alveolar bone of the implant area, ensuring the grayscale accuracy of subsequent finite element material assignment.

[0048] Image segmentation and region extraction:

[0049] After acquiring high-resolution DICOM data, specific anatomical structures need to be separated from the background.

[0050] S121, Define the region of interest:

[0051] The system defines a local bounding box in three-dimensional voxel space that includes the donor tooth, the implantation area, and the opposing dentition, eliminating irrelevant high-density structures and reducing computational load.

[0052] S122, Perform organization classification based on grayscale features:

[0053] Based on the difference in X-ray absorption between tooth hard tissue and surrounding soft tissue, the module classifies cells according to voxel grayscale values. The system iterates through voxels within the Region of Interest (ROI) and filters cells with grayscale values ​​falling within a preset hard tissue threshold range. Voxels were used to initially extract hard tissue masks containing teeth and alveolar bone.

[0054] S123, using a region growing algorithm to separate independent anatomical structures:

[0055] To address the adhesion problem caused by the similar density of the donor tooth root to the surrounding alveolar bone, the module employs a region growing algorithm to select a seed point at the anatomical center of the donor tooth. To achieve automated and precise seed point localization, the system first performs an Euclidean distance transformation on the initial hard tissue mask obtained in step S122, calculating the distance from each voxel within the mask to the nearest background voxel. The voxel with the largest distance value (i.e., the local maximum point) is defined as the anatomical center seed point. This method effectively avoids low-density areas within the pulp chamber or root canal, ensuring the seed point is located within the dentin solid. Considering... The donor tooth may be a multi-rooted tooth with root canal calcification causing the pulp chamber to be disconnected. The system performs local maximum search to identify all isolated maxima points in the distance transformation map whose distance values ​​are greater than a preset threshold (e.g., 80% of the maximum distance value). The set of these points is defined as the multi-source seed point set. Parallel region growth is performed, and finally the union of each growth region is taken to ensure that the separated tooth model contains all independent root branches. Based on the connectivity criterion and gray-level difference constraint, iterative growth is performed to neighboring voxels until the donor tooth is completely separated from the surrounding bone tissue.

[0056] S124, Morphological Operations and Mask Correction:

[0057] Morphological opening and closing operations are performed on the generated mask data to remove small non-anatomical connections and fill tiny pores on the root surface. In practice, spherical structural elements are used, with a radius set between 0.3 mm and 0.5 mm. The selection of this radius range is based on the fact that it is slightly larger than the size of common CT scan artifact noise, while it must be smaller than the minimum radius of curvature of the root apex (usually >0.6 mm). This ensures that while removing noise, the donor tooth model will not lose the key anatomical structure of the root apex due to excessive corrosion, thus ensuring the accuracy of subsequent root canal length measurements and obtaining a mask with a solid internal anatomical structure.

[0058] S125: Deep Learning-Based Semantic Segmentation Enhancement

[0059] To address the issue of traditional threshold segmentation failing in some cases due to extremely narrow periodontal ligament spaces, this embodiment introduces a 3D convolutional neural network based on a U-Net or V-Net architecture. This network is pre-trained on a labeled large-scale dental CBCT dataset and can extract high-dimensional features of tooth edges. In processing step S123, the system can call this trained network model to automatically predict the probability map of each voxel belonging to the donor tooth, alveolar bone, or periodontal ligament, thereby achieving more accurate boundary segmentation in low-contrast regions. During network training, to address the problem of a severe imbalance in the volume ratios of teeth, alveolar bone, and periodontal ligament (i.e., very few periodontal ligament voxels), this embodiment employs a hybrid loss function of weighted Dice coefficient and cross-entropy to assign weights to the periodontal ligament category. The weight is 5-10 times that of other categories, forcing the network to focus on the boundary prediction of fine structures. At the same time, in order to prevent the network from overfitting on small sample datasets, an online data augmentation module is introduced during training to randomly perform transformation operations on the input voxel blocks, including rotation ([-15°, +15°] range), mirror flip, isotropic scaling (0.8-1.2 times), and elastic deformation. Among them, elastic deformation is achieved by generating a random displacement field that follows a Gaussian distribution on the control points of the coarse grid and applying it to the image voxels through B-spline interpolation, thereby simulating the natural variation of tooth root morphology in different patients and improving the robustness of the model to segmentation of irregular tooth roots. In addition, the image data is Z-score normalized at the input end, and the gray values ​​are truncated in the range of [-1000, 3000] HU to eliminate scanning differences between different devices.

[0060] 3D Mesh Reconstruction and Optimization:

[0061] S130, 3D mesh reconstruction:

[0062] This embodiment utilizes isosurface extraction techniques (such as the moving cube algorithm) to transform discrete voxel mask data into a continuous triangular mesh model. In order to eliminate staircase artifacts, the algorithm uses linear interpolation to determine the precise position of the vertices of the triangular facets on the voxel edges, so that the surface resolution of the reconstructed model is better than the original scanned voxel size.

[0063] S140, Mesh Optimization and Smoothing:

[0064] The module automatically detects and repairs non-manifold structures in the mesh. To simulate realistic bone surface morphology, a volume-preserving smoothing algorithm (such as the HC-Laplacian algorithm) is employed. This algorithm introduces restoring force constraints while performing Laplacian smoothing on the vertex coordinates, ensuring that the volume deviation of the model before and after smoothing is controlled within a preset tolerance (e.g., 0.5%), preventing root volume shrinkage from causing excessively tight subsequent guide plate fit. Finally, a quadratic error metric algorithm is used to simplify the mesh, reducing the data volume while preserving high-curvature features such as cusps and incisal edges.

[0065] Virtual registration and biological adaptation analysis:

[0066] S210, Establishment of spatial coordinate system and model normalization:

[0067] The specific spindle calculation process is as follows: The system first calculates the geometric center of all vertices of the donor tooth mesh, and then translates the coordinate origin to this center; subsequently, it constructs... covariance matrix ,in Let be the vertex coordinate vector; for the matrix Eigenvalue decomposition is performed to obtain three eigenvalues ​​and their corresponding eigenvectors. Given the directional ambiguity (i.e., the 180-degree flip problem) of the eigenvectors calculated by principal component analysis, the system introduces a morphological gradient-based direction correction mechanism. The tooth mesh is divided into several layers along the calculated initial Z-axis, and the area of ​​each layer's cross-section is calculated. Based on anatomical principles, the average cross-sectional area of ​​the crown region is significantly larger than that of the root apex region. If the average cross-sectional area at the positive Z-axis end is detected to be smaller than that at the negative Z-axis end, the current coordinate system is determined to be inverted. The system automatically multiplies the Z-axis eigenvector by -1 to flip it, ensuring that the established local coordinate system's Z-axis always points towards the crown direction and the Y-axis always points towards the lingual side. The eigenvector corresponding to the largest eigenvalue is defined as the long axis direction (Z-axis) of the donor tooth, and the eigenvector corresponding to the smallest eigenvalue is defined as the buccolingual diameter direction (Y-axis), thus establishing a local intrinsic coordinate system that conforms to the anatomical morphology of the tooth. The system constructs a unified world coordinate system and establishes a local coordinate system for the donor tooth. The principal axis direction of the donor tooth is calculated through principal component analysis, and it is initialized to a standard upright posture, facilitating subsequent pose transformation matrix solving.

[0068] S220, Initial coarse registration based on feature points:

[0069] The module extracts homologous anatomical landmarks (such as the apex and mesial and distal points of the neck) on the donor tooth and implantation site model, and uses the least squares method to solve the rigid body transformation parameters to quickly move the donor tooth to the preset range of the alveolar socket in the implantation site, providing a good initial position for fine registration.

[0070] S230, Iterative Closest Point Fine Registration:

[0071] This embodiment employs an improved iterative nearest-point algorithm. To enhance the biological significance of registration, the system extracts only the root surface data (below the cementoenamel junction) of the donor tooth as the source point set and the inner surface data of the alveolar bone wall in the implantation area as the target point set. During the iteration process, a normal vector consistency constraint is introduced to eliminate mismatched points, and the optimal transformation matrix is ​​solved using a point-to-surface error metric until the mean square error converges. The specific criterion for determining the normal vector consistency constraint is: calculate the normal vector of a point in the source point set (donor tooth). The normal vector of the nearest corresponding point in the target point set (planted area) The angle between ,like If the match is invalid, it is discarded. The angle threshold is set based on the curvature change characteristics of the root surface, which can effectively prevent the buccal point of the root from being incorrectly matched to the lingual bone wall of the alveolar fossa, thereby improving the global convergence stability of the registration.

[0072] S310, calcaneal morphology compatibility analysis:

[0073] To meet the real-time interactive requirements of clinical design and address the issue of excessively long distance calculations between high-density meshes, the system pre-constructs hierarchical or directional bounding box trees as spatial index structures for the alveolar bone model of the implant area. When calculating the distance field, the system first traverses the bounding box tree to quickly remove mesh faces that are too far apart, then performs precise point-triangle distance calculations only on faces within intersecting or adjacent leaf nodes, thereby reducing the time complexity from... Reduce to The system quantitatively assesses the root-bone matching degree by calculating the signed distance field from the vertex of the donor tooth root surface to the bone wall surface of the recipient area. Positive values ​​represent periodontal ligament gaps, while negative values ​​represent bone interference. The system also statistically analyzes the average gap width and fitting uniformity, and generates an intuitive heat map using three-dimensional chromatographic mapping technology to assist doctors in determining the amount of bone removed and the feasibility of implantation.

[0074] S320, positional adjustment based on occlusal function:

[0075] To ensure good chewing function and avoid occlusal trauma after donor tooth implantation, the module constructs a pose optimization model under multiple constraints; multi-objective pose optimization function construction: the module constructs a comprehensive objective function. The aim is to find the optimal rigid body transformation increment. This minimizes the weighted total cost, which includes occlusal interference, adjacency relationships, and calcaneal position preservation:

[0076] ;

[0077] in, Indicates the depth of interlocking interference. Indicates the adjacency distance. The preset target adjacency distance, This is a penalty term for the amount of displacement of the tooth root from its matched position; , , These are the corresponding weighting coefficients. In this embodiment, they are assigned... The highest weight is given priority to avoid occlusal trauma; the specific weighting strategy is as follows: The value ranges from 0.6 to 0.8. The value ranges from 0.15 to 0.25. The value ranges from 0.05 to 0.15, and satisfies... This cascaded weighting ensures that non-invasive engagement is a hard constraint, while adjacency is a secondary optimization objective; simultaneously, the preset target adjacency distance... The distance is set to 1.5mm to 2.0mm, which is designed to allow sufficient biological width for the regeneration of the gingival papilla after tooth transplantation and to prevent the formation of black triangles.

[0078] Solution and Update: The above function is solved using the nonlinear least squares method. Based on the calculated optimal transformation parameters, the indentation, elongation, or tilt angle of the donor tooth is automatically fine-tuned until all constraints meet the clinically permissible range.

[0079] Biomechanical simulation and closed-loop optimization:

[0080] S330, Finite Element Model Construction and Solution:

[0081] The system uses 10-node tetrahedral elements (Tet10) for mesh generation. Compared to 4-node elements, Tet10 can more accurately fit the geometric features of the curved surface of the tooth root and alleviate shear self-locking. The average thickness of the generated simulated periodontal ligament layer elements is set to 0.2mm to 0.25mm, and at least one layer of elements is generated in the thickness direction. Before solving, the system automatically performs a mesh quality check to ensure that the Jacobian ratio of all elements is greater than 0.6. For distorted elements that do not meet the quality requirements, the system automatically performs local mesh reconstruction or Laplacian smoothing to ensure the convergence of the simulation results. The system converts the geometric model into a volume mesh and automatically identifies the gap region between the tooth root and the bone wall to generate simulated periodontal ligament layer elements. Material property mapping based on gray values: Unlike the traditional method of assigning uniform material properties, this embodiment dynamically maps the material properties of bone tissue based on the gray values ​​of CBCT data. For each bone tissue element, according to empirical formulas... (in , The system calculates the individual elastic modulus of the bone tissue region (using constants as a constant). To prevent abrupt changes in the elastic modulus of local elements due to noise or artifacts in CBCT images, which could lead to numerical singularities in the finite element calculation, the system first performs three-dimensional Gaussian filtering smoothing on the voxel grayscale field of the bone tissue region before performing material property mapping. While preserving the macroscopic bone density gradient, microscopic noise is smoothed out, ensuring that the generated stiffness matrix has a good condition number.

[0082] In this embodiment, to accurately simulate the mechanical properties of the mandible, the value range of constant a is set to 2000 to 3000 MPa, and the value range of constant b is set to 1.5 to 2.0. When processing the cortical bone region, the elastic modulus calculated based on its average density is usually distributed in the range of 10-15 GPa; while the elastic modulus of the cancellous bone region is distributed in the range of 0.5-2 GPa. The Poisson's ratio is uniformly set to 0.3. This material assignment method based on grayscale gradient can continuously simulate the transition from dense bone to porous bone, avoiding stress concentration artifacts caused by traditional homogenization models. It should be noted that for the unit simulating the periodontal ligament layer, due to its nonlinear viscoelasticity and near-incompressible biological characteristics, the system does not use the above grayscale mapping formula, but sets it as a homogeneous material, assigning an elastic modulus of 0.68MPa to 1.2MPa and a Poisson's ratio of 0.45 to 0.49, accurately reflecting its lateral expansion effect and buffering function under pressure. This allows the simulation model to realistically reflect the differences in mechanical properties of different parts of the implantation area (such as cortical bone and cancellous bone), improving prediction accuracy.

[0083] Subsequently, boundary constraints were applied and chewing loads were simulated to solve the global stiffness equation and calculate the displacement field and von Mises stress distribution at each node. Regarding boundary constraints, the bottom and lateral boundary nodes of the alveolar bone model in the implant area were set to full constraints (i.e., all six degrees of freedom were zero). Regarding load application, two loading conditions were simulated: under vertical loading, an axial force of 100-150 N was applied to the central fossa of the donor tooth's occlusal surface; under lateral loading, a lateral force of 50-80 N was applied to the buccal apex and lingual slope at a 45-degree angle to the major axis. The system calculated the peak stress under both conditions and used the maximum value as the optimization criterion.

[0084] S340, parameter iteration based on stress feedback:

[0085] To obtain the optimal scheme for long-term biomechanical stability, the system performs closed-loop optimization; a biomechanical adaptability evaluation is constructed: the system extracts stress data from key interfaces in the recipient area (such as the cortical bone region of the neck) and constructs an evaluation function that includes a stress peak factor and a stress distribution uniformity factor; stress-weighted centroid and pose correction are calculated: if the evaluation results do not meet the standards, the module calculates the stress-weighted centroid. Determine the direction of adjustment:

[0086] ;

[0087] in, These are the nodal stress values. For node coordinates; the specific rotation correction logic is as follows: calculate the reverse vector of the high stress concentration. With the donor tooth's current major axis vector vector cross product ,Will The unit vector is defined as the rotation axis, and the rotation angle step size (usually 0.5° to 2.0°) is calculated based on the proportion of stress peak exceeding the threshold. The rotation operation is performed according to the right-hand rule, thereby realizing the torque unloading deflection with the geometric center of the tooth as the fulcrum in three-dimensional space. The system calculates the vector from the stress-weighted centroid to the geometric center of the donor tooth, which indicates the opposite direction of high stress concentration. The module generates translation and rotation corrections based on this vector, automatically fine-tunes the donor tooth in the direction away from the stress concentration area, and redistributes the torque by changing the tilt angle.

[0088] Iterative Loop: The system applies the corrected pose parameters to the model, automatically triggering a new round of finite element calculations until the evaluation function converges or the maximum stress falls below the bone resorption threshold. In this embodiment, the bone resorption threshold is set based on Frost's bone remodeling theory (bone homeostasis theory). For the cortical bone region, the von Mises stress threshold is set to 25 MPa to 35 MPa (approximately 20%-30% of the cortical bone yield strength); for the cancellous bone region, the threshold is set to 3 MPa to 5 MPa. When the local stress exceeds this threshold, the system determines that there is a risk of pathological bone resorption and forcibly triggers the next round of pose adjustment until the stress values ​​of all nodes fall back below the safe threshold.

[0089] Surgical guide design and manufacturing:

[0090] S410, base surface design and recessed treatment:

[0091] The guide plate coverage area is determined based on the implantation area and adjacent tooth surface model. The weighted average direction of the normal vector of the surface patch within the coverage area is calculated as the optimal placement path direction. Using the principle of ray projection along the placement path direction, the monotonic non-decreasing height field surface is calculated to fill the undercut, and a preset gap is offset along the normal direction to accommodate the adhesive. The size of the preset gap is directly related to the placement accuracy and retention force of the guide plate. In this embodiment, the normal offset of the base adapter surface (i.e., the buffer gap) is set to 30μm to 50μm. This value range is slightly larger than the manufacturing tolerance of the 3D printing equipment 300 (usually 20μm). This ensures that the guide plate is placed without resistance during surgery and leaves an extremely thin flow layer for tissue fluid or temporary adhesive, generating capillary adsorption and enhancing the intraoperative stability of the guide plate.

[0092] S420, Guiding Structure Construction:

[0093] Based on the planned implant (donor tooth) axis and platform position, a guide cylinder is parametrically constructed. The system automatically calculates the vertical height of the top surface of the guide cylinder based on the drill bit length and planned depth. For the fluid dynamics optimization design of the cooling channel: when generating the sidewall cooling windows, this embodiment does not simply use Boolean subtraction, but instead constructs a fluid dynamics-compliant guide channel structure. The center path of the streamlined surface is generated by constructing a cubic Bézier curve, which starts from the starting point... (Cooling inlet center), end point (Cooling outlet center) and two intermediate control points Definition; The system automatically optimizes the position of control points so that the vector Parallel to the external water injection direction, vector The angle between the channel and the tangent direction of the drill bit cutting zone meets the above-mentioned set range, thereby ensuring that the water flow in the channel has no sudden turbulence and achieves a laminar flow acceleration effect. The inlet of the channel faces the buccal injection direction, and the inner wall of the channel is designed with a streamlined curved surface to guide the cooling water flow to accurately converge to the cutting zone where the drill bit contacts the bone under the action of centrifugal force, thereby minimizing the risk of osteonecrosis caused by heat generation during surgery. The inlet diameter of the guide channel is designed to be 2.0mm to 3.0mm, which is compatible with the cooling nozzle of the standard implant handpiece. The channel gradually narrows inside, and the outlet diameter is 1.0mm to 1.5mm. The Venturi effect is used to increase the flow rate. The angle between the channel outlet axis and the long axis of the drill bit is set to 15 degrees to 25 degrees to ensure that the high-speed water flow can directly flush the bone cutting interface 2mm to 5mm below the drill bit tip and remove the cutting heat.

[0094] S430, Multi-component Fusion and Topology Repair:

[0095] To balance the structural rigidity of the guide plate with surgical field visibility, the cross-section of the connecting rod is designed as a trapezoid or rounded rectangle, with a width of 4.0mm to 6.0mm and a thickness of 2.5mm to 3.5mm. This size range has been verified through topology optimization, enabling it to resist the torque generated during drilling (typically less than 30 N·cm) while controlling the elastic deformation of the guide plate to within 0.05mm, ensuring the accuracy of drill bit guidance. To address the shortcomings of traditional mesh Boolean operations in handling complex anatomical surfaces, which easily produce self-intersecting or non-manifold edges, this embodiment employs an implicit modeling technique based on a distance field. The system first connects the guide plate base and the guide... The cylinder and connecting rod are converted into high-resolution signed distance fields. The union operation of the distance fields is performed in voxel space (i.e., the minimum value of the distance value of the corresponding voxel is taken), and a smooth minimum function is applied to automatically generate a smooth transition chamfer at the connection. Finally, the fused distance fields are re-isosurfaced into triangular meshes using a polyhedron extraction algorithm. This method naturally ensures that the generated guide plate model is a watertight and topologically correct simply connected entity. The guide plate base, guide cylinder, and connecting rod are fused into a simply connected entity through Boolean union operation. A transition surface is generated at the connection to disperse stress, and non-manifold edges and holes are detected and repaired to ensure that the model meets the watertightness requirement.

[0096] S440, Additive Manufacturing Data Processing:

[0097] The model was discretized into STL format, the optimal build orientation was calculated, and auxiliary supports were generated, taking into account the anisotropic shrinkage characteristics of photosensitive resin. The strategy for generating auxiliary supports strictly followed the principle of tissue surface protection: the system automatically identified the inner surface of the surgical guide (i.e., the base adaptation surface in contact with teeth and alveolar bone) as a no-go zone, forcing the 3D printing build orientation to be set with the inner surface facing upwards (away from the printing platform); all auxiliary support structures were only allowed to be generated on the outer surface of the guide and the outer wall of the guide cylinder, and the contact diameter of the support points was set to a pinhead contact of 0.3mm to 0.5mm; this strategy ensured that the inner surface of the guide did not need to be mechanically polished after printing, avoiding artificial changes to the positioning accuracy of the guide due to the removal of support residues; a compensation scaling matrix was constructed to pre-deform the model, and finally output to the 3D printing equipment 300 for manufacturing; the compensation scaling matrix was obtained by pre-printing a standard cube calibration block with a side length of 20mm, and measuring its actual dimensions in the X, Y, and Z printing directions using a micrometer. Calculate the scaling factor for each axis. Then construct a diagonal matrix. When generating guide plate data, multiply the coordinates of all vertices of the guide plate mesh by this matrix. This counteracts the volume shrinkage during resin curing, ensuring that the average gap error between the inner surface of the guide plate and the tooth surface is controlled within ±10μm.

[0098] Specific application examples:

[0099] This embodiment provides a method for digital design and fabrication of autologous tooth transplants. This method is executed using a digital design and fabrication system for autologous tooth transplants. This system includes a data acquisition and reconstruction module 10, a virtual registration module 20, a transplantation plan optimization module 30, and a guide plate generation module 40, and is connected to a CBCT imaging device 200 and a 3D printing device 300. The detailed execution steps are as follows:

[0100] S100: High-precision data acquisition and reconstruction

[0101] First, the data acquisition and reconstruction module 10 receives patient oral cavity tomographic scan data from the CBCT imaging device 200.

[0102] Image data acquisition parameters and precision control (S111-S114):

[0103] The operator performs a cone-beam computed tomography (CBCT) scan of the oral and maxillofacial region and sets the voxel size. Within the range of 0.125mm to 0.2mm, the geometric resolution is ensured to meet the recognition requirements of the average physiological thickness of the periodontal ligament (approximately 0.25mm). The system automatically verifies DICOM data and uses a normalized metal artifact correction algorithm to eliminate artifacts, ensuring the grayscale accuracy of subsequent finite element material assignments.

[0104] Image segmentation and region extraction (S121-S125):

[0105] The system reads DICOM data, defines the region of interest, and uses a 3D convolutional neural network based on the U-Net architecture for semantic segmentation enhancement, combined with grayscale threshold ranges. To identify high-density hard tissue and address root adhesion issues, a region growing algorithm is applied to establish seed points at local maxima in the distance transformation map, separating donor tooth monomers. Subsequently, the mask is corrected through morphological opening and closing operations (structural element radius 0.3-0.5mm).

[0106] 3D mesh reconstruction (S130-S140):

[0107] The voxels were converted into triangular meshes using the moving cube algorithm, and the volume-preserving HC-Laplacian smoothing algorithm was applied to generate digital surface models of the donor tooth, the implantation area, and the opposing dentition.

[0108] S200: Virtual Registration and Spatial Alignment

[0109] A unified spatial coordinate system is established using the virtual registration module 20.

[0110] Coordinate system establishment (S210):

[0111] Calculate the covariance matrix of the donor tooth mesh. The principal axis direction (Z-axis points to the crown, Y-axis points to the lingual side) is determined by eigenvalue decomposition to achieve model normalization.

[0112] Registration operation (S220-S230):

[0113] An initial rigid body transformation is performed based on anatomical feature points, followed by an iterative nearest-point algorithm to calculate the Euclidean distance between the donor tooth root surface and the recipient bone cavity surface. A normal vector consistency constraint (threshold) is then introduced. Eliminate false matches, minimize mean square error, and determine the optimal geometric matching position.

[0114] S300: Transplantation Protocol Optimization and Biomechanical Analysis

[0115] Core morphological and mechanical simulations are performed using the transplantation scheme optimization module 30.

[0116] Root bone morphology adaptation and posture adjustment (S310-S320):

[0117] Using bounding box trees to accelerate the calculation of the signed distance field, quantifying areas of insufficient bone or obstruction, and constructing a multi-objective pose optimization function. :

[0118] ;

[0119] Solved using nonlinear least squares method, with priority given to weights. (0.6-0.8) is used to avoid occlusal trauma and to set the target adjoint distance. It is 1.5-2.0mm.

[0120] Finite element analysis and stress distribution optimization (S330-S340):

[0121] The target location is divided into tetrahedral meshes (Tet10 elements), and the elastic modulus of bone tissue is dynamically mapped based on the grayscale values. (in Simulated chewing loads (100-150N vertically, 50-80N laterally) are applied, and von Mises stress is calculated.

[0122] In this step, a stress distribution curve is generated based on the biomechanical simulation results, such as... Figure 3 As shown, Figure 3 The horizontal axis represents the root depth from the neck (0 mm) to the apex (12 mm), and the vertical axis represents the von Mises stress.

[0123] The black dashed line in the figure represents the original solution, showing that the stress value in the cervical region (0-2mm) is as high as about 60MPa, which far exceeds the safe threshold for bone resorption (35MPa, as shown by the dotted line in the figure), indicating an extremely high risk of alveolar bone resorption.

[0124] The dark gray solid line in the figure represents the optimized solution after adjustment by step S340 of the present invention. The maximum stress value is reduced to about 28 MPa, and the overall curve is flat and completely below the safe threshold of 35 MPa. This indicates that by adjusting the implantation angle and depth through the algorithm, the occlusal stress is effectively dispersed.

[0125] If the preliminary analysis results do not meet the requirements, the system calculates the stress-weighted centroid. :

[0126] ;

[0127] Using cross product Determine the rotation axis and automatically fine-tune the donor tooth position until all nodal stress values ​​return to normal. Figure 1 Below the indicated safety thresholds (cortical bone <35MPa, cancellous bone <5MPa).

[0128] S400: Surgical Guide Design and Manufacturing

[0129] The surgical guide geometry is constructed using the guide generation module 40.

[0130] Substrate design and guiding structure (S410-S420):

[0131] Extract the base fitting surface and set a buffer gap of 30-50μm. Generate a hollow guide sleeve along the long axis of the donor tooth.

[0132] In this step, a fluid dynamics structural design is performed for the cooling channel: a cubic Bézier curve path is constructed connecting the center of the water inlet and the center of the cooling outlet to form a smooth streamlined trajectory to reduce turbulence; the inner diameter of the channel is set to gradually narrow from the inlet end (diameter 2.0-3.0mm) to the outlet end (diameter 1.0-1.5mm) along this path, utilizing the Venturi effect to create a high-velocity jet of water at the outlet; the axial direction of the cooling outlet center is set to form an angle of 15-25 degrees with the long axis of the drill bit, so that the jet of water is precisely directed to the core heat-generating area of ​​bone cutting below the drill bit tip.

[0133] Multi-component integration and manufacturing (S430-S440):

[0134] Using distance field implicit modeling techniques, Boolean operations are used to fuse the base, sleeve, and connecting rod (section 4.0-6.0mm), and a transition chamfer is generated. Finally, the model is discretized into STL data, and a compensation scaling matrix is ​​constructed. To counteract resin shrinkage, the resin is transferred to 3D printing equipment 300 for manufacturing.

[0135] Implementation effect verification:

[0136] The distribution of root bone compatibility error in autologous tooth transplantation surgical planning and guide fabrication using the method described in this embodiment is as follows: Figure 4 As shown, Figure 4 The horizontal axis represents the root-bone gap error (mm), and the vertical axis represents the sample distribution probability. The gray dashed line in the figure corresponds to the traditional manual operation group, which has a low and wide peak, with the peak value mainly distributed around 0.8 mm, indicating that the traditional method has a large error and high dispersion. The black solid line in the figure corresponds to the method of this invention, which shows a high and narrow peak, with the peak center closely surrounding the ideal physiological gap of 0.25 mm. This proves that the method of this invention can control the root-bone gap in most cases within the optimal range (0.2-0.3 mm) that is conducive to periodontal ligament healing, which is superior to the traditional technique.

Claims

1. A method for digital design and fabrication of autologous tooth transplants, characterized in that, Includes the following steps: The system receives oral cavity tomographic scan data from patients, performs image segmentation processing, and constructs three-dimensional models of the donor tooth, the alveolar bone of the implantation area, and the opposing dentition, respectively. A unified spatial coordinate system is established, rigid body transformation and iterative nearest point calculation are performed to determine the geometric matching position of the donor tooth three-dimensional model in the alveolar bone three-dimensional model of the implantation area. The interference volume between the three-dimensional model of the donor tooth and the three-dimensional model of the alveolar bone in the implantation area is calculated. The axial angle and implantation depth of the three-dimensional model of the donor tooth are adjusted according to the three-dimensional model of the opposing dentition. A simulated occlusal load is applied to the adjusted proposed implantation position for finite element analysis. The interface stress distribution is calculated and the implantation pose parameters are finely adjusted according to the interface stress distribution to determine the final implantation pose. The three-dimensional model of the alveolar bone in the implantation area and the surface features of adjacent teeth are extracted as the base adapter surface. Based on the final implantation pose, a hollow guide sleeve structure is generated along the long axis of the donor tooth three-dimensional model, and a connecting support structure and a cooling channel structure are generated to connect the base adapter surface and the hollow guide sleeve structure. The three-dimensional model of the surgical guide is generated by Boolean operation and then 3D printed to obtain the surgical guide entity.

2. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of constructing the three-dimensional model of the donor tooth includes: Perform grayscale thresholding on the voxels within the region of interest defined by the oral cavity tomographic scan data to obtain an initial hard tissue mask; Perform Euclidean distance transformation on the initial hard tissue mask to generate a distance transformation map, and calculate the distance value from the voxel inside the mask to the nearest background voxel; The local maximum point with the largest distance value in the distance transformation graph is selected as the seed point; Starting from the seed point, a region growing algorithm is executed to separate the independent three-dimensional model of the donor tooth based on the connectivity criterion and gray-level difference constraint.

3. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of establishing a unified spatial coordinate system includes: Calculate the covariance matrix of the vertices of the 3D model mesh of the donor tooth, and perform eigenvalue decomposition on the covariance matrix to obtain eigenvectors; The eigenvector corresponding to the largest eigenvalue is defined as the Z-axis, and the eigenvector corresponding to the smallest eigenvalue is defined as the Y-axis. The donor tooth 3D model is divided into multiple cross sections along the Z-axis, and the area of ​​each cross section is calculated. Compare the average cross-sectional areas of the positive and negative ends of the Z-axis. If the average cross-sectional area of ​​the positive end is smaller than that of the negative end, then perform a flip operation on the Z-axis feature vector so that the Z-axis points in the direction of the tooth crown.

4. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of performing iterative nearest point calculation includes: The root surface point set of the donor tooth three-dimensional model is extracted as the source point set; The point set of the inner surface of the alveolar bone of the implantation area three-dimensional model is extracted as the target point set; During the iterative calculation, the angle between the normal vector of the source point set and the normal vector of the corresponding point in the target point set is calculated; Remove matching point pairs whose included angle is greater than a preset angle threshold; The transformation matrix is ​​calculated based on the retained matching point pairs until the mean square error converges.

5. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of applying simulated occlusal load to the adjusted proposed implantation location and performing finite element analysis involves finite element mesh generation, including: Extract the geometric position data of the adjacent teeth contained in the three-dimensional model of the donor tooth, the three-dimensional model of the alveolar bone in the implantation area, and the three-dimensional model of the opposing dentition. Calculate a multi-objective pose optimization function that includes occlusal interference depth, adjacency distance, and root deviation. The multi-objective pose optimization function is solved using the nonlinear least squares method to obtain the rigid body transformation increment; The axial angle and implantation depth of the donor tooth three-dimensional model are adjusted according to the rigid body transformation increment.

6. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of generating a finite element mesh for the proposed implantation location includes: The three-dimensional model of the donor tooth and the three-dimensional model of the alveolar bone in the implantation area are converted into tetrahedral mesh elements; Identify the gap region between the root of the donor tooth in the three-dimensional model and the bone wall of the alveolar bone in the implantation area, and generate a simulated periodontal ligament layer unit. For the grid cells of the bone tissue region, a mapping relationship between gray values ​​and elastic modulus is established. Based on the voxel gray values ​​of the oral cavity tomographic scan data, the elastic modulus of each cell is calculated using a power function formula. The simulated periodontal ligament unit is endowed with homogeneous material properties.

7. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of fine-tuning the implantation pose parameters according to the stress distribution further includes performing closed-loop optimization: Determine whether the calculated maximum von Mises stress value exceeds the preset bone resorption threshold; If the bone resorption threshold is exceeded, calculate the stress-weighted centroid of the nodes in the high-stress region; Calculate the reverse vector from the stress-weighted centroid to the geometric center of the donor tooth three-dimensional model; Calculate the product of the reverse vector and the current longitudinal axis of the donor tooth 3D model to determine the rotation axis; Based on the rotation axis and stress over-limit ratio, a rotation correction amount is generated, the pose of the donor tooth three-dimensional model is adjusted, and a new round of finite element calculation is triggered.

8. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of generating the cooling channel structure includes: Construct a cubic Bézier curve path connecting the center of the water inlet and the center of the cooling outlet; The inner diameter of the cooling channel structure gradually narrows from the inlet end to the outlet end along the path. The axial direction of the cooling outlet center is set to form a preset angle with the central axis of the hollow guide sleeve structure, so that the axis of the cooling outlet center points to a preset cutting area located in the extension direction of the central axis.

9. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, The step of generating a three-dimensional model of the surgical guide through Boolean operations includes: The base adapter surface, the hollow guide sleeve structure, the connecting support structure, and the cooling channel structure are respectively converted into a signed distance field; Perform a union operation on each of the signed distance fields in voxel space; A smooth minimum function is applied to generate a transition chamfer at the connection point; The fused distance field is reconstructed into a triangular mesh using a polyhedron extraction algorithm.

10. The method for digital design and fabrication of autologous tooth transplants according to claim 1, characterized in that, Prior to the step of 3D printing to obtain the surgical guide plate entity, the procedure also includes: The ratio of the actual size to the design size of the standard calibration block in three printing directions is used to obtain the scaling factor for each axis. Construct a diagonal matrix containing the scaling factor; Multiply the coordinates of all vertices of the 3D model of the surgical guide by the diagonal matrix to generate pre-deformation compensated manufacturing data, and use the manufacturing data to perform the 3D printing manufacturing.

Images