False tooth edge line determination method and device, computer equipment and storage medium

Abstract

The invention provides a false tooth edge line determination method and device, computer equipment and a storage medium, and relates to the technical field of false tooth design, the method comprises the following steps: obtaining first oral data and second oral data corresponding to a target object; wherein the first oral cavity data is used for representing the oral cavity of the target object in a first pressure state, the second oral cavity data is used for representing the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is smaller than that in the first pressure state; performing false tooth reference line identification on the first oral cavity data according to the trained AI model to obtain a first false tooth reference line; and mapping the first denture reference line to the second oral cavity data to obtain a target denture edge line. The first oral cavity data and the second oral cavity data in different pressure states are combined, the false tooth edge line of the second oral cavity data with high precision is automatically and accurately recognized, and meanwhile, subsequent digital storage and design are facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of denture design, and in particular to a method, device, computer equipment and storage medium for determining a denture edge line. Background Art

[0002] In dentistry, especially when making complete dentures or removable partial dentures, identifying the static line is crucial to ensure the retention and function of the denture. The static zone refers to the area related to the activity of the masticatory muscles that needs to be considered during denture restoration. The mucosa in these areas does not move during physiological activities such as chewing and speaking. The dividing line between the static zone and the dynamic zone is the static line. Based on the static line, the denture edge line can be determined, so the corresponding denture can be designed. Therefore, correctly identifying the static line helps to design a more stable and comfortable denture.

[0003] Currently, myostatic lines are usually obtained by manual marking, for example, by manually marking the myostatic lines directly on the plaster model, or by manually marking the myostatic lines directly on the oral digital impression or the scanned plaster model based on digital drawing software tools; wherein the oral digital impression is obtained by scanning the inside of the mouth with a scanner; the plaster model is a physical model made of plaster obtained by traditional molding, and the scanned plaster model is obtained by scanning the plaster model with a scanner.

[0004] However, the above-mentioned manual marking of myostatic lines relies on the clinical experience of designers and other relevant personnel, is easily affected by subjective factors, is difficult to achieve consistency and standardization, and requires relevant personnel to invest a lot of time and energy, with high labor costs and low efficiency. Among them, in the manual marking of myostatic lines based on plaster models, plaster models need physical storage, and sharing and transmission are not as convenient as oral digital impressions. In addition, the shrinkage of impression materials, the expansion of plaster or other physical factors will cause deviations between plaster models and actual oral conditions, and the accuracy of plaster models is not as good as oral digital impressions; the manual marking of myostatic lines based on digital drawing software tools requires the support of relevant marking tools and software, has a certain learning cost, and is inconvenient to operate; in addition, oral digital impressions are impressions close to a pressure-free state, and the patient's mouth is basically in a static state during intraoral scanning, so it is impossible to distinguish between the myostatic area and the myodynamic area, and it is impossible to obtain accurate denture edge lines. Summary of the invention

[0005] The object of the present invention is to provide a method, device, computer equipment and storage medium for determining the edge line of a denture, so as to automatically and accurately identify the edge line of a denture, and facilitate subsequent digital storage and design.

[0006] In a first aspect, an embodiment of the present invention provides a method for determining a denture edge line, comprising: Acquire first oral cavity data and second oral cavity data corresponding to the target object; wherein the first oral cavity data is used to represent the oral cavity of the target object in a first pressure state, and the second oral cavity data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; Perform denture reference line recognition on the first oral cavity data according to the trained AI model to obtain the first denture reference line; According to the registration information between the first oral cavity data and the second oral cavity data, the first denture reference line is mapped onto the second oral cavity data to obtain the target denture edge line.

[0007] Further, the denture reference line is identified on the first oral data according to the trained AI model to obtain the first denture reference line, including: Preprocessing the first oral cavity data to obtain preprocessed first oral cavity data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, sparse coding representation, and surface feature fusion; Input the preprocessed first oral cavity data into the AI ​​model to obtain the denture reference line recognition result output by the AI ​​model; wherein the AI ​​model is obtained by training a three-dimensional deep learning model based on a plurality of sample oral cavity data with denture reference line annotation data, and the sample oral cavity data is used to represent the oral cavity of the sample object in the first pressure state; According to the denture reference line recognition result, the first denture reference line is determined.

[0008] Further, the first denture reference line includes a first myostatic line, the denture reference line annotation data includes myostatic annotation data, and the myostatic annotation data includes myostatic line data or myostatic area data; or, The first denture reference line includes a first denture edge line, and the denture reference line annotation data includes myostatic annotation data or denture edge line annotation data.

[0009] Further, according to the registration information between the first oral cavity data and the second oral cavity data, the first denture reference line is mapped onto the second oral cavity data to obtain the target denture edge line, including: Using a preset rigid region, registering the first oral cavity data and the second oral cavity data to obtain registration information; wherein the rigid region includes the alveolar ridge and / or the upper jaw; According to the registration information, the first denture reference line is mapped onto the second oral cavity data to obtain the second denture reference line; Based on the second denture reference line, determine the target denture edge line.

[0010] Further, according to the registration information, the first denture reference line is mapped onto the second oral cavity data to obtain the second denture reference line, including: Obtain multiple sampling points from the first denture reference line; Determine matching points of each sampling point on the second oral cavity data according to the registration information; Perform three-dimensional curve fitting on each matching point to obtain the second denture reference line.

[0011] Further, determining the target denture edge line according to the second denture reference line includes: Performing smoothing on the second denture reference line to obtain a smoothed denture reference line; Determine the target denture edge line based on the denture reference line after smoothing.

[0012] Furthermore, the above method also includes: Acquire multiple sample oral cavity data with denture reference line annotation data; The three-dimensional deep learning model is trained according to the oral data of each sample to obtain an AI model; wherein the three-dimensional deep learning model includes any of the following: PointNet model, graph convolutional neural network model, MeshCNN model, multi-view model that parses three-dimensional information from multiple perspectives, convolutional neural network model based on graph theory, and diffusion network DiffusionNet model.

[0013] Furthermore, a plurality of sample oral cavity data with denture reference line annotation data is obtained, including: Acquire a plurality of initial oral cavity data with denture reference line annotation data; wherein the initial oral cavity data is used to represent the oral cavity of the sample object in a first pressure state; Preprocessing is performed on a plurality of initial oral data to obtain a plurality of sample oral data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, multi-scale data enhancement, sparse coding representation, and surface feature fusion.

[0014] Further, the first oral cavity data is obtained by three-dimensionally scanning a plaster model obtained by impression of the oral cavity of the target object under the first pressure, or by three-dimensionally scanning the historical denture of the target object, or by scanning the oral cavity of the target object under the state of blowing pressure; The second oral cavity data is obtained by directly scanning the oral cavity of the target object without physical pressure or under standard atmospheric pressure.

[0015] Furthermore, after obtaining the first oral cavity data and the second oral cavity data corresponding to the target object, the method further includes: Performing data integrity judgment on the first oral cavity data and the second oral cavity data to obtain a degree of integrity; Based on the degree of completeness, determine the accuracy of the target denture edge line.

[0016] Further, the first oral cavity data and the second oral cavity data are subjected to data integrity judgment to obtain the integrity degree, including: Determine whether the first oral data and / or the second oral data include a preset key area, and obtain a determination result; wherein the key area corresponding to the first oral data includes a first designated area, and the key area corresponding to the second oral data includes a second designated area; Determine the degree of completeness based on the inclusion judgment results; After determining the data integrity of the first oral cavity data and the second oral cavity data to obtain the integrity level, the method further includes: Determining whether the integrity level is less than a preset integrity threshold; If it is less than the completeness threshold, a reminder of incomplete data will be issued.

[0017] Furthermore, according to the registration information between the first oral cavity data and the second oral cavity data, after mapping the first denture reference line to the second oral cavity data to obtain the target denture edge line, the method further includes: When a first instruction is detected, determining a first pose change of the target denture edge line relative to the initial position, wherein the first instruction is an instruction for the user to adjust the pose of the target denture edge line in the second oral cavity data; Based on the first posture change and the preset reference condition, determine the first adjusted posture of the target denture edge line, and update and display the posture of the target denture edge line based on the first adjusted posture of the target denture edge line, wherein the preset reference condition includes one or more of the target denture edge line fitting with the outer surface of the second oral data, the target denture edge line conforming to the simulated physiological movement, or the target denture edge line being located in the denture coverage area; Wherein, when it is detected that the first pose change makes the target denture edge line not meet the preset reference condition, the pose of the target denture edge line is adaptively updated according to the preset reference condition, or, the pose of the target denture edge line is not updated and the result of user reconfirmation is obtained and the pose of the target denture edge line is adjusted according to the result of user reconfirmation; and / or When it is detected that the first adjusted posture causes the target denture edge line to fail to meet the preset reference conditions, the posture of the target denture edge line is adaptively updated according to the preset reference conditions, or the posture of the target denture edge line is not updated and the result of user re-confirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user re-confirmation.

[0018] Furthermore, according to the registration information between the first oral cavity data and the second oral cavity data, after mapping the first denture reference line to the second oral cavity data to obtain the target denture edge line, the method further includes: Generate a three-dimensional denture model according to the target denture edge line and the second oral cavity data, and send the three-dimensional denture model to a 3D printing device for printing the three-dimensional denture model; When a user request for modifying the three-dimensional denture model is received, a revised target denture margin line and / or a revised three-dimensional denture model is output according to the user request, the target denture margin line and the second oral data.

[0019] In a second aspect, an embodiment of the present invention further provides a device for determining a denture edge line, comprising: A data acquisition module, used to acquire first oral cavity data and second oral cavity data corresponding to the target object; wherein the first oral cavity data is used to represent the oral cavity of the target object in a first pressure state, and the second oral cavity data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; A denture reference line recognition module, used to perform denture reference line recognition on the first oral data according to the trained AI model to obtain a first denture reference line; The denture edge line mapping module is used to map the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data to obtain the target denture edge line.

[0020] In a third aspect, an embodiment of the present invention further provides a computer device, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and when the processor executes the computer program, the denture edge line determination method of the first aspect is implemented.

[0021] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the denture edge line determination method of the first aspect is executed.

[0022] The denture edge line determination method, device, computer equipment and storage medium provided by the embodiment of the present invention can obtain the first oral data and the second oral data corresponding to the target object when determining the denture edge line; wherein the first oral data is used to represent the oral cavity of the target object in a first pressure state, and the second oral data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; the denture reference line is identified on the first oral data according to the trained AI model to obtain the first denture reference line; according to the registration information between the first oral data and the second oral data, the first denture reference line is mapped to the second oral data to obtain the target denture edge line. Since the pressure at the time of impression corresponding to the first oral data is greater than the pressure at the time of impression corresponding to the second oral data, the accuracy of the second oral data is higher than that of the first oral data, and the first denture reference line corresponding to the first oral data is not easily affected by the diversity of the tooth and jaw morphology and the quality of the mesh surface. The first denture reference line corresponding to the first oral data can be accurately identified through artificial intelligence learning, and then the first denture reference line is mapped to the second oral data with higher accuracy, which can facilitate subsequent data storage and design. In this way, the first oral data under different pressure states are combined with the second oral data to achieve the effect of edge trimming, automatically and accurately identify the denture edge line of the second oral data, and facilitate subsequent digital storage and design. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A schematic flow chart of a method for determining a denture edge line provided by an embodiment of the present invention; Figure 2 A schematic diagram of the appearance of an oral digital impression provided by an embodiment of the present invention; Figure 3 A schematic diagram of the appearance of a target three-dimensional plaster model provided by an embodiment of the present invention; Figure 4 A schematic diagram of a target three-dimensional plaster model with marked myostatic lines provided by an embodiment of the present invention; Figure 5 A schematic diagram of a muscle static line mapping provided by an embodiment of the present invention; Figure 6 A schematic diagram of displaying a target denture edge line on an interactive interface provided by an embodiment of the present invention; Figure 7 A schematic structural diagram of a denture edge line determination device provided by an embodiment of the present invention; Figure 8 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] Currently, the myostatic lines are usually manually marked on the oral digital impression or the scanned plaster model based on digital drawing software tools to obtain the denture edge line. However, this method has the following disadvantages: a) clinical experience is required to ensure that there is room for design; b) a lot of time and effort is required; c) support from related marking tools and software is required, and there is a learning cost; d) compared with the traditional direct marking on the plaster model, marking and editing on the three-dimensional model in the digital drawing software tool is a considerable challenge for designers; e) it is affected by human subjective factors and it is difficult to achieve consistency and standardization.

[0027] With the development of image processing technology, artificial intelligence technology can be used to automatically identify myostatic lines for oral digital impressions (such as digital impressions of edentulous jaws). However, since oral digital impressions are impressions close to being made in a pressure-free state, the method of automatically identifying myostatic lines directly on oral digital impressions cannot obtain accurate myostatic boundaries, and thus cannot obtain accurate denture edge lines, nor can it be used to pressurize the production of the posterior dyke area. The shrinkage of the material, the expansion of the plaster, or other physical factors can cause the plaster model obtained by the traditional method of taking impressions to deviate from the actual oral condition, and the accuracy is not as good as that of oral digital impressions.

[0028] Based on this, a denture edge line determination method, device, computer equipment and storage medium provided in an embodiment of the present invention can fully automatically identify denture edge lines on three-dimensional oral digital impressions based on the application of deep learning technology in three-dimensional grids. This method is highly robust and will not be affected by the diversity of dental and maxillary morphology and the quality of mesh surfaces, providing solid technical support for the subsequent design and manufacturing process of dentures. Exemplarily, when obtaining denture edge lines, oral digital impressions and traditional molding schemes can be combined, and a plaster model (such as an edentulous plaster model) obtained by a traditional molding method is scanned by a scanner, and the myostatic line or denture edge line is identified on the scanned plaster model by artificial intelligence learning, and then the myostatic line or denture edge line is mapped to the oral digital impression obtained by the intraoral scanner.

[0029] To facilitate understanding of this embodiment, a method for determining a denture edge line disclosed in an embodiment of the present invention is first introduced in detail.

[0030] The embodiment of the present invention provides a method for determining a denture edge line, which can be executed by a computer device with data processing capability. Figure 1 The flowchart of a method for determining a denture edge line is shown in FIG. 1 , and the method mainly includes the following steps S110 to S140: Step S110, obtaining first oral cavity data and second oral cavity data corresponding to the target object.

[0031] The first oral cavity data is used to represent the oral cavity of the target object in a first pressure state, and the second oral cavity data is used to represent the oral cavity of the target object in a second pressure state, where the pressure in the second pressure state is less than the pressure in the first pressure state.

[0032] Optionally, the first oral data may be obtained by three-dimensionally scanning a plaster model obtained by impression of the oral cavity of the target object under a first pressure, or by three-dimensionally scanning a historical denture (old denture) of the target object, or by scanning the oral cavity of the target object under a state of blowing pressure. The first pressure state may refer to the pressure state of the impression on the oral cavity, the pressure state of the historical denture on the oral cavity, or the pressure state of blowing on the oral cavity. The second oral data may be obtained by directly scanning the oral cavity of the target object without physical pressure or at standard atmospheric pressure (the second oral data may be referred to as an oral digital impression). The second pressure state may refer to no physical pressure or standard atmospheric pressure.

[0033] Specifically, the target object may be a person who needs to install dentures. The first oral data may be obtained by performing a three-dimensional scanning of a plaster model corresponding to the target object using scanning technology. In this case, the first oral data may be referred to as a target three-dimensional plaster model (that is, a three-dimensional plaster model corresponding to the target object). The first oral data and the second oral data may also be obtained by performing an oral scanning of the target object using an intraoral scanner. The difference lies in the different pressures on the target object's oral cavity during scanning.

[0034] The oral cavity of the target object can be scanned using an intraoral scanner, an extraoral scanner, or other scanning equipment. Figure 2 The oral digital impression shown can be scanned using an extraoral scanner, intraoral scanner, desktop scanner, etc. Figure 3 The target three-dimensional plaster model is obtained by using the plaster model shown in FIG. Figure 2 and Figure 3 The digital oral impression and the target three-dimensional plaster model corresponding to a target object are shown respectively. Both the digital oral impression and the target three-dimensional plaster model can be gridded data obtained by three-dimensional reconstruction, point cloud registration and gridding after three-dimensional scanning. The specific acquisition process of the digital oral impression and the target three-dimensional plaster model can refer to the relevant existing technology and will not be repeated here.

[0035] Oral digital impressions can be obtained through high-precision optical oral scanning technology. The acquisition process does not involve physical pressure, which makes it difficult to accurately record the characteristics of the myostatic line under the muscle functional state (i.e. muscle movement state), thereby affecting the denture edge line. Relatively speaking, the traditional plaster model is affected by the functional pressure of the oral soft tissue during the production process, which can more accurately reflect the anatomical morphology of the myostatic line, so that the denture edge line obtained is more accurate.

[0036] Step S120, performing denture reference line recognition on the first oral data according to the trained AI model to obtain a first denture reference line.

[0037] The first denture reference line may be the first myostatic line on the first oral cavity data, or the first denture edge line on the first oral cavity data. Exemplarily, the myostatic line / denture edge line in the scanned target three-dimensional plaster model may be identified by using three-dimensional deep learning technology to obtain the first myostatic line / first denture edge line, such as Figure 4 As shown in the figure, the line indicated by the arrow is the first muscle static line.

[0038] In some possible embodiments, the AI ​​model may be a denture reference line recognition model, which is obtained by training a three-dimensional deep learning model based on a plurality of sample oral data with denture reference line annotation data, and the sample oral data is used to represent the oral cavity of the sample object in a first pressure state, and the training process of the AI ​​model will be described in detail later. Based on this, step S120 may include: using the denture reference line recognition model to recognize the first oral data to obtain a first denture reference line.

[0039] In some other possible embodiments, considering that specific anatomical points in the oral cavity can be used to locate denture reference lines such as myostatic lines or denture edge lines, the AI ​​model can be an anatomical point recognition model, which is obtained by training a three-dimensional deep learning model based on multiple sample oral data with anatomical point annotation data, and the sample oral data is used to represent the oral cavity of the sample object in a first pressure state. Based on this, step S120 may include: using the anatomical point recognition model to recognize the first oral data to obtain the target anatomical point; based on the target anatomical point, determining the first denture reference line. In one possible implementation, the target anatomical point can be firstly curve-fitted to obtain the anatomical point curve, and then the anatomical point curve can be preset and adjusted to obtain the first denture reference line. In another possible implementation, the reference point corresponding to the target anatomical point can be firstly determined, and then the reference point can be curve-fitted to obtain the first denture reference line.

[0040] Step S130, mapping the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data, to obtain a target denture edge line.

[0041] Taking the first oral data as the target three-dimensional plaster model and the second oral data as the oral digital impression as an example, the common rigid areas such as the alveolar ridge and / or the upper jaw can be used to obtain the alignment information between the oral digital impression and the target three-dimensional plaster model. The aligned oral digital impression and the target three-dimensional plaster model are in the same coordinate system, so that the first denture reference line can be mapped to the oral digital impression to obtain the target denture edge line.

[0042] The method for determining the denture edge line provided by the embodiment of the present invention can obtain the first oral data and the second oral data corresponding to the target object when determining the denture edge line; wherein the first oral data is used to represent the oral cavity of the target object in a first pressure state, and the second oral data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; the denture reference line is identified on the first oral data according to the trained AI model to obtain the first denture reference line; according to the registration information between the first oral data and the second oral data, the first denture reference line is mapped to the second oral data to obtain the target denture edge line. Since the pressure at the time of impression corresponding to the first oral data is greater than the pressure at the time of impression corresponding to the second oral data, the accuracy of the second oral data is higher than that of the first oral data, and the first denture reference line corresponding to the first oral data is not easily affected by the diversity of the tooth and jaw morphology and the quality of the mesh surface. The first denture reference line corresponding to the first oral data can be accurately identified through artificial intelligence learning, and then the first denture reference line is mapped to the second oral data with higher accuracy, which can facilitate subsequent data storage and design. In this way, the first oral data under different pressure states are combined with the second oral data to achieve the effect of edge trimming, automatically and accurately identify the denture edge line of the second oral data, and facilitate subsequent digital storage and design.

[0043] For ease of understanding, the specific contents involved in the above-mentioned denture edge line determination method are introduced in detail below.

[0044] In order to improve the calculation efficiency and recognition accuracy, the step S120 of performing denture reference line recognition on the first oral data according to the trained AI model to obtain the first denture reference line may include the following sub-steps: Sub-step S121, preprocessing the first oral cavity data to obtain preprocessed first oral cavity data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, sparse coding representation and surface feature fusion. These preprocessing processes will be introduced in detail later.

[0045] In sub-step S122, the preprocessed first oral data is input into the AI ​​model to obtain the denture reference line recognition result output by the AI ​​model; wherein the AI ​​model is obtained by training a three-dimensional deep learning model based on multiple sample oral data with denture reference line annotation data, and the sample oral data is used to represent the oral cavity of the sample object in a first pressure state.

[0046] The first denture reference line can be the first myostatic line. In this case, the denture reference line annotation data can be myostatic annotation data, which includes myostatic line data or myostatic area data. The denture reference line recognition result output by the AI ​​model is a myostatic recognition result, which can be an identified myostatic line or an identified myostatic area.

[0047] The first denture reference line may also be the first denture edge line. In this case, the denture reference line annotation data may be myostatic annotation data or denture edge line annotation data, and the denture reference line recognition result output by the AI ​​model may be myostatic recognition result or denture edge line recognition result.

[0048] Sub-step S123, determining the first denture reference line according to the denture reference line recognition result.

[0049] When the first denture reference line is the first myostatic line and the denture reference line recognition result is a myostatic recognition result, if the myostatic recognition result is an identified myostatic line, the identified myostatic line can be directly used as the first myostatic line; if the myostatic recognition result is an identified myostatic area, the boundary line of the identified myostatic area can be directly determined as the first myostatic line.

[0050] When the first denture reference line is the first denture edge line, and the denture reference line recognition result is the myostatic recognition result, if the myostatic recognition result is the recognized myostatic line, the recognized myostatic line can be used as the first myostatic line, and the first denture edge line can be determined based on the first myostatic line; if the myostatic recognition result is the recognized myostatic area, the boundary line of the recognized myostatic area can be first determined as the first myostatic line, and then the first denture edge line can be determined based on the first myostatic line. Optionally, the step of determining the first denture edge line based on the first myostatic line can include: determining the equidistant line of about 2mm-5mm from the boundary of the first myostatic line toward the gum direction as the first denture edge line.

[0051] When the first denture reference line is the first denture edge line and the denture reference line recognition result is the denture edge line recognition result, the denture edge line in the denture edge line recognition result can be directly determined as the first denture edge line.

[0052] This will give you the first denture reference line.

[0053] In order to improve the accuracy of mapping, the step S130 of mapping the first denture reference line onto the second oral data to obtain the target denture edge line according to the registration information between the first oral data and the second oral data may include the following sub-steps: Sub-step S131, using a preset rigid body region, registering the first oral cavity data and the second oral cavity data to obtain registration information; wherein the rigid body region includes the alveolar ridge and / or the upper jaw.

[0054] The specific registration process can refer to the relevant existing technology and will not be described in detail here.

[0055] Sub-step S132, mapping the first denture reference line onto the second oral data according to the registration information to obtain the second denture reference line.

[0056] Sampling points can be obtained from the first denture reference line, and adjacent points (i.e., matching points) can be found on the second oral data based on the registration information, and the adjacent points can be fitted into a three-dimensional curve, which is the second denture reference line. Based on this, sub-step S132 can be implemented through the following process: obtaining multiple sampling points from the first denture reference line; determining matching points of each sampling point on the second oral data based on the registration information; performing three-dimensional curve fitting on each matching point to obtain the second denture reference line.

[0057] When the first denture reference line is the first myostatic line, the second denture reference line is the second myostatic line; when the first denture reference line is the first denture edge line, the second denture reference line is the second denture edge line.

[0058] Sub-step S133, determining the target denture edge line according to the second denture reference line.

[0059] In order to improve data quality and enhance data processing precision and accuracy, the second denture reference line can be smoothed to obtain a smoothed denture reference line; based on the smoothed denture reference line, the target denture edge line is determined.

[0060] Optionally, the above-mentioned smoothing of the second denture reference line to obtain the smoothed denture reference line may include: smoothing the second denture reference line using a target smoothing algorithm to obtain the smoothed denture reference line; wherein the target smoothing algorithm is determined based on preset denture morphological features and processing requirements, and the target smoothing algorithm includes one or more of a Laplace smoothing algorithm, a least squares smoothing algorithm, and a curvature-based feature smoothing algorithm.

[0061] Optionally, determining the target denture edge line based on the denture reference line after smoothing may include: when the second denture reference line is the second myostatic line, the target denture edge line may be determined based on the second myostatic line after smoothing; when the second denture reference line is the second denture edge line, the second denture edge line after smoothing may be directly determined as the target denture edge line. Optionally, the step of determining the target denture edge line based on the second myostatic line after smoothing may include: determining the equidistant line of about 2mm-5mm from the boundary of the second myostatic line after smoothing toward the gum direction as the target denture edge line.

[0062] In some possible embodiments, the above AI model can be trained through the following two steps: A1. Obtain multiple sample oral data with denture reference line annotation data.

[0063] The scanned three-dimensional plaster model and other initial oral data can be manually marked with myostatic lines / myostatic areas / denture edge lines, and then preprocessed such as mesh optimization to reduce the amount of calculation while retaining the geometric features. The preprocessed data is then sent to the constructed three-dimensional deep learning model for training.

[0064] In one possible implementation, multiple sample oral data can be obtained in the following manner: acquiring multiple initial oral data with denture reference line annotation data; wherein the initial oral data is used to represent the oral cavity of the sample object in a first pressure state; preprocessing the multiple initial oral data to obtain multiple sample oral data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, multi-scale data enhancement, sparse coding representation and surface feature fusion.

[0065] The following is a detailed introduction to the above preprocessing method: (a) Data denoising: The buccal and lingual data can be denoised and then smoothed to improve the robustness of the algorithm.

[0066] (b) Data alignment: By referring to oral features such as dental arch curves, the three-dimensional tooth grid is preliminarily aligned to reduce overfitting of training data and facilitate the improvement of the generalization ability of the network. The initial oral data such as the three-dimensional plaster model can be aligned with the preset reference model for oral features such as dental arch curves. In specific implementation, the three-dimensional plaster model can be projected at multiple angles to obtain multiple two-dimensional depth images, and the dental arch curve can be obtained by fitting the multiple two-dimensional depth images. The dental arch curve can be an isoline obtained by curve fitting the highest point of the tooth based on the depth data in the multiple two-dimensional depth images, which is in the shape of a curved crescent.

[0067] (c) Mesh simplification: The number of mesh vertices in flat areas is reduced through optimization algorithms while preserving their geometric features as much as possible to improve processing efficiency and reduce computational complexity. Flat areas refer to areas in the image with relatively consistent visual characteristics, that is, areas with small changes in curvature. These areas may lack obvious changes in color, texture, or shape and appear relatively uniform. Flat areas do not contain much visual detail or information and are relatively less useful for image analysis and understanding, so the number of mesh vertices can be appropriately reduced.

[0068] (d) Laplace feature calculation (obtaining data with Laplace features): The Laplace operator of the grid is used to capture and extract the characteristics of the grid in different domains, providing key information for further analysis. The different domains refer to high frequencies (better grid details) and low frequencies (better overall shape) in the frequency domain.

[0069] (e) Multi-scale data enhancement: Data is enhanced at different scales (i.e., data augmentation is performed by scaling the size) to increase the number of samples, which helps the model learn details at different levels and enhances its generalization ability.

[0070] (f) Sparse coding representation: Sparse coding techniques are used to represent the grid to eliminate redundant features, highlight important information, and potentially improve the interpretability of the model.

[0071] (g) Surface feature fusion: Calculate and fuse geometric features such as the surface curvature and normal vector of the mesh to enrich the model’s input information and enhance its understanding of complex structures. Through surface feature fusion, data with geometric features such as surface curvature and normal vector can be obtained, which facilitates the 3D deep learning network to better extract features.

[0072] Optionally, preprocessing can be performed in the order of (a) (b) (c) (d) (g), which can ensure high processing efficiency and good training effect. Preprocessing can also be performed in the order of (a) to (g), which can improve the accuracy of the trained recognition model and enhance the robustness and generalization ability of the recognition model. It should be noted that the specific preprocessing method can be selected according to actual needs, and this embodiment does not limit this.

[0073] A2. Train the three-dimensional deep learning model according to the oral data of each sample to obtain an AI model; wherein the three-dimensional deep learning model includes any one of the following: PointNet model, graph convolutional neural network model, MeshCNN model, multi-view model that parses three-dimensional information from multiple perspectives, convolutional neural network model based on graph theory, and DiffusionNet model.

[0074] The above-mentioned 3D deep learning model can be a deep learning model that specifically processes 3D mesh data, and can include any of the following: PointNet model that directly operates point cloud data, Graph CNNs model (i.e., graph convolutional neural network model) and MeshCNN model that use graph convolutional networks to process mesh topology, multi-view model that parses 3D information from multiple perspectives, Spectral CNN model based on spectral theory (i.e., convolutional neural network model based on graph spectrum theory), and DiffusionNet model designed for geometric data, etc. The 3D deep learning model can be comprehensively considered and decided based on the needs of actual application scenarios, such as computing resource consumption, real-time requirements, and other factors.

[0075] For both complete dentures and removable partial dentures, multiple sample oral data corresponding to complete dentures and multiple sample oral data corresponding to removable partial dentures can be obtained respectively, and these two types of sample oral data can be used to jointly train an AI model, which can be applicable to the recognition of myostatic lines / myostatic areas / denture edge lines in both complete dentures and removable partial dentures; multiple sample oral data corresponding to complete dentures can also be used alone to train a first AI model corresponding to complete dentures, which can be applicable to the recognition of myostatic lines / myostatic areas / denture edge lines in the case of complete dentures, and multiple sample oral data corresponding to removable partial dentures can be used alone to train a second AI model corresponding to removable partial dentures, which can be applicable to the recognition of myostatic lines / myostatic areas / denture edge lines in the case of removable partial dentures.

[0076] For ease of understanding, the following takes the first oral data as the target three-dimensional plaster model and the second oral data as the oral digital impression as an example to introduce in detail the method for determining the denture edge line when the first denture reference line is the first myostatic line and when the first denture reference line is the first denture edge line.

[0077] 1. The reference line of the first denture is the first muscle static line The above-mentioned method for determining the denture edge line may include the following steps: Step a1, obtaining a target three-dimensional plaster model and a digital oral impression corresponding to the target object.

[0078] Step a2, preprocessing the target three-dimensional plaster model to obtain a preprocessed target three-dimensional plaster model. The preprocessing may include one or more of the following: data denoising, data alignment, mesh simplification, Laplace feature calculation, sparse coding representation, and surface feature fusion.

[0079] Step a3, input the preprocessed target three-dimensional plaster model into the first recognition model to obtain the myostatic recognition result output by the first recognition model; wherein the first recognition model is obtained by training a three-dimensional deep learning model based on multiple first sample three-dimensional plaster models with myostatic annotation data, and the myostatic annotation data includes myostatic line data or myostatic area data.

[0080] The above-mentioned myostatic force recognition result corresponds to the labeled data used in the training of the first recognition model, and can be an identified myostatic force line or an identified myostatic force area.

[0081] Step a4, determining the first static muscle line according to the static muscle force recognition result.

[0082] In a possible implementation, when the static muscle force recognition result is an identified static muscle line, the identified static muscle line can be directly used as the first static muscle line; when the static muscle force recognition result is an identified static muscle area, the boundary line of the identified static muscle area (i.e., the dividing line between the static muscle area and the dynamic muscle area) can be directly determined as the first static muscle line.

[0083] In another possible implementation, the static force recognition result can be adjusted in an interactive manner: the static force recognition result is displayed on an interactive interface to facilitate the user to view and adjust the static force recognition result; in response to the user's adjustment operation on the static force recognition result, the adjusted static force recognition result is determined, and the first static force line is determined according to the adjusted static force recognition result. When the static force recognition result is an identified static force line, the adjusted static force recognition result is an adjusted static force line, and the adjusted static force line can be determined as the first static force line; when the static force recognition result is an identified static force area, the adjusted static force recognition result is an adjusted static force area, and the boundary line of the adjusted static force area can be determined as the first static force line.

[0084] Step a5, using a preset rigid area, registering the oral digital impression and the target three-dimensional plaster model to obtain registration information; wherein the rigid area includes the alveolar ridge and / or the upper jaw.

[0085] Step a6: Map the first myostatic line onto the oral digital impression according to the registration information to obtain the second myostatic line.

[0086] See also Figure 5 A schematic diagram of a myostatic line mapping is shown. After obtaining the oral digital impression and the first myostatic line on the target three-dimensional plaster model, the oral digital impression and the target three-dimensional plaster model can be aligned according to the alveolar ridge area, and the myostatic line (i.e., the first myostatic line) can be mapped to the digital impression (i.e., the oral digital impression).

[0087] Sampling points can be obtained from the first myostatic line, and adjacent points (i.e., matching points) can be found on the oral digital impression based on the registration information, and the adjacent points can be fitted into a three-dimensional curve, which is the second myostatic line. Based on this, sub-step S132 can be implemented through the following process: obtaining multiple sampling points from the first myostatic line; determining matching points of each sampling point on the oral digital impression based on the registration information; performing three-dimensional curve fitting on each matching point to obtain the second myostatic line.

[0088] In this way, the myostatic lines identified on the three-dimensional plaster model corresponding to the traditional impression scheme are mapped to the oral digital impression, ensuring a high-precision digital model while obtaining accurate myostatic boundary lines.

[0089] Step a7, determining the target denture edge line according to the second myostatic line.

[0090] In order to improve data quality and enhance data processing precision and accuracy, the second muscle static line can be smoothed to obtain the smoothed second muscle static line; based on the smoothed second muscle static line, the target denture edge line is determined.

[0091] Optionally, the above-mentioned smoothing of the second muscle static line to obtain the second muscle static line after smoothing may include: using a target smoothing algorithm to smooth the second muscle static line to obtain the second muscle static line after smoothing; wherein the target smoothing algorithm is determined based on the preset morphological characteristics and processing requirements of the denture, and the target smoothing algorithm includes one or more of a Laplace smoothing algorithm, a least squares smoothing algorithm and a curvature-based feature smoothing algorithm.

[0092] In one possible implementation, the target denture edge line can be determined directly based on the second muscle static line after smoothing. For example, the boundary of the second muscle static line after smoothing is toward an equidistant line of about 2mm-5mm in the gingival direction as the target denture edge line. In another possible implementation, the second muscle static line after smoothing can be adjusted in an interactive manner: the second muscle static line after smoothing is displayed on the interactive interface to facilitate the user to view and adjust the second muscle static line after smoothing; in response to the user's adjustment operation on the second muscle static line after smoothing, the adjusted target muscle static line is determined, and the target denture edge line is determined according to the adjusted target muscle static line. For example, the boundary of the second muscle static line after smoothing is toward an equidistant line of about 2mm-5mm in the gingival direction as the target denture edge line.

[0093] 2. The first denture reference line is the first denture edge line The above-mentioned method for determining the denture edge line may include the following steps: Step b1, obtaining a target three-dimensional plaster model and a digital oral impression corresponding to the target object.

[0094] Step b2, preprocessing the target three-dimensional plaster model to obtain a preprocessed target three-dimensional plaster model. The preprocessing may include one or more of the following: data denoising, data alignment, mesh simplification, Laplace feature calculation, sparse coding representation, and surface feature fusion.

[0095] Step b3, input the preprocessed target three-dimensional plaster model into the first recognition model, obtain the myostatic recognition result output by the first recognition model, and determine the first denture edge line according to the first myostatic line corresponding to the myostatic recognition result; or, input the preprocessed target three-dimensional plaster model into the second recognition model, obtain the denture edge line recognition result output by the second recognition model, and determine the denture edge line in the denture edge line recognition result as the first denture edge line.

[0096] Among them, the first recognition model is obtained by training a three-dimensional deep learning model based on multiple first sample three-dimensional plaster models with myostatic annotation data, and the myostatic annotation data includes myostatic line data or myostatic area data; the second recognition model is obtained by training a three-dimensional deep learning model based on multiple second sample three-dimensional plaster models with denture edge line annotation data.

[0097] Step b4, using a preset rigid area, registering the oral digital impression and the target three-dimensional plaster model to obtain registration information; wherein the rigid area includes the alveolar ridge and / or the upper jaw.

[0098] Step b5, mapping the first denture edge line onto the oral digital impression according to the registration information to obtain the second denture edge line.

[0099] Step b6, determining the target denture edge line based on the second denture edge line.

[0100] The acquisition process and mapping method of the above-mentioned registration information can refer to the corresponding content of the above-mentioned embodiment, which will not be repeated here. The step of determining the target denture edge line according to the second denture edge line may include: smoothing the second denture edge line to obtain the second denture edge line after smoothing; determining the target denture edge line according to the second denture edge line after smoothing.

[0101] In a possible implementation, the second denture edge line after smoothing can be obtained in the following manner: the second denture edge line is smoothed by using a target smoothing algorithm to obtain the second denture edge line after smoothing; wherein the target smoothing algorithm is determined based on the preset morphological characteristics and processing requirements of the denture, and the target smoothing algorithm includes one or more of the Laplace smoothing algorithm, the least squares smoothing algorithm, and the curvature-based feature smoothing algorithm.

[0102] In one possible implementation, the second denture edge line after the smoothing process can be directly determined as the target denture edge line; the target denture edge line can also be determined in the following manner: displaying the second denture edge line after the smoothing process on the interactive interface; in response to the user's adjustment operation on the second denture edge line after the smoothing process, determining the adjusted denture edge line, and determining the adjusted denture edge line as the target denture edge line.

[0103] In this embodiment, based on the traditional impression method, three-dimensional deep learning technology is used to accurately mark / identify the myostatic line / denture edge line, ensuring better robustness for different forms of jaws; the plaster model obtained by the traditional impression scheme is combined with the oral digital impression to ensure a high-precision digital model while obtaining accurate myostatic boundary lines, and then obtaining accurate denture edge lines.

[0104] In an optional embodiment, after the step of obtaining the first oral data and the second oral data corresponding to the target object, the method further includes: performing data integrity judgment on the first oral data and the second oral data to obtain a degree of integrity; and determining the accuracy of the target denture edge line based on the degree of integrity.

[0105] Optionally, the above-mentioned degree of integrity can be determined in the following manner: determine whether the first oral data and / or the second oral data contain a preset key area to obtain a inclusion judgment result; wherein the key area corresponding to the first oral data includes a first designated area, and the key area corresponding to the second oral data includes a second designated area; determine the degree of integrity based on the inclusion judgment result.

[0106] In a possible implementation, the first designated area and the second designated area may both include at least two of the lip frenulum, the cheek frenulum, the maxillary tuberosity, the pterygomaxillary notch, and the area 2 mm behind the maxillary fossa.

[0107] In another possible implementation, considering that the oral digital impression needs to confirm that the key areas such as the lip frenulum, cheek frenulum, vestibular mucosa fold, zygomatic lower edge line and maxillary tuberosity buccal side are completely scanned, so as to ensure that the scan data already contains the complete myostatic line; the target three-dimensional plaster model needs to completely cover the alveolar ridge, jaw structure, palate, labial, buccal and lingual mucosa, frenulum and salivary gland opening and other key areas, to ensure accurate occlusal relationship and clear edges, so as to reflect the detailed structure of the target oral cavity and ensure accurate extraction of myostatic line or denture edge line. Based on this, the first designated area can include at least two of the alveolar ridge, jaw structure, palate, labial, buccal and lingual mucosa, frenulum and salivary gland opening; the second designated area can include at least two of the lip frenulum, cheek frenulum, vestibular mucosa fold, zygomatic lower edge line and maxillary tuberosity buccal side.

[0108] The judgment of whether a preset key area is included can be realized through AI, and the inclusion judgment result can include the inclusion of different key areas, such as whether the oral digital impression includes other key areas except the lip frenulum; based on the inclusion judgment result, the inclusion percentage corresponding to the first oral data and / or the inclusion percentage corresponding to the second oral data can be determined, wherein the inclusion percentage corresponding to the first oral data and the inclusion percentage corresponding to the second oral data are calculated separately; based on the inclusion percentage corresponding to the first oral data and / or the inclusion percentage corresponding to the second oral data, the degree of completeness is determined.

[0109] When determining the inclusion percentage of key areas, the proportion of each key area can be preset, and the inclusion percentage can be determined based on the proportion and inclusion of each key area. The proportions of different key areas can be the same or different; when the proportions of different key areas are the same, the quotient of the number of key areas included divided by the total number of key areas can be used as the inclusion percentage; when the proportions of different key areas are different, the proportions of the included key areas can be summed to obtain the inclusion percentage. If there is only the inclusion percentage corresponding to the first oral data or the inclusion percentage corresponding to the second oral data, the inclusion percentage corresponding to the first oral data or the inclusion percentage corresponding to the second oral data is directly used as the degree of completeness; if there is an inclusion percentage corresponding to the first oral data and an inclusion percentage corresponding to the second oral data, the two (i.e., two inclusion percentages) can be weighted and summed to obtain the degree of completeness, where the weights of the two can be set according to actual needs and are not limited here.

[0110] It should be noted that the above-mentioned key areas can be selected according to actual needs, and this embodiment does not limit this.

[0111] Optionally, when determining the accuracy of the target denture edge line according to the degree of completeness, the accuracy of the target denture edge line can be obtained by searching for the corresponding degree of completeness in the preset correspondence between the degree of completeness and the degree of accuracy. The accuracy can be divided into multiple levels according to actual needs, such as two levels of accurate and inaccurate, and three levels of high, medium and low, etc. The accuracy can also be expressed in percentage, for example, the accuracy of the target denture edge line is 80%.

[0112] Further optionally, after performing data integrity judgment on the first oral data and the second oral data and obtaining the degree of integrity, the method further includes: judging whether the degree of integrity is less than a preset integrity threshold; if it is less than the integrity threshold, issuing a data incompleteness reminder.

[0113] The integrity threshold can be set according to actual needs and is not limited here. When reminding of incomplete data, only the incomplete data can be given, for example, the first oral data is incomplete; it can also be further given which part of the data is incomplete, for example, the buccal side of the maxillary tuberosity of the second oral data is incomplete. This makes it convenient for users to understand the incomplete data in a timely manner so that they can replace the data in a timely manner.

[0114] In an optional embodiment, the target denture edge line will be displayed on the interactive interface, and the user can adjust the target denture edge line on the interactive interface. During the adjustment process, since the user adjusts the movement of the target denture edge line on the two-dimensional display interface (i.e., the interactive interface), if the target denture edge line is adjusted directly according to the user's two-dimensional instructions, it is easy for the target denture edge line to move too much, deviate from the muscle static area or be too close to the muscle dynamic area. The total denture (full denture) or removable partial denture designed according to the target denture edge line that is deviated from the muscle static area or too close to the muscle dynamic area is easy to shift during the patient's wearing process, or it will collide during physiological movements such as chewing, causing the patient to be uncomfortable. At the same time, since the patient's oral environment is not a regular cylinder, if the target denture edge line is simply translated or enlarged / reduced directly according to the user's two-dimensional instructions, the target denture edge line will be separated from the model surface (i.e., the surface of the second oral data), causing the target denture edge line to be inconsistent with the patient's actual oral condition, and subsequent denture design operations cannot be performed.

[0115] To solve the above problem, after the step of mapping the first denture reference line to the second oral data according to the registration information between the first oral data and the second oral data to obtain the target denture edge line, the method further includes: when a first instruction is monitored, determining a first pose change of the target denture edge line relative to the initial position, wherein the first instruction is an instruction for the user to adjust the pose of the target denture edge line in the second oral data; based on the first pose change and preset reference conditions, determining a first adjusted pose of the target denture edge line, and updating and displaying the pose of the target denture edge line based on the first adjusted pose of the target denture edge line, wherein the preset reference conditions include the target denture edge line fitting with the outer surface of the second oral data, the target denture edge line symbol One or more of the following: simulating physiological movement or the target denture edge line being located within the denture coverage area; wherein, when it is detected that the first posture change makes the target denture edge line not meet the preset reference condition, the posture of the target denture edge line is adaptively updated according to the preset reference condition, or, the posture of the target denture edge line is not updated and the result of user reconfirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user reconfirmation; and / or when it is detected that the first adjusted posture makes the target denture edge line not meet the preset reference condition, the posture of the target denture edge line is adaptively updated according to the preset reference condition, or, the posture of the target denture edge line is not updated and the result of user reconfirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user reconfirmation.

[0116] It can be understood that not updating the pose of the target denture edge line means not moving or offsetting the target denture edge line, and not updating the pose of the target denture edge line after the move or offset; updating the pose of the target denture edge line can be understood as continuing to move or offset the target denture edge line, and updating the pose of the target denture edge line after the move or offset. Adjusting the pose of the target denture edge line according to the result of the user's reconfirmation can be based on the user's choice not to update the pose of the target denture edge line or to update the pose of the target denture edge line. When the pose of the target denture edge line is not updated, a prompt box or other prompt information can be issued to give the user certain feedback.

[0117] Therefore, in this embodiment, the first adjustment posture can be recalculated based on the first posture change of the target denture edge line relative to its initial posture and / or the offset angle of the target denture edge line relative to its initial posture included in the user's first instruction, combined with the outer surface of the second oral data, and the preset reference conditions (medical reference requirements) such as the determined denture coverage area or simulated physiological movement, and some adjustment instructions in the first posture change are ignored and some adjustment instructions in the first posture change are executed, that is, the instructions in the first posture change that do not meet the medical reference requirements are ignored, and the instructions in the first posture change that meet the medical reference requirements are executed, so that the target denture edge line posture (first adjustment posture) on the two-dimensional display interface always meets the preset reference conditions, that is, meets the basic medical requirements, such as the target denture edge line should be within the denture coverage area, or conform to simulated physiological movement, that is, during chewing or physiological movements such as mandibular movement, the target denture edge line will not move or the movement amplitude of the target denture edge line is within a preset range, which can be the range corresponding to the denture coverage area.

[0118] Among them, the first pose change can be decomposed according to the preset reference conditions, and some or all instructions in the first pose change that meet the preset reference conditions are determined as the first adjustment pose of the target denture edge line, and some or all instructions in the first pose change that do not meet the preset reference conditions are ignored; wherein the preset reference conditions may include: the target denture edge line should fit the outer surface of the second oral data, the target denture edge line should be within the denture coverage area, or the target denture edge line should conform to one or more of the simulated physiological movements.

[0119] For example, the first instruction includes the X distance of the target denture edge line moving horizontally or obliquely by the user clicking the mouse. If the X distance makes the target denture edge line not fit the outer surface of the second oral data, not within the denture coverage area, or not conforming to the simulated physiological movement, then the instruction of the X distance of the target denture edge line moving horizontally or obliquely will be ignored and not executed. This indicates that the first pose change does not meet the preset reference conditions. At this time, there are two operation modes: the first one is that the computer automatically enlarges or reduces the first pose change according to the preset ratio so that it conforms to the simulated physiological movement or is within the denture coverage area, and determines it as the first adjustment pose, and displays the updated target denture edge line; the second one is that the adjusted target denture edge line is not displayed first, and a prompt box is issued to prompt the user that there may be problems with such adjustment, and the user is asked to confirm again. According to the result of the user's reconfirmation, the first adjustment pose is determined; if the user indicates "confirmation", this adjustment method is used as the first adjustment pose, and the pose of the target denture edge line is updated according to the user's instructions; if the user indicates "abandon modification", the first adjustment pose is 0, and the pose of the target denture edge line is not updated according to the user's instructions.

[0120] For example, the first instruction includes the X distance that the user moves diagonally by clicking the target denture edge line with the mouse, and decomposing the X distance into vertical (such as Figure 6 The Z direction in the vertical direction refers to the up and down movement along the gingival surface) and the horizontal direction (such as Figure 6 In the X direction, the horizontal movement refers to the forward and backward movement along the normal of the outer surface of the second oral data) moving C distance, where the horizontal movement C distance will be ignored and not executed, and the vertical movement B distance will be executed; this is equivalent to projecting the X distance of the oblique movement onto the vertical plane (such as Figure 6 The first adjustment posture of the target denture edge line is that the target denture edge line moves up and down along the gingival surface by a distance B, and the latest target denture edge line is automatically calculated based on the target denture edge line after the vertical movement of the distance B and the outer surface of the second oral data, so that the target denture edge line fits the outer surface of the model.

[0121] For example, after determining the first adjustment posture, it is possible to determine again whether the target denture edge line conforms to the simulated physiological movement or is located in the denture coverage area. If so, it means that the first adjustment posture conforms to the medical reference conditions, and the target denture edge line can be adjusted according to the first adjustment posture and displayed in the interactive interface after the adjustment; if not, it means that the first adjustment posture does not conform to the medical reference conditions. At this time, there are two operation methods: first, the computer automatically enlarges or reduces the first adjustment posture according to a preset ratio to make it conform to the simulated physiological movement or be located in the denture coverage area; second, the adjusted target denture edge line is not displayed first, and a prompt box is issued to prompt the user that such adjustment may have problems, and the user is asked to confirm again "whether to use this adjustment method to modify"; if the user indicates "confirm", this adjustment method is used to modify, and the posture of the target denture edge line is updated according to the user's instructions; if the user indicates "abandon modification", the posture of the target denture edge line is not updated according to the user's instructions.

[0122] In one embodiment, based on the first instruction, the target denture edge line can be controlled to rotate to the position indicated by the first instruction based on the current coordinate point of the mouse. At this time, the interactive interface shows that the target denture edge line falls at the position where the mouse is clicked. At the same time, the angle or distance of the target denture edge line rotating around the target denture edge line is the projection of the mouse movement distance onto the vertical plane (such as Figure 6 The XZ plane in the image is determined by the

[0123] It should be noted that the myostatic area can be divided into an upper and lower part by the target denture edge line, and the denture coverage area can refer to the entire myostatic area, or directly refer to the upper part of the myostatic area divided by the target denture edge line, the upper part includes rigid areas such as gums or teeth, and when wearing a total denture (full denture) or a removable partial denture, the upper part is just covered by the denture. Simulated physiological movement can refer to the simulated physiological movement automatically generated by inputting the second oral data into the motion simulation deep learning model, and the simulated physiological movement can be selectively displayed on the interactive interface according to the user's instructions. When the user chooses to display the simulated physiological movement on the interactive interface, the second oral data on the interactive interface will change the simulated physiological movement, and the user can observe with the naked eye whether the target denture edge line is appropriate, and can move up and down or adjust the target denture edge line. When the user chooses not to display the simulated physiological movement on the interactive interface, the interactive interface can directly output the result of the computer judging whether the target denture edge line conforms to the simulated physiological movement.

[0124] The simulated physiological motion automatically generated by the motion simulation deep learning model can be generated based on the average motion of multiple patient training samples, or based on the mandibular motion trajectory and chewing motion trajectory obtained by CBCT (Cone Beam Computed Tomography) equipment, facial scanner, extraoral scanner or intraoral scanner.

[0125] In an optional embodiment, a three-dimensional denture model can be obtained based on the target denture edge line, and the three-dimensional denture model can be quickly printed to obtain a physical denture for user trial wearing. Based on the user trial wearing results, the target denture edge line or the three-dimensional denture model can be further adjusted to obtain a denture that better meets the user's needs.

[0126] Based on this, after the step of mapping the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data to obtain the target denture edge line, the method further includes: Generate a three-dimensional denture model according to the target denture edge line and the second oral cavity data, and send the three-dimensional denture model to a 3D printing device for printing the three-dimensional denture model; When a user request for modifying the three-dimensional denture model is received, a revised target denture margin line and / or a revised three-dimensional denture model is output according to the user request, the target denture margin line and the second oral data.

[0127] Corresponding to the above-mentioned method for determining the edge line of a denture, an embodiment of the present invention further provides a device for determining the edge line of a denture, see Figure 7 A schematic structural diagram of a denture edge line determination device is shown, the device comprising: The data acquisition module 701 is used to acquire first oral cavity data and second oral cavity data corresponding to the target object; wherein the first oral cavity data is used to represent the oral cavity of the target object in a first pressure state, and the second oral cavity data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; The denture reference line recognition module 702 is used to perform denture reference line recognition on the first oral data according to the trained AI model to obtain a first denture reference line; The denture edge line mapping module 703 is used to map the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data to obtain the target denture edge line.

[0128] The denture edge line determination device provided by the embodiment of the present invention can obtain the first oral data and the second oral data corresponding to the target object when determining the denture edge line; wherein the first oral data is used to represent the oral cavity of the target object in a first pressure state, and the second oral data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; the denture reference line is identified on the first oral data according to the trained AI model to obtain the first denture reference line; according to the registration information between the first oral data and the second oral data, the first denture reference line is mapped to the second oral data to obtain the target denture edge line. Since the pressure at the time of impression corresponding to the first oral data is greater than the pressure at the time of impression corresponding to the second oral data, the accuracy of the second oral data is higher than that of the first oral data, and the first denture reference line corresponding to the first oral data is not easily affected by the diversity of the tooth and jaw morphology and the quality of the mesh surface. The first denture reference line corresponding to the first oral data can be accurately identified through the artificial intelligence learning method, and then the first denture reference line is mapped to the second oral data with higher accuracy, which can facilitate the subsequent data storage and design. In this way, the first oral data under different pressure states are combined with the second oral data to achieve the effect of edge trimming, automatically and accurately identify the denture edge line of the second oral data, and facilitate subsequent digital storage and design.

[0129] Furthermore, the above-mentioned denture reference line recognition module 702 is specifically used for: preprocessing the first oral data to obtain preprocessed first oral data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, sparse coding representation and surface feature fusion; inputting the preprocessed first oral data into the AI ​​model to obtain the denture reference line recognition result output by the AI ​​model; wherein the AI ​​model is obtained by training a three-dimensional deep learning model based on multiple sample oral data with denture reference line annotation data, and the sample oral data is used to represent the oral cavity of the sample object in a first pressure state; according to the denture reference line recognition result, determining the first denture reference line.

[0130] Further, the first denture reference line includes a first myostatic line, the denture reference line annotation data includes myostatic annotation data, and the myostatic annotation data includes myostatic line data or myostatic area data; or, The first denture reference line includes a first denture edge line, and the denture reference line annotation data includes myostatic annotation data or denture edge line annotation data.

[0131] Furthermore, the above-mentioned denture edge line mapping module 703 is specifically used for: utilizing a preset rigid body region to align the first oral data and the second oral data to obtain alignment information; wherein the rigid body region includes the alveolar ridge and / or the upper jaw; according to the alignment information, mapping the first denture reference line onto the second oral data to obtain the second denture reference line; and determining the target denture edge line according to the second denture reference line.

[0132] Furthermore, the denture edge line mapping module 703 is also used to: obtain multiple sampling points from the first denture reference line; determine the matching points of each sampling point on the second oral data according to the registration information; perform three-dimensional curve fitting on each matching point to obtain the second denture reference line.

[0133] Furthermore, the denture edge line mapping module 703 is also used to: perform smoothing on the second denture reference line to obtain a smoothed denture reference line; and determine a target denture edge line based on the smoothed denture reference line.

[0134] Furthermore, the above-mentioned device also includes a training module, which is used to: obtain multiple sample oral data with denture reference line annotation data; train the three-dimensional deep learning model according to each sample oral data to obtain an AI model; wherein the three-dimensional deep learning model includes any one of the following: PointNet model, graph convolutional neural network model, MeshCNN model, multi-view model that parses three-dimensional information from multiple perspectives, convolutional neural network model based on graph theory and diffusion network DiffusionNet model.

[0135] Furthermore, the above-mentioned training module is specifically used to: obtain multiple initial oral data with denture reference line annotation data; wherein the initial oral data is used to represent the oral cavity of the sample object in a first pressure state; preprocess the multiple initial oral data to obtain multiple sample oral data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, multi-scale data enhancement, sparse coding representation and surface feature fusion.

[0136] Furthermore, the first oral cavity data is obtained by three-dimensionally scanning a plaster model obtained by impression of the oral cavity of the target object under the first pressure, or by three-dimensionally scanning the historical denture of the target object, or by scanning the oral cavity of the target object under the state of blowing pressure; The second oral cavity data is obtained by directly scanning the oral cavity of the target object without physical pressure or under standard atmospheric pressure.

[0137] Furthermore, the above device also includes: A first judgment module is used to perform data integrity judgment on the first oral data and the second oral data to obtain a degree of integrity; The accuracy determination module is used to determine the accuracy of the target denture edge line according to the degree of completeness.

[0138] Furthermore, the above-mentioned first judgment module is specifically used to: judge whether the first oral data and / or the second oral data contain a preset key area, and obtain a inclusion judgment result; wherein the key area corresponding to the first oral data includes a first specified area, and the key area corresponding to the second oral data includes a second specified area; determine the degree of integrity based on the inclusion judgment result.

[0139] Furthermore, the above device also includes: A second judgment module is used to judge whether the integrity level is less than a preset integrity threshold; The reminder module is used to remind you if the data is incomplete if it is less than the integrity threshold.

[0140] Furthermore, the above-mentioned device also includes a user adjustment module, which is used to: when a first instruction is monitored, determine a first posture change of the target denture edge line relative to the initial position, wherein the first instruction is an instruction for the user to adjust the posture of the target denture edge line in the second oral data; based on the first posture change and a preset reference condition, determine a first adjusted posture of the target denture edge line, and based on the first adjusted posture of the target denture edge line, update and display the posture of the target denture edge line, the preset reference condition includes that the target denture edge line is in contact with the outer surface of the second oral data, the target denture edge line conforms to simulated physiological movement, or the target denture edge line is located within the denture coverage area. or more; wherein, when it is detected that the first posture change makes the target denture edge line not meet the preset reference condition, the posture of the target denture edge line is adaptively updated according to the preset reference condition, or, the posture of the target denture edge line is not updated and the result of user reconfirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user reconfirmation; and / or, when it is detected that the first adjusted posture makes the target denture edge line not meet the preset reference condition, the posture of the target denture edge line is adaptively updated according to the preset reference condition, or, the posture of the target denture edge line is not updated and the result of user reconfirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user reconfirmation.

[0141] Furthermore, the above-mentioned device also includes: a generation module, which is used to generate a three-dimensional denture model based on the target denture edge line and the second oral data, and send the three-dimensional denture model to a 3D printing device for printing the three-dimensional denture model; a modification module, which is used to output a revised target denture edge line and / or a revised three-dimensional denture model based on the user request, the target denture edge line and the second oral data when receiving a user request for modifying the three-dimensional denture model.

[0142] The implementation principle and technical effects of the device provided in this embodiment are the same as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0143] like Figure 8 As shown, a computer device 800 provided in an embodiment of the present invention includes: a processor 801, a memory 802 and a bus, the memory 802 stores machine-readable instructions executable by the processor 801, when the computer device 800 is running, the processor 801 and the memory 802 communicate through the bus, and the processor 801 executes the machine-readable instructions to perform the steps of the denture edge line determination method as described above.

[0144] Specifically, the above-mentioned memory 802 and processor 801 can be general-purpose memory and processor, which are not specifically limited here. When the processor 801 runs the computer program stored in the memory 802, the above-mentioned denture edge line determination method can be executed.

[0145] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the denture edge line determination method in the above method embodiment is executed. The computer-readable storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a RAM, a magnetic disk, or an optical disk, etc., which can store program codes.

[0146] The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can represent including any one or more elements selected from the set consisting of A, B, and C.

[0147] In all examples shown and described herein, any specific values ​​should be interpreted as merely exemplary and not as limiting, and thus other examples of the exemplary embodiments may have different values.

[0148] The flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and a part of a module, a program segment or a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0149] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or modules can be electrical, mechanical or other forms.

[0150] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0152] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining a denture edge line, characterized in that: include: Acquire first oral cavity data and second oral cavity data corresponding to the target object; wherein the first oral cavity data is used to represent the oral cavity of the target object in a first pressure state, and the second oral cavity data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; Perform denture reference line recognition on the first oral data according to the trained AI model to obtain a first denture reference line; According to the registration information between the first oral data and the second oral data, the first denture reference line is mapped onto the second oral data to obtain a target denture edge line.

2. The method according to claim 1, characterized in that The step of performing denture reference line recognition on the first oral data according to the trained AI model to obtain a first denture reference line includes: Preprocessing the first oral cavity data to obtain preprocessed first oral cavity data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, sparse coding representation, and surface feature fusion; Inputting the preprocessed first oral cavity data into the AI ​​model to obtain a denture reference line recognition result output by the AI ​​model; wherein the AI ​​model is obtained by training a three-dimensional deep learning model based on a plurality of sample oral cavity data with denture reference line annotation data, and the sample oral cavity data is used to represent the oral cavity of the sample object in a first pressure state; A first denture reference line is determined according to the denture reference line recognition result.

3. The method according to claim 2, characterized in that The first denture reference line includes a first myostatic line, the denture reference line annotation data includes myostatic annotation data, and the myostatic annotation data includes myostatic line data or myostatic area data; or, The first denture reference line includes a first denture edge line, and the denture reference line annotation data includes myostatic annotation data or denture edge line annotation data.

4. The method according to claim 1, characterized in that: The step of mapping the first denture reference line onto the second denture data according to the registration information between the first oral data and the second oral data to obtain a target denture edge line includes: Using a preset rigid region, registering the first oral cavity data and the second oral cavity data to obtain registration information; wherein the rigid region includes the alveolar ridge and / or the upper jaw; According to the registration information, mapping the first denture reference line to the second oral data to obtain a second denture reference line; The target denture edge line is determined according to the second denture reference line.

5. The method according to claim 4, characterized in that Mapping the first denture reference line onto the second oral data according to the registration information to obtain a second denture reference line includes: Acquire a plurality of sampling points from the first denture reference line; Determining matching points of each of the sampling points on the second oral data according to the registration information; Perform three-dimensional curve fitting on each of the matching points to obtain the second denture reference line.

6. The method according to claim 4, characterized in that Determining the target denture edge line according to the second denture reference line includes: Performing smoothing on the second denture reference line to obtain a smoothed denture reference line; The target denture edge line is determined according to the denture reference line after the smoothing process.

7. The method according to claim 1, characterized in that The method further comprises: Acquire multiple sample oral cavity data with denture reference line annotation data; The three-dimensional deep learning model is trained according to the oral data of each sample to obtain the AI ​​model; wherein the three-dimensional deep learning model includes any one of the following: PointNet model, graph convolutional neural network model, MeshCNN model, multi-view model that parses three-dimensional information from multiple perspectives, convolutional neural network model based on graph theory, and diffusion network DiffusionNet model.

8. The method according to claim 7, characterized in that The step of obtaining a plurality of sample oral cavity data with denture reference line annotation data includes: Acquire a plurality of initial oral cavity data with denture reference line annotation data; wherein the initial oral cavity data is used to represent the oral cavity of the sample object in a first pressure state; The multiple initial oral data are preprocessed to obtain the multiple sample oral data; wherein the preprocessing includes one or more of the following: data denoising, data alignment, grid simplification, Laplace feature calculation, multi-scale data enhancement, sparse coding representation and surface feature fusion.

9. The method according to claim 1, characterized in that: The first oral cavity data is obtained by three-dimensionally scanning a plaster model obtained by impression of the oral cavity of the target object under the first pressure, or by three-dimensionally scanning a historical denture of the target object, or by scanning the oral cavity of the target object under a blowing pressure state; The second oral cavity data is obtained by directly scanning the oral cavity of the target object without physical pressure or under standard atmospheric pressure.

10. The method according to claim 1, characterized in that After obtaining the first oral cavity data and the second oral cavity data corresponding to the target object, the method further includes: Performing data integrity judgment on the first oral data and the second oral data to obtain a degree of integrity; Based on the degree of completeness, the accuracy of the target denture edge line is determined.

11. The method according to claim 10, characterized in that The performing data integrity judgment on the first oral data and the second oral data to obtain the degree of integrity includes: Determine whether the first oral data and / or the second oral data include a preset key area, and obtain a determination result; wherein the key area corresponding to the first oral data includes a first designated area, and the key area corresponding to the second oral data includes a second designated area; Determining the degree of integrity according to the inclusion judgment result; After determining the data integrity of the first oral data and the second oral data to obtain the degree of integrity, the method further includes: Determining whether the integrity level is less than a preset integrity threshold; If it is less than the integrity threshold, a data incompleteness reminder is issued.

12. The method according to claim 1, characterized in that After mapping the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data to obtain the target denture edge line, the method further includes: When a first instruction is detected, determining a first pose change of the target denture edge line relative to the initial position, wherein the first instruction is an instruction for the user to adjust the pose of the target denture edge line in the second oral data; Based on the first posture change and preset reference conditions, determining a first adjusted posture of the target denture edge line, and updating and displaying the posture of the target denture edge line based on the first adjusted posture of the target denture edge line, wherein the preset reference conditions include one or more of the following: the target denture edge line fits the outer surface of the second oral data, the target denture edge line conforms to simulated physiological movement, or the target denture edge line is located in the denture coverage area; Wherein, when it is detected that the first posture change makes the target denture edge line not meet the preset reference condition, the posture of the target denture edge line is adaptively updated according to the preset reference condition, or, the posture of the target denture edge line is not updated and the result of user reconfirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user reconfirmation; and / or When it is detected that the first adjustment posture causes the target denture edge line to fail to meet the preset reference condition, the posture of the target denture edge line is adaptively updated according to the preset reference condition, or the posture of the target denture edge line is not updated and the result of user re-confirmation is obtained and the posture of the target denture edge line is adjusted according to the result of user re-confirmation.

13. The method according to any one of claims 1 to 12, characterized in that After mapping the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data to obtain the target denture edge line, the method further includes: Generate a three-dimensional denture model according to the target denture edge line and the second oral data, and send the three-dimensional denture model to a 3D printing device for printing the three-dimensional denture model; When a user request for modifying the three-dimensional denture model is received, a revised target denture margin line and / or a revised three-dimensional denture model is output according to the user request, the target denture margin line and the second oral data.

14. A denture edge line determination device, characterized in that: include: A data acquisition module, used to acquire first oral cavity data and second oral cavity data corresponding to a target object; wherein the first oral cavity data is used to represent the oral cavity of the target object in a first pressure state, and the second oral cavity data is used to represent the oral cavity of the target object in a second pressure state, and the pressure in the second pressure state is less than the pressure in the first pressure state; A denture reference line recognition module, used to perform denture reference line recognition on the first oral data according to the trained AI model to obtain a first denture reference line; The denture edge line mapping module is used to map the first denture reference line onto the second oral data according to the registration information between the first oral data and the second oral data to obtain a target denture edge line.

15. A computer device, characterized in that: It comprises a memory and a processor; the memory stores a computer program that can be run on the processor, and the processor implements the denture edge line determination method described in any one of claims 1 to 13 when executing the computer program.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the denture margin line determination method according to any one of claims 1 to 13 is executed.

Images