An automated method for generating dental crown preparations using four-quadrant zoning control and dynamic skeleton guidance
By using a four-quadrant partitioning control and dynamic skeleton guidance method, a dental crown preparation model is automatically generated, which solves the problems of reliance on manual interaction and insufficient accuracy in existing technologies, and achieves efficient and accurate generation of dental crown preparations.
Patent Information
- Application Number
- CN202611627206.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-21
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies struggle to automatically generate clinically compliant preparatory models based on 3D scanning models of dental crowns, and rely on manual interaction with insufficient precision.
Using a four-quadrant partitioning control and dynamic skeleton guidance method, the cervical and non-cervical margin vertices are determined by analyzing the three-dimensional mesh model of the crown, anatomical skeleton paths are constructed, a polar coordinate system and a two-dimensional weight field are established, the amount of tooth removal is decomposed and iterative deformation is performed, and combined with Laplacian smoothing, a preparative model that meets clinical requirements is generated.
It enables automatic and adaptive generation of crown preparation models, improving the efficiency and accuracy of digital tooth preparation, reducing reliance on the operator's professional experience, ensuring no misalignment of the cervical margin, preventing mesh tearing and wrinkling, and enhancing the fine control of the amount of tooth removal.
Smart Images

Figure CN123118977A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital oral medicine technology, specifically to an automatic method for generating dental crown preparations using four-quadrant zoning control and dynamic skeleton guidance. Background Technology
[0002] In the field of fixed dental restorations, tooth preparation is a core operational step before fabricating restorations such as crowns, veneers, and inlays. Its goal is to precisely remove an appropriate amount of tooth structure from the affected tooth to create a prepared body that meets specific geometric requirements. Typically, this requires: approximately 1.5–2.0 mm reduction at the incisal edge, approximately 1.2–1.5 mm at the labial / buccal surfaces, approximately 0.8–1.0 mm at the lingual surface, and approximately 1.0–1.2 mm at the mesial / distal surfaces, while simultaneously creating a clear shoulder at the cervical margin, with an axial convergence typically between 2° and 8°.
[0003] With the development of digital technology in oral medicine, computer-aided tooth preparation design based on digital models of triangular mesh surfaces reconstructed from three-dimensional scanning of the crown has gradually become a research hotspot. However, existing CAD software (such as 3Shape, Cerec, etc.) still heavily relies on manual interaction in generating preparation models: operators need to manually annotate the cervical margin and specify the amount of grinding on each surface, which is cumbersome and has a high learning threshold; moreover, the amount of grinding is usually applied in a globally uniform offset manner, which fails to adaptively adjust according to the geometric characteristics of different anatomical surfaces of the tooth, which easily produces wrinkles in the cervical margin shoulder area, and the accuracy of axial convergence control is limited.
[0004] Existing methods for generating digital dental preparations in the field of oral medicine can be broadly categorized into the following technical approaches: First, the parametric sweep reconstruction method, which semi-automatically extracts the cervical margin, selects several cross-sections along the insertion direction, calculates feature points based on preset parameters such as shoulder width and shoulder angle, and sweeps them into a surface. This method relies on manual annotation, and when based on discrete cross-section interpolation, it is difficult to achieve continuous and precise anatomical adaptive control. Second, the statistical shape model matching method, which uses statistical modeling from clinical databases to infer the shape of the missing tooth crown from the morphology of adjacent teeth. However, the research objective of this method is mainly the generation of restorations, rather than the grinding and deformation of the preparation. Third, the robotic laser hardware processing method, which focuses on driving a robotic arm to control laser layer-by-layer ablation. The shape of the preparation depends on pre-designed CAD data input. The research focus of this method is on the accuracy of processing execution, rather than the automatic generation algorithm of the 3D model of the preparation.
[0005] Therefore, the aforementioned existing technologies all have corresponding limitations: cross-sectional scanning methods rely on manual annotation of the neck margin line and lack sufficient precision; statistical shape model methods are geared towards the generation of restorations and require the establishment and training of a large-scale database; hardware fabrication methods fail to provide an automatic generation algorithm for anatomically adaptive 3D models of prepared restorations.
[0006] Therefore, designing a method that can automatically and adaptively generate a prepared model that conforms to clinical standards from the STL model of the crown to be prepared is of great significance for improving the efficiency and accuracy of digital tooth preparation and promoting the standardization and intelligentization of related clinical workflows. Summary of the Invention
[0007] This invention provides an automatic generation method for crown preparations using four-quadrant zoning control and dynamic skeleton guidance. This method addresses the problem in existing technologies where it is difficult to automatically generate clinically compliant preparation models based on intraoral scan crown models. The goal is to achieve automatic and adaptive generation of crown preparation models, thereby improving the efficiency and accuracy of digital tooth preparation.
[0008] This invention is achieved through the following technical solution:
[0009] A method for automatically generating dental crown preparations using four-quadrant zoning control and dynamic skeleton guidance includes the following steps:
[0010] Step S1: Obtain the three-dimensional mesh model of the crown to be prepared, analyze the vertices and faces, and configure the grinding parameters for each face;
[0011] Step S2: Determine the neck edge vertices and non-neck edge vertices; keep the coordinate positions of each neck edge vertex fixed;
[0012] Step S3: Construct the anatomical framework path along the long axis of the tooth crown;
[0013] Step S4: Establish a polar coordinate system, divide the four quadrants based on the lip surface, tongue surface, mesial surface and distal surface, and construct a two-dimensional weight field;
[0014] Step S5: Decompose the target amount of grinding into several small steps, apply a mixed displacement field of the fusion axial normal grinding component, the axial compression component of the cut end and the convergence component to each non-neck edge vertex, and perform several iterations to obtain the preliminary deformation preparation model.
[0015] Step S6: Perform Laplace physical smoothing on the preliminary deformation preparation model;
[0016] Step S7: Perform radial tapered polymerization finishing on the axial surface area; perform directional smoothing on the lip surface area;
[0017] Step S8: Construct triangulations for the final mesh point set and patch connection table, and export the preparatory volume 3D mesh model file.
[0018] To address the problem in existing technologies of the difficulty in automatically generating clinically compliant crown preparation models based on intraoral scan models of the crowns to be prepared, this invention proposes an automatic crown preparation generation method based on four-quadrant zoning control and dynamic skeleton guidance. This method first obtains a three-dimensional mesh model of the crown to be prepared using any existing method, and analyzes the vertices and faces in the three-dimensional mesh model. Simultaneously, it configures the grinding parameters for each face according to specific working conditions (e.g., different cases, restorative material requirements, etc.). Then, based on the three-dimensional mesh model of the crown, the cervical margin vertices and non-cervical margin vertices are determined; and the coordinate positions of each cervical margin vertex are kept fixed throughout all subsequent calculations, serving as a positional invariant constraint for the entire process. Next, an anatomical skeleton path is constructed along the long axis of the crown; those skilled in the art should understand that the anatomical skeleton path represents the long axis of tooth anatomy and serves as a spatial reference for subsequent aggregation deformation and height normalization. Because the required amount of tooth removal varies across different anatomical surfaces, and the cervical margin requires positional locking while the incisal edge needs sufficient free deformation, this method also establishes a polar coordinate system and constructs a two-dimensional weighted field based on the labial, lingual, mesial, and distal surfaces, dividing the space into four quadrants. This two-dimensional weighted field is continuously differentiable, enabling precise control over the displacement of each vertex. The target amount of tooth removal is then decomposed into several small steps. While maintaining the fixed coordinate positions of each cervical margin vertex, a mixed displacement field is applied to each non-cervical margin vertex, incorporating the axial normal tooth removal component, the incisal edge axial compression component, and the convergence component. This process is iterated several times (the number of iterations is the same as the decomposition steps) to prevent the mesh from self-intersecting or tearing under large displacements, thus obtaining a preliminary deformable preparation model. The preliminary deformable preparation model is then subjected to Laplace physical smoothing, followed by radial conical convergence refinement of the axial region and directional smoothing of the labial region to obtain the crown preparation model. Finally, triangulation is constructed from the mesh point set and facet connection table in the crown preparation model, and the three-dimensional mesh model file of the preparation is exported.
[0019] This application uses a 3D mesh model of the crown to be prepared as input, eliminating the need for manual annotation of the cervical margin or interactive adjustments. A single run completes the entire conversion from the original crown to the prepared model, significantly reducing reliance on the operator's professional experience. This application ensures that the coordinates of the cervical margin vertices remain unchanged throughout the entire process, resulting in a perfect, unbiased fit between the final cervical margin line and the original intraoral scan model. This application achieves automatic and adaptive generation of the crown preparation model, significantly improving the efficiency and accuracy of digital tooth preparation.
[0020] Furthermore, in step S1, the grinding parameters for each surface include any one or more of the following: cutting edge grinding amount, lip surface grinding amount, tongue surface grinding amount, mesial surface grinding amount, distal surface grinding amount, axial surface convergence degree, and neck edge shoulder buffer width.
[0021] The parameters mentioned above are all professional terms in the field of dental crown preparation and are not difficult for those skilled in the art to understand. In practice, these parameters can be flexibly adjusted based on the requirements of different cases and restorative materials.
[0022] Furthermore, in step S2, the methods for determining the neck edge vertex and non-neck edge vertices include:
[0023] Step S201: Traverse all triangular faces in the three-dimensional mesh model of the crown and define the edge referenced by only one triangular face as a free edge;
[0024] Step S202: Take all the vertices of the free edges as cervical vertices; take the remaining vertices in the three-dimensional mesh model of the crown as non-cervical vertices.
[0025] The 3D mesh model of a dental crown is generally an STL model obtained from an oral cavity scan. It has an open boundary at the cervical margin, meaning free edges exist. This method utilizes this feature to accurately locate the cervical margin vertices, thus obtaining the set of cervical margin vertices, which is used to maintain their coordinates unchanged in all subsequent iterative calculations. Vertices in the 3D mesh model that do not belong to the cervical margin vertex set are defined as non-cervical margin vertices.
[0026] Furthermore, step S3 specifically includes:
[0027] Step S301: Divide the three-dimensional mesh model of the tooth crown into layers along the longitudinal direction, so that the number of layers is within the set range;
[0028] Step S302: Obtain the centroid of each layer;
[0029] Step S303: Apply a low-pass filter to the planar coordinates of the centroids of each layer;
[0030] Step S304: Connect the centroid coordinates of each layer in sequence along the longitudinal direction to obtain the anatomical skeleton path.
[0031] This scheme first dynamically calculates the number of layers based on the total crown height and the preset target layer thickness, and limits the number of layers within a set range to ensure the algorithm's adaptability to crown models of different sizes. Then, it calculates the centroid coordinates of each layer, applies a low-pass filter to the planar coordinates of each layer's centroid, and connects the centroid coordinates of each layer sequentially along the longitudinal direction to obtain the anatomical skeleton path. The low-pass filtering can utilize existing techniques, and the filter window width is positively correlated with the number of layers.
[0032] Furthermore, step S302 specifically includes:
[0033] Step S3021: Calculate the center height of each floor;
[0034] Step S3022: Based on the center height, set a buffer zone between floors within a specified height range for each floor;
[0035] Step S3023: Extract all vertices within the interlayer buffer and calculate the mean of the planar coordinates of all vertices to obtain the centroid plane coordinates;
[0036] Step S3024: Combine the center height of each layer with the centroid plane coordinates to obtain the centroid coordinates of each layer;
[0037] Step S3025: Extract the layer located in the specified area at the top of the crown as the top layer, perform a weighted calculation on the centroid plane coordinates of each top layer and the reference centroid plane coordinates, and correct the centroid coordinates of the top layer.
[0038] Wherein, the reference centroid plane coordinates are the centroid plane coordinates of the preset layer, the height of the preset layer is greater than 50% of the total crown height; and the weight of the centroid plane coordinates of the top layer is less than the weight of the reference centroid plane coordinates.
[0039] This scheme sets up an inter-layer buffer to ensure that a sufficient number of vertices are collected when selecting the centroid of a layer, avoiding situations where some layers fail to collect enough vertices due to uneven vertex distribution. The average X and Y coordinates of all vertices within the inter-layer buffer are calculated; the resulting average X and Y values are the X and Y coordinates of the centroid of that layer. The Z coordinate of the centroid is then the Z coordinate value at the center height of that layer.
[0040] This scheme also sets up a top layer and introduces a reference centroid plane. Based on the reference centroid plane, the centroids within the top layer are weighted and corrected, which can effectively solve the problem of lateral drift of the skeleton caused by sparse point clouds at the cusps, and thus effectively prevent the skeleton at the incisal edge from collapsing laterally. The "specified area at the top of the crown," "preset layer," and the "weights of the centroid plane coordinates of the top layer" and "weights of the reference centroid plane coordinates" can all be adaptively set within a limited range according to different tooth positions and crown model sizes, etc., and are not specifically limited here. Of course, the sum of the "weights of the centroid plane coordinates of the top layer" and the "weights of the reference centroid plane coordinates" is 1.
[0041] Furthermore, step S4 specifically includes:
[0042] Step S401: Establish a planar polar coordinate system using the average coordinates of all neck edge vertices as the origin of the polar coordinates;
[0043] Step S402: Calculate the polar angle of each vertex in the three-dimensional mesh model of the tooth crown, and divide the plane into four quadrants based on the four basic directions of labial, lingual, mesial, and distal surfaces to obtain the transverse weight field; in the four quadrants of the plane, such that:
[0044] Polar angle = 90°, facing the lip surface, i.e., directly opposite the center of the lip surface; polar angle = -90°, facing the tongue surface, i.e., directly opposite the center of the tongue surface; polar angle = 0°, facing the mesial surface, i.e., directly opposite the center of the mesial surface; polar angle = ±180°, facing the distal surface, i.e., directly opposite the center of the distal surface.
[0045] Step S403: For each vertex in the three-dimensional mesh model of the crown, calculate the relative height with the nearest cervical margin vertex, and divide the three-dimensional mesh model of the crown into the locked cervical margin area, the transition area, and the free deformation area.
[0046] Step S404: Establish a longitudinal weight field: Set the longitudinal weight of the locked neck edge region to 0, the longitudinal weight of the free deformation region to 1, and the longitudinal weight of the transition region to be obtained by interpolation.
[0047] Since different anatomical surfaces have different requirements for the amount of material removed, such as the neck margin requiring position locking while the cut edge requires sufficient free deformation, this application constructs a continuously differentiable two-dimensional weight field to achieve precise control over the amount of material removed at each vertex.
[0048] The two-dimensional weight field constructed in this scheme includes a horizontal weight field and a vertical weight field.
[0049] in:
[0050] The lateral direction refers to the angular direction. The lateral weight field is constructed based on a planar four-quadrant system to distinguish the four anatomical directions: the labial surface, the lingual surface, the mesial surface, and the distal surface, allowing the amount of material removed in different anatomical directions to be configured independently. These four quadrants exist side by side in the horizontal cross-section.
[0051] The longitudinal direction refers to the height direction, which is another dimension independent of the horizontal direction. It describes the question of "whether each vertex can move and how much it can move" in the Z-axis direction from the neck edge line to the cut end: the top of the locked neck edge area is locked in the longitudinal direction, the vertices of the transition area are smoothly connected, and the vertices of the free deformation area can be fully deformed.
[0052] By superimposing the horizontal and vertical dimensions mentioned above, a complete four-quadrant two-dimensional weighted field is formed. Its physical meaning is that for any vertex (grid point), there is both the amount of wear determined by the angle (horizontal) and the degree of freedom of deformation determined by the height (vertical). Only by combining the two can the final actual displacement of the vertex be obtained.
[0053] The longitudinal weight of the transition zone can be obtained by any existing interpolation method that can be achieved by those skilled in the art, and no specific limitation is made here, as long as the morphology between the neck edge zone and the grinding zone is continuously and smoothly transitioned.
[0054] Furthermore, in step S5, the mixed displacement field is obtained by the following method:
[0055] Step S501: Based on the weight values of the direction corresponding to the target vertex in the transverse weight field, and combined with the polar angle θ of the target vertex, determine the axial normal grinding component of the target vertex:
[0056] 0°≤θ<90°, axial normal grinding component = cos²θ×near mid-surface grinding amount + sin²θ×lip surface grinding amount;
[0057] 90°≤θ<180°, axial normal grinding component = cos²θ×distal surface grinding amount + sin²θ×lip surface grinding amount;
[0058] -90°≤θ<0°, axial normal grinding component = cos²θ×near mid-surface grinding amount + sin²θ×tongue surface grinding amount;
[0059] -180°≤θ<-90°, axial normal grinding component = cos²θ×distal surface grinding amount + sin²θ×tongue surface grinding amount;
[0060] Step S502: Take the opposite direction of the horizontal projection of the normal vector of the target vertex as the grinding direction of the axial normal grinding component;
[0061] Step S503: Calculate the normalized height of the target vertex;
[0062] Step S504: Using the nonlinear power function of the normalized height as the cutting edge weight, determine the axial compression component of the cutting edge of the target vertex: Axial compression component of cutting edge = cutting edge wear amount × cutting edge weight.
[0063] Step S505: Based on the two-dimensional weighted field, the axial normal grinding component and the axial compression component of the cut end are fused to obtain a mixed displacement field;
[0064] Step S506: For the vertices of the locked cervical margin region and the transition region, the displacement vector obtained through the hybrid displacement field is projected onto the radial direction, and the projection result is used as the amount of material removed from the vertices of the locked cervical margin region and the transition region; wherein, the radial direction is a horizontal direction perpendicular to the long axis of the tooth, pointing from the target vertex or away from the long axis;
[0065] Step S507: After each iteration, for the target vertex within a preset range from the cutting end, shrink it inward by a preset amount in the opposite direction of its own normal.
[0066] Step S508: Apply a light Laplacian smoothing to the non-neck edge vertices.
[0067] This application employs a multi-iteration micro-step cumulative deformation strategy, decomposing the target grinding amount into multiple small steps to avoid self-intersection or tearing of the mesh under large displacement. In each iteration, steps S501 to S508 are executed sequentially for each non-neck vertex.
[0068] The axial normal grinding component is the amount of material removed from the target vertex inward on the horizontal cross section. The axial compression component of the cut end applied in each iteration = the set amount of material removed from the cut end × the cut end weight ÷ the total number of iterations.
[0069] This scheme linearly fuses the axial normal grinding component and the axial compression component of the cutting edge at a height to obtain a hybrid displacement field. During the research process, the inventors discovered that for the vertices of the crown shoulder region (i.e., the locked cervical margin region and transition region), the above fusion may contain components along the tangential direction of the model surface. This tangential displacement can cause relative sliding between adjacent vertices, resulting in mesh stacking and wrinkling in the shoulder region. Therefore, for the vertices in this type of region, an anti-wrinkling mechanism with radial constraints on the shoulder region is applied: the displacement vector is projected into the radial direction, retaining only the inward contraction component, while eliminating all tangential components. This ensures that the movement near the cervical margin is strictly radially centripetal, preventing tangential misalignment.
[0070] In addition, after each iteration, this method shrinks the target vertex within a preset range from the cut end inward by a preset amount in the opposite direction of its own normal, ensuring that a gradually narrowing conical profile is formed from the neck edge to the cut end, and ensuring that the cut end repair gap meets clinical requirements.
[0071] This scheme also applies a light Laplacian smoothing to all non-neck vertices at the end of any iteration cycle. This is used to suppress localized minor fluctuations and high-frequency noise accumulated over multiple iterations. Simultaneously, the light Laplacian smoothing effectively preserves the displacement iteration results of the current cycle without disrupting the applied grinding effect. The smoothing intensity of the light Laplacian smoothing in this scheme can be adaptively set according to specific working conditions, preferably lower than the smoothing intensity of the physical Laplacian smoothing treatment in step S6.
[0072] Furthermore, step S6 specifically includes:
[0073] Step S601: Based on the longitudinal weight field, the preliminary deformation preparation model is divided into: the neck edge sensitive area, the cut end and the axial free area;
[0074] Step S602: Perform Laplace physical smoothing:
[0075] For the sensitive area at the neck margin, a first relaxation factor is configured; for the cut end and the free area on the axial surface, a second relaxation factor is configured; and the first relaxation factor is greater than the second relaxation factor.
[0076] The previous steps applied radial constraints to the shoulder area of the neck margin. Although this prevented tangential misalignment, the cumulative radial compression from multiple rounds could still leave small local wrinkles and unevenness in the shoulder area. Therefore, this solution eliminates these by using Laplacian physical smoothing. The core idea is to use different smoothing intensities for different areas to precisely eliminate wrinkles while preserving as much of the anatomical grinding morphology generated in the previous steps as possible. The specific partitioning in this solution is based on the longitudinal weight field; that is, areas with longitudinal coordinates below a preset threshold are considered the neck margin sensitive area; the remaining areas are the cut edge and axial free area.
[0077] This approach uses a relatively higher relaxation factor for the sensitive cervical margin area, resulting in greater smoothing force to eliminate wrinkles in the shoulder area; for the incisal edge and axial free area, a relatively lower relaxation factor is used, resulting in less smoothing force, which is more conducive to preserving the amount of material removed and the surface morphology obtained in the previous steps. In other words, this approach focuses on restoring the crown shoulder in the post-processing stage, effectively ensuring the geometric quality of the shoulder area.
[0078] Furthermore, in step S7, the method for radial tapered polymerization finishing of the axial surface region includes:
[0079] Step S701: Using the line connecting the start and end points of the anatomical skeleton path as the long axis of the crown, calculate the radial distance from each vertex to the long axis of the crown;
[0080] Step S702: Based on the radial distance, calculate the refinement target position corresponding to each vertex:
[0081] r_target = r_current×(1-shrink_factor×z_norm);
[0082] In the formula: r_target is the target position for refinement; r_current is the radial distance from the vertex to the long axis of the crown; shrink_factor is the shrinkage ratio coefficient; z_norm represents the normalized height of the vertex;
[0083] Step S703: Move each vertex toward its corresponding refinement target position along the radial direction, and the moving distance is less than the distance between the vertex and the refinement target position.
[0084] During the iterative process of obtaining the preliminary deformation preparation model, due to the multi-round small-step superposition, there is a certain cumulative error, which may cause the actual convergence angle to deviate slightly from the set value. In order to overcome this problem, this scheme performs deterministic radial shrinkage correction on each vertex to eliminate systematic errors and local surface defects, ensuring that the axial convergence degree is highly consistent with the set value, and providing a reliable guarantee for the placement of the repair body.
[0085] Furthermore, in step S7, the method for directional smoothing of the lip surface region includes:
[0086] Step S704: Calculate the dot product of the normal vector of each vertex and the lip reference direction vector. If the dot product is greater than a preset threshold and the current vertex meets the preset height condition, then the corresponding vertex is identified as a lip vertex.
[0087] Step S705: For each vertex of the lip surface, calculate the difference vector between its neighborhood mean position and its own position;
[0088] Step S706: Extract the projection component of the difference vector onto the normal of the lip vertex, and use it as the actual movement of the lip vertex, so that the lip vertex contracts inward along the normal.
[0089] Step S707: In the transition area between the lip surface and the rest of the area, perform omnidirectional smoothing with a preset intensity.
[0090] During the iterative process of obtaining the preliminary deformable model, due to the multiple rounds of small-step stacking, localized minor bulges may appear in the upper and middle segments of the labial surface, i.e., local areas protrude outwards, making the crown slightly thicker in that area. To overcome the above problem, this scheme applies normal-guided directional smoothing to the upper and middle segments of the labial surface to specifically eliminate these bulges without disrupting the overall anatomical curvature of the labial surface.
[0091] Compared with the prior art, the present invention has at least the following advantages and beneficial effects:
[0092] 1. The present invention provides an automatic generation method for tooth crown preparation body with four-quadrant partition control and dynamic skeleton guidance. It uses the three-dimensional mesh model of the tooth crown to be prepared as input, and does not require manual annotation of the cervical margin or interactive adjustment. The complete conversion from the original tooth crown to the preparation body can be completed in a single run, which significantly reduces the dependence on the operator's professional experience and realizes the automatic and adaptive generation of tooth crown preparation body model, which significantly improves the efficiency and accuracy of digital tooth preparation.
[0093] 2. This invention can ensure that the coordinates of the neck edge vertex remain unchanged throughout the entire process, so that the final neck edge line completely matches the original oral scan model without any offset.
[0094] 3. The present invention sets up a top layer and introduces a reference centroid plane. Based on the reference centroid plane, the centroid in the top layer is weighted and corrected, which can effectively solve the problem of lateral drift of the skeleton caused by sparse point cloud at the tooth cusp, and thus effectively prevent the skeleton at the incisal edge from collapsing laterally.
[0095] 4. This invention constructs a continuously differentiable two-dimensional weight field to achieve precise control over the amount of material removed from each vertex.
[0096] 5. The present invention adopts a multi-iteration micro-step cumulative deformation strategy to decompose the target grinding amount into multiple small steps, thereby avoiding self-intersection or tearing of the mesh under large displacement.
[0097] 6. The present invention applies a radial constraint to the apex of the crown shoulder area to prevent wrinkling, thereby ensuring that the movement near the cervical margin is strictly radially centripetal and does not produce tangential misalignment.
[0098] 7. After each iteration, the present invention shrinks the target vertex within a preset range from the cut end inward by a preset amount in the opposite direction of its own normal, ensuring that a gradually narrowing conical profile is formed from the neck edge to the cut end, and ensuring that the cut end repair gap meets clinical requirements.
[0099] 8. The present invention performs Laplace physical smoothing processing on the preliminary deformation preparation model with different smoothing intensities in different sections, which can effectively preserve the anatomical grinding morphology generated in the steps while accurately eliminating wrinkles.
[0100] 9. This invention performs deterministic radial shrinkage correction on each vertex to eliminate systematic errors and local surface defects, ensuring that the axial surface cohesion is highly consistent with the set value, thus providing a reliable guarantee for the placement of the repair body.
[0101] 10. The present invention applies normal-guided directional smoothing to the upper part of the lip surface, eliminating local micro-bulges caused by multiple rounds of small step iterations, while not destroying the overall anatomical curvature of the lip surface. Attached Figure Description
[0102] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:
[0103] Figure 1 This is a flowchart illustrating a specific embodiment of the present invention;
[0104] Figure 2 This is a schematic diagram of the neck edge vertex in a specific embodiment of the present invention;
[0105] Figure 3 This is a schematic diagram of the anatomical skeleton path in a specific embodiment of the present invention;
[0106] Figure 4 This is a schematic diagram of the four quadrants of the plane in a specific embodiment of the present invention;
[0107] Figure 5 This is a schematic diagram of the vertical weight field established in a specific embodiment of the present invention;
[0108] Figure 6 This is a schematic diagram showing the superposition and comparison of the final crown preparation and the original crown model in a specific embodiment of the present invention. Detailed Implementation
[0109] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments and accompanying drawings. The illustrative embodiments and descriptions of this invention are for explaining the invention only and are not intended to limit the invention. In the description of this application, it should be understood that terms such as "front," "rear," "left," "right," "upper," "lower," "vertical," "horizontal," "high," "low," "inner," and "outer," indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the scope of protection of this application.
[0110] Example 1:
[0111] like Figure 1 The method for automatically generating a crown preparation body using four-quadrant zoning control and dynamic skeleton guidance, as shown, includes the following steps:
[0112] Step S1: Obtain the three-dimensional mesh model of the crown to be prepared, analyze the vertices and faces, and configure the grinding parameters for each face.
[0113] In this embodiment, the STL file of the complete 3D mesh model of the tooth crown with the gingival margin after 3D restoration processing, obtained from the oral cavity scan, is read and parsed into a vertex set V and a face connection table F.
[0114] In this embodiment, the grinding parameters for each surface include any one or more of the following: cutting edge grinding amount, lip surface grinding amount, tongue surface grinding amount, mesial surface grinding amount, distal surface grinding amount, axial surface convergence degree, and neck edge shoulder buffer width.
[0115] In this embodiment, the default settings are: incision abrasion amount = 1.8mm, labial abrasion amount = 1.5mm, lingual abrasion amount = 1.0mm, mesial abrasion amount = distal abrasion amount = 1.2mm, axial convergence = 6°, and shoulder buffer width = 1.0mm. The operator can flexibly adjust the abrasion amount parameters of each surface according to the specific case and / or the needs of the repair materials.
[0116] Step S2: Determine the neck edge vertices and non-neck edge vertices; keep the coordinate positions of each neck edge vertex fixed.
[0117] Specifically, it includes:
[0118] Step S201: Traverse all triangular faces in the three-dimensional mesh model of the crown and define the edge referenced by only one triangular face as a free edge.
[0119] Step S202: Take all the vertices of the free edges as cervical vertices; take the remaining vertices in the three-dimensional mesh model of the crown as non-cervical vertices.
[0120] Please refer to Figure 2 , Figure 2 The red vertices in the diagram represent the obtained neck edge vertices. Connecting these vertices sequentially yields a complete closed neck edge line. Throughout all subsequent calculation stages, the coordinates of each neck edge vertex remain constant to ensure that the final neck edge shape of the prepared model perfectly matches the original oral scan model.
[0121] Step S3: Construct the anatomical skeleton path along the long axis of the tooth crown.
[0122] Specifically, it includes:
[0123] Step S301: Divide the three-dimensional mesh model of the crown into layers along the longitudinal direction, so that the number of layers is within the set range.
[0124] In this embodiment, the height z of the lowest point of the neck edge is obtained along the Z-axis. min The height z of the highest point of the cut end max and total crown height height =z max -z min The number of layers is dynamically calculated with a target layer thickness of 0.35mm, and the number of layers is verified to be within the set range of [20, 50]. If not, the target layer thickness is adjusted until the number of layers is within the set range.
[0125] Step S302: Obtain the centroids of each layer. The specific process is as follows:
[0126] Step S3021: Calculate the center height of each floor;
[0127] Step S3022: Using the center height as a reference, set a buffer zone between layers within a specified height range for each layer. For example, the center height of the kth layer is z. steps (k), specifying a height range of ±20% of the center height, then the interlayer buffer of the k-th layer is [0.8z]. steps (k), 1.2z steps (k)].
[0128] Step S3023: Extract all vertices within the interlayer buffer and calculate the mean of the planar coordinates of all vertices to obtain the centroid plane coordinates;
[0129] Step S3024: Combine the center height of each layer with the centroid plane coordinates to obtain the centroid coordinates of each layer.
[0130] Step S3025: Extract the layer located in the specified area at the top of the crown as the top layer, perform a weighted calculation on the centroid plane coordinates of each top layer and the reference centroid plane coordinates, and correct the centroid coordinates of the top layer.
[0131] In this embodiment, the region located at 15% of the crown's height is defined as the crown's top area; that is, 0.85 × total height Layers above the crown height are considered the top layer. The reference centroid plane is taken as the plane at 70% of the total crown height; therefore, the coordinates of the reference centroid plane are the centroid coordinates of the layer where 70% of the total crown height is located.
[0132] In this embodiment, the weight of the reference centroid plane coordinates is 70%, and the weight of the centroid plane coordinates of the top layer is 30%. Based on this, the XY coordinates of each top layer are weighted and calculated to obtain the corrected centroid coordinates of each top layer.
[0133] Step S303: Apply a low-pass filter to the planar coordinates of the centroids of each layer. In this embodiment, a LOWESS low-pass filter is used, with a filter window width of one-quarter of the total number of layers, to eliminate scanning noise.
[0134] Step S304: Connect the centroid coordinates of each layer longitudinally to obtain the anatomical skeleton path, as shown below. Figure 3 As shown by the middle red line.
[0135] Step S4: Establish a polar coordinate system, divide the four quadrants based on the lip surface, tongue surface, mesial surface and distal surface, and construct a two-dimensional weighted field.
[0136] Specifically, it includes:
[0137] Step S401: Establish a planar polar coordinate system using the average coordinates of all neck edge vertices as the origin of the polar coordinates.
[0138] Step S402: Calculate the polar angle of each vertex in the 3D mesh model of the tooth crown, and divide it based on the four basic directions of labial surface, lingual surface, mesial surface, and distal surface, as follows: Figure 4 The transverse weight field is obtained by dividing the plane into four quadrants as shown. The four quadrants defined in this embodiment must satisfy the following conditions:
[0139] Polar angle = 90°, facing the lip surface;
[0140] Polar angle = -90°, facing the tongue surface;
[0141] Polar angle = 0°, facing near the mid-surface;
[0142] Polar angle = ±180°, facing the far middle surface.
[0143] Step S403: For each vertex in the three-dimensional mesh model of the crown, calculate the relative height with the nearest cervical margin vertex, and divide the three-dimensional mesh model of the crown into the locked cervical margin area, the transition area, and the free deformation area.
[0144] In this embodiment, the preset locking height and transition height are:
[0145] If the relative height of a vertex is less than the locking height, then the vertex belongs to the locking neckline area;
[0146] If the relative height of a vertex is greater than or equal to the locked height and less than the transition height, then the vertex belongs to the transition zone.
[0147] If the relative height of a vertex is greater than or equal to the transition height, then the vertex belongs to the free deformation zone.
[0148] Preferably, the locking height is 0.1 mm, and the transition height is configured according to the four quadrants: 2.0 mm for the lip surface, 0.8 mm for the tongue surface, 0.5 mm for the mesial surface, and 0.5 mm for the distal surface.
[0149] Step S404: Establish a longitudinal weight field: Set the longitudinal weight of the locked neck edge region to 0, the longitudinal weight of the free deformation region to 1, and the longitudinal weight of the transition region to be obtained by interpolation.
[0150] In this embodiment, the longitudinal weight of the transition region is preferably obtained by the following quadratic cosine function interpolation:
[0151] In the formula: w_rigid represents the vertical weight; t is the normalized transition parameter, calculated using the following formula: , where h rel The relative height between the target vertex and the nearest neck edge vertex.
[0152] In this embodiment, the quadratic cosine function is preferably used to interpolate the longitudinal weight of the transition zone. Compared with other interpolation methods such as traditional linear interpolation, the first derivative of the quadratic cosine function is zero at both endpoints t=0 and t=1. That is, the rate of change of the longitudinal weight at the junction of the locked neck edge region and the transition zone, as well as at the junction of the transition zone and the free deformation region, is smoothly reduced to zero. This avoids step-like morphological abrupt changes, thus ensuring a more continuous and smooth transition between the neck edge region and the grinding region, and effectively preventing visible steps or ridges from appearing on the surface of the prepared body.
[0153] The vertical weight field established in this embodiment is as follows: Figure 5 As shown, in Figure 5In the diagram: the bottom blue area is the locked neckline area, and the vertical weight of the vertices in this area is 0; the middle colored area is the transition area, and the vertical weight of the vertices in this area is between 0 and 1; the upper reddish-brown area is the free deformation area, and the vertical weight of the vertices in this area is 1; the legend represents the weights, which are dimensionless.
[0154] Step S5: Decompose the target amount of grinding into several small steps, apply a mixed displacement field of the fusion axial normal grinding component, the axial compression component of the cut end and the convergence component to each non-neck edge vertex, and perform several iterations to obtain the preliminary deformation preparation model.
[0155] Step S6: Perform Laplace physical smoothing on the preliminary deformation preparation model.
[0156] Specifically, it includes:
[0157] Step S601: Based on the longitudinal weight field and the preset longitudinal weight threshold, the preliminary deformation preparation model is divided into: the neck edge sensitive area, the cut end and the axial free area.
[0158] In this embodiment, the region with a longitudinal weight less than 0.95 is divided into the neck edge sensitive region, and the region with a longitudinal weight greater than or equal to 0.95 is divided into the cut end and axial free region.
[0159] Step S602: Perform Laplace physical smoothing:
[0160] For the sensitive area at the neck margin, a first relaxation factor is configured; for the cut end and the free area on the axial surface, a second relaxation factor is configured; and the first relaxation factor is greater than the second relaxation factor.
[0161] In this embodiment, the first relaxation factor is set to 0.8 and the second relaxation factor is set to 0.2.
[0162] The engineering iterative formula for Laplace physical smoothing is as follows: In the formula: V(i) represents the current three-dimensional coordinates of the i-th vertex; V_ 邻域均值(i) It is the average of the coordinates of all vertices directly adjacent to the i-th vertex via grid edges; It is a relaxation factor.
[0163] In this embodiment, the Laplacian physical smoothing process is iterated 50 times. Of course, the Laplacian physical smoothing process is skipped entirely for the neck vertices, keeping the coordinates of each neck vertex constant.
[0164] Step S7: Perform radial tapered polymerization finishing on the axial surface area; perform directional smoothing on the lip surface area.
[0165] Methods for radial tapered polymer finishing of axial surface regions include:
[0166] Step S701: Using the line connecting the start and end points of the anatomical skeleton path as the long axis of the crown, calculate the radial distance from each vertex to the long axis of the crown;
[0167] Step S702: Based on the radial distance, calculate the refinement target position corresponding to each vertex:
[0168] ;
[0169] In the formula: r_target is the target position for refinement; r_current is the radial distance from the vertex to the long axis of the crown; shrink_factor is the shrinkage ratio coefficient; z_norm represents the normalized height of the vertex;
[0170] In this embodiment, shrink_factor = 0.25. The normalized height of each vertex is calculated using the following formula:
[0171] ;
[0172] In the formula, z min The height of the lowest point of the neck margin, z max This is the height of the highest point of the cut end.
[0173] Of course, each cervical margin vertex skips this radial tapering finishing step entirely. In a more preferred embodiment, vertices within a region of 5% of the total crown height above the cervical margin are also excluded from this radial tapering finishing step.
[0174] Step S703: Move each vertex toward its corresponding refinement target position along the radial direction, and the moving distance is less than the distance between the vertex and the refinement target position.
[0175] Methods for directional smoothing of the lip surface region include:
[0176] Step S704: Calculate the dot product of the normal vector of each vertex and the lip reference direction vector. If the dot product is greater than a preset threshold and the current vertex meets the preset height condition, then the corresponding vertex is identified as a lip vertex.
[0177] In this embodiment, if the dot product is greater than 0.25 and the height of the current vertex is greater than 40% of the total crown height, then the corresponding vertex is identified as a labial vertex.
[0178] Step S705: For each vertex of the lip face, calculate the difference vector between its neighborhood mean position and its own position.
[0179] Step S706: Extract the projection component of the difference vector onto the normal of the lip vertex, and use it as the actual movement of the lip vertex, so that the lip vertex contracts inward along the normal.
[0180] Step S707: In the transition area between the lip surface and the tongue surface, mesial surface, and distal surface, perform omnidirectional smoothing with a preset intensity. The purpose of this step is to make the lines and angles between the lip surface and the tongue surface, mesial surface, and distal surface naturally rounded, avoiding overly sharp edges, and meeting the aesthetic requirements of clinical placement of the prosthesis.
[0181] In this embodiment, the transition region is preferably the region corresponding to the vertices whose dot product is in the range of 0.1 to 0.4 in step S704.
[0182] Step S8: Construct triangulations for the final mesh point set and patch connection table, and export the preparatory volume 3D mesh model file.
[0183] This embodiment constructs a triangulation object based on the final vertex set and the original facet connection table, and writes it to an STL format file with the timestamp as the filename, thus completing the output of the preliminary 3D mesh model file. This file can be directly used for CNC machining or further CAD design of the repair body.
[0184] Example 2:
[0185] A method for automatically generating crown preparations using four-quadrant zoning control and dynamic skeleton guidance, based on Example 1: The mixed displacement field in step S5 is obtained through the following method:
[0186] Step S501: Based on the weight values of the direction corresponding to the target vertex in the transverse weight field, and combined with the polar angle θ of the target vertex, determine the axial normal grinding component of the target vertex:
[0187] 0°≤θ<90°, axial normal grinding component = cos²θ×near mid-surface grinding amount + sin²θ×lip surface grinding amount;
[0188] 90°≤θ<180°, axial normal grinding component = cos²θ×distal surface grinding amount + sin²θ×lip surface grinding amount;
[0189] -90°≤θ<0°, axial normal grinding component = cos²θ×near mid-surface grinding amount + sin²θ×tongue surface grinding amount;
[0190] -180°≤θ<-90°, axial normal grinding component = cos²θ×distal surface grinding amount + sin²θ×tongue surface grinding amount.
[0191] This step demonstrates the application of the horizontal weight field.
[0192] For example:
[0193] For the vertex directly opposite the lip surface, θ=90°, cos²θ=0, sin²θ=1, therefore its corresponding axial normal grinding component = 0×1.2+1×1.5=1.5mm;
[0194] For the vertex directly opposite the adjacent face, θ=0°, cos²θ=1, sin²θ=0, therefore its corresponding axial normal grinding component = 1×1.2+0×1.5=1.2mm.
[0195] Step S502: Take the opposite direction of the horizontal projection of the normal vector of the target vertex as the grinding direction of the axial normal grinding component.
[0196] Step S503: Calculate the normalized height z_norm of the target vertex; the calculation method is the same as in step S702 of Example 1, and will not be repeated here.
[0197] Step S504: Using the nonlinear power function of the normalized height as the cutting edge weight, determine the axial compression component of the target vertex: Axial compression component of the cutting edge = cutting edge wear amount × cutting edge weight.
[0198] In this embodiment, the cut-off weight = z_norm 2.5 .
[0199] Step S505: Based on the two-dimensional weighted field, the axial normal grinding component and the axial compression component of the cut end are fused to obtain a mixed displacement field.
[0200] In this embodiment, the axial surface normal grinding component and the cut end axial compression component are superimposed and fused in the following manner:
[0201] Total amount of wear at the target vertex: total_move = {vec_side × side_red × (1 - z_norm)} 2.5 )+[0, 0,-1]×prep_params.incisal×z_norm 2.5}×w_rigid(i); where: vec_side represents the grinding direction of the axial normal grinding component; side_red represents the axial normal grinding component; prep_params.incisal represents the amount of cutting edge grinding; w_rigid(i) represents the longitudinal weight of the i-th node. "[0, 0, -1]×prep_params.incisal×z_norm 2.5 This whole represents the axial compression component of the cut end.
[0202] The physical meaning of the above superposition and fusion formula is as follows:
[0203] At the bottom of the neck margin, z_norm=0, and the component weights of the mesial and distal surfaces are (1 z_norm 2.5 =1, z_norm of the tangent component 2.5 =0, the displacement comes entirely from the lateral normal wear;
[0204] At the vertex of the tangent, z_norm=1, the weights of the near and far surfaces are 0, the weight of the tangent component is 1, and the displacement comes entirely from the axial compression of the tangent.
[0205] In the middle region, taking z_norm=0.5 as an example, the weights of the near-mid surface and far-mid surface components are 0.823, and the weight of the tangent component is 0.177. Both play a role simultaneously and are superimposed proportionally.
[0206] The sum of the weights of the axial normal grinding component and the axial compression component at the cut end is always equal to 1 at any height, ensuring that the fusion process is continuous and without jumps.
[0207] Step S506: For the vertices of the locked cervical margin region and the transition region, the displacement vector obtained through the mixed displacement field is projected to the radial direction, and the projection result is used as the amount of grinding of the vertices of the locked cervical margin region and the transition region; wherein, the radial direction is a horizontal direction perpendicular to the long axis of the tooth, pointing from the target vertex or away from the long axis.
[0208] Step S507: After each iteration, for the target vertices within a preset range from the cutting edge, shrink them inward by a preset amount in the opposite direction of their own normal. In this embodiment, the preset range is 2mm. The preset shrinkage amount is 0.025mm.
[0209] Step S508: Apply a light Laplacian smoothing to the non-neck edge vertices.
[0210] The method described in this embodiment can obtain a complete final crown preparation, and its superposition and comparison with the original crown model are shown below. Figure 6 As shown.
[0211] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
[0212] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 process, method, article, or apparatus.
Claims
1. A four-quadrant partition control and dynamic skeleton-guided automatic generation method of dental crown preparation, characterized by, Includes the following steps: Step S1: Obtain the three-dimensional mesh model of the crown to be prepared, analyze the vertices and faces, and configure the grinding parameters for each face; Step S2: Determine the neck edge vertices and non-neck edge vertices; keep the coordinate positions of each neck edge vertex fixed; Step S3: Construct the anatomical framework path along the long axis of the tooth crown; Step S4: Establish a polar coordinate system, divide the four quadrants based on the lip surface, tongue surface, mesial surface and distal surface, and construct a two-dimensional weight field; Step S5: Decompose the target amount of grinding into several small steps, apply a mixed displacement field of the fusion axial normal grinding component, the axial compression component of the cut end and the convergence component to each non-neck edge vertex, and perform several iterations to obtain the preliminary deformation preparation model. Step S6: Perform Laplace physical smoothing on the preliminary deformation preparation model; Step S7: Perform radial tapered polymerization finishing on the axial surface area; Targeted smoothing is applied to the lip area; Step S8: Construct triangulations for the final mesh point set and patch connection table, and export the preparatory volume 3D mesh model file.
2. The method of claim 1, wherein the method is a four-quadrant partition control and dynamic skeleton-guided automatic generation of a dental crown preparation. In step S1, the grinding parameters of each surface include any one or more of the following: cutting edge grinding amount, lip surface grinding amount, tongue surface grinding amount, mesial surface grinding amount, distal surface grinding amount, axial surface convergence degree, and neck edge shoulder buffer width.
3. The method of claim 1, wherein the method further comprises: In step S2, the methods for determining the neck edge vertex and non-neck edge vertices include: Step S201: Traverse all triangular faces in the three-dimensional mesh model of the crown and define the edge referenced by only one triangular face as a free edge; Step S202: Take all the vertices of the free edges as cervical vertices; take the remaining vertices in the three-dimensional mesh model of the crown as non-cervical vertices.
4. The method of claim 1, wherein, Step S3 specifically includes: Step S301: Divide the three-dimensional mesh model of the tooth crown into layers along the longitudinal direction, so that the number of layers is within the set range; Step S302: Obtain the centroid of each layer; Step S303: Apply a low-pass filter to the planar coordinates of the centroids of each layer; Step S304: Connect the centroid coordinates of each layer in sequence along the longitudinal direction to obtain the anatomical skeleton path.
5. The method of claim 4, wherein the four-quadrant partitioning control and dynamic skeleton-guided automatic generation of a dental crown preparation is characterized by, Step S302 specifically includes: Step S3021: Calculate the center height of each floor; Step S3022: Based on the center height, set a buffer zone between floors within a specified height range for each floor; Step S3023: Extract all vertices within the interlayer buffer and calculate the mean of the planar coordinates of all vertices to obtain the centroid plane coordinates; Step S3024: Combine the center height of each layer with the centroid plane coordinates to obtain the centroid coordinates of each layer; Step S3025: Extract the layer located in the specified area at the top of the crown as the top layer, perform a weighted calculation on the centroid plane coordinates of each top layer and the reference centroid plane coordinates, and correct the centroid coordinates of the top layer. Wherein, the reference centroid plane coordinates are the centroid plane coordinates of the preset layer, the height of the preset layer is greater than 50% of the total crown height; and the weight of the centroid plane coordinates of the top layer is less than the weight of the reference centroid plane coordinates.
6. The method of claim 1, wherein, Step S4 specifically includes: Step S401: Establish a planar polar coordinate system using the average coordinates of all neck edge vertices as the origin of the polar coordinates; Step S402, calculate the polar angle of each vertex in the three-dimensional mesh model of the dental crown, and divide the plane into four quadrants based on the four basic directions of the labial surface, the lingual surface, the mesial surface and the distal surface to obtain a transverse weight field; in the plane four quadrants, make: Polar angle = 90°, towards the labial surface; Polar angle = -90°, towards the lingual surface; Polar angle = 0°, towards the mesial surface; Polar angle = ±180°, towards the distal surface; Step S403, for each vertex in the three-dimensional mesh model of the dental crown, calculate the relative height with the nearest cervical margin vertex, and divide the three-dimensional mesh model of the dental crown into a locking cervical margin area, a transition area and a free deformation area; Step S404, establish a longitudinal weight field: set the longitudinal weight of the locking cervical margin area to 0, the longitudinal weight of the free deformation area to 1, and the longitudinal weight of the transition area to be obtained by interpolation.
7. The method of claim 6, wherein the method further comprises: In step S5, the mixed displacement field is obtained by the following method: Step S501, according to the weight value of the target vertex corresponding direction in the transverse weight field, combined with the polar angle θ of the target vertex, determine the axial surface normal grinding component of the target vertex: 0°≤θ<90°, axial surface normal grinding component = cos²θ × mesial surface grinding amount + sin²θ × labial surface grinding amount; 90°≤θ<180°, axial surface normal grinding component = cos²θ × distal surface grinding amount + sin²θ × labial surface grinding amount; -90°≤θ<0°, axial surface normal grinding component = cos²θ × mesial surface grinding amount + sin²θ × lingual surface grinding amount; -180°≤θ<-90°, axial surface normal grinding component = cos²θ × distal surface grinding amount + sin²θ × lingual surface grinding amount; Step S502, take the opposite direction of the horizontal projection of the normal vector of the target vertex as the grinding direction of the axial surface normal grinding component; Step S503, calculate the normalized height of the target vertex; Step S504, take the nonlinear power function of the normalized height as the incisal weight to determine the incisal axial compression component of the target vertex: incisal axial compression component = incisal grinding amount × incisal weight; Step S505, based on the two-dimensional weight field, fuse the axial surface normal grinding component and the incisal axial compression component to obtain a mixed displacement field; Step S506, for the vertices of the locking cervical margin area and the transition area, project the displacement vector obtained by the mixed displacement field to the radial direction, and take the projection result as the grinding amount of the vertices of the locking cervical margin area and the transition area; wherein the radial direction is a horizontal direction perpendicular to the dental body long axis, pointing to or away from the long axis; Step S507, after each iteration, the target vertex within a preset range from the incisal end is inwardly contracted by a preset contraction amount in the opposite direction of its normal; Step S508, apply a slight Laplace smoothing to the non-cervical margin vertex.
8. The method of claim 6, wherein the method further comprises: Step S6 specifically includes: Step S601, based on the longitudinal weight field, divide the preliminary deformation preparation body model into a cervical margin sensitive area, an incisal end and an axial surface free area; Step S602, perform Laplace physical smoothing processing: For the sensitive area at the neck margin, a first relaxation factor is configured; for the cut end and the free area on the axial surface, a second relaxation factor is configured; and the first relaxation factor is greater than the second relaxation factor.
9. The method of claim 7, wherein the method further comprises: In step S7, the method for radial tapered polymerization finishing of the axial surface region includes: Step S701: Using the line connecting the start and end points of the anatomical skeleton path as the long axis of the crown, calculate the radial distance from each vertex to the long axis of the crown; Step S702: Based on the radial distance, calculate the refinement target position corresponding to each vertex: r_target = r_current×(1-shrink_factor×z_norm); In the formula: r_target is the target position for refinement; r_current is the radial distance from the vertex to the long axis of the crown; shrink_factor is the shrinkage ratio coefficient; z_norm represents the normalized height of the vertex; Step S703: Move each vertex toward its corresponding refinement target position along the radial direction, and the moving distance is less than the distance between the vertex and the refinement target position.
10. The method of claim 9, wherein, In step S7, the method for directional smoothing of the lip surface region includes: Step S704: Calculate the dot product of the normal vector of each vertex and the lip reference direction vector. If the dot product is greater than a preset threshold and the current vertex meets the preset height condition, then the corresponding vertex is identified as a lip vertex. Step S705: For each vertex of the lip surface, calculate the difference vector between its neighborhood mean position and its own position; Step S706: Extract the projection component of the difference vector onto the normal of the lip vertex, and use it as the actual movement of the lip vertex, so that the lip vertex contracts inward along the normal. Step S707: In the transition area between the lip surface and the rest of the area, perform omnidirectional smoothing with a preset intensity.
