Method, program and system for processing the "black triangle" of jaw segmentation based on AI oral scanning and tooth segmentation

Abstract

The present invention relates to a method for processing "black triangles" in tooth segmentation based on AI dental scanning and tooth segmentation: S1: obtaining main tooth mesh data; S2: obtaining the maximum connected domain mesh of the S1 mesh data; S3: obtaining a circle of point set data on each tooth position automatically identified by AI tooth segmentation; S4: sequentially determining whether each tooth position in S3 is normal; S5: locating key points at the edges of two tooth positions; S6: obtaining a hole-filling point set for the black triangle area of the two abnormal tooth positions based on the four key points of the two tooth positions; S7: triangulating the obtained hole-filling point set of the black triangle area to obtain a facet of the black triangle area; S8: combining the facet in S7 with the facet after segmentation as the final tooth segmentation mesh. The method can automatically perform tooth segmentation and process the "black triangle" problem. The automated operation has high efficiency and is not limited by skill level. It is highly accurate and generates an accurate restoration plan. The digital processing ensures the accuracy and quality of the restoration.

Description

Technical Field

[0001] The present invention relates to the field of stomatology, and in particular to a method, program and system for processing a "black triangle" of tooth segmentation based on AI oral scanning and tooth segmentation. Background Art

[0002] The "black triangle" refers to the missing space between two teeth caused by tooth misalignment or tooth decay. This phenomenon allows dirt to easily accumulate between teeth, which not only affects oral aesthetics but can also lead to problems such as bad breath and tooth decay. Traditional methods typically require manual cutting and processing, which requires a long time and highly precise skills. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to design a method, program and system for automatic post-processing of the "black triangle" of jaw segmentation based on AI oral scanning and tooth segmentation. Based on AI oral scanning and tooth segmentation technology, a high-precision tooth model can be obtained, and the processing of jaw segmentation is more accurate and convenient. Then, computer graphics knowledge can be used to analyze and process the obtained tooth model to identify the "black triangle" phenomenon between teeth, thereby solving existing technical problems.

[0004] To solve the above technical problems, the method for processing the "black triangle" of jaw segmentation based on AI mouth scanning and tooth segmentation of the present invention includes the following steps:

[0005] Step S1: Obtain the subject's dental mesh data (maxillary / mandibular);

[0006] Step S2: Obtain the grid of the largest connected domain of the grid data in step S1;

[0007] Step S3: Obtain a circle of point sets on each tooth position automatically identified by AI tooth separation {P FDI} data, FDI corresponds to the tooth position number, point set {P FDI}Used for cutting and restoring tooth models;

[0008] Step S4: sequentially determining whether each tooth position in step S3 is normal, including the following steps:

[0009] Step S41: Calculate each circle of point sets {P FDI}A collection of attached triangular mesh faces;

[0010] Step S42: Expanding the triangular mesh face set in step S41 to obtain an expanded triangular mesh face set;

[0011] Step S43: Eliminate the triangular mesh facet set in step S42 from the maximum connected domain mesh in step S2;

[0012] Step S44: Determine whether the mesh connectivity domain after step S43 is equal to 1. If it is equal to 1, it is an abnormal situation. The mesh surface is connected, that is, the cutting line point set generated by AI cannot separate a single tooth from the entire dental model.

[0013] Step S45: If the result of step S44 is abnormal, the abnormal two tooth point sets are pressed into the set {P sp1}、{P sp2 If there is no common point set between the two tooth positions, the situation of no common point set should be treated as the situation of having a common point set;

[0014] Step S46: further determine the abnormal tooth position in step S45: whether there are two point sets with close spatial positions of the tooth positions, and set a distance threshold;

[0015] Step S47: Post-process the point sets of the two tooth position circles that are closer in step S46 to obtain a new dividing line point set {P new1}、{P new2}, the two teeth are based on the new dividing line point set {P new1}、{P new2} for segmentation;

[0016] Step S5: Locate the key points at the edges of the two teeth, with a total of four key points P on the two teeth. keystart1 、P keyend1 、P keystart2 、P keyend2 ;

[0017] Step S6: Obtaining a set of filling points for the black triangle areas of abnormal tooth position 1 and abnormal tooth position 2 based on the four key points of the two tooth positions in step S5;

[0018] Step S7: triangulate the black triangle area based on the hole filling point set obtained in step S6 to obtain a facet of the black triangle area;

[0019] Step S8: Combine the face patch in step S7 and the face patch segmented in step S47 as the final segmented mesh of the tooth.

[0020] Furthermore, in step S45, processing the non-common point set as the existing common point set includes the following steps:

[0021] Step S451: Find the two nearest points of the tooth position dividing line point set, passing through the midpoint P of the two points mid , perpendicular to the direction of the line connecting the two points and the direction of tooth root growth, symmetrically generate ray beams at equal intervals along the opposite direction of tooth root growth, and generate point sets {P com}, until the distance between the ray source and the intersection point satisfies the convex function characteristic, and the intersection point with the smallest distance is in the middle;

[0022] Step S452: traverse the cutting line point set {P sp1}、{P sp2} and the midpoint P of the two nearest points in step S451 mid Construct a vector and select the point where the angle between the vector and the buccal and lingual sides is closest to 90 degrees. Generate points P1 and P3 on the labial side of the two tooth positions and points P2 and P4 on the lingual side. Remove the point set of the dividing line between the two points belonging to the tooth position and closer to the midpoint.

[0023] Step S453: The two points P1 and P3 on the labial side of the tooth are connected to the starting point P generated on the same side. start Connect the lines, take points at equal intervals, and adsorb them to the nearest vertex on the mesh surface; similarly, connect the two points P2 and P4 on the lingual side of the tooth to the end point P end Connect the lines and take points at equal distances, adsorbing them to the nearest vertex on the mesh surface; merge the point set of the cutting line processed in step S452 with the point set generated by the connecting lines, and finally transform the point set without a common cutting line into a point set with a common cutting line, and then proceed to the subsequent operations;

[0024] Furthermore, in step S451, the point set must satisfy the requirement that there are 5 intersection points on the left and 5 intersection points on the right of the minimum point, {P com The point set is sorted from labial to lingual.

[0025] Furthermore, step S47 specifically includes the following steps:

[0026] Step S471: Find the common point set in the two tooth position circle point sets {P com};

[0027] Step S472: Obtain the starting point P of the common point set start and the end point P end ;

[0028] Step S473: Set a reference vector, and take the opposite direction of the z-axis direction of the dental model, which is the direction of root growth;

[0029] Step S474: traverse the abnormal tooth position point set {P sp1}、{P sp2} Construct a vector with the starting point in step S472, calculate the projection length of the vector in the opposite direction of root growth, and obtain the maximum value h max ;

[0030] Step S475: Starting from the starting point P in step S472 start As the initial point, search for points connected to the starting point along the opposite direction of the tooth root, calculate the root mean square curvature of all connected points, and take the point corresponding to the maximum root mean square curvature as the next point of the starting point; and so on, iterate and search until the initial starting point Pstart The projection length of the connected point after iteration in the root growth direction exceeds h max , the iteration ends;

[0031] Step S476: Similarly, the end point P in step S472 is used. end As the starting point, search along the opposite direction of the tooth root to find the end point P end Connected points, calculate the root mean square curvature of all connected points, take the point corresponding to the maximum root mean square curvature as the next point of the starting point; and so on, iterate and find it until the initial starting point P end The projection length of the connected point after iteration in the root growth direction exceeds h max , the iteration ends;

[0032] Step S477: Merge the two sets of points found on the lingual side and labial side in steps S476 to S477 to form a common point set of two adjacent teeth. newcom}, the order is consistent;

[0033] Step S478: Based on the initial tooth position point set {P sp1}、{P sp2} and the new common point set of the two tooth positions {P newcom} are merged to obtain a new set of cutting line points {P new1}、{P new2}, the two teeth are based on the new dividing line point set {P new1}、{P new2} for splitting.

[0034] Furthermore, step S5 specifically includes the following steps:

[0035] Step S51: Use the new common point set {P newcom} Fitted into a ray, the direction is along the growth direction of the tooth root;

[0036] Step S52: Calculate the intersection point P between the ray and the dental model cross ;

[0037] Step S53: Calculate the new edge point set of the tooth position {P new1}、{P new2} The distance from the intersection point P in step S52 cross The closest point and record the shortest distance d min ;

[0038] Step S54: traverse the new edge point set P of the tooth position new1}、{P new2 Each point in} and the intersection point P in step S52 crossConstruct a vector and calculate the distance between the two points; if the distance is the shortest distance d in step S53 min Within the weighted coefficient range, calculate the positive and negative value of the multiplication of the constructed vector and the model reference Y-axis vector point. Here, the Y-axis is the labial and lingual direction vector. Calculate the angle with the model reference Z-axis. Here, the Z-axis is the root growth direction. Take the point with the angle closest to 90 degrees as the key point. There are four key points P in the two tooth positions. keystart1 、P keyend1 、P keystart2 、P keyend2 .

[0039] Furthermore, in step S6, connect P start With P keystart1 , P end With P keyend1 , take the point closest to the initial jaw model grid as the intersection point, and use it together with the four points as the filling point set of the black triangle area of abnormal tooth position 1; connect P start With P keystart2 , P end With P keyend2 , and similarly take the point closest to the initial jaw model grid as the intersection point, and use it together with the four points as the filling point set for the black triangle area of abnormal tooth position 2.

[0040] Furthermore, AI tooth segmentation includes the following steps:

[0041] Step 1: Obtain the patient's three-dimensional digital dental model, which can be obtained by directly scanning the patient's mouth or by scanning a physical model (such as a plaster model);

[0042] Step 2: 3D digital dental preprocessing: Use open source software such as MeshLab to crop the 3D dental data, retain the teeth and surrounding areas, and repair the model to remove unreasonable surfaces (such as self-intersections and small channels).

[0043] Step 3: 3D digital jaw segmentation: This application uses deep learning to segment the 3D digital jaw. Each vertex, vertex normal, and vertex curvature of the 3D facet are used as deep learning inputs. After convolution and fully connected layers, the classification information of each vertex is finally output. The classification information of each facet is determined based on the classification information of the three vertices of each facet, and the maximum connected domain of each tooth is found. The boundary of the maximum connected domain is found as the tooth segmentation line;

[0044] Step 4: Tooth segmentation: Perform Gaussian smoothing on the segmentation line, remove the facets within the segmentation line, and replace the edge points with the smoothed edge line. This completes the AI tooth segmentation process.

[0045] The present invention also provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, the computer device executes the aforementioned method for processing the "black triangle" of tooth segmentation based on AI oral scanning.

[0046] The present invention also provides a system for processing the "black triangle" of tooth segmentation based on AI mouth scanning and tooth segmentation, comprising:

[0047] at least one processor; and

[0048] at least one memory in communication with the processor;

[0049] 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 execute the aforementioned method for processing the "black triangle" of tooth segmentation based on AI oral scanning and tooth segmentation.

[0050] Beneficial effects of the present invention:

[0051] 1. Automation: This technology can quickly and accurately scan the patient's oral cavity and generate a high-precision tooth model, eliminating the steps of manual cutting and processing and improving work efficiency.

[0052] 2. High accuracy: Based on computer vision technology, the tooth model can be analyzed and processed to identify the "black triangle" problem, thereby generating an accurate restoration plan.

[0053] 3. Digital processing: This technology can use digital processing technology to process and produce restorations, thereby ensuring the accuracy and quality of the restorations.

[0054] 4. Not restricted by skill level: Traditional manual processing requires a high level of skill and operating experience, while the processing method based on AI oral scanning and tooth segmentation technology allows people who do not have relevant skills to perform this operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The specific embodiments of the present invention will be further explained below with reference to the accompanying drawings.

[0056] Figure 1 This is a schematic diagram of the comparison before and after three-dimensional dental pretreatment;

[0057] Figure 2 This is an example diagram of tooth dividing lines;

[0058] Figure 3 This is a schematic diagram of the tooth segmentation results;

[0059] Figure 4 This is the "black triangle" processing flow chart;

[0060] Figure 5 This is a schematic diagram of the original data of the dental jaw;

[0061] Figure 6 A schematic diagram of the jaw with dividing line point data;

[0062] Figure 7 Schematic diagram of the dental jaw with part of the mesh deleted;

[0063] Figure 8 This is a diagram of the tooth dividing line data on both sides of the "black triangle";

[0064] Figure 9 Schematic diagram of treating no common point set between two teeth as having a common point set;

[0065] Figure 10 Schematic diagram of the common point set above the two teeth (labial side);

[0066] Figure 11 Schematic diagram of the common point set above the two teeth (lingual side);

[0067] Figure 12 Schematic diagram of the new dividing line of the tooth;

[0068] Figure 13 is the intersection point of the ray generated along the tooth root direction and the dental model;

[0069] Figure 14 Schematic diagram of the four key points of the jaw. DETAILED DESCRIPTION

[0070] In this embodiment, AI tooth separation includes the following steps:

[0071] Step 1: Obtain the patient's three-dimensional digital dental model, which can be obtained by directly scanning the patient's mouth or by scanning a physical model (such as a plaster model);

[0072] Step 2: 3D digital jaw preprocessing: Use open source software such as MeshLab to crop the 3D jaw data, retain the teeth and surrounding areas, and repair the model to remove unreasonable surfaces (such as self-intersections, small channels, etc.). Compare the before and after repair. Figure 1 As shown, Figure 1 The left side is the original data, and the right side is the repaired data;

[0073] Step 3: 3D digital jaw segmentation: This application uses deep learning to segment the 3D digital jaw. Each vertex, vertex normal and vertex curvature of the 3D facet are used as deep learning inputs. After convolution and full connection layers, the classification information of each vertex is finally output. The classification information of each facet is determined based on the classification information of the three vertices of each facet, and the maximum connected domain of each tooth is found. The boundary of the maximum connected domain is found as the tooth dividing line. The dividing line is as follows: Figure 2 As shown;

[0074] Step 4: Teeth segmentation: Gaussian smoothing is performed on the segmentation line, the facets within the segmentation line are taken out, and the edge points are replaced with the smoothed edge lines. This completes the AI teeth segmentation process. The teeth segmentation results are as follows: Figure 3 shown.

[0075] Combine Figure 4 The method for processing the "black triangle" of jaw segmentation based on AI mouth scanning and tooth segmentation of the present invention comprises the following steps:

[0076] Step S1: Obtain the subject's jaw mesh data (maxillary / mandibular), such as Figure 5 As shown;

[0077] Step S2: Obtain the grid of the largest connected domain of the grid data in step S1;

[0078] Step S3: Obtain a circle of point sets on each tooth position automatically identified by AI tooth separation {P FDI} data, FDI corresponds to the tooth position number, point set {P FDI}Used for cutting and restoring tooth models, such as Figure 6 As shown;

[0079] Step S4: sequentially determining whether each tooth position in step S3 is normal, including the following steps:

[0080] Step S41: Calculate each circle of point sets {P FDI}A collection of attached triangular mesh faces;

[0081] Step S42: Expand the triangular mesh face set in step S41 by 1 to 2 circles to obtain an expanded triangular mesh face set;

[0082] Step S43: Eliminate the triangular mesh facet set in step S42 from the maximum connected domain mesh in step S2, such as Figure 7 As shown;

[0083] Step S44: Determine whether the mesh connectivity domain after step S43 is equal to 1. If it is equal to 1, it is an abnormal situation. The mesh surface is connected, that is, the cutting line point set generated by AI cannot separate a single tooth from the entire dental model.

[0084] Step S45: If the result of step S44 is abnormal, the abnormal two tooth point sets are pressed into the set {P sp1}、{P sp2}; If there is no common point set between the two teeth, such as Figure 8 As shown, the case where there is no common point set needs to be treated as the case where there is a common point set, such as Figure 9 As shown; processing a set without common points into a set with common points includes the following steps:

[0085] Step S451: Find the two nearest points of the tooth position dividing line point set, passing through the midpoint P of the two points mid , perpendicular to the direction of the line connecting the two points and the direction of tooth root growth, symmetrically generate ray beams at equal intervals along the opposite direction of tooth root growth, and generate point sets {P com}, until the distance between the ray source point and the intersection point satisfies the convex function characteristic, and the intersection point with the smallest distance is in the middle; the point set needs to satisfy the requirement that there are 5 intersection points on the left and 5 intersection points on the right of the point with the smallest distance, {P com The point set is sorted from labial to lingual, P start For the point set {P com} starting point, P end For the point set {P com}end point.

[0086] Step S452: traverse the cutting line point set {P sp1}、{P sp2} and the midpoint P of the two nearest points in step S451 mid Construct a vector and select the point where the angle between the vector and the buccal and lingual sides is closest to 90 degrees. Generate points P1 and P3 on the labial side of the two tooth positions and points P2 and P4 on the lingual side. Remove the point set of the dividing line between the two points belonging to the tooth position and closer to the midpoint.

[0087] Step S453: The two points P1 and P3 on the labial side of the tooth are connected to the starting point P generated on the same side. start Connect the lines, take points at equal intervals, and adsorb them to the nearest vertex on the mesh surface; similarly, connect the two points P2 and P4 on the lingual side of the tooth to the end point P end Connect the lines and take points at equal distances, adsorbing them to the nearest vertex on the mesh surface; merge the point set of the cutting line processed in step S452 with the point set generated by the connecting lines, and finally transform the point set without a common cutting line into a point set with a common cutting line, and then proceed to the subsequent operations;

[0088] Step S46: further determine the abnormal tooth position in step S45: whether there are two point sets with close spatial positions of the tooth positions, and set a distance threshold. In this embodiment, the distance threshold is set to 20 mm;

[0089] Step S47: Post-process the point sets of the two tooth position circles that are closer in step S46 to obtain a new dividing line point set {P new1}、{P new2}, the two teeth are based on the new dividing line point set {P new1}、{P new2} for segmentation; specifically including the following steps:

[0090] Step S471: Find the common point set in the two tooth position circle point sets {P com};

[0091] Step S472: Obtain the starting point P of the common point set start and the end point P end ;

[0092] Step S473: Set a reference vector, and take the opposite direction of the z-axis direction of the dental model, which is the direction of root growth;

[0093] Step S474: traverse the abnormal tooth position point set {P sp1}、{P sp2} Construct a vector with the starting point in step S472, calculate the projection length of the vector in the opposite direction of root growth, and obtain the maximum value h max ;

[0094] Step S475: Starting from the starting point P in step S472 start As the initial point, search along the opposite direction of the tooth root to find the starting point P start Connected points, calculate the root mean square curvature of all connected points, take the point corresponding to the maximum root mean square curvature as the next point of the starting point; and so on, iterate and find it until the initial starting point P start The projection length of the connected point after iteration in the root growth direction exceeds h max , then the iteration ends, if Figure 10 As shown;

[0095] Step S476: Similarly, the end point P in step S472 is used. end As the starting point, search along the opposite direction of the tooth root to find the end point P end Connected points, calculate the root mean square curvature of all connected points, take the point corresponding to the maximum root mean square curvature as the next point of the starting point; and so on, iterate and find it until the initial starting point P end The projection length of the connected point after iteration in the root growth direction exceeds h max , then the iteration ends, if Figure 11 As shown;

[0096] Step S477: Merge the two sets of points found on the lingual side and labial side in steps S475 to S476 to form a common point set of two adjacent teeth. newcom}, the order is consistent;

[0097] Step S478: Based on the initial tooth position point set {P sp1}、{P sp2} and the new common point set of the two tooth positions {P newcom} are merged to obtain a new set of cutting line points {P new1}、{P new2},like Figure 12 As shown, the two tooth positions are based on the new dividing line point set {P new1}、{P new2} for splitting.

[0098] Step S5: Locate the key points at the edges of the two teeth, with a total of four key points P on the two teeth. keystart1 、P keyend1 、P keystart2 、P keyend2 ; Specifically include the following steps:

[0099] Step S51: Use the new common point set {P newcom} Fitted into a ray, the direction is along the growth direction of the tooth root;

[0100] Step S52: Calculate the intersection point P between the ray and the dental model cross ,like Figure 13 As shown;

[0101] Step S53: Calculate the new edge point set of the tooth position {P new1}、{P new2} The distance from the intersection point P in step S52 cross The closest point and record the shortest distance d min ;

[0102] Step S54: traverse the new edge point set P of the tooth position new1}、{P new2 Each point in} and the intersection point P in step S52 cross Construct a vector and calculate the distance between the two points; if the distance is the shortest distance d in step S53 min Within the weighted coefficient range, in this embodiment, the weighted coefficient is 2, and subsequent operations are performed to calculate the positive and negative value of the multiplication of the constructed vector and the model reference Y-axis vector point. Here, the Y-axis is the labial and lingual direction vector, and the angle with the model reference Z-axis is calculated. Here, the Z-axis is the root growth direction. The point with the angle closest to 90 degrees is taken as the key point. There are four key points P in the two tooth positions. keystart1 、P keyend1、P keystart2 、P keyend2 ,like Figure 14 shown.

[0103] Step S6: Obtain the hole filling point set of the black triangle area of abnormal tooth position 1 and abnormal tooth position 2 according to the four key points of the two tooth positions in step S5: Connect P start With P keystart1 , P end With P keyend1 , take the point closest to the initial jaw model grid as the intersection point, and use it together with the four points as the filling point set of the black triangle area of abnormal tooth position 1; connect P start With P keystart2 , P end With P keyend2 , and similarly take the point closest to the initial jaw model grid as the intersection point, and use it together with the four points as the filling point set for the black triangle area of abnormal tooth position 2.

[0104] Step S7: triangulate the black triangle area based on the hole filling point set obtained in step S6 to obtain a facet of the black triangle area;

[0105] Step S8: Combine the face patch in step S7 and the face patch segmented in step S47 as the final segmented mesh of the tooth.

[0106] This embodiment also provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, the computer device executes the aforementioned method for processing the "black triangle" of tooth segmentation based on AI oral scanning.

[0107] This embodiment also provides a system for processing the "black triangle" of tooth segmentation based on AI oral scanning and tooth segmentation, including:

[0108] at least one processor; and

[0109] at least one memory in communication with the processor;

[0110] 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 execute the aforementioned method for processing the "black triangle" of tooth segmentation based on AI oral scanning and tooth segmentation.

[0111] 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 processing the "black triangle" of jaw segmentation based on AI oral scanning and tooth segmentation, characterized by: The steps include: Step S1: Obtaining the subject's dental mesh data; Step S2: Obtain the grid of the largest connected domain of the grid data in step S1; Step S3: Obtain a circle of point sets on each tooth position automatically identified by AI tooth separation {P FDI } data, FDI corresponds to the tooth position number, point set {P FDI }Used for cutting and restoring tooth models; Step S4: sequentially determining whether each tooth position in step S3 is normal, including the following steps: Step S41: Calculate each circle of point sets {P FDI }A collection of attached triangular mesh faces; Step S42: Expanding the triangular mesh face set in step S41 to obtain an expanded triangular mesh face set; Step S43: Eliminate the triangular mesh facet set in step S42 from the maximum connected domain mesh in step S2; Step S44: Determine whether the connected domain of the mesh processed in step S43 is equal to 1. If it is equal to 1, it is an abnormal situation, and the mesh facets are connected, that is, the cutting line point set generated by AI cannot separate a single tooth from the entire dental model; Step S45: If the result of step S44 is abnormal, the abnormal two tooth point sets are pressed into the set {P sp1 }、{P sp2 If there is no common point set between the two tooth positions, the situation of no common point set should be treated as the situation of having a common point set; Step S46: further determine the abnormal tooth position in step S45: whether there are two point sets with close spatial positions of the tooth positions, and set a distance threshold; Step S47: Post-process the point sets of the two tooth position circles that are closer in step S46 to obtain a new dividing line point set {P new1 }、{P new2 }, the two teeth are based on the new dividing line point set {P new1 }、{P new2 } for segmentation; Step S5: Locate the key points at the edges of the two teeth, with a total of four key points P on the two teeth. keystart1 、P keyend1 、P keystart2 、P keyend2 ; Step S6: Obtaining a set of filling points for the black triangle areas of abnormal tooth position 1 and abnormal tooth position 2 based on the four key points of the two tooth positions in step S5; Step S7: triangulate the black triangle area based on the hole filling point set obtained in step S6 to obtain a facet of the black triangle area; Step S8: Combine the face patch in step S7 and the face patch segmented in step S47 as the final segmented mesh of the tooth.

2. The method for processing the "black triangle" of dental segmentation based on AI oral scanning and tooth segmentation according to claim 1 is characterized by: In step S45, processing the non-common point set as the existing common point set includes the following steps: Step S451: Find the two nearest points of the tooth position dividing line point set, passing through the midpoint P of the two points mid , perpendicular to the direction of the line connecting the two points and the direction of tooth root growth, symmetrically generate ray beams at equal intervals along the opposite direction of tooth root growth, and generate point sets {P com }, until the distance between the ray source and the intersection point satisfies the convex function characteristic, and the intersection point with the smallest distance is in the middle, {P com The point set is sorted from labial to lingual, P start For the point set {P com } starting point, P end For the point set {P com }end point; Step S452: traverse the cutting line point set {P sp1 }、{P sp2 } and the midpoint P of the two nearest points in step S451 mid Construct a vector and select the point where the angle between the vector and the buccal and lingual sides is closest to 90 degrees. Generate points P1 and P3 on the labial side of the two tooth positions and points P2 and P4 on the lingual side. Remove the point set of the dividing line between the two points belonging to the tooth position and closer to the midpoint. Step S453: The two points P1 and P3 on the labial side of the tooth are aligned with the starting point P start Connect the lines, take points at equal intervals, and adsorb them to the nearest vertex on the mesh surface; similarly, connect the two points P2 and P4 on the lingual side of the tooth to the end point P end Connect the lines and take points at equal distances, adsorbing them to the nearest vertex on the mesh surface; merge the cutting line point set processed in step S452 with the point set generated by the connecting lines, and finally transform the point set without a common cutting line into a point set with a common cutting line.

3. The method for processing the "black triangle" of dental segmentation based on AI dental scanning and segmentation according to claim 2 is characterized by: In step S451, the point set must have 5 intersection points on the left and 5 intersection points on the right of the minimum point, {P com The point set is sorted from labial to lingual.

4. The method for processing the "black triangle" of dental segmentation based on AI dental scanning and segmentation according to claim 1 is characterized by: Step S47 specifically includes the following steps: Step S471: Find the common point set in the two tooth position circle point sets {P com }; Step S472: Obtain the starting point P of the common point set start and the end point P end ; Step S473: Set a reference vector, and take the opposite direction of the z-axis direction of the dental model, which is the direction of root growth; Step S474: traverse the abnormal tooth position point set {P sp1 }、{P sp2 } Construct a vector with the starting point in step S472, calculate the projection length of the vector in the opposite direction of root growth, and obtain the maximum value h max ; Step S475: Starting from the starting point P in step S472 start As the initial point, search for points connected to the starting point along the opposite direction of the tooth root, calculate the root mean square curvature of all connected points, and take the point corresponding to the maximum root mean square curvature as the next point of the starting point; and so on, iterate and search until the initial starting point P start The projection length of the connected point after iteration in the root growth direction exceeds h max , the iteration ends; Step S476: Similarly, the end point P in step S472 is used. end For the initial point, search along the opposite direction of the tooth root to find the end point P end Connected points, calculate the root mean square curvature of all connected points, take the point corresponding to the maximum root mean square curvature as the next point of the starting point; and so on, iterate and find it until the initial starting point P end The projection length of the connected point after iteration in the root growth direction exceeds h max , the iteration ends; Step S477: Merge the two sets of points found on the lingual side and labial side in steps S475 to S476 to form a common point set of two adjacent teeth. newcom }, the order moves towards consistency; Step S478: Based on the initial tooth position point set {P sp1 }、{P sp2 } and the new common point set of the two tooth positions {P newcom } are merged to obtain a new set of cutting line points {P new1 }、{P new2 }, the two teeth are based on the new dividing line point set {P new1 }、{P new2 } for splitting.

5. The method for processing the "black triangle" of dental segmentation based on AI dental scanning and segmentation according to claim 1 is characterized by: Step S5 specifically includes the following steps: Step S51: Use the new common point set {P newcom } Fitted into a ray, the direction is along the growth direction of the tooth root; Step S52: Calculate the intersection point P between the ray and the dental model cross ; Step S53: Calculate the new edge point set of the tooth position {P new1 }、{P new2 } The distance from the intersection point P in step S52 cross The closest point and record the shortest distance d min ; Step S54: traverse the new edge point set P of the tooth position new1 }、{P new2 Each point in} and the intersection point P in step S52 cross Construct a vector and calculate the distance between the two points; if the distance is the shortest distance d in step S53 min Within the weighted coefficient range, calculate the positive and negative value of the multiplication of the constructed vector and the model reference Y-axis vector point. Here, the Y-axis is the labial and lingual direction vector. Calculate the angle with the model reference Z-axis. Here, the Z-axis is the root growth direction. Take the point with the angle closest to 90 degrees as the key point. There are four key points P in the two tooth positions. keystart1 、P keyend1 、P keystart2 、P keyend2 .

6. The method for processing the "black triangle" of jaw segmentation based on AI-assisted dental scanning according to claim 1, characterized in that: In step S6, connect P start With P keystart1 , P end With P keyend1 , take the point closest to the initial jaw model grid as the intersection point, and use it together with the four points as the filling point set of the black triangle area of abnormal tooth position 1; connect P start With P keystart2 , P end With P keyend2 , and similarly take the point closest to the initial jaw model grid as the intersection point, and use it together with the four points as the filling point set for the black triangle area of abnormal tooth position 2.

7. The method for processing the "black triangle" of dental segmentation based on AI dental scanning and segmentation according to claim 1 is characterized by: AI tooth segmentation includes the following steps: Step 1: Obtain the patient's three-dimensional digital dental model; Step 2: 3D digital dental preprocessing: crop the 3D dental data, retain the teeth and surrounding areas, repair the model, and remove unreasonable surfaces; Step 3: 3D digital jaw segmentation: Each vertex, vertex normal, and vertex curvature of the 3D facet is used as deep learning input. After passing through convolution and fully connected layers, the classification information of each vertex is finally output. The classification information of each facet is determined based on the classification information of the three vertices of each facet, and the maximum connected domain of each tooth is found. The boundary of the maximum connected domain is found as the tooth segmentation line; Step 4: Tooth segmentation: Perform Gaussian smoothing on the segmentation line, remove the facets within the segmentation line, and replace the edge points with the smoothed edge line. This completes the AI tooth segmentation process.

8. A computer program product, characterized in that: The computer program product includes computer instructions, which, when executed by a processor, enable a computer device to execute the method for processing the "black triangle" of tooth segmentation based on AI oral scanning and tooth segmentation as described in any one of claims 1 to 7.

9. A system for processing the "black triangle" of jaw segmentation based on AI-powered oral scanning and tooth segmentation, characterized by: include: at least one processor; as well as at least one memory in communication with the processor; Wherein, 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 execute the method for processing the "black triangle" of maxillary segmentation based on AI oral scanning and tooth segmentation as described in any one of claims 1-7.

Images