Method and system for automatically cutting out scanned dental models to extract regions of interest

Abstract

The present invention relates to a method and system for automatically cropping oral scanned dental models to extract regions of interest, comprising: step S1: converting batch mesh training data mesh1 into voxel data volum1; step S2: cropping mesh1 in step S1 by extending 3 units downward from the lowest point of the tooth crown, and filling holes in the corresponding cropped mesh data mesh2, traversing each voxel point in step S1, and setting the label value of the voxel point located inside or on the surface of mesh2 to 1; step S3: performing network training on the data processed by steps S1 and S2; and step S4: inferring the new mesh data mesh3. The method and system for automatically cropping oral scanned dental models to extract regions of interest automatically crop the oral scanned dental models through AI learning results, extract regions of interest, perform automatic cropping, and eliminate redundant regions of the initial dental model, thereby effectively improving the efficiency of tooth segmentation and tooth arrangement.

Description

Technical Field

[0001] The present invention relates to a field of technology, and in particular to a method and system for automatically cutting out scanned dental models to extract patches of regions of interest. Background Art

[0002] Currently, the data of oral scan dental models is large, and most of it is irrelevant, which affects the efficiency of tooth segmentation and tooth arrangement, and there is no better automated cutting solution. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to design a method and system for automatically cropping oral scan dental casts to extract patches of regions of interest, so as to solve the technical problem that the existing oral scan dental cast data is large and the key information is unclear.

[0004] To solve the above technical problems, the method of automatically cutting out a dental model and extracting a region of interest patch includes the following steps:

[0005] Step S1: Convert batch grid training data mesh1 into voxel data volum1;

[0006] Step S2: Extend mesh1 in step S1 downward by 3 units at the lowest point of the tooth crown and perform hole filling on the corresponding mesh data mesh2. Traverse each voxel point in step S1 and set the label value of the voxel point inside or on the surface of mesh2 to 1;

[0007] Step S3: Perform network training on the data processed in steps S1 and S2 above;

[0008] Step S4: Perform inference on the new mesh data mesh3.

[0009] Furthermore, step S4 specifically includes:

[0010] Step S41: converting new grid data into voxel data;

[0011] Step S42: Apply the trained model in step S3 to perform network inference to obtain the inference result, i.e., the segmented mask;

[0012] Step S43: meshing the mask result in step S42 to obtain mesh4;

[0013] Step S44: Based on the original mesh data mesh3 and the mesh mesh4 in step S43, the distances between all vertices of mesh4 and all vertices in mesh3 are calculated, and the vertex closest to each vertex of mesh4 in mesh3 is calculated to obtain the retained main mesh data mesh5;

[0014] Step S45: Calculate the maximum boundary line of mesh5;

[0015] Step S46: Based on the maximum boundary in step S45, curve cutting is performed on the original mesh data to obtain a cut mesh.

[0016] Furthermore, in step S45, by determining whether all half edges of mesh5 are boundaries, the corresponding half edge vertices are retrieved and the maximum boundaryline of mesh5 is calculated.

[0017] Furthermore, in step S46, the point closest to each vertex of the boundaryline in the original mesh is calculated, these points are combined into a curve, and Laplace smoothing is performed to implement curve cutting on the original mesh data.

[0018] The present invention also provides a system for automatically cutting out a scanned dental model to extract a region of interest patch, comprising:

[0019] at least one processor; and

[0020] at least one memory in communication with the processor;

[0021] The memory stores instructions that can be executed by the processor, and the instructions are executed by the processor to enable the electronic device to perform the aforementioned method of automatically cutting out the surface of the area of ​​interest from the dental model scan.

[0022] Beneficial effects of the present invention: The method and system of the present invention for automatically cropping oral scanned dental models to extract patches of regions of interest automatically crop the oral scanned dental models through AI learning results, extract patches of regions of interest, implement automatic cropping, and remove excessively redundant patches of the initial dental models, which can effectively improve the efficiency of tooth segmentation and tooth arrangement. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The specific embodiments of the present invention will be further explained below with reference to the accompanying drawings.

[0024] Figure 1 is the original data;

[0025] Figure 2 Convert the original data grid to voxel data;

[0026] Figure 3 Voxel data segmented for inference;

[0027] Figure 4 The maximum boundary point of the segmentation result voxel conversion grid;

[0028] Figure 5 This is the result of curve cutting on the original data. DETAILED DESCRIPTION

[0029] Example 1

[0030] Combine Figure 1-Figure 5 The method for automatically cutting out a dental model and extracting a region of interest patch of the present invention comprises the following steps:

[0031] Step S1: Convert the batch grid training data mesh1 into voxel data volum1, such as Figure 2 As shown;

[0032] Step S2: Extend mesh1 in step S1 downward by 3 units at the lowest point of the tooth crown and perform hole filling on the corresponding mesh data mesh2. Traverse each voxel point in step S1 and set the label value of the voxel point inside or on the surface of mesh2 to 1;

[0033] Step S3: Perform network training on the data processed in steps S1 and S2 above;

[0034] Step S4: Inferring the new mesh data mesh3, specifically including:

[0035] Step S41: converting new grid data into voxel data;

[0036] Step S42: Apply the trained model in step S3 to perform network inference and obtain the inference result, i.e. the segmented mask, such as Figure 3 As shown;

[0037] Step S43: Mesh the mask result in step S42 to obtain mesh4, such as Figure 4 As shown;

[0038] Step S44: Based on the original mesh data mesh3 and the mesh mesh4 in step S43, the distances between all vertices of mesh4 and all vertices in mesh3 are calculated, and the vertex closest to each vertex of mesh4 in mesh3 is calculated to obtain the retained main mesh data mesh5;

[0039] Step S45: By determining whether the vertex of mesh5 is a boundary, take out the corresponding half-edge vertex and calculate the maximum boundary line of mesh5, such as Figure 4 As shown;

[0040] Step S46: Based on the maximum boundary in step S45, calculate the point in the original mesh that is closest to each vertex of the boundaryline, combine these points into a curve, and perform Laplace smoothing on the original mesh data to perform curve cutting to obtain the cut mesh, such as Figure 5shown.

[0041] Example 2

[0042] The present invention also provides a system for automatically cutting out a scanned dental model to extract a region of interest patch, comprising:

[0043] at least one processor; and

[0044] at least one memory in communication with the processor;

[0045] The memory stores instructions that can be executed by the processor, and the instructions are executed by the processor to enable the electronic device to perform the method of automatically cutting the dental model and extracting the area of ​​interest patch in Example 1.

[0046] In the above description, many specific details are set forth in order to fully understand the present invention. However, the above description is only a preferred embodiment of the present invention. The present invention can be implemented in many other ways different from those described herein, so the present invention is not limited to the specific implementation disclosed above. At the same time, any person skilled in the art can make many possible changes and modifications to the technical solution of the present invention using the methods and technical contents disclosed above without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.

Claims

1. A method for automatically cutting out a scanned dental model to extract a region of interest patch, characterized by: The steps include: Step S1: Convert batch grid training data mesh1 into voxel data volum1; Step S2: Extend mesh1 in step S1 downward by 3 units at the lowest point of the tooth crown and perform hole filling on the corresponding mesh data mesh2. Traverse each voxel point in step S1 and set the label value of the voxel point inside or on the surface of mesh2 to 1; Step S3: Perform network training on the data processed in steps S1 and S2 above; Step S4: perform reasoning on the new mesh data mesh3; Step S4 specifically includes: Step S41: converting new grid data into voxel data; Step S42: Apply the trained model in step S3 to perform network inference to obtain the inference result, i.e., the segmented mask; Step S43: meshing the mask result in step S42 to obtain mesh4; Step S44: Based on the new mesh data mesh3 and the mesh mesh4 in step S43, the distances between all vertices of mesh4 and all vertices in mesh3 are calculated, and the vertex closest to each vertex of mesh4 in mesh3 is calculated to obtain the retained main mesh data mesh5; Step S45: Calculate the maximum boundary line of mesh5; Step S46: Based on the maximum boundary in step S45, curve cutting is performed on the original mesh data to obtain a cut mesh.

2. The method for automatically cutting out a dental model and extracting a region of interest patch according to claim 1, characterized in that: In step S45, by determining whether all half-edges of mesh5 are boundaries, the corresponding half-edge vertices are retrieved and the maximum boundaryline of mesh5 is calculated.

3. The method for automatically cutting out a dental model and extracting a region of interest patch according to claim 1, characterized in that: In step S46, the point closest to each vertex of the boundaryline in the original mesh is calculated, these points are combined into a curve, and Laplace smoothing is performed to implement curve cutting on the original mesh data.

4. A system for automatically cutting out regions of interest from scanned dental models, characterized by: include: at least one processor; as well as at least one memory in communication with the processor; The memory stores instructions that can be executed by the processor, and the instructions are executed by the processor to enable the electronic device to perform the method for automatically cutting and scanning dental models to extract regions of interest patches according to any one of claims 1 to 3.

Images