Method, device, storage medium and program for predicting z-axis of partial coordinate system of oral scan crown

Abstract

This invention relates to a method for predicting the Z-axis of the local coordinate system of dental crowns in intraoral scans: S1: Extract tooth structures from CT images and intraoral scan meshes to obtain TMESH and CMESH; S2: Calculate the weighted centroid of CMESH and iteratively shrink it, traversing all vertices of the surface mesh and storing them in a skeletonized point set P; S3: Fit the optimal tangent plane M and calculate its normal vector V; S4: Extend along the normal vector to generate a pipeline; S5: Train a UNet model using the pipeline data to infer and generate pipelines; S6: Input dental crown data to infer and generate pipeline Pi; S7: Skeletonize Pi using the method in S2 to obtain a point set Pr, and fit the points in Pr with the least squares method to find the optimal straight line as the z-axis; S8: Map the Z-axis coordinate values ​​of each dental crown in the CT image to the intraoral scan data using the correspondence between teeth to obtain the Z-axis direction of each dental crown in the intraoral scan. This invention realizes the conversion of the local coordinate system of dental crowns between different data sources, laying the foundation for more accurate data fusion; and improves the accuracy and efficiency of digital arch treatment plan design.

Description

Technical Field

[0001] This invention relates to the field of oral medical data processing technology, and in particular to a method, device, storage medium and program for predicting the Z-axis of the local coordinate system of a dental crown in an oral scan. Background Technology

[0002] Currently, digital dental arch treatment planning typically requires the combined use of CT and intraoral scan data. However, there is a certain image registration error between the two, affecting the accuracy and efficiency of the treatment plan. Manually constructing the local coordinate system of the crown requires technicians to spend a significant amount of time on measurement and calculation, and manual measurements are susceptible to subjective errors. Therefore, from the perspective of reducing the workload of dental technicians, improving prediction accuracy, and standardizing the restoration process, it is necessary to propose a method for predicting the Z-axis of the local coordinate system of the crown from intraoral scans. This would significantly reduce the workload of technicians, improve restoration efficiency, provide more accurate and reliable results, construct a standardized local coordinate system, guide technicians in a standardized crown restoration and fabrication process, and facilitate quality control. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to design a method, device, storage medium and program based on the Z-axis of the local coordinate system of the crown of the tooth with root bone for prediction, thereby solving the existing technical problems.

[0004] In this invention, TMesh is the mesh reconstruction result of the surface mesh of the CT segmented tooth voxel data; CMesh is the mesh of the dental crown segmentation result from intraoral scanning; and P is the vertex point set of the TMesh skeletonization result mesh of a single tooth.

[0005] To solve the above-mentioned technical problems, the method for predicting the Z-axis of the local coordinate system of the crown during oral scanning of the present invention includes the following steps:

[0006] Step S1: Extract tooth structure from CT images and oral scan mesh, and obtain the surface mesh reconstruction mesh result TMesh of tooth CT segmentation tooth voxel data and the mesh result CMesh of oral scan segmentation crown data respectively;

[0007] Step S2: Calculate the weighted centroid of CMESH and iteratively shrink it. For each tooth model of TMESH, calculate the average curvature of its mesh facets and use it as the facet weight. Calculate the weighted centroid C based on the facet weights, and then start the iterative shrinking process. After the iteration is completed, traverse all vertices of the surface mesh SMESH and store them in the skeletonized point set P.

[0008] Step S3: Fit the best tangent plane M: Using the points in P, fit the best tangent plane M using the least squares method, and calculate the normal vector V of M;

[0009] Step S4: Extend along the normal vector to generate the pipe: Extend the centroid C along the positive and negative directions of V by a distance of 5*|V| to obtain points Pl and Pu. Loft P, Pl, and Pu into the shape of a pipe.

[0010] Step S5: Train the UNet model: Train the UNet model using pipeline data for inference pipeline generation;

[0011] Step S6: Input the crown data, and UNet inference generates the pipeline Pi;

[0012] Step S7: Skeletonize the pipeline and fit a straight line: Skeletonize Pi using the method in step S2 to obtain a point set Pr. Use the points in Pr to fit the best straight line using the least squares method, which is used as the z-axis.

[0013] Step S8: Obtain root canal data: In the constructed spatial rectangular coordinate system, pip represents the root canal. Using the correspondence between teeth, the Z-axis coordinate values ​​of each crown in the CT image are mapped to the oral scan data to obtain the Z-axis direction of each crown in the oral scan.

[0014] Furthermore, in step S1, the tooth structure is segmented using an image segmentation method based on the intensity threshold of the CT image to obtain three-dimensional voxel data of each tooth.

[0015] Furthermore, in step S1, the tooth data is reconstructed using the marchingcube method to obtain the tooth TMESH.

[0016] Furthermore, in step S1, the 3D-UNet network is used to directly predict the crown region based on the features in the mesh, and the oral scan crown mesh data CMesh is segmented.

[0017] Furthermore, in step S2, the iterative shrinkage process includes the following steps:

[0018] Step S21: Set the maximum number of iterations (maxIter) and the minimum moving distance (minMove);

[0019] Step S22: Calculate the mean curvature (meanCurvature) of each vertex of the surface mesh;

[0020] Step S23: Initialize the contraction iteration: Initialize the total vertex movement distance totalMove = 0, and initialize the iteration count iter = 0;

[0021] Step S24: Vertex contraction iteration: Traverse each vertex i;

[0022] Step S25: Determine whether the stop condition is met: If totalMove < minMove and iter < maxIter, then stop and proceed to step S26; otherwise, return to step S24 to continue the iteration.

[0023] Step S26: The iteration ends. Traverse all vertices of the surface mesh SMesh and store them in the point set P.

[0024] Furthermore, in step S24, it specifically includes the following steps:

[0025] Step S241: Obtain the average curvature curv and the adjacent face list neighFaces of vertex i.

[0026] Step S242: Initialize the velocity vector, v x = v y = v z = 0;

[0027] Step S243: Traverse each adjacent face faceIndex: Obtain the normal vector n of the face faceIndex x , n y , n z , if curv > 0, this face is a convex face, and the velocity direction is the same as the normal vector direction. Add the normal vector to the velocity vector, v x = v x + n x , v y = v y + n y , v z = v z + n z ; if curv < 0, this face is a concave face, and the velocity direction is opposite to the normal vector direction. Subtract the normal vector from the velocity vector, v x = v x - n x , v y = v y - n y , v z = v z - n z ;

[0028] Step S244: Normalize the velocity vector and multiply it by the average curvature curv as the velocity magnitude:

[0029] v x = curv * v x / sqrt(v x * v x + v y * v y + v z * vz );

[0030] v y =curv*v y / sqrt(v x *v x +v y *v y +v z *v z );

[0031] v z =curv*v z / sqrt(v x *v x +v y *v y +v z *v z );

[0032] Step S245: Move the vertex position according to the velocity vector and update the total movement distance totalMove:

[0033] totalMove += sqrt(v x *v x +v y *v y +v z *v z The iteration count is incremented by 1.

[0034] Furthermore, in step S21, maxIter = 500, minMove = 0.0002 * TMesh length of the bounding box diagonal.

[0035] The present invention also provides an electronic device, comprising:

[0036] At least one processor; and

[0037] At least one memory communicatively connected to the processor;

[0038] The memory stores instructions that can be executed by a processor, which are then executed by the processor to cause the electronic device to perform the aforementioned method for predicting the Z-axis of the local coordinate system of the dental crown during oral scanning.

[0039] The present invention also provides a computer-readable storage medium storing computer-executable instructions thereon, which, when executed by a processor, implement the aforementioned method for predicting the Z-axis of the local coordinate system of the dental crown during oral scanning.

[0040] The present invention also provides a computer program product, including computer instructions, which, when executed by a processor, cause a computer device to perform the aforementioned method for predicting the Z-axis of the local coordinate system of the dental crown.

[0041] Compared with the prior art, the present invention has the following advantages:

[0042] (1) The Z-axis direction of each crown of the teeth is predicted by the oral scan, which provides a reference for the registration of the two data sources;

[0043] (2) It realizes the transformation of the local coordinate system of the crown between different data sources, laying the foundation for more accurate data fusion;

[0044] (3) Improved the accuracy and efficiency of digital dental arch treatment design. Attached Figure Description

[0045] The specific embodiments of the present invention will be further explained below with reference to the accompanying drawings.

[0046] Figure 1 A schematic diagram of the original CT tooth segmentation voxel data;

[0047] Figure 2 A schematic diagram of the TMesh mesh for reconstructing the surface mesh of CT tooth voxel data;

[0048] Figure 3 This is a schematic diagram of the raw data from the oral scan.

[0049] Figure 4 A schematic diagram of the CMesh grid for segmenting the dental crown.

[0050] Figure 5 A schematic diagram of the crown and centroid of a tooth during oral scanning;

[0051] Figure 6 A schematic diagram of the point set P and TMesh of the TMESH skeletonization result for a single tooth;

[0052] Figure 7 A schematic diagram of generating pipes and CMESH for the Z-axis. Detailed Implementation

[0053] Example 1

[0054] The method for predicting the Z-axis of the local coordinate system of the dental crown during intraoral scanning in this embodiment includes the following steps:

[0055] Step S1: Extract tooth structure from CT images and oral scan mesh, and obtain the surface mesh reconstruction mesh result TMesh of tooth CT segmentation tooth voxel data and the mesh result CMesh of oral scan segmentation crown data respectively;

[0056] In this preferred embodiment, in step S1, the tooth structure is segmented using an image segmentation method based on the intensity threshold of the CT image. This embodiment uses watershed image segmentation to obtain three-dimensional voxel data for each tooth, such as... Figure 1 As shown;

[0057] In this preferred embodiment, in step S1, the marchingcube method is used to reconstruct the tooth data into a surface mesh to obtain the tooth TMESH, such as... Figure 2 As shown;

[0058] In this preferred embodiment, in step S1, the mouth is scanned using a professional 3D scanner to generate an oral scanning mesh, and the original oral scanning data is as follows: Figure 3 As shown, using the 3D-UNet network, the crown region is directly predicted based on features in the mesh, and the oral cavity scan crown mesh data (CMesh) is segmented, as shown below. Figure 4 As shown;

[0059] Step S2: Calculate the weighted centroid of CMESH and iteratively shrink it. For each tooth model in TMESH, calculate the average curvature of its mesh facets, which is used as the facet weight. Calculate the weighted centroid C based on the facet weights, as follows: Figure 5 As shown, the iterative shrinkage process then begins. After the iteration is complete, all vertices of the surface mesh SMash are traversed and stored in the skeletonized point set P.

[0060] In this preferred embodiment, step S2, the iterative shrinkage process includes the following steps:

[0061] Step S21: Set the maximum number of iterations maxIter and the minimum movement distance minMove; in this embodiment, maxIter = 500, minMove = 0.0002 * the length of the diagonal of the TMesh bounding box;

[0062] Step S22: Calculate the mean curvature (meanCurvature) of each vertex of the surface mesh;

[0063] Step S23: Initialize the contraction iteration: Initialize the total vertex movement distance totalMove = 0, and initialize the iteration count iter = 0;

[0064] Step S24: Vertex contraction iteration: Traverse each vertex i; specifically including the following steps:

[0065] Step S241: Obtain the average curvature of vertex i and the list of adjacent faces, neighFaces;

[0066] Step S242: Initialize the velocity vector, v x =v y =v z=0;

[0067] Step S243: Traverse each adjacent face indices: Obtain the normal vector n of the face index. x n y n z If curv > 0, this surface is convex, and the velocity direction is in the same direction as the normal vector. Add the normal vector to the velocity vector, v. x =v x +n x v y =v y +n y v z =v z +n z If curv < 0, this surface is concave, and the velocity direction is opposite to the normal vector direction. Subtract the normal vector from the velocity vector, v x =v x -n x v y =v y -n y v z =v z -n z ;

[0068] Step S244: Standardize the velocity vector and multiply it by the mean curvature to obtain the velocity magnitude:

[0069] v x =curv*v x / sqrt(v x *v x +v y *v y +v z *v z );

[0070] v y =curv*v y / sqrt(v x *v x +v y *v y +v z *v z );

[0071] v z =curv*v z / sqrt(v x *v x +v y *v y +v z *v z );

[0072] Step S245: Move the vertex positions according to the velocity vectors, and update the total movement distance totalMove:

[0073] totalMove += sqrt(v x * v x + v y * v y + v z * v z ), and increment the iteration count iter by 1;

[0074] Step S25: Determine whether the stopping condition is reached: If totalMove < minMove and iter < maxIter, then the iteration stops and proceed to Step S26; otherwise, return to Step S24 to continue the iteration;

[0075] Step S26: The iteration ends. Traverse all vertices of the surface mesh SMesh and store them in the point set P, as Figure 6 shown;

[0076] Step S3: Fit the best cutting plane M: Use the points in P to fit the best cutting plane M by the least squares method, and calculate the normal vector V of M;

[0077] Step S4: Generate a pipeline by extending along the normal vector: Extend the centroid C along the positive and negative directions of V by a distance of 5 * |V| to obtain points Pl and Pu, and loft P, Pl, and Pu into a pipeline shape, as Figure 7 shown;

[0078] Step S5: Train the UNet model: Use the pipeline data to train the UNet model for inference to generate pipelines;

[0079] Step S6: Input the crown data, and the UNet infers and generates the pipeline Pi;

[0080] Step S7: Skeletonize the pipeline and fit a straight line: Use the method in Step S2 to skeletonize Pi to obtain the point set Pr, and use the points in Pr to fit the best straight line by the least squares method as the z-axis;

[0081] Step S8: Obtain the root canal data: In the constructed spatial rectangular coordinate system, pip represents the root canal. Using the corresponding relationship between teeth, map the Z-axis coordinate values of each crown in the CT image to the intraoral scan data to obtain the Z-axis direction of each crown in the intraoral scan.

[0082] The method for predicting the Z-axis of the dental crown in this embodiment uses a local coordinate system constructed from the axial projection contour of the dental arch. This method enables the prediction and conversion of the Z-axis direction of the dental crown between CT 3D images of the dental region and intraoral scan data, providing an important reference for image registration and data fusion. The main advantages are as follows:

[0083] (1) High efficiency. It can achieve semi-automatic or fully automatic coordinate system prediction without the need for technicians to perform a lot of manual measurement and calculation, and can quickly obtain prediction results.

[0084] (2) Accuracy. Based on 3D scanned point cloud data, computer image processing and artificial intelligence technologies can be used to provide high-precision prediction results.

[0085] (3) Easy to operate. It is simple to operate and requires no professional training. Technicians can easily obtain the prediction results and carry out repair work accordingly.

[0086] (4) Low cost. Apart from the relatively expensive 3D scanning equipment, the research and development and implementation costs of the software system are low, which is conducive to its promotion and application.

[0087] (5) It is conducive to standardization. A standardized coordinate system construction method can generate standardized repair plans, simplify the technician's decision-making process, and control the repair quality.

[0088] (6) Scalability. Coordinate system prediction algorithms and methods can be continuously optimized and improved, and the application of artificial intelligence technology is also continuously expanding. Overall, the invention has strong scalability.

[0089] Example 2

[0090] The electronic device in this embodiment includes:

[0091] At least one processor; and

[0092] At least one memory communicatively connected to the processor;

[0093] The memory stores instructions that can be executed by a processor, which are then executed by the processor to cause the electronic device to perform the method for predicting the Z-axis of the local coordinate system of the dental crown in Embodiment 1.

[0094] Example 3

[0095] The computer-readable storage medium of this embodiment stores computer-executable instructions. When the processor executes the computer-executable instructions, it implements the method for predicting the Z-axis of the local coordinate system of the dental crown in embodiment 1.

[0096] Example 4

[0097] The computer program product of this embodiment includes computer instructions that, when run by a processor, cause a computer device to perform the method for predicting the Z-axis of the local coordinate system of the dental crown in Embodiment 1.

[0098] Many specific details have been set forth in the foregoing description to provide a thorough understanding of the present invention. However, the above description is merely a preferred embodiment of the present invention, and the present invention can be implemented in many other ways different from those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed above. Furthermore, any person skilled in the art can make many possible variations and modifications to the technical solutions of the present invention, or modify them into equivalent embodiments, using the methods and techniques disclosed above, without departing from the scope of the present invention. Any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention, without departing from the content of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A method for predicting the Z-axis of the local coordinate system of a dental crown during oral scanning, characterized in that: Includes the following steps: Step S1: Extract tooth structure from CT images and oral scan mesh, and obtain the surface mesh reconstruction mesh result TMesh of tooth CT segmentation tooth voxel data and the mesh result CMesh of oral scan segmentation crown data respectively; Step S2: Calculate the weighted centroid of CMESH and iteratively shrink it. For each tooth model of TMESH, calculate the average curvature of its mesh facets and use it as the facet weight. Calculate the weighted centroid C based on the facet weights, and then start the iterative shrinking process. After the iteration is completed, traverse all vertices of the surface mesh SMESH and store them in the skeletonized point set P. Step S3: Fit the best tangent plane M: Using the points in P, fit the best tangent plane M using the least squares method, and calculate the normal vector V of M; Step S4: Extend along the normal vector to generate the pipe: Extend the centroid C along the positive and negative directions of V by a distance of 5*|V| to obtain points Pl and Pu. Loft P, Pl, and Pu into the shape of a pipe. Step S5: Train the UNet model: Train the UNet model using pipeline data for inference pipeline generation; Step S6: Input the crown data, and UNet inference generates the pipeline Pi; Step S7: Skeletonize the pipeline and fit a straight line: Skeletonize Pi using the method in step S2 to obtain a point set Pr. Use the points in Pr to fit the best straight line using the least squares method, which is used as the z-axis. Step S8: Obtain root canal data: In the constructed spatial rectangular coordinate system, pip represents the root canal. Using the correspondence between teeth, the Z-axis coordinate values ​​of each crown in the CT image are mapped to the oral scan data to obtain the Z-axis direction of each crown in the oral scan. In step S2, the iterative shrinkage process includes the following steps: Step S21: Set the maximum number of iterations (maxIter) and the minimum moving distance (minMove); Step S22: Calculate the mean curvature (meanCurvature) of each vertex of the surface mesh; Step S23: Initialize the contraction iteration: Initialize the total vertex movement distance totalMove=0, and initialize the iteration count iter=0; Step S24: Vertex contraction iteration: Traverse each vertex i; Step S25: Determine if the stopping condition has been met: If totalMove < minMove and iter < maxIter, then the stopping condition has been met, and proceed to step S26; otherwise, return to step S24 to continue iterating. Step S26: The iteration ends. Traverse all vertices of the surface mesh SMESH and store them in the point set P.

2. The method for predicting the Z-axis of the local coordinate system of the dental crown during oral scanning according to claim 1, characterized in that: In step S1, the tooth structure is segmented using an image segmentation method based on the intensity threshold of the CT image to obtain three-dimensional voxel data of each tooth.

3. The method for predicting the Z-axis of the local coordinate system of the dental crown during oral scanning according to claim 1, characterized in that: In step S1, the tooth data is reconstructed using the marchingcube method to obtain the tooth TMESH.

4. The method for predicting the Z-axis of the local coordinate system of the dental crown during oral scanning according to claim 1, characterized in that: In step S1, the 3D-UNet network is used to directly predict the crown region based on the features in the mesh, and the oral scan crown mesh data CMesh is segmented.

5. The method for predicting the Z-axis of the local coordinate system of a dental crown during oral scanning according to claim 1, characterized in that: Step S24 specifically includes the following steps: Step S241: Obtain the average curvature of vertex i and the list of adjacent faces, neighFaces; Step S242: Initialize the velocity vector, v x =v y =v z =0; Step S243: Traverse each adjacent face indices: Obtain the normal vector n of the face index. x n y n z If curv > 0, this surface is convex, and the velocity direction is in the same direction as the normal vector. Add the normal vector to the velocity vector, v. x =v x +n x v y =v y +n y v z =v z +n z If curv < 0, this surface is concave, and the velocity direction is opposite to the normal vector direction. Subtract the normal vector from the velocity vector, v x =v x -n x v y =v y -n y v z =v z -n z ; Step S244: Standardize the velocity vector and multiply it by the mean curvature to obtain the velocity magnitude: v x =curv*v x / sqrt(v x *v x +v y *v y +v z *v z ); v y =curv*v y / sqrt(v x *v x +v y *v y +v z *v z ); v z =curv*v z / sqrt(v x *v x +v y *v y +v z *v z ); Step S245: Move the vertex position according to the velocity vector and update the total movement distance totalMove: totalMove+=sqrt(v x *v x +v y *v y +v z *v z The iteration count is incremented by 1.

6. The method for predicting the Z-axis of the local coordinate system of a dental crown during oral scanning according to claim 1, characterized in that: In step S21, maxIter=500, minMove = 0.0002 * the length of the diagonal of the TMesh bounding box.

7. An electronic device, characterized in that: include: At least one processor; as well as At least one memory communicatively connected to the processor; The memory stores instructions that can be executed by a processor, which are then executed by the processor to cause the electronic device to perform the method for predicting the Z-axis of the local coordinate system of the dental crown as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that: It stores computer execution instructions, which, when executed by the processor, implement the method for predicting the Z-axis of the local coordinate system of the dental crown as described in any one of claims 1-6.

9. A computer program product, characterized in that: Includes computer instructions that, when executed by a processor, cause a computer device to perform a method for predicting the Z-axis of the local coordinate system of the dental crown as described in any one of claims 1-6.

Images