Three-dimensional tooth model local coordinate system prediction method, program, storage medium, system and device

Abstract

The invention relates to a three-dimensional tooth model local coordinate system prediction method. The method comprises the following steps: S1, acquiring a three-dimensional digital model of a single tooth; s2, sampling a single three-dimensional tooth to obtain a three-dimensional point set of the tooth model; s3, predicting a three-dimensional vertex by using the trained deep learning model; s4, fitting a linear vector and generating a local coordinate system vector; s5, carrying out vector cross multiplication operation on the x vector and the z vector so as to construct a y vector of the tooth local coordinate system; and S6, carrying out cross multiplication on the y vector and the z vector to obtain a new x vector, and completing the construction of the tooth local coordinate system. According to the method, the local coordinate system of the teeth can be automatically and accurately generated, the digital processing precision in the fields of tooth restoration, orthodontic treatment and the like is improved, meanwhile, the time for making and implementing a treatment plan is shortened, and the working efficiency is improved. Manual rule setting is not needed, the limitation of a traditional algorithm is avoided, and the processing capacity of a complex tooth model is improved.

Description

Technical Field

[0001] The present invention relates to the field of oral computer-aided design technology, and in particular to a method, program, storage medium, system and device for predicting a local coordinate system of a three-dimensional tooth model. Background Art

[0002] In the dental computer-aided design system, the local coordinate system of the tooth can represent the posture of the tooth in three-dimensional space. The local coordinate system is particularly important for key diagnosis and treatment operations such as computer-aided orthodontic planning.

[0003] Manually setting the local coordinate system is limited by large subjective differences, low efficiency, and time-consuming and labor-intensive. The problem is particularly serious when dealing with complex cases or large-scale data, and the cost of optimization and adjustment is high. Summary of the invention

[0004] The technical problem to be solved by the present invention is to design a method, program, storage medium, system and device for predicting the local coordinate system of a three-dimensional tooth model, which are used to automatically generate the local coordinate system of the tooth and solve the existing technical problems.

[0005] In order to solve the above technical problems, the present invention provides a method for predicting a local coordinate system of a three-dimensional tooth model, which specifically includes the following steps:

[0006] Step S1: Obtain a three-dimensional digital model of a single tooth.

[0007] Step S2: point sampling and tooth point set acquisition: sampling a single three-dimensional tooth to obtain a three-dimensional point set of the tooth model.

[0008] Step S3: Predict three-dimensional vertices using the trained deep learning model.

[0009] Step S4: Fitting the straight line vector and generating the local coordinate system vector: further fitting the straight line vector through the predicted three-dimensional vertices, and generating the x vector and z vector of the local coordinate system accordingly.

[0010] Step S5: construct a complete local coordinate system by cross product: perform a vector cross product operation on the x vector and the z vector to construct the y vector of the local coordinate system of the tooth.

[0011] Step S6: Use the y vector to cross-multiply the z vector to obtain a new x vector, completing the construction of the tooth local coordinate system.

[0012] Furthermore, in step S3, the PointNet++ network is used to extract point set features, and a multi-layer perceptron is added after the feature extraction module of PointNet++, so that the model can directly regress the coordinates of 2N three-dimensional points.

[0013] Furthermore, in step S4, the first N points are used to fit the x vector of the local coordinate system, and the last N points are used to fit the z vector of the local coordinate system. The fitting method includes the following steps:

[0014] Step 1: Select the first 3D point as the reference point, denoted as P0;

[0015] Step 2: Initialize a three-dimensional vector to store the accumulated direction vector;

[0016] Step 3: Calculate the offset vectors of the remaining N-1 points relative to the center point P0, and add them up to get the sum of the offset vectors of all points;

[0017] Step 4: Divide the accumulated offset vector by N-1 to obtain the average direction vector of each point relative to the center point;

[0018] Step 5: Normalize the average offset vector to obtain a unit vector with consistent direction and length of 1.

[0019] Furthermore, the fitting of the x vector includes the following steps:

[0020] Step 1: Select the first 3D point from the first N points as the reference point, denoted as P a =(x a ,y a , z a );

[0021] Step 2: Initialize a three-dimensional vector is (0,0,0), used to store the accumulated direction vector;

[0022] Step 3: For each point P i , where i = 1, 2, 3, ... N-1, calculate its relative to the center point P a The offset vector is added to the vector After the loop, the sum of all point offset vectors is obtained.

[0023]

[0024] Step 4: Calculate the average direction vector of each point relative to the center point

[0025] Step 5: Obtain The unit vector represents the average direction of the remaining N-1 points, ‖·‖ is the norm.

[0026] Furthermore, the fitting of the z vector includes the following steps:

[0027] Step 1: Select the first 3D point from the last N points as the reference point, denoted as P b =(x b ,y b , z b );

[0028] Step 2: Initialize a three-dimensional vector is (0,0,0), used to store the accumulated direction vector;

[0029] Step 3: For each point P i , where i = 1, 2, 3, ... N-1, calculate its relative to the center point P b The offset vector is added to the vector After the loop, the sum of all point offset vectors is obtained.

[0030]

[0031] Step 4: Calculate the average direction vector of each point relative to the center point

[0032] Step 5: Obtain The unit vector represents the average direction of the remaining N-1 points, ‖·‖ is the norm.

[0033] The present invention also provides a three-dimensional tooth model local coordinate system prediction system, comprising:

[0034] Tooth model acquisition module: used to obtain the three-dimensional digital model of a single tooth as the basis for subsequent processing.

[0035] Sampling and point set acquisition module: used to perform vertex sampling on the three-dimensional tooth digital model to obtain a set of three-dimensional vertices on the tooth surface, namely the three-dimensional tooth point set.

[0036] Deep learning module: used to input a 3D tooth point set and predict 3D vertices.

[0037] Local coordinate system fitting module: used to further fit the straight line vector through the predicted three-dimensional vertices, and generate the x vector and z vector of the local coordinate system based on this.

[0038] Local coordinate system construction module: used to perform vector cross multiplication operation on the x vector and the z vector to construct the y vector of the local coordinate system of the tooth; and to cross multiply the y vector with the z vector to obtain a new x vector to complete the construction of the local coordinate system of the tooth.

[0039] Furthermore, it also includes a control module, which controls the tooth model acquisition module, the sampling and point set acquisition module, the deep learning module, the local coordinate system fitting module and the local coordinate system construction module to execute each step according to the aforementioned three-dimensional tooth model local coordinate system prediction method to realize the functions of the corresponding modules.

[0040] The present invention also provides a three-dimensional tooth model local coordinate system prediction device, comprising:

[0041] at least one processor; and

[0042] at least one memory in communication with the processor;

[0043] The memory stores instructions executable by the processor, and the instructions are executed by the processor to enable the device to perform the aforementioned three-dimensional tooth model local coordinate system prediction method.

[0044] The present invention also provides a computer-readable storage medium having computer-executable instructions stored thereon. When a processor executes the computer-executable instructions, the above-mentioned method for predicting the local coordinate system of a three-dimensional tooth model is implemented.

[0045] The present invention also provides a computer program product, comprising computer instructions, which, when executed by a processor, enable a computer device to execute the above-mentioned method for predicting a local coordinate system of a three-dimensional tooth model.

[0046] Beneficial effects of the present invention:

[0047] (1) Improving the accuracy and efficiency of tooth model processing: By designing an innovative three-dimensional tooth model local coordinate system prediction method, the present invention can automatically and accurately generate the local coordinate system of the tooth, thereby improving the accuracy of digital processing in the fields of tooth restoration, orthodontic treatment, etc., while shortening the time for treatment plan formulation and implementation, and improving work efficiency.

[0048] (2) Simplify the operation process and lower the technical threshold: Compared with the traditional method of manually setting or relying on complex algorithms to generate the local coordinate system of the tooth, the present invention greatly simplifies the operation process through an intelligent prediction program, reduces the professional skill requirements for operators, and enables more medical institutions to adopt and benefit from this technology.

[0049] (3) Strong versatility: The present invention realizes intelligent prediction of the local coordinate system of teeth through deep learning neural networks, can make adaptive adjustments based on the morphological characteristics of teeth, and automatically learns features from data through multi-layer neural networks. It does not require manual setting of rules, thus avoiding the limitations of traditional algorithms and improving the processing capabilities of complex tooth models. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The specific implementation of the present invention will be further explained below in conjunction with the accompanying drawings.

[0051] Figure 1 It is a flow chart of the method for predicting the local coordinate system of a three-dimensional tooth model of the present invention;

[0052] Figure 2 is a schematic diagram of a three-dimensional tooth model in the present invention;

[0053] Figure 3 is a schematic diagram of a tooth point set in the present invention;

[0054] Figure 4 A schematic diagram of three-dimensional vertices predicted by the deep learning model in the present invention;

[0055] Figure 5 A network structure diagram of the deep learning model in the present invention;

[0056] Figure 6 Schematic diagram of the x-axis and z-axis obtained by fitting in the present invention;

[0057] Figure 7 The local coordinate system finally generated by the present invention;

[0058] Figure 8 It is a block diagram of the local coordinate system prediction system of the three-dimensional tooth model of the present invention. DETAILED DESCRIPTION

[0059] Example 1

[0060] Combination Figure 1 The method for predicting the local coordinate system of the three-dimensional tooth model in this embodiment specifically includes the following steps:

[0061] Step S1: Obtain a three-dimensional digital model of a single tooth as a basis for subsequent processing. Figure 2 shown.

[0062] Step S2: point sampling and tooth point set acquisition: sampling a single three-dimensional tooth to obtain a three-dimensional point set of the tooth model.

[0063] Specifically, in this embodiment, vertex sampling is performed on the three-dimensional tooth digital model to obtain a set of three-dimensional vertices on the tooth surface, that is, a three-dimensional tooth point set.

[0064] Specifically, in this embodiment, a random sampling method is used for a single tooth. In this embodiment, 3000 points are sampled, and the tooth point set is as follows: Figure 3 As shown in the figure, the red part is the acquired tooth point set.

[0065] Step S3: Predicting three-dimensional vertices using the trained deep learning model. Using the trained deep learning artificial neural network, a series of three-dimensional vertices are predicted based on the above tooth point set. The three-dimensional vertices predicted by the deep learning model are as follows: Figure 4 As shown, the two red points in the figure are the predicted three-dimensional vertices.

[0066] Preferably, in step S3 of this embodiment, a PointNet++ network is used to extract point set features, and a multi-layer perceptron is added after the feature extraction module of PointNet++ so that the model can directly regress the coordinates of 2N three-dimensional points.

[0067] Specifically, this embodiment is based on PointNet++, and is modified in the network back propagation stage to directly implement the regression task through a multi-layer perceptron (MLP). This embodiment directly adds a multi-layer perceptron (fully connected layer) after the feature extraction module of PointNet++, so that the model can directly regress the coordinates of 40 three-dimensional points. The network model is as follows Figure 5 As shown, a vector with a dimension of 120 is finally output, which represents the coordinates of 40 three-dimensional points.

[0068] Step S4: Fitting the line vector and generating the local coordinate system vector: The line vector is further fitted through the predicted three-dimensional vertices, and the x-vector and z-vector of the local coordinate system are generated accordingly. The x-axis and z-axis obtained by fitting are as follows Figure 6 shown.

[0069] In this embodiment, preferably, in step S4, the first N points are used to fit the x vector of the local coordinate system, and the last N points are used to fit the z vector of the local coordinate system. The fitting method includes the following steps:

[0070] Step 1: Select the first 3D point as the reference point, denoted as P0;

[0071] Step 2: Initialize a three-dimensional vector to store the accumulated direction vector;

[0072] Step 3: Calculate the offset vectors of the remaining N-1 points relative to the center point P0, and add them up to get the sum of the offset vectors of all points;

[0073] Step 4: Divide the accumulated offset vector by N-1 to obtain the average direction vector of each point relative to the center point;

[0074] Step 5: Normalize the average offset vector to obtain a unit vector with consistent direction and length of 1.

[0075] Specifically, in this embodiment, the first 20 points are used to fit the x vector of the local coordinate system, and the last 20 points are used to fit the z vector of the local coordinate system.

[0076] The fitting of the x vector includes the following steps:

[0077] Step 1: Select the first 3D point as the reference point, denoted as P a =(x a ,y a , z a );

[0078] Step 2: Initialize a three-dimensional vector is (0,0,0), used to store the accumulated direction vector;

[0079] Step 3: For each point P i , where i = 1, 2, 3, ... 19, calculate its relative to the center point P a The offset vector is added to the vector After the loop, the sum of all point offset vectors is obtained.

[0080]

[0081] Step 4: Calculate the average direction vector of each point relative to the center point

[0082] Step 5: Obtain The unit vector represents the average direction of the remaining 19 points, ‖·‖ is the norm.

[0083] The fitting of the z vector includes the following steps:

[0084] Step 1: Select the first 3D point as the reference point, denoted as P b =(x b ,y b , z b );

[0085] Step 2: Initialize a three-dimensional vector is (0,0,0), used to store the accumulated direction vector;

[0086] Step 3: For each point P i , where i = 1, 2, 3, ... 19, calculate its relative to the center point P b The offset vector is added to the vector After the loop, the sum of all point offset vectors is obtained.

[0087]

[0088] Step 4: Calculate the average direction vector of each point relative to the center point

[0089] Step 5: Obtain The unit vector represents the average direction of the remaining 19 points, ‖·‖ is the norm.

[0090] The first of the 20 points is selected as the starting point, and the average offset of the remaining 19 points relative to this center point is calculated. Finally, the average offset vector is normalized to obtain a unit vector with a consistent direction and a length of 1. This unit vector describes the average direction of the remaining 19 points relative to the center point.

[0091] Step S5: construct a complete local coordinate system by cross product: perform a vector cross product operation on the x vector and the z vector to construct the y vector of the local coordinate system of the tooth.

[0092] At this time, the x vector and z vector are not perpendicular to each other, so the following steps are required:

[0093] Step S6: Use the y vector and the z vector to cross-multiply to obtain a new x vector, and complete the construction of the tooth local coordinate system. The final constructed local coordinate system is as follows Figure 7 shown.

[0094] Example 2

[0095] Combination Figure 8 This embodiment is a three-dimensional tooth model local coordinate system prediction system, which specifically includes the following modules:

[0096] Tooth model acquisition module: used to obtain the three-dimensional digital model of a single tooth as the basis for subsequent processing.

[0097] Sampling and point set acquisition module: used to perform vertex sampling on the three-dimensional tooth digital model to obtain a set of three-dimensional vertices on the tooth surface, namely the three-dimensional tooth point set.

[0098] Deep learning module: used to input a 3D tooth point set and predict 3D vertices.

[0099] Local coordinate system fitting module: used to further fit the straight line vectors through the predicted three-dimensional vertices, and generate the x vector and z vector of the local coordinate system based on them.

[0100] Local coordinate system construction module: used to perform vector cross multiplication operation on the x vector and the z vector to construct the y vector of the local coordinate system of the tooth; and to cross multiply the y vector with the z vector to obtain a new x vector to complete the construction of the local coordinate system of the tooth.

[0101] Preferably, this embodiment also includes a control module, which controls the tooth model acquisition module, the sampling and point set acquisition module, the deep learning module, the local coordinate system fitting module and the local coordinate system construction module to execute each step according to the three-dimensional tooth model local coordinate system prediction method in Example 1 to realize the functions of the corresponding modules.

[0102] Example 3

[0103] This embodiment is a device for predicting a local coordinate system of a three-dimensional tooth model, the device comprising:

[0104] at least one processor; and

[0105] at least one memory in communication with the processor;

[0106] The memory stores instructions that can be executed by the processor, and the instructions are executed by the processor to enable the device to execute the three-dimensional tooth model local coordinate system prediction method in Example 1.

[0107] Example 4

[0108] This embodiment is a computer-readable storage medium on which computer execution instructions are stored. When a processor executes the computer execution instructions, the three-dimensional tooth model local coordinate system prediction method in Embodiment 1 is implemented.

[0109] Example 5

[0110] This embodiment is a computer program product, comprising computer instructions, which, when executed by a processor, enable a computer device to execute the method for predicting a local coordinate system of a three-dimensional tooth model as in Embodiment 1.

[0111] Many specific details are described in the above description to facilitate a full understanding of 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 familiar with 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 embodiments according to 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 predicting a local coordinate system of a three-dimensional tooth model, characterized in that: The steps include: Step S1: obtaining a three-dimensional digital model of a single tooth; Step S2: point sampling and tooth point set acquisition: sampling a single three-dimensional tooth to obtain a three-dimensional point set of the tooth model; Step S3: predicting three-dimensional vertices using the trained deep learning model; Step S4: Fitting the straight line vector and generating the local coordinate system vector: further fitting the straight line vector through the predicted three-dimensional vertices, and generating the x vector and z vector of the local coordinate system accordingly; Step S5: constructing a complete local coordinate system by cross multiplication: performing a vector cross multiplication operation on the x vector and the z vector, thereby constructing the y vector of the local coordinate system of the tooth; Step S6: Use the y vector to cross-multiply the z vector to obtain a new x vector, completing the construction of the tooth local coordinate system.

2. The method for predicting a local coordinate system of a three-dimensional tooth model according to claim 1, characterized in that: In step S3, the PointNet++ network is used to extract point set features, and a multi-layer perceptron is added after the feature extraction module of PointNet++ so that the model can directly regress the coordinates of 2N three-dimensional points.

3. The method for predicting a local coordinate system of a three-dimensional tooth model according to claim 2, characterized in that: In step S4, the first N points are used to fit the x vector of the local coordinate system, and the last N points are used to fit the z vector of the local coordinate system. The fitting method includes the following steps: Step 1: Select the first 3D point as the reference point, denoted as P0; Step 2: Initialize a three-dimensional vector to store the accumulated direction vector; Step 3: Calculate the offset vectors of the remaining N-1 points relative to the center point P0, and add them up to get the sum of the offset vectors of all points; Step 4: Divide the accumulated offset vector by N-1 to obtain the average direction vector of each point relative to the center point; Step 5: Normalize the average offset vector to obtain a unit vector with consistent direction and length of 1.

4. The method for predicting a local coordinate system of a three-dimensional tooth model according to claim 3, characterized in that: The fitting of the x vector includes the following steps: Step 1: Select the first 3D point from the first N points as the reference point, denoted as P a =(x a ,y a , z a ); Step 2: Initialize a three-dimensional vector is (0,0,0), used to store the accumulated direction vector; Step 3: For each point P i , where i = 1, 2, 3, ... N-1, calculate its relative to the center point P a The offset vector is added to the vector After the loop, the sum of all point offset vectors is obtained. Step 4: Calculate the average direction vector of each point relative to the center point Step 5: Obtain The unit vector represents the average direction of the remaining N-1 points, ‖·‖ is the norm.

5. The method for predicting a local coordinate system of a three-dimensional tooth model according to claim 3, characterized in that: The fitting of the z vector includes the following steps: Step 1: Select the first 3D point from the last N points as the reference point, denoted as P b =(x b ,y b , z b ); Step 2: Initialize a three-dimensional vector is (0,0,0), used to store the accumulated direction vector; Step 3: For each point P i , where i = 1, 2, 3, ... N-1, calculate its relative to the center point P b The offset vector is added to the vector After the loop, the sum of all point offset vectors is obtained. Step 4: Calculate the average direction vector of each point relative to the center point Step 5: Obtain The unit vector represents the average direction of the remaining N-1 points, ‖·‖ is the norm.

6. A three-dimensional tooth model local coordinate system prediction system, characterized by: include: Tooth model acquisition module: used to obtain a three-dimensional digital model of a single tooth; Sampling and point set acquisition module: used to perform vertex sampling on the three-dimensional tooth digital model to obtain a three-dimensional tooth point set; Deep learning module: used to input 3D tooth point set and predict 3D vertices; Local coordinate system fitting module: used to further fit the straight line vector through the predicted three-dimensional vertices, and generate the x vector and z vector of the local coordinate system accordingly; Local coordinate system construction module: used to perform vector cross multiplication operation on the x vector and the z vector to construct the y vector of the local coordinate system of the tooth; and to cross multiply the y vector with the z vector to obtain a new x vector to complete the construction of the local coordinate system of the tooth.

7. The three-dimensional tooth model local coordinate system prediction system according to claim 6, characterized in that: It also includes a control module, which controls the tooth model acquisition module, the sampling and point set acquisition module, the deep learning module, the local coordinate system fitting module and the local coordinate system construction module to perform each step according to the three-dimensional tooth model local coordinate system prediction method described in any one of claims 1-5.

8. A device for predicting a local coordinate system of a three-dimensional tooth model, characterized in that: include: at least one processor; as well as at least one memory in communication with the processor; The memory stores instructions executable by the processor, and the instructions are executed by the processor so that the device executes the three-dimensional tooth model local coordinate system prediction method according to any one of claims 1-5.

9. A computer-readable storage medium, characterized in that: Computer-executable instructions are stored thereon, and when the processor executes the computer-executable instructions, the three-dimensional tooth model local coordinate system prediction method according to any one of claims 1 to 5 is implemented.

10. A computer program product, characterized in that: The method comprises computer instructions, which, when executed by a processor, enable a computer device to execute the method for predicting a local coordinate system of a three-dimensional tooth model according to any one of claims 1 to 5.

Images