Three-dimensional dental model segmentation method and apparatus

Abstract

The application relates to a three-dimensional tooth model segmentation method and device. The method comprises the following steps: acquiring a gum line of a three-dimensional tooth model and a tooth region of the three-dimensional tooth model; extracting a wave peak point of the gum line, and pairing the wave peak point to obtain a wave peak point combination; determining a segmentation path between teeth in the tooth region based on the wave peak point combination; and performing segmentation processing on the tooth region according to the segmentation path to obtain target segmented teeth. The application solves the technical problem that a three-dimensional tooth model cannot be accurately segmented.

Description

Technical Field

[0001] This application relates to the field of orthodontic technology, and in particular to a method and apparatus for segmenting a three-dimensional tooth model. Background Technology

[0002] With the development of science and technology, manufacturing technology, digital modeling technology, materials science, CNC technology, and other technologies have rapidly grown and integrated. Computer technology has increasingly permeated all aspects of teaching, research, and clinical applications in the medical field, enabling effective collaboration. Furthermore, with the development and popularization of measurement technology, people can easily obtain digital dental models, which play a crucial role in oral clinical diagnosis and treatment. The emergence and development of 3D printing technology has become a hot topic, and its application in the medical field is increasingly common. 3D printing has been used in the medical field for over twenty years, widely applied in dental implantology, orthopedics, neurosurgery, and other procedures.

[0003] The inventors discovered that in current orthodontic treatment applications, it's crucial to quickly present results to users to enhance their confidence in the treatment. This can be achieved through tooth target position simulation. This simulation involves inputting the patient's scan data and, through simple operation and automated system execution, obtaining and displaying the "corrected" effect. It can also demonstrate a continuous "animation" of the change from the original state to the target state, facilitating better communication between doctors and patients and increasing the likelihood of a successful outcome. As a preliminary step in orthodontics, it plays a key role in supporting doctor-patient communication, and the output of the target position simulation is also part of the content generated in subsequent processes, with tooth separation being a crucial component.

[0004] Currently, among related technologies, three-dimensional tooth model segmentation technology is still in its early stages and cannot meet the needs of efficient processing. Summary of the Invention

[0005] This application provides a method and apparatus for segmenting a three-dimensional tooth model to solve the technical problem of being unable to accurately segment a three-dimensional tooth model.

[0006] According to one aspect of the embodiments of this application, a three-dimensional tooth model segmentation method is provided, comprising: obtaining the gingival line of the three-dimensional tooth model and the tooth region of the three-dimensional tooth model; extracting the peak points of the gingival line and pairing the peak points to obtain a peak point combination; determining the segmentation path between teeth in the tooth region based on the peak point combination; and segmenting the tooth region according to the segmentation path to obtain a target segmented tooth.

[0007] According to another aspect of the embodiments of this application, a method for manufacturing a dental instrument is also provided, comprising:

[0008] The target segmented tooth can be obtained using any of the methods described above;

[0009] A shaped tooth model is obtained by printing teeth based on target segmentation; the shaped tooth model is used to obtain dental instruments.

[0010] According to another aspect of the embodiments of this application, a three-dimensional tooth model segmentation device is also provided, comprising:

[0011] The memory and the processor, the memory stores the computer program, and the computer program is executed by the processor to perform any of the methods mentioned above.

[0012] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, which stores a computer program that is executed by a processor to perform any of the above methods.

[0013] Compared with related technologies, the technical solutions provided in this application have the following advantages:

[0014] After obtaining the gingival line and tooth region of the 3D tooth model, this application can extract the peak points of the gingival line, pair the peak points to obtain peak point combinations, determine the segmentation path based on the peak point combinations, and segment the tooth region according to the segmentation path to obtain the target segmented tooth. This achieves the effect of accurately segmenting the 3D tooth model and solves the problem of low accuracy in segmenting 3D tooth models in the prior art. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the hardware environment for an optional three-dimensional tooth model segmentation method provided according to an embodiment of this application;

[0018] Figure 2 This is a flowchart of an optional three-dimensional tooth model segmentation method provided according to an embodiment of this application;

[0019] Figure 3 This is a schematic diagram of the peaks and troughs of an optional three-dimensional tooth model provided according to an embodiment of this application;

[0020] Figure 4 This is a schematic diagram of tooth segmentation for an optional three-dimensional tooth model provided according to an embodiment of this application;

[0021] Figure 5 A flowchart illustrating an optional three-dimensional tooth model provided according to an embodiment of this application;

[0022] Figure 6 This is a schematic diagram of an optional combination of peak points in a three-dimensional tooth model according to an embodiment of this application;

[0023] Figure 7 This is a segmentation curve diagram of an optional three-dimensional tooth model provided according to an embodiment of this application;

[0024] Figure 8 This is a flowchart of an optional gingival line extraction method provided according to an embodiment of this application;

[0025] Figure 9 This is a schematic diagram of a polygonal patch of an optional gingival line extraction method provided according to an embodiment of this application;

[0026] Figure 10 This is a diagram illustrating an optional gingival line extraction method provided according to an embodiment of this application.

[0027] Figure 11 This is a schematic diagram of a convolutional pooling operation based on edges in a target tooth model, according to an embodiment of this application, for an optional gingival line extraction method.

[0028] Figure 12 This is a block diagram of an optional three-dimensional tooth model segmentation device provided according to an embodiment of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] In the following description, the use of suffixes such as "module," "part," or "unit" to denote elements is solely for the purpose of illustration and has no specific meaning in itself. Therefore, "module" and "part" may be used interchangeably.

[0031] To address the problems mentioned in the background art, according to one aspect of the embodiments of this application, an embodiment of a three-dimensional tooth model segmentation method is provided.

[0032] Optionally, in the embodiments of this application, the above method can be applied to, for example, Figure 1 The hardware environment shown consists of terminal 101 and server 103. Figure 1 As shown, server 103 is connected to terminal 101 via a network and can be used to provide services to the terminal or clients installed on the terminal. Database 105 can be set up on the server or independently of the server to provide data storage services for server 103. The network mentioned above includes, but is not limited to, wide area network, metropolitan area network or local area network. Terminal 101 can be a device for acquiring three-dimensional dental models, such as an oral model scanner, or a terminal for receiving three-dimensional dental models, such as a mobile phone or computer.

[0033] The three-dimensional tooth model segmentation method in this embodiment can be executed by server 103, or it can be executed jointly by server 103 and terminal 101.

[0034] like Figure 2 As shown, the method may include the following steps:

[0035] Step S202: Obtain the gingival line and tooth region of the three-dimensional tooth model;

[0036] Step S204: Extract the peaks of the gingival line and pair the peaks to obtain peak combinations;

[0037] Step S206: Based on the combination of peak points, determine the segmentation path between teeth in the tooth region;

[0038] Step S208: The tooth region is segmented according to the segmentation path to obtain the target segmented tooth.

[0039] Optionally, the purpose of the above-mentioned three-dimensional tooth model segmentation method is to accurately segment each tooth on the three-dimensional tooth model into individual teeth.

[0040] The aforementioned 3D dental model can be an intraoral scan model, which refers to a 3D model including the user's teeth generated by scanning the inside of the user's mouth. The data of the 3D dental model is stored on a computer or server, and its shape or style can be displayed on a screen for the doctor's convenience.

[0041] Optionally, the aforementioned gingival line can be the dividing line between the tooth region and the gingival region on a three-dimensional tooth model. The gingival line can be determined in various ways, such as by recognition through a neural network model or by recognizing a two-dimensional projection image of the three-dimensional tooth model.

[0042] Optionally, the peak point of the above-mentioned 3D tooth model can be a point on the gingival line of the 3D tooth model that is higher than the adjacent points. For example, on the gingival line of the 3D tooth model, if a straight line connecting multiple points continuously rises and then begins to decline at one of the points, then that point is a peak point. A peak point can be understood as the highest point of the gingiva located at the gap between two teeth.

[0043] In this embodiment, the pairing of peak points allows each pair of peak points to be defined as a peak point combination. The two peak points in a peak point combination are used to determine the boundary between two teeth. Each peak point combination contains two peak points.

[0044] In this embodiment, after determining the combination of peak points, the segmentation path between two teeth in the tooth region can be determined based on the two peak points.

[0045] The tooth region can be segmented according to the segmentation path to obtain individual teeth. For example... Figure 3 As shown, the gingival line between the tooth and the gum in the 3D tooth model has peaks and troughs. Figure 3 (The unshown portion on the other side of the middle tooth also exists). The peak point is the highest point on the gum line between two teeth (it may be the highest point, or not the highest point, but located on the gum line). The trough point is the lowest point on the gum line. By pairing peak points, the peak points at the gap between every two teeth can be combined into a peak point combination. Using this peak point combination, the segmentation path between the two teeth can be determined. Finally, the teeth are segmented using this path to obtain individual teeth. The segmentation result can be as follows: Figure 4 As shown.

[0046] This application obtains the gingival line and tooth region of a 3D tooth model, extracts the peak points of the gingival line, pairs the peak points to obtain peak point combinations, determines the segmentation path based on the peak point combinations, and segments the tooth region according to the segmentation path to obtain the target segmented tooth. This method can accurately segment each tooth of the 3D tooth model, achieving the effect of accurately segmenting teeth in the 3D tooth model.

[0047] As an optional example, obtaining the tooth region of a three-dimensional tooth model includes: segmenting the three-dimensional tooth model based on the gingival line to obtain the tooth region of the three-dimensional tooth model.

[0048] Optionally, in this embodiment, before segmenting the teeth in the 3D tooth model, the 3D tooth model can be segmented using the gingival line. After segmentation, the tooth regions of the 3D tooth model are obtained. Segmenting the 3D tooth model can be considered as hiding or not processing the gingival region of the 3D tooth model. This can prevent the gingival region from affecting the segmentation accuracy when segmenting the teeth in the subsequent 3D tooth model.

[0049] As an optional example, before obtaining the gingival line of the three-dimensional tooth model, the method further includes: determining the initial orientation of the three-dimensional tooth model; and adjusting the three-dimensional tooth model from the initial orientation to the target orientation.

[0050] Optionally, in this embodiment, after obtaining the user's oral scan data to obtain the user's 3D tooth model, the orientation of the 3D tooth model can be adjusted. The purpose of adjusting the orientation of the 3D tooth model is to "align" it, so that all 3D tooth models can be adjusted to a fixed orientation, making curvature recognition more accurate. For example, the initial orientation of all 3D tooth models may be different. If the tabletop is taken as a horizontal plane, the direction of the upward normal of the tabletop can be taken as the target orientation. Then, after the 3D tooth model is adjusted to the target orientation, it can be in a tooth-upward state. After obtaining the 3D tooth model, its orientation can be adjusted to the target orientation, thereby unifying the orientation of all acquired 3D tooth models.

[0051] As an optional example, determining the initial orientation of a 3D tooth model includes: determining the minimum directed bounding box of the 3D tooth model when the 3D tooth model is a non-closed model; and determining the initial orientation of the 3D tooth model based on the minimum directed bounding box.

[0052] Optionally, in this embodiment, the three-dimensional tooth model can be divided into two types: a closed model and a non-closed model. The closed model is formed by stretching the edges and adding a plane as the bottom surface.

[0053] For a non-closed 3D tooth model, a minimum directed bounding box can be used to enclose it. The two faces with the largest areas of this minimum directed bounding box represent the tooth-facing side and the side of the gum region furthest from the tooth. The direction of one of the two normal vectors from the two largest faces of the minimum directed bounding box can be chosen as the target orientation.

[0054] As an optional example, determining the initial orientation of a 3D tooth model based on the minimum directed bounding box includes: identifying a preset axis direction based on the two faces with the largest areas of the minimum directed bounding box; obtaining the first average coordinate value of the boundary edges of the 3D tooth model and the second average coordinate value of the vertices of the polygonal facets of the 3D tooth model based on the preset axis direction; and determining the initial orientation based on the first average coordinate value and the second average coordinate value.

[0055] Optionally, in this embodiment, the aforementioned preset axis can be a preset direction, such as a coordinate axis in three-dimensional space or other directions. Any direction can be used as the preset direction, and the direction of the preset axis is the same as the preset direction.

[0056] After determining the preset axes, the first average coordinate values ​​of the boundary edges of the 3D tooth model and the second average coordinate values ​​of all vertices can be obtained. The boundary edges mark the division between the teeth and gums. The first average coordinate value of the boundary edges can be understood as the average value of the dividing points between the teeth and gums, and the second average coordinate value can be understood as the centroid of the 3D tooth model. The centroid of the 3D tooth model is biased towards the tooth region. Therefore, among the two largest faces of the smallest directed bounding box enclosing the 3D tooth model, the face closer to the second average coordinate value is the side containing the teeth. Thus, the initial orientation of the 3D tooth model can be determined.

[0057] As an optional example, determining the initial orientation of a 3D tooth model includes: when the 3D tooth model is a closed model, determining the polygon patch group to which the polygon patches on the surface of the 3D tooth model belong, wherein the included angle between the planes of any two polygon patches in the same polygon patch group is less than a first threshold; and determining the initial orientation of the 3D tooth model based on the sum of the areas of the polygon patches in the polygon patch group.

[0058] In this embodiment, the closed model can be obtained by filling a non-closed model, or it can be obtained directly by intraoral scanning during an intraoral scan of the oral cavity. The initial orientation of the closed three-dimensional tooth model can be determined by calculating the area of ​​the polygonal patch group composed of polygonal patches.

[0059] As an optional example, when the 3D tooth model is a closed model, determining the polygon patch group to which the polygon patches on the surface of the 3D tooth model belong includes: selecting any first patch from all patches not yet assigned to a polygon patch group as a patch in a polygon patch group; determining polygon patches whose plane normal to the first patch is less than a first threshold as patches in the same polygon patch group as the first patch; wherein, initially, all polygon patches on the 3D tooth model are not assigned to polygon patch groups; continuing to select any target patch from all patches not yet assigned to a polygon patch group as a patch in another polygon patch group; and determining polygon patches whose plane normal to the target patch is less than a first threshold as patches in the same polygon patch group as the target patch, until all polygon patches are assigned to polygon patch groups.

[0060] As an optional example, when the 3D tooth model is a closed model, determining the polygon patch group to which the polygon patches on the surface of the 3D tooth model belong includes: selecting any one first patch from all patches not yet assigned to a polygon patch group as a patch in a polygon patch group, and determining polygon patches whose normals are at an angle less than a first threshold as patches in the same polygon patch group as the first patch. Initially, none of the polygon patches on the 3D tooth model are assigned to a polygon patch group. Then, selecting any one target patch from all patches not yet assigned to a polygon patch group as a patch in another polygon patch group, and determining polygon patches whose normals are at an angle less than a first threshold as patches in the same polygon patch group as the target patch, until all polygon patches are assigned to polygon patch groups.

[0061] In this embodiment, the polygonal facets can be grouped to obtain different polygonal facet groups, and the face angle between the polygonal facets in the polygonal facet group is less than a first threshold.

[0062] When dividing polygonal facets into groups, any polygonal facet can be considered as a polygonal facet within a group. Then, using this polygonal facet as the base facet, all polygonal faces whose face angle with the base facet is less than a first threshold, or whose normal angle with the base facet is less than a first threshold, are added to this polygonal facet group. Then, for the remaining polygonal faces, the face angle with the base facet is greater than or equal to the first threshold. This operation is repeated until all polygonal faces belong to a single polygonal facet group. The polygonal faces within each polygonal facet group can be considered to lie on the same plane. The sum of the areas of the polygonal faces in each polygonal facet group is calculated. The polygonal facet in the group with the largest sum of areas is considered to close the gingival side of the model, and the other side is the tooth side, thus determining the initial orientation of the 3D tooth model.

[0063] As an optional example, obtaining the gingival line of a three-dimensional tooth model includes: extracting features from the three-dimensional tooth model; identifying the features to obtain the gingival line.

[0064] In this embodiment, a neural network model can be used to extract the gingival line from a 3D tooth model. After the 3D tooth model is input into the neural network model, the neural network model extracts the features of the 3D tooth model, performs convolution and pooling on the features, and outputs the identified gingival line.

[0065] As an optional example, extracting the peak point of the gingival line includes: determining the coordinate value of each point on the gingival line based on the target orientation; and determining the peak point from the gingival line based on the coordinate value, wherein the coordinate value of the peak point is greater than the coordinate value of the adjacent point of the peak point on the gingival line.

[0066] In this embodiment, since the orientation of the three-dimensional tooth model is adjusted to the target orientation, for a point on the gingival line, if the coordinate value in the target direction is greater than that of the adjacent point, then that point is taken as the peak point.

[0067] If two adjacent peaks are too close, the point with the larger coordinate value can be used as the peak, and the other point can be excluded. This is because the peak is the gap between two teeth, so the two peaks should not be too close. If they are too close, it means that the peak may have been determined incorrectly.

[0068] As an optional example, pairing peak points to obtain peak point combinations includes: dividing peak points into a first peak point group and a second peak point group according to the positional relationship between the peak points and the three-dimensional tooth model; taking each peak point in the first peak point group as the first peak point, and taking the first peak point and the second peak point in the second peak point group as a peak point combination, wherein the second peak point is the peak point in the second peak point group that is closest to the first peak point, and the angle between the line connecting the first peak point and the second peak point and the tooth midline of the three-dimensional tooth model is greater than a third threshold.

[0069] Optionally, in this embodiment, when pairing peak points, two peak points that are closest to each other and whose connecting line passes through the dental midline are paired as a peak point combination. The dental midline is the line connecting the midpoints of teeth in the three-dimensional dental model. Peak points passing through the connecting line indicate that peak points on both sides of the tooth are paired, rather than peak points on one side of the tooth.

[0070] As an optional example, after taking each peak in the first peak group as the first peak, and combining the first peak with the second peak in the second peak group as a peak combination, the method further includes: deleting the successfully paired peaks from the first peak group and the second peak group; taking each peak in the second peak group as the third peak, and combining the third peak with the fourth peak in the first peak group as a peak combination, wherein the fourth peak is the peak closest to the third peak in the first peak group, and the angle between the line connecting the third peak and the fourth peak and the tooth midline of the three-dimensional tooth model is greater than a third threshold.

[0071] In this embodiment, we can start from the peak points on one side of the tooth in the three-dimensional tooth model and pair them one by one. For each peak point on one side of the tooth in the three-dimensional tooth model, we select the peak point with the smallest distance from the other side of the tooth and the angle between the line connecting the two peak points and the tooth midline meets the preset threshold condition. We then combine the two peak points into a peak point combination. After the pairing is completed, we delete all the successfully paired peak points and then start from the other side of the tooth to pair them again, so as to achieve the pairing of all peak points.

[0072] Furthermore, if the minimum distance peak point is found based solely on distance, it is easy to make incorrect pairings in the case of oblique teeth or incisors. The pairing method described above can avoid pairing errors in such scenarios by using whether the included angle meets a preset threshold condition.

[0073] There may be several reasons why pairing may fail, such as the distance between the two peaks being too large, or one peak on one side being able to pair with both peaks on the other side.

[0074] As an optional example, determining the segmentation path between teeth in a tooth region based on peak point combinations includes: determining each peak point combination as the current combination; and determining the shortest distance between two peak points in the current combination on the 3D tooth model as a segmentation path.

[0075] Once the combination of peak points is determined, the segmentation path can be determined based on this combination. The segmentation path can be the shortest path between two peak points on the tooth.

[0076] As an optional example, determining the segmentation path between teeth in a tooth region based on peak point combinations includes: determining each peak point combination as the current combination; determining the shortest distance between two peak points in the current combination on the 3D tooth model as the first path; correcting the first path by superimposing curvature to obtain the target path; and determining the target path as the segmentation path.

[0077] In this embodiment, the segmentation path can be used as the path for segmenting teeth. Alternatively, the segmentation path can be adjusted. After obtaining the shortest path between two peaks on a tooth, the path can be corrected by superimposing curvature. The corrected path can then be used as the segmentation path.

[0078] As an optional example, the first path is corrected by superimposing curvature to obtain the target path, which includes: taking each point on the first path as the current point, determining the replacement point of the current point within a preset range on the three-dimensional tooth model; wherein the replacement point is a point that meets the curvature threshold requirement; and taking the line connecting the replacement points as the target path.

[0079] In this embodiment, the shortest path is corrected using a superimposed curvature method. This involves locally adjusting points on the path based on curvature, correcting each point on the shortest path. Each point is treated as the current point. If a point meeting the curvature threshold exists within a preset range near the current point on the 3D tooth model, that point replaces the current point. The replaced points are then connected to form the segmentation path.

[0080] The above three-dimensional tooth model segmentation method will be illustrated with an example below.

[0081] The three-dimensional tooth model segmentation method of this application can be mainly applied in orthodontic demonstration scenarios. After tooth segmentation, the tooth position is adjusted according to the orthodontic plan to generate and display the corrected teeth. Furthermore, tooth segmentation technology can also be applied to the design of orthodontic guides, the design of dental models, and the design of temporary teeth.

[0082] In this embodiment, if a user's teeth need to be examined or orthodontized, the user's oral cavity can be scanned first to obtain three-dimensional dental model data. This data can be displayed on a screen to both the user and the doctor, allowing the doctor to explain the condition of the teeth to the user.

[0083] The 3D dental model obtained from an intraoral scan can be either closed or open. The 3D dental model includes tooth and gingival regions, but the boundary between these regions is not marked. Therefore, to segment teeth, the dividing line between the teeth and gingiva (gingival line) and the segmentation path between the teeth must be determined.

[0084] In other words, in this solution, the scanning data output by the 3D scanner is acquired; the model is automatically aligned based on the identified model features; based on the automatically aligned model, the feature values ​​of the 3D dental model are extracted using a curvature geometric algorithm, and the extracted feature values ​​are filtered and noise-reduced to obtain the gingival line and its peak points, and the tooth part and the gingival part are separated to obtain the pure tooth part; the peak points of the gingival line are paired, and the segmentation path between teeth is determined by the curvature and shortest path method to obtain all the segmented teeth.

[0085] In one embodiment, the above-described three-dimensional tooth model segmentation method can be applied to a fully automated orthodontic production process. First, a digital three-dimensional model of the patient is acquired. The segmentation method described above is then applied to this digital three-dimensional model to obtain a single tooth model. Misaligned teeth are then automatically aligned, generating multiple sets of orthodontic tooth models. Simultaneously, positioning elements and placement markings are set for these multiple sets of orthodontic tooth models. These multiple sets of orthodontic tooth models are printed, and after printing, a preheated polymer film is pressed onto the dental model to form a shell-like film. The bottom of the dental model (with the film attached to its surface) is photographed using an image acquisition terminal (CCD). Then, using OCR recognition technology, the marking information is identified and sent to the server. The server retrieves the corresponding marking or cutting instructions from its database based on the marking information. When executing the marking instructions, the 3D label to be printed is determined; the boundary constraints of the 3D label are determined, and a label area is determined from the base plate based on the boundary constraints. This label area is used to print the 3D label; the 3D label is then cut through to the label area. When executing a cutting instruction, the image acquisition terminal captures an image of the bottom of the product to be cut; it identifies the identification code in the image captured by the image acquisition terminal and sends the identification code or the identification information contained in the identification code to the receiving terminal; the receiving terminal matches the corresponding cutting instruction, and the cutting execution terminal performs the cutting operation on the product to be cut according to the cutting instruction. Each product to be cut has a corresponding identification code on its bottom surface, and different products correspond to different identification codes. The product to be cut is placed above the image acquisition terminal, and by capturing an image of the bottom of the product to be cut, the identification code in the image can be identified. The identification code corresponds to the corresponding cutting instruction, and according to the cutting instruction, the cutting execution terminal performs the corresponding cutting operation to remove the excess parts and finally obtain the product.

[0086] It should be noted that the above-mentioned scheme for automatically arranging misaligned teeth can be as follows: establish a tooth alignment coordinate system; define the feature points of a single tooth and establish a local coordinate system for the tooth; based on this, analyze the position and orientation of each tooth in the dentition from a low-dimensional perspective, and use a weighted fitting optimization method to calculate the coordinate translation and local coordinate axis rotation of the teeth respectively, forming a correlation constraint between the tooth pose and the spatial dentition curve. Combined with the collision detection method of the rectangular bounding box, an iterative algorithm based on the steepest descent method is designed to adjust the tooth pose within the constraint range of the spatial dentition curve, thereby completing the automatic alignment of teeth.

[0087] Figure 5 This is an optional flowchart of this embodiment. The segmentation of the three-dimensional tooth model in this embodiment mainly includes the following steps:

[0088] 1. Import the model

[0089] Importing a model refers to importing a 3D dental model. A 3D dental model is a model obtained from an intraoral scan. Once imported into the system, the system can display the structure of the 3D dental model on a screen.

[0090] The 3D tooth model can be a scanned model (non-closed model) output by a 3D scanner, and the input 3D tooth model can be in any orientation. In this solution, other types of tooth models with planes do not affect the implementation of this invention.

[0091] A three-dimensional tooth model is a digital three-dimensional body composed of a series of polygonal facets.

[0092] 2. Automatic model alignment

[0093] After a 3D tooth model is imported into the modeling system, it may have different orientations. Therefore, you can first adjust the orientation of the 3D tooth model to the target orientation to align the 3D tooth model correctly.

[0094] For the imported 3D tooth model, its features are identified. The purpose of feature identification is to find the alignment angle of the 3D tooth model. Since different tooth models are used for different purposes, their features are inconsistent. Furthermore, based on different tooth model types, feature extraction allows for classification. During alignment, it is necessary to determine the initial orientation of the 3D tooth model and then adjust it to the target orientation.

[0095] There are several methods for determining the initial orientation. For example,

[0096] 1) If the model is an intraoral scan model, which is a non-closed model, the Oriented Bounding Box (OBB) technique is first applied to determine the bounding box of the bounding model. The direction of a normal to the two faces with the largest area among the six faces of the bounding box is then determined as the final tooth orientation. This method determines the size and orientation of the box based on the geometry of the object itself; the box does not need to be perpendicular to the coordinate axes. This allows for the selection of the most suitable and compact bounding box.

[0097] After obtaining the containing box, based on the preset axis direction, the first average coordinate value of the boundary edge of the 3D tooth model and the second average coordinate value of the vertices of the polygonal facets of the 3D tooth model are obtained. Then, the side containing the face with the second average coordinate value closer to the first average coordinate value is determined as the side of the tooth orientation, thus determining the initial orientation of the 3D tooth model.

[0098] Based on the target's angle, a rotation matrix is ​​used to rotate the containing box, thereby rotating the model to face the target.

[0099] 2) If the model is a closed model with a flat bottom, it is aligned by aligning the model with its largest flat bottom. The method for detecting the largest plane of the tooth is as follows: Define a polygonal facet, superimpose the defined polygonal facet onto the 3D tooth model composed of polygonal facets, and set an error threshold e. When e is greater than a certain value, the defined polygonal facet is considered to be uneven with the facet on the 3D tooth model; otherwise, they are considered to be on the same plane. When the defined polygonal facet and a facet on the tooth model are on the same plane, superimpose them and continue to find the next polygonal facet and judge the error threshold. Repeat the above steps until the largest plane of the tooth model is obtained. Determine the current normal vector of the largest plane of the tooth model and obtain the target normal vector of the largest plane of the tooth model.

[0100] According to the cross product operation method, the rotation angle and rotation axis are solved based on the vector values ​​before and after rotation (the rotation angle and rotation axis are obtained by solving the current normal vector and the target normal vector using the cross product operation method). The cross product operation method is a binary operation of vectors in vector space, and its result is a vector rather than a scalar. With the above rotation angle and rotation axis, any model can be rotated to the desired spatial position.

[0101] 3. Extracting the gingival line based on feature recognition

[0102] The purpose of identifying and extracting the gingival line from a 3D tooth model is to segment the teeth based on this line. The gingival line can be determined in several ways. For example, it can be determined using a neural network model. The 3D tooth model can be input into the neural network model, which will then annotate the gingival line. Alternatively, the 3D tooth model can be projected onto a 2D plane, and the tooth regions on the 2D plane can be identified and mapped onto the 3D tooth model to mark the boundary between the teeth and the gum line.

[0103] Taking the recognition of the gingival line using a neural network model as an example, after placing the model in the designated position, the feature values ​​extracted from the tooth model are processed using the curvature geometric calculation method, and then filtered and denoised to obtain the three-dimensional tooth model outline. The curvature calculation method is the rotation rate of the tangent direction angle with respect to the arc length of a certain surface on the dental model, indicating the degree of concavity and convexity of the surface, which is also referred to as a feature in this scheme. Therefore, the feature values ​​of the concave and convex areas of the dental model can be obtained through the curvature method (the actual gingival line on the dental model is also represented by concavity and convexity).

[0104] After obtaining the initial dental model feature values, noise needs to be removed. Noise removal is necessary because other areas of the teeth besides the gingival line region also have irregularities, which are also considered features. These irregularities will affect the subsequent fitting with the contour line, so they need to be filtered or removed to finally obtain the optimal feature values.

[0105] The eigenvalues ​​are simplified into a contour line, i.e., an initial gingival line. Since the initial gingival line is not a smooth curve and may have local folds and deviations, principal component analysis (PCA) is used to process the initial gingival line and obtain the points of the principal direction. The consistency of the principal direction is used to determine this; for example, peaks and troughs indicate changes in the curve's direction, and these locations can be used to identify the points of the principal direction. To address this, this solution uses PCA to obtain the main contour shape, determine the points of the principal direction, avoid uneven areas, and obtain the final smooth gingival line.

[0106] 4. Dividing the gingiva and teeth based on the gingival line

[0107] After identifying the gingival line, the tooth and gingival portions of the 3D tooth model can be segmented according to the position of the gingival line. The purpose of segmenting the tooth and gingival portions is to make tooth division more accurate.

[0108] 5. Obtain and match wave peaks based on the gingival line.

[0109] The points mentioned above that determine the main direction of the contour are also the peak and trough points. This method determines the insertion control points based on the curvature of the gingival line (peak and trough points are all curved areas). Since the trough point is always lower than the peak point, the peak point can be determined by the vertical distance (the difference in height in the target direction), and the trough points can be filtered out to obtain the complete peak point.

[0110] After obtaining the peak points, pair the two peak points that are closest to each other and whose connecting line passes through the dental midline, and use them as the start and end points for tooth segmentation. For example, as shown... Figure 6 As shown, Figure 6 In the middle, the peaks on both sides of the teeth are paired. Figure 6 A group of wave crests connected by a dashed line (not showing all dashed lines).

[0111] 6. Based on point pairing, obtain the tooth segmentation curve according to the curvature threshold and minimum path.

[0112] Since teeth are connected and in contact with each other, a curve is used to cut the teeth and form independent single teeth. This scheme first uses the shortest distance method to determine the initial cutting curve. The shortest distance shown is the distance from the paired starting peak point, along the surface of the dental model triangular facet, to the ending peak point. Since the shortest distance is not the optimal cutting line, this scheme uses curvature superposition to correct the shortest distance. Because the contact between teeth is uneven, the shortest distance dividing line must be within a range of high curvature. The final dividing line is established by superimposing the two methods. The determined dividing path is as follows: Figure 7 As shown.

[0113] 7. Divide teeth according to the segmentation curve

[0114] All teeth are divided according to the dividing curve to obtain segmented teeth.

[0115] Optional, such as Figure 8 As shown, this method for extracting the gingival line may include the following steps:

[0116] Step S222: Obtain the target tooth model, wherein the target tooth model includes multiple polygonal facets;

[0117] Step S224: Determine the edge categories of the edges of multiple polygonal patches, where the edge categories include tooth edges and gum edges;

[0118] Step S226: Determine the gingival line of the target tooth model based on the edge category.

[0119] The target tooth model mentioned above can be an intraoral scan model, which refers to a digital 3D model including the user's teeth and gums generated by scanning the inside of the user's mouth; alternatively, it can be obtained by taking impressions of the teeth and gums first, and then scanning the impressions to obtain a digital 3D model. The 3D tooth model data is stored on a computer or server, and the shape or style of the 3D tooth model can be displayed on a screen for the doctor's convenience.

[0120] The aforementioned polygonal facets are the facets that make up the surface of the target tooth model (both the tooth and non-tooth parts, the entire surface of the model). Specifically, they can be triangular, quadrilateral, pentagonal, etc. Each polygonal facet lies on a plane, and different polygonal facets can lie on the same or different planes. The surface of the target tooth model is composed of multiple polygonal facets, each containing multiple edges. Adjacent polygonal facets share an edge. The endpoints of the edges are the vertices of the corresponding polygonal facets, and a single vertex can be shared by multiple polygonal facets.

[0121] It should be noted that the target tooth model includes a tooth portion and a gingival portion, and the gingival line is the boundary between the tooth portion and the gingival portion. The edges that make up the polygonal facets can be located in either the tooth portion or the gingival portion, and thus can be classified as tooth edges and gingival edges. Based on this, in this embodiment, the gingival line of the target tooth model can be determined according to the edge category of the polygonal facets, and the determined gingival line is used to segment the tooth portion and the gingival portion of the target tooth model.

[0122] In this embodiment, after obtaining the target tooth model, the edge category can be determined by determining the edge category of the polygonal facet on the target tooth model, thereby determining the gingival line of the target tooth model. The edges, as components of the polygonal facet, can further refine the target tooth model, thereby improving the accuracy of gingival line extraction of the tooth model.

[0123] As an optional example, determining the edge category of multiple polygonal patches includes: obtaining target features of the edges of the multiple polygonal patches; and determining the edge category of the edges of the multiple polygonal patches based on the target features.

[0124] In this embodiment, the target features include geometric features. Of course, in other embodiments, the target features may also include non-geometric features; this is not specifically limited here. In this embodiment, the edge category can be determined by extracting the target features of the edges of polygonal patches and identifying the target features. It should be noted that when the target features include geometric features, the topological relationship of the space will be considered when classifying the edges, thereby improving the accuracy of gingival line recognition.

[0125] In this embodiment, target features of all edges on the target tooth model can be extracted, or target features of a subset of edges can be selected. The edge category is determined by identifying the extracted target features of the edges.

[0126] As an optional example, the target features of an edge include at least one of the following: a dihedral angle between a first polygonal patch and a second polygonal patch sharing the edge; a first curvature value of a first vertex of the edge; a second curvature value of a second vertex of the edge; a first distance from a vertex of the first polygonal patch furthest from the edge to the edge; a second distance from a vertex of the second polygonal patch furthest from the edge to the edge; a first diagonal of the edge in the first polygonal patch; and a second diagonal of the edge in the second polygonal patch. In one application scenario, the target features of an edge include the above seven items.

[0127] As another alternative example, the target features of an edge may include at least one of the following features: a dihedral angle between a first polygonal patch and a second polygonal patch sharing the edge, a first curvature value of a first vertex of the edge, a second curvature value of a second vertex of the edge, a first spatial coordinate of a first vertex, a second spatial coordinate of a second vertex, the length of the edge, the angle between the edge and an adjacent edge, a first distance from a vertex of the first polygonal patch away from the edge to the edge, a second distance from a vertex of the second polygonal patch away from the edge to the edge, a first diagonal of the edge in the first polygonal patch, a second diagonal of the edge in the second polygonal patch, a first normal of the first vertex, a second normal of the second vertex, a normal of the first polygonal patch, and a normal of the second polygonal patch.

[0128] The dihedral angle between the first and second polygonal faces sharing a common edge is the angle between the two polygonal faces sharing a common edge. The first and second vertices are the two endpoints of the edge, and the curvature values ​​are the curvature of the face corresponding to the vertex at that vertex. Curvature is the rate of rotation of the tangent direction angle about the arc length at a point on a curve, defined by differentiation, indicating the degree to which the curve deviates from a straight line. It is a numerical value indicating the degree of curvature of the curve at a point. The greater the curvature, the more curved the curve at that point. The spatial coordinates are the coordinates in the three-dimensional Cartesian coordinate system where the vertex is located. The three-dimensional Cartesian coordinate system can be the coordinate system of the target tooth model. For example, the plane containing the base plate of the target tooth model can be defined as the xOy plane, and the normal direction of this plane can be defined as the z-axis. The directions of the coordinate axes can be predetermined. For example, the gingival surface of the target tooth model can be used as the plane containing the X and Y axes, and the tooth direction can be used as the Z-axis direction. The length of a side is the distance between its two endpoints. Since a side may have multiple adjacent sides, the angle between two adjacent sides can be multiple, ranging from 0 to 180 degrees. When the polygon is quadrilateral or more polygonal, the distance from a vertex furthest from the side to the side can be more than one, and each vertex's distance to the side can be different. However, when the polygonal face is triangular, there is only one vertex furthest from the side, and therefore only one distance from that vertex to the side. Similarly, a side in a polygonal face can have one or more opposite angles, and when there are multiple angles, each opposite angle can be different. For example, as... Figure 9 As shown, the polygonal patches are triangular patch 302, quadrilateral patch 304 and pentagonal patch 306, and each side of the three polygonal patches corresponds to at least one or at least two of the above features.

[0129] Specifically, during feature extraction, the polygonal faces corresponding to the edges of the target tooth model are identified. The cross product of the normals of the polygonal faces yields the angle between the normals of two polygonal faces. The dihedral angle between two polygonal faces is complementary to the angle between their normals; subtracting the angle between the normals from 180° gives the dihedral angle between the two polygonal faces. The curvature values ​​of the two vertices can be obtained using curvature geometry calculations or other methods. The distance from the vertex furthest from the edge of a polygonal face to the edge can be obtained using the point-to-line distance method. Two vectors are obtained based on the orientation of the two edges; the dot product of these two vectors yields the angle between the two edges at the vertex of the polygonal face, i.e., the diagonal of the edge in each polygonal face. It should be noted that the spatial coordinates of each vertex and the proximity relationships between points, edges, and faces are all known, allowing for the derivation of each target feature.

[0130] As an optional example, determining the edge category based on the target features includes: increasing the dimensionality of the target features in the initial dimension of the edge to obtain a first feature in the first dimension; reducing the dimensionality of the first feature to obtain a second feature in the target dimension; and determining whether the edge is a tooth edge or a gingival edge based on the value of the second feature. That is, in this embodiment, the target dimension can be two-dimensional, corresponding to tooth edges and gingival edges respectively. Of course, in other embodiments, it can also be multi-dimensional, as long as the edge category of the corresponding edge can be determined; this is not limited here.

[0131] Optionally, in this embodiment, after extracting the target features of the edges, the initial dimension of the target features can be determined according to the type and quantity of the extracted target features. First, the target features are increased in dimensionality to the first dimension to obtain the first feature. Then, the first feature is reduced in dimensionality to the target dimension. However, since dimensionality is increased first and then reduced, the first feature after dimensionality reduction is no longer the target feature, but rather the second feature of the target dimension. Based on this, the edge category of the corresponding edge can be determined.

[0132] Taking an initial dimension of seven as an example, the seven-dimensional target features of the corresponding edge are input into the edge classification network to classify the corresponding edge. Specifically, the input dimension is first seven dimensions, corresponding to seven geometric features. After one convolution operation, the dimension of the target feature increases to N dimensions. Then, after k identical convolution operations, the dimension of the target feature remains unchanged. Finally, a pooling operation is performed to reduce the number of edges. Then repeat the above process, i.e., one dimensionality-increasing convolution + k same-dimensional convolutions + pooling, so that the dimension of the target feature becomes 2N, 4N, and 8N respectively. It should be noted that in the last dimensionality-increasing loop, after k same-dimensional convolutions, an upper pooling operation is performed to gradually restore the number of edges, and then a dimensionality-reducing convolution is performed to reduce the feature dimension to 4N. After k more identical convolution operations, the feature dimension remains unchanged, and then the above process is repeated, i.e., upper pooling + one dimensionality-reducing convolution + k dimensionality-increasing convolutions, so that the feature dimension becomes 4N, 2N, and N respectively. It should be noted that after the last dimensionality-reducing loop, the feature dimension can be further reduced directly to 2 dimensions. Figure 10 The dashed arrow in the middle indicates the feature fusion operation. The final values ​​of the two dimensions are the probability values ​​of whether the edge is a tooth edge or a gingival edge. The category with the larger probability value can be selected as the edge category. At this point, each edge of the model is classified as a tooth edge or a gingival edge.

[0133] It should be noted that the above are merely examples and not intended to be limiting. For instance, the initial dimension is not limited to 7 dimensions, nor is the number of iterations during dimensionality increase and decrease. N is an integer not less than 7, such as 7, 8, 9, 10, etc. These can be adjusted according to the actual situation in the specific implementation plan.

[0134] It should be further noted that the convolution and pooling operations in this step differ from those in the convolution and pooling operations within the image neighborhood, such as... Figure 11 As shown, its operation is based on the edges in the target tooth model.

[0135] In the convolution operation, to eliminate the influence of different edge orders (e.g., (a,b,c,d) and (c,d,a,b)), the original edge features are processed according to the following equation (1), where f(a) and f'(a) represent the original feature and the processed feature of edge a, respectively. The definition of the convolution operation for the processed edge e is given in equation (2), where {w k |k=0,1,2,3,4} represents the weight parameters that need to be trained. The pooling operation transforms 5 edges (e,a,b,c,d) into 2 edges (h,i), as defined in equation (3). The upper pooling operation is the opposite of the pooling operation, transforming 2 edges (h,i) into (e',a',b',c',d'), as defined in equation (4).

[0136]

[0137] C(e)=w0f'(e)+w1f'(a)+w2f'(b)+w3f'(c)+w4f'(d) (2)

[0138]

[0139]

[0140] In addition, before performing edge classification, the classification network needs to be trained to obtain network weight values ​​and some hyperparameter values, such as k and N. Specifically, multiple sets of geometric feature data and the corresponding edge classification values ​​(e.g., 0 for gingiva, 1 for teeth) are input into the edge classification network for training. For example, 500, 1000, or 2000 sets of pre-classified tooth model data can be input, each set containing geometric feature data and classification data corresponding to 5000, 10000, or 20000 edges, enabling the classification network to classify the edge based on its geometric feature data. It should be noted that this embodiment combines deep learning technology, and the edge classification method can automatically identify gingival line data, which can greatly improve the accuracy and efficiency of recognition, improve the efficiency of the entire dental treatment process such as invisible orthodontics, and enhance the user's experience during treatment.

[0141] As an optional example, determining the gingival line of the target tooth model based on the edge category includes: determining the facet category of the polygon facet based on the edge category, wherein the facet category includes tooth facet and gingival facet; and extracting the gingival line based on the shared edges of adjacent polygon facets with different facet categories.

[0142] Optionally, in this embodiment, after determining the edge category, the polygonal patch category can be determined based on the edge category. Understandably, a tooth patch is a patch located on the tooth portion of a tooth model, and a gingival patch is a patch located on the gingival portion. Determining the patch category based on the edge category can be done by determining the polygonal patch category based on all edges of the polygonal patch, or by determining the polygonal patch category based on at least one edge of the polygonal patch.

[0143] As an optional example, determining the polygonal facet category based on the edge category includes: when the number of gingival edges is greater than the number of tooth edges in the polygonal facet, the polygonal facet is determined to be a gingival facet; when the number of gingival edges is less than the number of tooth edges in the polygonal facet, the polygonal facet is determined to be a tooth facet.

[0144] Optionally, in this embodiment, the polygonal patch category is determined based on the edge categories of all its edges. The polygonal patch is classified as either a tooth edge or a gingival edge based on the number of tooth edges and gingival edges within it.

[0145] Traverse all edges of the target tooth model. If the two facets corresponding to an edge are of different categories, the edge is marked as a shared edge.

[0146] As an optional example, extracting the gingival line based on the shared edges of adjacent polygonal faces of different face types includes: determining a first feature point based on the vertices of the shared edges; determining a target feature point from the first feature point according to the curvature relationship; and determining the gingival line based on the target feature point.

[0147] Optionally, in this embodiment, since the polygonal facet category can be determined based on the edge category, after determining the facet category, the shared edges of polygonal faces with different categories serve as the boundary between the tooth facet and the gingival facet. Therefore, the gingival line can be determined based on all the shared edges of the tooth facet and the gingival facet.

[0148] In this embodiment, a first feature point can be determined from the vertices of the shared edge, then a target feature point can be determined from the first feature point, and the gingival line can be determined based on the target feature point. The target feature point can be calculated based on the curvature.

[0149] In this embodiment, starting from a vertex of a shared edge, connected shared edges are found, and the vertices of each connected shared edge are recorded as feature points in a counter-clockwise or clockwise order. The first feature point can be all or part of the recorded feature points. Specifically, the first feature point is located on the initial tooth model, thus improving the accuracy of the extracted gingival line. The target feature point is a portion of the first feature points that possess specific characteristics.

[0150] As an optional example, determining the first feature point based on the vertices of the shared edge includes: when the target tooth model is a model obtained by edge shrinking the initial tooth model, determining the point on the shared edge that is closest to the initial tooth model as the first feature point; when the target tooth model is the initial tooth model, determining the vertices on the shared edge as the first feature point.

[0151] Optionally, the initial tooth model mentioned above is the tooth model initially obtained. The initial tooth model can be processed by edge contraction to obtain the target tooth model, or the initial tooth model can be directly used as the target tooth model. Whether or not edge contraction is performed on the initial tooth model affects whether the vertices of the shared edges are directly used as the first feature points. In this embodiment, the first feature point is either the vertex of the shared edge or the point on the initial tooth model where the shared edge is closest.

[0152] In other words, in this embodiment, starting from the vertex of a shared edge, the connected shared edges are found, and the vertices are recorded as feature points in counterclockwise or clockwise order. Then, the recorded feature points can be projected back to the original model, and the point on the original model closest to the recorded feature point is found as the new feature point.

[0153] Specifically, a KDTree can be established using the vertices of the initial tooth model. A KDTree is a tree-like data structure that stores instance points in k-dimensional space for fast retrieval. In this embodiment, all recorded feature points can be traversed, and the nearest point in the KDTree can be found as the new feature point. This new feature point is the first feature point.

[0154] As an optional example, determining the target feature point from the first feature points based on the curvature relationship includes: identifying the first feature points whose curvature is greater than that of the neighboring first feature points as the target feature points.

[0155] After determining the first feature points based on whether the initial tooth model has undergone edge shrinkage processing, target feature points can be determined from the first feature points based on their curvature. Each first feature point corresponds to a curvature. In this embodiment, the first feature point with a curvature greater than that of its neighboring first feature points is selected as the target feature point. In other words, the point in the curve that is more curved is selected as the target feature point, which better reflects the actual trend of the curve.

[0156] After the recorded feature points are projected back to the original model, the curvature of the discrete points can be estimated using a geometric algorithm of curvature, and feature points with large curvature can be retained as target feature points. The reason for retaining feature points with large curvature is that points with large curvature are often inflection points, and inflection points have strong characteristics and are representative.

[0157] As an optional example, determining the gingival line based on target feature points includes: performing an interpolation operation on the target feature points to obtain interpolated target feature points; and connecting the interpolated target feature points in sequence to form the gingival line.

[0158] In this embodiment, interpolation is performed on the target feature points to compensate for the loss of feature points caused by determining the target feature points from the first feature points, and the interpolation operation can combine the target feature points into a closed curve. The closed curve is the gingival line.

[0159] When interpolating target feature points, B-spline curve interpolation can be performed on adjacent target feature points; connecting all points in sequence forms the gingival line. It should be noted that B-spline curves are not limited to interpolation of target feature points; other spline curves can also be used, and no specific limitation is made here.

[0160] As an optional example, obtaining the target tooth model includes: obtaining an initial tooth model, wherein the initial tooth model includes multiple polygonal patches, and the number of edges in the initial tooth model is greater than the number of edges in the target tooth model; reducing the number of edges in the initial tooth model to a preset number according to a preset method to obtain the target tooth model.

[0161] Optionally, in this embodiment, the preset method can be edge contraction or edge collapse. Performing edge contraction on the initial tooth model aims to reduce the number of edges in the initial tooth model, thereby obtaining the target tooth model, simplifying the tooth model, and improving the recognition rate of the gingival line.

[0162] There are several ways to reduce the number of edges in the initial tooth model. You can reduce one edge at a time or in batches, such as reducing one edge at a time until a preset number is reached or a preset number of times is reduced. Alternatively, you can reduce edges in batches until a preset number is reached or a preset number of times is reduced.

[0163] Taking edge collapse as an example, the imported initial tooth model can be a tooth model in any direction. The initial tooth model is a digital 3D volume composed of a series of vertices and polygonal faces. After importing the initial tooth model, the tooth model is downsampled using the edge collapse method. After downsampling, the polygonal faces of the tooth model become sparser, the number of polygonal faces decreases, and the number of edges also decreases, until the specified number of edges is reached before downsampling stops.

[0164] The imported initial tooth model often has many edges with varying numbers. However, the required number of edges is usually specific when classifying the tooth using a classification network in subsequent steps. Therefore, it is necessary to downsample the imported initial tooth model, deleting some edges and retaining a specified number. The determination of the number of edges can be based on both classification speed (fewer edges mean faster speed) and the accuracy of the final gingival line (more edges mean higher accuracy and less information loss from the original model). The main process is as follows:

[0165] (1) For each vertex of the initial tooth model, record the sum of its distances to the adjacent face as the error value, with an initial value of 0; where the adjacent face refers to the face where the vertex is located; the initial error value of 0 means that before the edge shrinkage is performed, the sum of the distances from the vertex to its adjacent face is 0.

[0166] (2) Calculate the error value of the vertex after deleting the edge (here, it refers to the sum of the distances from the vertex to the adjacent face when the edge was not deleted). Sort them according to the error value from smallest to largest, and delete the edge with the smallest error first. When an edge shrinks into a vertex, delete the two adjacent face, and delete one vertex and three edges. There are multiple cases for the position of an edge shrinking into a point. For example, it can be moved from the two vertices of the edge to any point on the edge for deletion. The specific position of the point is selected in this step according to the smallest error value.

[0167] (3) Repeat the above steps, continuously calculating the error value and deleting the corresponding edges, until the number of edges reaches the specified number. At this point, the adjustment of the number of edges in the initial tooth model is complete.

[0168] As an optional example, reducing the number of edges in the initial tooth model to a preset number according to a preset method to obtain the target tooth model includes: for each edge of the tooth model to be processed, calculating the sum of distances from the vertex corresponding to each edge to the adjacent polygonal facets in the tooth model to be processed after edge shrinkage processing according to each of a variety of shrinkage methods; wherein, in the first processing, the tooth model to be processed is the initial tooth model; determining the edge corresponding to the minimum value among the calculated sums of multiple distances and the shrinkage method; performing edge shrinkage processing on the edge corresponding to the minimum value according to the shrinkage method corresponding to the minimum value to obtain the target tooth model.

[0169] As an optional example, the edge corresponding to the minimum value is shrunk according to the shrinkage method corresponding to the minimum value to obtain the target tooth model. This includes: after the edge shrinkage process, if the number of edges in the obtained tooth model is not greater than a preset number, then the obtained tooth model is the target tooth model; if the number of edges is greater than the preset number, then the obtained tooth model is determined as the tooth model to be processed, so that the edge shrinkage process can continue.

[0170] Optionally, in this embodiment, when performing edge shrinkage processing on the initial tooth model, it is also necessary to determine which edge to shrink. This process can be a cyclical process. That is, shrinking one edge at a time and repeating, or shrinking a batch of edges at a time and repeating, until the shrunken tooth model meets the conditions. Therefore, in each shrinkage, it is necessary to determine which edge or which batch of edges to shrink.

[0171] In this embodiment, the initial model can be used as the model to be processed, and the edges to be shrunk in the model to be processed are determined. The determination method can be either simulation or calculation. Calculate the sum of distances from each vertex corresponding to that edge in the shrunk model to the adjacent polygonal faces of the corresponding vertex in the tooth model to be processed, if a certain edge is shrunk. If this sum is minimized, then the edge is shrunk according to this strategy. Each time an edge is shrunk, the calculation is repeated, and the edge shrunk is repeated until the number of edges in the tooth model to be processed meets the preset requirement. At this point, the edge shrunking is complete, and the target tooth model is obtained.

[0172] According to another aspect of the embodiments of this application, a method for manufacturing a dental instrument is also provided, which may include the following steps:

[0173] Step 1: Obtain a digital model of the teeth.

[0174] The tooth model can be the target tooth model in the aforementioned gingival line extraction method embodiment. Accordingly, the method of obtaining the tooth model can be the same as that of obtaining the target tooth model, that is, it can be obtained by intraoral scanning or by traditional impression taking, and there is no limitation here.

[0175] Step two involves preprocessing the digitized tooth model.

[0176] In one embodiment, the preprocessing operation may include steps of a three-dimensional tooth model segmentation method to obtain the target segmented tooth. The preprocessing operation may also include steps from a gingival line extraction method to identify the gingival line, and may further include converting the gingival line into a cutting line for subsequent steps to cut initial instruments to obtain dental instruments. Alternatively, in some applications, it may be applied to automated tooth segmentation, gingival-crown separation, etc. The gingival line extraction method may be the same as in the aforementioned gingival line extraction method embodiments; please refer to the foregoing content for relevant details, which will not be repeated here.

[0177] In one embodiment, the three-dimensional tooth model segmentation method in the preprocessing can also employ the following steps:

[0178] Step S242: Obtain a two-dimensional projection image of the three-dimensional tooth model, and identify the two-dimensional projection image to obtain multiple tooth regions;

[0179] Step S244: In the three-dimensional tooth model, identify the original seed points corresponding to multiple tooth regions;

[0180] Step S246: Expand the original seed points within a preset range to obtain the target seed points of the teeth in the three-dimensional tooth model.

[0181] Step S248: Segment the three-dimensional tooth model based on the target seed points of each tooth to obtain segmented teeth.

[0182] Optionally, the purpose of the above-mentioned three-dimensional tooth model segmentation method is to accurately segment each tooth on the three-dimensional tooth model into individual teeth.

[0183] The aforementioned 3D dental model can be an intraoral scan model, which refers to a 3D model including the user's teeth generated by scanning the inside of the user's mouth. The data of the 3D dental model is stored on a computer or server, and its shape or style can be displayed on a screen for the doctor's convenience.

[0184] The aforementioned two-dimensional projection image can be an image obtained by projecting a three-dimensional tooth model onto a plane. The two-dimensional projection image includes both tooth and non-tooth regions. By recognizing the two-dimensional projection image, multiple tooth regions can be identified, thereby segmenting the teeth on the three-dimensional tooth model. The recognition method can employ machine vision or artificial intelligence to identify the tooth regions from the two-dimensional projection image.

[0185] The surface of a 3D tooth model is composed of polygonal patches, such as triangular patches. A triangular patch contains three vertices, and all vertices of the triangular patches can be considered seed points. The aforementioned original seed points are the seed points among all seed points in the 3D tooth model that correspond to the tooth region in the 2D projection image. In other words, the tooth region can be initially determined on the 3D tooth model using the original seed points.

[0186] Since the tooth region determined by the original seed point may not be accurate, the original seed point can be expanded to obtain the target seed point. The area covered by the target seed point can be regarded as a tooth on the 3D tooth model. Through the target seed point, the 3D tooth model can be accurately segmented.

[0187] After identifying the tooth region, it can be understood that the location and area of ​​the tooth have been determined on the 2D projection image. Then, this is mapped onto the 3D tooth model to determine the initial seed points on the 3D tooth model. Since the seed points are the vertices of the triangular facets on the surface of the 3D tooth model, the initial seed points can be understood as covering the area of ​​the tooth on the 3D tooth model. To ensure accuracy, the initial seed points are further expanded to obtain target seed points. Based on the target seed points, the 3D tooth model is segmented to obtain segmented teeth. Thus, the teeth in the 3D tooth model can be divided into individual teeth.

[0188] The above method identifies the original seed points on the three-dimensional tooth model by recognizing the two-dimensional projection image of the three-dimensional tooth model, and expands the original seed points to obtain target seed points. The range of teeth on the three-dimensional tooth model is marked by the target seed points. Furthermore, the three-dimensional tooth model can be segmented according to the target seed points to obtain segmented teeth, thereby achieving the effect of accurately segmenting teeth in the three-dimensional tooth model.

[0189] Optionally, in this embodiment, after obtaining the original seed point, the original seed point can be expanded to obtain the target seed point. The process of expanding the original seed point can be divided into one or more stages.

[0190] For example, in one stage, the original seed point can be expanded according to a preset curvature threshold to obtain the target seed point. The preset curvature threshold can be understood as a constraint used when expanding the original seed point to prevent the expansion from exceeding the limit. The preset curvature threshold can include one or more curvature values. If the preset curvature threshold includes one curvature value, the original seed point can be expanded according to that single curvature value to obtain the target seed point. If the preset curvature threshold includes multiple curvature values, the first curvature value can be used to expand the original seed point, then the second curvature value can be used to expand the result of the expansion of the first curvature value, and the third curvature value can be used to expand the result of the expansion of the second curvature value, and so on, until all curvature values ​​have been used once.

[0191] As an optional example, expanding the original seed point according to a preset curvature threshold to obtain the target seed point includes: expanding the original seed point according to an initial curvature threshold to obtain a first seed point; expanding the first seed point according to a target curvature threshold to obtain the target seed point, wherein the target curvature threshold is obtained based on the initial curvature threshold.

[0192] Taking a preset curvature threshold consisting of an initial curvature threshold and a target curvature threshold as an example, the original seed point is first expanded using the initial curvature threshold to obtain the first seed point. Then, the first seed point is expanded using the target curvature threshold to obtain the target seed point. The initial curvature threshold and the target curvature threshold can be the same or different. Through these two expansions, the target seed point is obtained.

[0193] As an optional example, expanding the original seed point according to the initial curvature threshold to obtain the first seed point includes: taking the seed point adjacent to the original seed point as the current seed point; and taking the current seed point and the original seed point as the first seed point if the curvature of the current seed point is less than or equal to the initial curvature threshold.

[0194] In this embodiment, when expanding the original seed point using an initial curvature threshold, the curvature of the seed points adjacent to the original seed point can be obtained. Curvature is the rate of rotation of the tangent direction angle with respect to the arc length at a point on the curve, defined by differentiation, indicating the degree to which the curve deviates from a straight line. It is a numerical value indicating the degree of curvature of the curve at a point. Each seed point corresponds to a curvature. By comparing the curvature with the initial curvature threshold, it is determined whether the seed points adjacent to the original seed point can be used as the first seed point. The original seed point can be used as the first seed point without comparing its curvature.

[0195] Here, "adjacent" as mentioned below can be understood as two vertices that form the same edge of the same triangular facet as the seed point. For example, "adjacent" to the original seed point means two vertices that form the same edge of the same triangular facet as the original seed point.

[0196] As an optional example, expanding the first seed point according to the target curvature threshold to obtain the target seed point includes: taking the seed point adjacent to the first seed point as the current seed point; and taking the first seed point and the current seed point as the target seed point if the curvature of the current seed point is less than or equal to the target curvature threshold.

[0197] After expanding the original seed point using an initial curvature threshold to obtain the first seed point, the first seed point can be expanded using a target curvature threshold to obtain the target seed point. The curvature of the seed points adjacent to the first seed point is determined, and then this curvature is compared with the target curvature threshold. By comparing their magnitudes, it is determined whether a seed point adjacent to the first seed point is a target seed point. This completes the process of expanding the original seed point to obtain the target seed point using the initial curvature threshold and the target curvature threshold.

[0198] As an optional example, before expanding the first seed point according to the target curvature threshold to obtain the target seed point, the above method further includes: using the sum of the initial curvature threshold and a preset value as the target curvature threshold, wherein the preset value is a positive number.

[0199] Optionally, in this embodiment, the initial curvature threshold and the target curvature threshold can be empirical values, or the initial curvature threshold can be an empirical value, and the target curvature threshold can be determined based on the initial curvature threshold. For example, the sum of the initial curvature threshold and a preset value can be used as the target curvature threshold, that is, the target curvature threshold is obtained based on the initial curvature threshold, and the target curvature threshold is greater than the initial curvature threshold. The preset value is a pre-set value that can be modified according to different three-dimensional tooth models.

[0200] As an optional example, after expanding the original seed point according to the initial curvature threshold to obtain the first seed point, or expanding the first seed point according to the target curvature threshold to obtain the target seed point, the above method further includes: if the corresponding point of the first seed point on the two-dimensional projection image does not fall into the corresponding tooth region, adjusting the first seed point to a non-first seed point; or if the corresponding point of the target seed point on the two-dimensional projection image does not fall into the corresponding tooth region, adjusting the target seed point to a non-target seed point.

[0201] Optionally, in this embodiment, when expanding the original seed point according to the initial curvature threshold or expanding the first seed point according to the target curvature threshold, it is also necessary to check whether the expanded seed point meets the requirements, that is, whether the expansion exceeds the range, and whether a seed point in a non-tooth region is used as the first seed point or a seed point in a non-tooth region is used as the target seed point. For example, when expanding the original seed point according to the initial curvature threshold, if the curvature of a seed point adjacent to the original seed point is less than or equal to the initial curvature threshold, it is also necessary to determine the region where the corresponding point of the seed point is located in the two-dimensional projection image. If it is not located in the tooth region, it means that the seed point has been removed from the tooth region of the three-dimensional tooth model, and therefore, the seed point should be regarded as a non-first seed point. When expanding the first seed point according to the target curvature threshold, if the curvature of a seed point adjacent to the first seed point is less than or equal to the target curvature threshold, it is also necessary to determine the region where the corresponding point of the seed point is located in the two-dimensional projection image. If it is not located in the tooth region, it means that the seed point has been removed from the tooth region of the three-dimensional tooth model, and therefore, the seed point should be regarded as a non-target seed point.

[0202] As an optional example, after expanding the original seed points according to the curvature threshold to obtain the target seed points, and before segmenting the three-dimensional tooth model based on the target seed points of each tooth to obtain the segmented teeth, the above method further includes: expanding the regions obtained by expanding multiple tooth regions to obtain the target region; and expanding the target seed points in the target region according to the initial curvature threshold and height.

[0203] After expanding the original seed points to obtain target seed points, the teeth of the 3D tooth model can be segmented according to the target seed points. Furthermore, the target seed points can be expanded again before segmenting the teeth. That is, in addition to the first stage of expansion using initial and target curvature thresholds to obtain target seed points, a second stage of expansion can be performed.

[0204] In the second stage of expansion, the tooth region on the 2D projection image can be adjusted first to expand the tooth region and obtain the target region. Then, using the target region as a constraint, the target seed points are expanded in the second stage using the initial curvature threshold and height. The purpose of expanding the tooth region into the target region is to ensure that all seed points located on the teeth in the 3D tooth model are marked as target seed points, avoiding omissions.

[0205] As an optional example, when expanding the target seed point according to the initial curvature threshold and height, the seed point adjacent to the target seed point can be used as the current seed point; if the height of the current seed point is greater than the preset standard height and the curvature of the current seed point is less than or equal to the target curvature threshold, the current seed point can be used as the target seed point.

[0206] The height mentioned above can be the numerical value of the seed point in a preset direction of the 3D tooth model. For example, the preset direction of the 3D tooth model can be used as the Z-axis, and the coordinate value of the seed point on the Z-axis can be used as the height of the seed point.

[0207] The preset direction can be any direction. After obtaining the 3D tooth model, the orientation of the 3D tooth model can be adjusted to the preset direction, thereby unifying the orientation of all obtained 3D tooth models.

[0208] When expanding a target seed point using an initial curvature threshold and height, a seed point adjacent to the target seed point can be used as the current seed point. If the curvature and height values ​​of the current seed point meet the requirements of the initial curvature threshold and height, then the current seed point can be used as the target seed point, thereby completing the expansion of the target seed point.

[0209] As an optional example, when the height of the current seed point is greater than a preset standard height and the curvature of the current seed point is less than or equal to a target curvature threshold, using the current seed point as a target seed point includes: when the height of the current seed point is greater than the standard height, the curvature of the current seed point is less than or equal to the target curvature threshold, and the corresponding point of the current seed point on the two-dimensional projection image is located within the target region, the current seed point is used as a target seed point, where the target region is the region obtained after enlarging multiple tooth regions; when the corresponding point of the current seed point on the two-dimensional projection image is located outside the target region, the current seed point is used as a non-target seed point.

[0210] Optionally, when expanding the target seed point using the initial curvature threshold and height, it is essential to ensure that the expanded target seed point does not exceed the target region. The target region is the area obtained after expanding the tooth region. The purpose of expanding the tooth region is to include a portion of the non-tooth region near the tooth region within the target region. This allows the target seed point to extend into the non-tooth region near the tooth region when expanding the target seed point using the initial curvature threshold and height. This ensures that the expanded target seed point covers the entire tooth region.

[0211] As an optional example, after expanding the target seed point according to the initial curvature threshold and height, the above method further includes: taking the seed point adjacent to the expanded target seed point as the current seed point; and taking the current seed point as the target seed point as well.

[0212] After expanding the target seed point using the initial curvature threshold and height, the expanded target seed point can be expanded again, with adjacent seed points also used as target seed points. The purpose of this expansion is also to ensure that the expanded target seed point covers the entire tooth area, so that when segmenting the tooth according to the expanded target seed point, the tooth remains intact.

[0213] As an optional example, before expanding the original seed points within a preset range, the method further includes: expanding multiple tooth regions to obtain a target region; marking points in the target region other than those on the three-dimensional tooth model as third seed points; and expanding the third seed points according to a curvature threshold.

[0214] In this embodiment, the target region is first expanded to include all tooth areas, as well as non-tooth areas near the teeth. Therefore, when points outside the target region on the 3D tooth model are marked as third seed points, these third seed points are not points on the teeth. Expanding the third seed points allows points near the teeth in non-tooth areas (gingival areas or gaps between teeth) to be marked as third seed points. This brings the boundary between the non-tooth and tooth areas closer to the teeth, reducing the size of the target region. While the target region includes non-tooth areas, it contains fewer non-tooth areas.

[0215] After expanding the third seed point, the dividing line between the tooth part and the part outside the tooth on the 3D tooth model will become "thinner".

[0216] As an optional example, expanding the third seed point according to the curvature threshold includes: taking the seed point adjacent to the third seed point as the current seed point; and taking the current seed point as the third seed point if the curvature of the current seed point is greater than the curvature threshold.

[0217] In this embodiment, when expanding the third seed point, it can be expanded according to curvature and a curvature threshold. The curvature of the third seed point can be calculated by differentiating the rate of rotation of the tangent direction angle with respect to the arc length at that point. The curvature threshold can be a preset threshold, and the curvature threshold can be different for different three-dimensional tooth models.

[0218] As an optional example, after taking the seed point adjacent to the third seed point as the current seed point, the method further includes: taking the current seed point as a non-third seed point if the point corresponding to the current seed point in the two-dimensional projection image is located within the target area.

[0219] After expanding the third seed point, it's necessary to check if it has expanded into the target region. Since the target region includes the tooth area, if the third seed point expands into the target region, it may have expanded into the tooth area. Therefore, any third seed points located within the target region should be reverted and treated as non-third seed points.

[0220] As an optional example, after expanding the third seed point according to the curvature threshold, the above method further includes: determining a first region composed of the third seed point; determining a sub-region from the first region; and determining the seed point in the sub-region as the target seed point if the sub-region is surrounded by the target seed point.

[0221] Optionally, in this embodiment, after expanding the third seed point, the third seed point forms a first region. The first region can be divided into multiple sub-regions. Each of the multiple sub-regions can be understood as a region composed of multiple third seed points. If a certain sub-region is surrounded by a target seed point, it means that the sub-region is located within the tooth part in the three-dimensional tooth model, but the sub-region may not be a tooth, that is, the hole part on the tooth. Therefore, the seed point of the sub-region should be used as the target seed point and divided into the tooth part.

[0222] As an optional example, obtaining a two-dimensional projection image of a three-dimensional tooth model includes: adjusting the orientation of the three-dimensional tooth model from an initial orientation to a target orientation; and projecting the three-dimensional tooth model with the target orientation onto a target surface to obtain a two-dimensional projection image.

[0223] Optionally, in this embodiment, when projecting the 3D tooth model into a 2D projection image, the orientation of the 3D tooth model can be adjusted to a target orientation. The target orientation can be a pre-set orientation. The purpose of adjusting the orientation of the 3D tooth model is to make the tooth area in the 2D projection image more complete and less obstructed when projecting the 3D tooth model into a 2D projection image. For example, the horizontal plane can be used as the X and Y axes of the 3D coordinate system, and the upward direction of the horizontal plane can be used as the target orientation. After obtaining the 3D tooth model, the orientation of the teeth in the 3D tooth model can be aligned with the target orientation.

[0224] After adjusting the orientation of the 3D tooth model, it can be projected onto a target surface, which can be a horizontal plane.

[0225] As an optional example, recognizing multiple tooth regions from a two-dimensional projection image includes: inputting the two-dimensional projection image into a recognition model, and having the recognition model mark multiple tooth regions on the two-dimensional projection image.

[0226] Optionally, in this embodiment, the recognition model can identify the two-dimensional projection image to identify the tooth region. After the two-dimensional projection image is input into the recognition model, the recognition model extracts features and identifies the tooth region, outputting multiple tooth regions.

[0227] As an optional example, in a 3D tooth model, identifying the original seed point corresponding to multiple tooth regions includes: taking the vertex of each triangle face in the 3D tooth model as the current vertex; and taking the current vertex as an original seed point if the corresponding point of the current vertex in the 2D projection image is located in multiple tooth regions.

[0228] Optionally, in this embodiment, the surface of the three-dimensional tooth model is covered by polygonal patches, which can be triangular patches. Taking a triangular patch as an example, each triangular patch has three vertices. Two adjacent triangular patches share one edge. Among the vertices of each triangular patch, the point corresponding to the two-dimensional projection image is located within the tooth region, and this vertex can be used as the original seed point.

[0229] As an optional example, after segmenting the 3D tooth model based on the target seed points of each tooth to obtain segmented teeth, the above method also includes: sorting the segmented teeth.

[0230] As an optional example, sorting the segmented teeth includes: using the average value of the midpoints of all teeth as the starting point and the midpoint of each tooth as the ending point to form a vector for each tooth; using the line containing the two teeth with the largest distance between their midpoints as the target line; and sorting all teeth according to the angle between the vector and the target line.

[0231] Optionally, after the teeth are segmented in this embodiment, they can be sorted. The line connecting the midpoint of each tooth to the midpoint of the tooth itself is used as a vector, and the angle between this vector and the target line formed by the two teeth with the largest distance can be used to sort the teeth. That is, the sorting can start from the first tooth on one side and proceed to the last tooth on the other side.

[0232] As an optional example, after segmenting the 3D tooth model based on the target seed points of the teeth to obtain segmented teeth, the above method also includes: smoothing the edges of the teeth to obtain smoothed edges.

[0233] As an optional example, smoothing the edge of a tooth to obtain a smoothed edge includes: sorting the vertices on the edge of the tooth; taking the second vertex in the sorted vertices as the current vertex, and performing the following operations on the current vertex until the current vertex does not include the next vertex: taking the center point of the previous and next vertices of the current vertex as the smoothing point; for each smoothing point obtained, taking the next vertex of the current vertex as the new current vertex, and taking the obtained smoothing point as the previous vertex of the new current vertex; connecting the first vertex on the edge of the tooth with the obtained smoothing points in sequence to obtain the smoothed edge of the tooth.

[0234] In this embodiment, the purpose of smoothing the edges of the teeth is to make the angles of the dividing line between the teeth and the parts outside the teeth smaller, so as to avoid the tooth model being too sharp after being cut, which would cause discomfort to the user's gums after installation.

[0235] During smoothing, the tooth edge passes through a series of seed points that form a curve. Starting from the initial seed point, the average value is taken with the second seed point to obtain the first midpoint. The average value of the first midpoint is taken with the third seed point to obtain the second midpoint. The average value of the second midpoint is taken with the fourth seed point to obtain the third midpoint. This process is repeated until the last seed point on the tooth edge. Connecting all the midpoints sequentially yields the smoothed tooth edge.

[0236] The above three-dimensional tooth model segmentation method will be illustrated with an example below.

[0237] After obtaining the user's oral scan data to create a 3D dental model, the orientation of the 3D dental model can be adjusted. The purpose of adjusting the orientation of the 3D dental model is to ensure that, when projected into a 2D projection image, as much of the tooth portion of the 3D dental model as possible appears on the 2D projection image.

[0238] When adjusting the orientation, a 3D tooth model can be input and aligned towards the positive Z-axis. Any alignment method in this field can be used, without further limitations. Taking placing the teeth of the 3D tooth model upwards on a table as an example, the table can be considered the X and Y axes of a 3D Cartesian coordinate system, and the normal to the side of the table facing the teeth is the Z-axis. After aligning the 3D tooth model in this direction, the teeth will face upwards, resulting in a 2D projection image of the 3D tooth model onto the table.

[0239] After projecting a 3D tooth model onto a plane to obtain a 2D projection image, a recognition model can be used to identify the 2D projection image, thereby identifying the tooth regions within it. The identified tooth regions can be marked with bounding boxes. These boxes can be the smallest axial bounding box containing the masked region or oblique boxes. The identified tooth regions can be regular or irregular in shape, for example… Figure 4 Box 402 in the image represents the identified tooth region. The tooth region includes the teeth and some non-tooth areas. After identifying the tooth region, the 2D projection image allows us to determine the actual tooth area in the 3D model. The tooth region in the 2D projection image corresponds to the tooth portion in the 3D tooth model. All points on this tooth portion are filtered, removing points whose normal directions intersect with the original model. The remaining points after filtering are used as the original seed points.

[0240] For example, the surface of a 3D tooth model (including both tooth and non-tooth parts, the entire surface of the model) is composed of triangular facets (or polygonal facets such as quadrilaterals, pentagons, etc.). These triangular facets are not entirely on the same plane; they have angles between them. The vertices of the triangular facets are seed points. Alternatively, after removing noise from the vertices of the triangular facets, the remaining points are used as seed points. If the corresponding point of a seed point on the 3D tooth model is located in the tooth region in the 2D projection image, then that seed point is used as the original seed point.

[0241] In this embodiment, the original seed point can be expanded to obtain the target seed point. The purpose of expanding the original seed point is that the original seed point may not include all the tooth parts of the 3D tooth model. Therefore, the expansion is used to cover the tooth parts of the 3D tooth model.

[0242] The expansion of the original seed point is divided into multiple stages.

[0243] The first stage involves expanding the original seed point using curvature. Expanding the original seed point essentially means checking if there are any seed points adjacent to the original seed point that can be used together as the target seed point.

[0244] To determine whether seed points are adjacent, we can determine whether seed points lie on the same straight line.

[0245] When expanding the original seed point using curvature, two expansions can be performed sequentially. First, expansion is performed using an initial curvature threshold, followed by expansion using a target curvature threshold. The target curvature threshold is the sum of the initial curvature threshold and a preset value. Therefore, for seed points adjacent to the original seed point, it is first determined whether their curvature is less than the initial curvature threshold. If the curvature is less than the initial curvature threshold, it is used as the first seed point along with the original seed point. Then, for seed points adjacent to the first seed point, if their curvature is less than or equal to the target curvature threshold, they are used as target seed points along with the first seed point. This completes the two-stage expansion of the original seed point. At this point, the first stage of expansion is not yet finished. Because the original seed point has undergone two expansions in the first stage, there may be too many expanded points, exceeding the tooth area. Therefore, it is necessary to determine whether the corresponding position of the expanded point on the 2D projection image falls within the tooth area. If it falls within the tooth area, it means that the expansion of the original seed point on the 3D tooth model has not exceeded the tooth area. If it does not fall within the tooth area, the corresponding seed point is no longer used as a target seed point. This concludes the first phase of expansion.

[0246] In the second stage, the target seed points from the first stage can be expanded using an initial curvature threshold and height. In this stage, seed points adjacent to the expanded target seed points from the first stage can be used as additional seed points. These seed points must have a curvature less than or equal to the initial curvature threshold and a height greater than a preset standard height. If both conditions are met, these seed points are also considered target seed points. If either condition is not met, they are not considered target seed points. In a specific example, the preset standard height can be the height of the current point.

[0247] In addition to meeting the curvature and height requirements, these seed points must also be located within the target region on the 2D projection image. The target region is the enlarged area of ​​the tooth region. That is, the tooth region is slightly enlarged on the 2D projection image to obtain the target region. Then, during the second stage of expansion, if the expanded seed points extend beyond the target region on the 2D projection image, these seed points are not considered target seed points; that is, the expansion is reversed. At this point, the second stage of expansion is complete.

[0248] In other words, the seed points of each tooth are first expanded according to a given curvature threshold, ensuring that they do not exceed the tooth region on the two-dimensional projection image. These points are used as the initial points of each tooth. Based on the initial points, the expansion is performed again according to a given (curvature threshold + 0.2 to obtain the target curvature threshold), ensuring that it does not exceed the bounding box of the tooth region.

[0249] The expansion is performed based on a given curvature threshold and height, ensuring that it does not exceed the bounding box of the tooth region plus the expanded area (i.e., the target region); each tooth boundary point is extended outward once (to the vicinity of the interdental spaces and gingival line), and the extension distance can be preset.

[0250] The initial seed point does not represent the entire area of ​​the tooth; it needs to be continuously expanded. This scheme uses a 2D projection to form the diffusion range. The bounding box of the tooth region plus the expanded region refers to the area that the seed point can diffuse to. The bounding box of the tooth region plus the expanded region forms the diffusion range based on the 2D projection image, and the diffusion range can be controlled through algorithm parameters.

[0251] The reason for multiple stages of tooth region expansion is that the separation between adjacent teeth in some dental models is not obvious. Limiting the height at the end of the tooth region expansion is to reduce the possibility of teeth extending downwards into the gum line. Extending the tooth boundary outwards once is because directly dividing according to the curvature threshold results in teeth that are significantly smaller than the original, due to the low curvature near the gum line, thus requiring further extension.

[0252] In addition to the expansion in the first and second stages, the area outside the teeth can also be expanded. That is, points outside the target area on the 3D tooth model and corresponding to the 2D projection image are designated as third seed points. These third seed points are points outside the teeth, such as points in the gum line or gaps between teeth. These seed points are then expanded towards the teeth. Expansion can be based on curvature. If the curvature of a seed point adjacent to the third seed point is greater than a required curvature threshold, then this seed point is included along with the third seed point as a third seed point, i.e., a point outside the teeth. However, it's important to note that if the expanded third seed point is located within the target area on the 2D projection image, it may have expanded into the teeth. Therefore, a rollback is necessary, and third seed points located within the target area are treated as non-third seed points.

[0253] The boundary between the target seed point obtained from the first and second stages of expansion, and the third seed point obtained after expansion, can be used as the boundary of the tooth.

[0254] If the tooth contains a cavity, there may be an enclosing relationship between the target seed point and the third seed point. For example, if the target seed point encloses part of the third seed point, this is because the cavity is identified as the third seed point. In this case, points within the region enclosed by the target seed point are also considered target seed points.

[0255] After expansion, the teeth are segmented according to the target seed point and the third seed point to obtain individual teeth. Alternatively, the boundaries of the teeth can be smoothed before segmentation, resulting in gentler boundaries.

[0256] After segmenting the teeth into individual teeth, they can be sorted. The average midpoint of the teeth in the 3D tooth model is used as the starting point, and the midpoint of each tooth is used as the midpoint, connecting to multiple vectors. The line connecting the two teeth with the largest midpoints is taken as the target line. The angle between the vector and the target line is then examined, and the teeth are sorted according to the size of the angle.

[0257] It should be noted that the above-mentioned scheme for automatically arranging misaligned teeth can be as follows: establish a tooth alignment coordinate system; define the feature points of a single tooth and establish a local coordinate system for the tooth; based on this, analyze the position and orientation of each tooth in the dentition from a low-dimensional perspective, and use a weighted fitting optimization method to calculate the coordinate translation and local coordinate axis rotation of the teeth respectively, forming a correlation constraint between the tooth pose and the spatial dentition curve. Combined with the collision detection method of the rectangular bounding box, an iterative algorithm based on the steepest descent method is designed to adjust the tooth pose within the constraint range of the spatial dentition curve, thereby completing the automatic alignment of teeth.

[0258] Step 3: Based on the pre-processed digital tooth model, a shaping process is performed to obtain a solid tooth model.

[0259] In one embodiment, a shaped tooth model is obtained by printing based on the target segmented teeth, wherein the shaped tooth model is used to obtain dental instruments. That is, the target segmented teeth are arranged together to obtain a digital tooth model, and a solid tooth model is obtained based on a molding process (also called printing). This 3D printing can employ photopolymerization 3D printing, such as SLA, DLP, LCD, or other 3D printing methods, such as 3DP, MJF, FDM, Polyjet, etc.

[0260] In some embodiments, based on the aforementioned preprocessing operations, the resulting solid dental model may have a hollowed-out base plate, fixing accessories, identification information, etc.

[0261] Furthermore, 3D printing also includes post-processing steps, which can be selected depending on the 3D printing method used. For example, when using photopolymer 3D printing technology, post-processing steps can include post-curing, cleaning, etc.

[0262] Of course, in other embodiments, the molding method is not limited to 3D printing, and other molding methods, such as injection molding, can also be used, which are not limited here.

[0263] Step 4: Press the solid dental model into a mold.

[0264] After obtaining the solid dental model, a preheated polymer film can be pressed onto the solid dental model to obtain a shell-shaped film covering the solid dental model, which is the initial instrument that has not yet been cut. Specifically, the shell-shaped film covers at least the tooth portion of the solid dental model.

[0265] In one embodiment, an image of a physical dental model can be acquired through an image acquisition terminal, such as a CCD image sensor, and then the identification information can be identified through recognition technologies such as OCR. Based on the identification information, the corresponding pressing instruction can be retrieved from the database, and pressing processing can be performed based on the dental model instruction. The pressing instruction may include pressure parameters, diaphragm preheating time, pressing temperature, etc.

[0266] Step 5: Mark the shell-shaped membrane.

[0267] This step mainly involves marking the shell-shaped diaphragm based on the identification information added in the pre-processing operation, so as to form a mark on the shell-shaped diaphragm and obtain an initial device with the mark.

[0268] Specifically, similar to the steps described above, an image of the dental model can be acquired using an image acquisition terminal, such as a CCD image sensor. Then, the identification information can be identified using recognition technologies such as OCR. Based on the identification information, the corresponding marking instruction can be invoked to mark the shell-shaped diaphragm based on the parameter information contained in the marking instruction.

[0269] It should be noted that the manufacturing method of dental instruments does not limit the execution of this step; that is, whether or not a marking operation is required can be chosen according to actual needs.

[0270] Step 6: Cut the initial instrument to obtain dental instruments.

[0271] This step primarily involves cutting the initial instrument using a cutting line derived from the extracted gingival margin. This process removes the useless parts of the initial instrument, retaining only the useful components, thus obtaining a dental instrument. Specifically, this dental instrument can be an invisible aligner.

[0272] It should be noted that the acquisition of the cutting line in this step can be the same as in the aforementioned embodiment of the dental instrument manufacturing method. For details, please refer to the aforementioned implementation method, which will not be repeated here. In addition, the order of steps five and six is ​​not limited in this embodiment.

[0273] In one embodiment, similar to the steps described above, an image of a dental model can be acquired using an image acquisition terminal, such as a CCD image sensor. Then, the identification information can be identified using recognition technologies such as OCR. Based on the identification information, the corresponding cutting command can be retrieved to cut the initial instrument based on the parameter information contained in the cutting command.

[0274] According to another aspect of the embodiments of this application, such as Figure 12 As shown, a three-dimensional tooth model segmentation device is provided, including: a memory 801 and a processor 803. The memory stores a computer program, and the computer program is executed by the processor to perform the above-described three-dimensional tooth model segmentation method.

[0275] It should be noted that the examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should also be noted that the above modules, as part of a device, can operate in environments such as... Figure 1 The hardware environment shown can be implemented either through software or hardware. The device may also include other components, such as the communication interface 805 and the connecting cable 807.

[0276] According to another aspect of the embodiments of this application, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps of any of the above embodiments.

[0277] According to another aspect of the embodiments of this application, an electronic device is provided, including a memory, a processor, a communication interface and a communication bus. The memory stores a computer program that can run on the processor. The memory and the processor communicate with each other through the communication bus and the communication interface. When the processor executes the computer program, it implements the steps of the above-described three-dimensional tooth model segmentation method.

[0278] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0279] In specific implementation, the embodiments of this application can be referred to the above embodiments and have corresponding technical effects.

[0280] It is understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described herein, or combinations thereof.

[0281] For software implementation, the techniques described herein can be implemented by units that perform the functions described herein. The software code can be stored in memory and executed by a processor. The memory can be implemented in the processor or external to the processor.

[0282] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0283] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0284] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0285] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0286] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0287] If the aforementioned function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks. It should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. In the absence of further restrictions, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0288] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for segmenting a three-dimensional tooth model, characterized in that, include: Obtain the gingival line and the tooth region of the three-dimensional tooth model; Extract the peaks of the gingival line and pair the peaks to obtain peak combinations; Based on the combination of wave crests, the segmentation path between teeth in the tooth region is determined; The tooth region is segmented according to the segmentation path to obtain the target segmented tooth; The acquisition of the gingival line from the three-dimensional tooth model includes: - The gingival line is obtained by recognizing a two-dimensional projection image of the three-dimensional tooth model; or - Determine the edge categories of the edges of multiple polygonal facets of the three-dimensional tooth model, and determine the gingival line of the three-dimensional tooth model based on the edge categories; wherein, the three-dimensional tooth model includes multiple polygonal facets, and the edge categories include tooth edges and gingival edges; The extraction of the peak points of the gingival line includes: - Based on the target orientation, determine the coordinates of each point on the gingival line, and based on the coordinates, determine the peak point on the gingival line, wherein the coordinates of the peak point are greater than the coordinates of the adjacent points on the gingival line; and / or, - Determine the insertion control point based on the curvature of the gingival line, and determine the peak and trough points based on the height difference in the target direction; filter out the trough points to obtain the final peak points; The method further includes, prior to obtaining the gingival line of the three-dimensional tooth model: - If the three-dimensional tooth model is a non-closed model, determine the minimum directed bounding box of the three-dimensional tooth model; - Determine the initial orientation of the three-dimensional tooth model based on the minimum directed bounding box; - Adjust the three-dimensional tooth model from the initial orientation to the target orientation; or - In the case that the three-dimensional tooth model is a closed model, determine the polygon patch group to which the polygon patches on the surface of the three-dimensional tooth model belong, wherein the included angle between any two polygon patches in the same polygon patch group is less than a first threshold. - The initial orientation of the three-dimensional tooth model is determined based on the sum of the areas of the polygonal facets in the polygonal facet group; - Adjust the three-dimensional tooth model from the initial orientation to the target orientation.

2. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, Obtaining the tooth region of the three-dimensional tooth model includes: The three-dimensional tooth model is segmented based on the gingival line to obtain the tooth region of the three-dimensional tooth model.

3. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, Determining the initial orientation of the three-dimensional tooth model based on the minimum directed bounding box includes: The preset axis direction is determined based on the two faces with the largest areas of the minimum oriented bounding box; Based on the preset axial direction, the first average coordinate value of the boundary edge of the three-dimensional tooth model and the second average coordinate value of the vertices of the polygonal facet of the three-dimensional tooth model are obtained. The initial orientation is determined based on the first average coordinate value and the second average coordinate value.

4. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, When the three-dimensional tooth model is a closed model, determining the polygon patch group to which the polygon patches on the surface of the three-dimensional tooth model belong includes: Select any one first facet from all faces that have not been assigned to a polygon facet group as a facet in a polygon facet group. Determine the polygon facets whose plane normal angle with the first facet is less than a first threshold as facets in the same polygon facet group as the first facet. In the initial case, all polygon facets on the three-dimensional tooth model have not been assigned to polygon facet groups. Continue selecting any target face from all faces that have not been assigned to a polygon face group as a face in another polygon face group, and determine polygon faces whose plane angle with the target face is less than a first threshold as faces in the same polygon face group as the target face, until all polygon faces are assigned to polygon face groups.

5. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, When the three-dimensional tooth model is a closed model, determining the polygon patch group to which the polygon patches on the surface of the three-dimensional tooth model belong includes: Select any one first facet from all faces that have not been assigned to a polygon facet group as a facet in a polygon facet group. Determine the polygon facets whose normals are at an angle less than a first threshold as facets in the same polygon facet group as the first facets. In the initial case, all polygon facets on the three-dimensional tooth model are not assigned to polygon facet groups. In all faces not assigned to a polygon patch group, select any target patch as a patch in another polygon patch group, and determine the polygon patches whose normals are at an angle less than a first threshold as patches in the same polygon patch group as the target patch, until all polygon patches are assigned to polygon patch groups.

6. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, The gingival line for obtaining the three-dimensional tooth model includes: Extract features from the three-dimensional tooth model; The gingival line is obtained by identifying the aforementioned features.

7. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, The peak points are paired to obtain the peak point combinations, including: Based on the positional relationship between the peak points and the three-dimensional tooth model, the peak points are divided into a first peak point group and a second peak point group. Each peak in the first peak point group is taken as the first peak point, and the first peak point and the second peak point in the second peak point group are taken as a peak point combination. The second peak point is the peak point in the second peak point group that is closest to the first peak point. The angle between the line connecting the first peak point and the second peak point and the tooth midline of the three-dimensional tooth model is greater than a third threshold.

8. The three-dimensional tooth model segmentation method according to claim 7, characterized in that, After taking each peak point in the first peak point group as the first peak point, and combining the first peak point with the second peak point in the second peak point group as a peak point combination, the method further includes: Remove the successfully paired peak points from the first peak point group and the second peak point group; Each peak in the second peak group is taken as the third peak, and the third peak and the fourth peak in the first peak group are taken as a peak combination. The fourth peak is the peak closest to the third peak in the first peak group, and the angle between the line connecting the third peak and the fourth peak and the tooth midline of the three-dimensional tooth model is greater than a third threshold.

9. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, The step of determining the segmentation path between teeth in the tooth region based on the combination of wave crests includes: Each of the aforementioned peak point combinations is determined as the current combination; The shortest distance between the two peaks in the current combination on the three-dimensional tooth model is determined as a segmentation path.

10. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, The step of determining the segmentation path between teeth in the tooth region based on the combination of wave crests includes: Each of the aforementioned peak point combinations is determined as the current combination; The shortest distance between the two peak points in the current combination on the three-dimensional tooth model is determined as the first path; The first path is corrected by superimposing curvature to obtain the target path; The target path is determined as the segmentation path.

11. The three-dimensional tooth model segmentation method according to claim 10, characterized in that, The step of correcting the first path by superimposing curvature to obtain the target path includes: Each point on the first path is taken as the current point, and a replacement point for the current point is determined within a preset range on the three-dimensional tooth model; wherein, the replacement point is a point that meets the curvature threshold requirement; The line connecting the replacement points is taken as the target path.

12. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, The determination of the edge categories of the edges of the plurality of polygonal facets includes: Obtain the target features of the edges of the plurality of polygonal patches, wherein the target features include geometric features; Based on the target features, determine the edge category of the edge.

13. The three-dimensional tooth model segmentation method according to claim 12, characterized in that, Determining the edge category based on the target features includes: The target feature of the initial dimension of the edge is increased in dimensionality to obtain the first feature of the first dimension; The first feature is reduced in dimensionality to obtain the second feature of the target dimension. The edge is determined to be a tooth edge or a gingival edge based on the magnitude of the value of the second feature.

14. The three-dimensional tooth model segmentation method according to claim 1, characterized in that, Determining the gingival line of the target tooth model based on the edge category of the edge includes: Based on the edge category, the face category of the polygonal face is determined, wherein the face category includes tooth face and gingival face; The gingival line is extracted based on the shared edges of adjacent polygonal faces of different face types.

15. A method for manufacturing a dental instrument, characterized in that, include: The method described in any one of claims 1-14 yields the target segmented tooth; A shaped tooth model is obtained by printing based on the target segmented tooth; wherein, the shaped tooth model is used to obtain dental instruments.

16. The method for manufacturing a dental instrument according to claim 15, characterized in that, Also includes: Convert the gingival line into a cutting line; Based on the cutting line, the initial instrument is cut to obtain the dental instrument; wherein the initial instrument is related to the formed tooth model.

17. The method for manufacturing a dental instrument according to claim 16, characterized in that, Before converting the gingival line into a cutting line, the process also includes: At least one of smoothing and repositioning processing is performed on the gingival line.

18. The method for manufacturing a dental instrument according to claim 15, characterized in that, Also includes: The molded tooth model is subjected to a pressing process to obtain a shell-shaped film covering the solid tooth model; The shell-shaped diaphragm is marked to obtain an initial device with markings.

19. A three-dimensional tooth model segmentation device, characterized in that, include: A memory and a processor, the memory storing a computer program that, when executed by the processor, performs the method according to any one of claims 1 to 14 or 15-18.

20. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is run by the processor, it performs the method described in any one of claims 1-14 or 15-18.

Images