Method, device and equipment for generating abutment edge line and storage medium

Abstract

The invention discloses an abutment edge line generation method and device, equipment and a storage medium. The abutment edge line generation method comprises the steps that oral cavity scanning data are acquired, and the oral cavity scanning data comprise point cloud data of a plurality of nodes; performing classification processing on each node based on the point cloud data of each node to obtain classification information of each node, the node types including an abutment tooth type and a background type; an undirected graph corresponding to the oral cavity scanning data is constructed, the undirected graph comprises a plurality of nodes in the oral cavity scanning data and edges among the nodes, and weights corresponding to the edges among the nodes are determined based on point cloud data of the nodes and / or classification information of the nodes; and performing graph segmentation on the undirected graph based on the weights of the edges in the undirected graph to obtain an abutment sub-graph and a background sub-graph, determining abutment edge nodes based on the abutment sub-graph and the background sub-graph, and generating abutment edge lines based on the abutment edge nodes. And the accuracy of generating the abutment edge line is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tooth restoration, and in particular to a method, device, equipment and storage medium for generating an abutment edge line. Background Art

[0002] In the field of dental restoration, the study of abutment margins has important scientific significance and clinical value. The abutment margin is the boundary where the denture contacts the natural teeth. Its accuracy is directly related to the accuracy of the restoration, the comfort of the patient and the oral health. Inaccurate margins may lead to problems such as food impaction, gingival inflammation and restoration shedding, which in turn affect the patient's quality of life and oral function. Therefore, accurate detection and design of abutment margins are the key to improving the quality of restorations and extending their service life. In the traditional denture production process, the acquisition of margins is usually completed through physical impression taking and visual inspection. First, a physical impression of the teeth is directly obtained in the patient's mouth using impression materials, and then the impression is marginally shaped to ensure the precise fit of the denture. However, this process is not only time-consuming and labor-intensive, but also requires a high level of technical skills from the doctor, which limits the efficiency of restoration and consistency of quality.

[0003] With the development of computer-aided design and manufacturing technology, dental technicians have begun to use digital means to design and manufacture dentures. However, when processing three-dimensional models of teeth, existing systems often require users to manually draw edge lines, which is also time-consuming and requires high operator proficiency. Summary of the invention

[0004] The present invention provides a method, device, equipment and storage medium for generating an abutment edge line, so as to solve the problem of low accuracy in generating an abutment edge line.

[0005] According to one aspect of the present invention, a method for generating an abutment edge line is provided, the method comprising:

[0006] Acquire oral scan data, where the oral scan data includes point cloud data of multiple nodes;

[0007] Classify each node based on the point cloud data of each node to obtain classification information of each node, where the node type includes abutment type and background type;

[0008] Constructing an undirected graph corresponding to the oral scan data, wherein the undirected graph includes a plurality of nodes in the oral scan data and edges between the nodes, and weights corresponding to the edges between the nodes are determined based on point cloud data of the nodes and / or classification information of the nodes;

[0009] The undirected graph is segmented based on the weights of the edges in the undirected graph to obtain abutment subgraph and background subgraph, and abutment edge nodes are determined based on the abutment subgraph and background subgraph, and abutment edge lines are generated based on the abutment edge nodes.

[0010] According to another aspect of the present invention, there is provided a device for generating an abutment edge line, the device comprising:

[0011] An oral scan data acquisition module, used to acquire oral scan data, where the oral scan data includes point cloud data of multiple nodes;

[0012] A classification module is used to classify each node based on the point cloud data of each node to obtain classification information of each node, and the types of nodes include abutment types and background types;

[0013] An undirected graph construction module, used to construct an undirected graph corresponding to the oral scan data, wherein the undirected graph includes a plurality of nodes in the oral scan data and edges between the nodes, and weights corresponding to the edges between the nodes are determined based on point cloud data of the nodes and / or classification information of the nodes;

[0014] The abutment edge line generation module is used to segment the undirected graph based on the weights of the edges in the undirected graph to obtain an abutment subgraph and a background subgraph, determine the abutment edge nodes based on the abutment subgraph and the background subgraph, and generate the abutment edge line based on the abutment edge nodes.

[0015] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:

[0016] at least one processor; and

[0017] a memory communicatively connected to at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by at least one processor. The computer program is executed by at least one processor so that the at least one processor can execute the abutment edge line generating method according to any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for generating an abutment edge line according to any embodiment of the present invention when executed.

[0020] The technical solution of the embodiment of the present invention obtains oral scanning data including point cloud data of multiple nodes, and classifies each node to obtain classification information of each node, constructs an undirected graph corresponding to the oral scanning data, and segments the undirected graph based on the weights of the edges in the undirected graph to obtain abutment subgraph and background subgraph, determines abutment edge nodes based on the abutment subgraph and the background subgraph, and generates abutment edge lines based on the abutment edge nodes, thereby solving the problem of low accuracy in generating abutment edge lines and improving the accuracy in generating abutment edge lines.

[0021] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 is a flow chart of a method for generating an abutment edge line provided in Embodiment 1 of the present invention;

[0024] Figure 2 is a schematic diagram of the structure of a point conversion module provided by an embodiment of the present invention;

[0025] Figure 3 is a schematic diagram of the structure of a downward transition module provided in an embodiment of the present invention;

[0026] Figure 4 is a flow chart of another method for generating abutment edge lines provided in the second embodiment of the present invention;

[0027] Figure 5 is a flow chart of a method for training a point cloud recognition model provided in Embodiment 3 of the present invention;

[0028] Figure 6 It is a structural schematic diagram of a device for generating an abutment edge line provided by a fourth embodiment of the present invention;

[0029] Figure 7 It is a structural schematic diagram of an electronic device provided in Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not 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 should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] Embodiment 1

[0033] Figure 1 It is a flow chart of a method for generating abutment edge lines provided in the first embodiment of the present invention. This embodiment is applicable to the case of automatically generating abutment edge lines. The method can be executed by an abutment edge line generating device. The abutment edge line generating device can be implemented in the form of hardware and / or software. The abutment edge line generating device can be configured in the electronic device provided in the embodiment of the present invention. The electronic device can be a server, a computer, a mobile terminal or a medical device, etc. For example, the mobile terminal can be a mobile phone, a tablet computer, etc. For example, the medical device can be a device equipped with a processor, such as an oral medical device, and the oral medical device can be equipped with a point cloud scanning device and / or an oral surgical instrument, etc. Figure 1 As shown, the method includes:

[0034] S110, obtaining oral scan data, where the oral scan data includes point cloud data of multiple nodes.

[0035] In this embodiment, the oral scan data may be data obtained by scanning the oral cavity through an intraoral scanning device, for example, oral 3D model data, which may be used in oral medicine fields such as orthodontics, implantation, and restoration. Optionally, the oral scan device may be an oral scanner.

[0036] The oral scan data includes point data of multiple nodes. The nodes in the oral scan data may include nodes located in the abutment area and nodes located in the background area, wherein the background area may include but is not limited to the gum area and the gum area. The point cloud data of each node may include feature information of the node, for example, one or more of the position information, color information and normal feature information of the node, wherein the position information of the node may be the three-dimensional coordinate information of the node, the color information of the node may be the RGB information of the node, and the normal feature information of the node represents the orientation of the node. For example, the oral scan data may be marked as M=(P,F), and the oral scan data is composed of point cloud data x of multiple nodes. i =(p i ,f i ), P represents the three-dimensional coordinates of all nodes of the point cloud, that is, the positions on the x, y, and z axes; F represents the features of all nodes, that is, the normal feature information and color information on the x, y, and z axes; p i represents the three-dimensional coordinates of node i; f i Indicates the feature of node i. The oral cavity is scanned by an intraoral scanning device to obtain oral scanning data including point cloud data of multiple nodes, which can provide a data basis for subsequent processing.

[0037] After the oral scan data is acquired, it is necessary to perform data preprocessing on the oral scan data, for example, one or more of alignment processing, centroid centering processing, scale normalization processing, and resampling processing.

[0038] Optionally, after acquiring the oral scan data, one or more of the following is also included: aligning the point cloud data of multiple nodes in the oral scan data with a predefined tooth coordinate system; centering the point cloud data of multiple nodes in the oral scan data; scale normalizing the point cloud data of multiple nodes in the oral scan data; and resampling the point cloud data of multiple nodes in the oral scan data.

[0039] Specifically, the point cloud data of multiple nodes in the oral scan data are aligned with the predefined tooth coordinate system, which can be done by aligning the coordinate system of the point cloud data of multiple nodes in the oral scan data with the predefined tooth coordinate system through an iterative closest point algorithm, wherein the predefined tooth coordinate system may include standard oral data, and the standard oral data includes the point cloud data of multiple nodes. The specific process is as follows: obtaining the point cloud data of a node among multiple nodes in the oral scan data; performing the closest point matching step, by finding the corresponding relationship between each node in the oral scan data and the closest point in the predefined tooth coordinate system, establishing a matching pair between the node in the oral scan data and the node in the predefined tooth coordinate system, and aligning the oral scan data in the predefined tooth coordinate system through the above matching pair. Specifically, the gradient descent method is used to calculate the error between the matching node pairs, and the optimal transformation parameters are determined to minimize the error; according to the calculated optimal transformation parameters, the position and posture of the point cloud data of multiple nodes in the oral scan data are adjusted to better align with the predefined tooth coordinate system. This process is repeated in the iterative process until the preset convergence condition is reached to ensure the accuracy and stability of the registration result. Finally, the point cloud data of multiple nodes in the registered oral scan data are output to achieve precise alignment with the predefined tooth coordinate system.

[0040] Specifically, the centroid processing of the point cloud data of multiple nodes in the oral scan data can be performed by moving the centroid of the upper and lower jaws to the origin of the coordinate system. The specific process is as follows: Calculate the coordinate average of all nodes in the oral point cloud data, expressed as where p i represents the three-dimensional coordinates of point i, and n represents the number of nodes in the point cloud; the coordinates of each node are subtracted from the average coordinate value to achieve centroid centralization, that is, where p′ i is the coordinate vector centered around the centroid.

[0041] Specifically, the scale normalization processing is performed on the point cloud data of multiple nodes in the oral scan data. The minimum and maximum values ​​of the Y axis are calculated for the point cloud data after centroiding, and the coordinate values ​​of each node are normalized so that they fall within the range of [-1, 1]. The formula is expressed as:

[0042]

[0043] Among them, p″ i is the normalized coordinate vector; min(·) represents the minimum function; max(·) represents the maximum function.

[0044] Specifically, the point cloud data of multiple nodes in the oral scan data are resampled, and the farthest point sampling algorithm can be used to evenly distribute the point cloud data and keep the number of point clouds fixed at N, thereby optimizing the distribution and quality of the point cloud data of multiple nodes in the oral scan data. The specific process is as follows: a point is randomly selected from the point cloud data of multiple nodes in the oral scan data as the starting point set, denoted as P0; the shortest distance from each remaining point in the point cloud data of multiple nodes in the oral scan data to the starting point set P0 is calculated, and the distance is determined by calculating the Euclidean distance from the point to all points in the point set and taking the minimum value; among all the unselected points, a point farthest from the existing point set is selected and added to the point set. Repeat the distance calculation and farthest point selection steps until the size of the point set reaches the predetermined N points. Through resampling processing, the number of nodes obtained by resampling is reduced, the amount of calculation in the process of generating the edge line of the abutment tooth is reduced, and the layout uniformity of the nodes is obtained by optimizing the resampling.

[0045] S120 , classifying each node based on the point cloud data of each node to obtain classification information of each node, where the node type includes abutment type and background type.

[0046] In the process of scanning the oral cavity with an intraoral scanning device, other locations, such as gums, may be scanned. The point cloud data includes point cloud data located in the abutment area and point cloud data in other areas (for example, areas outside the abutment area can be background areas). In order to reduce the interference in the generation of the abutment edge line, each node needs to be classified to identify the type of the node, providing a basis for the generation of the abutment edge line. Among them, the type of the node located in the abutment area can be the abutment type, and the type of the node located in any area outside the abutment area can be the background area.

[0047] In this embodiment, the classification information may be probability data of each node belonging to an abutment. The probability data of each node belonging to an abutment is obtained by performing classification processing on the point cloud data of each node.

[0048] Among them, the classification processing can be implemented by using a machine learning model and a neural network model, for example, the neural network model can be a convolutional neural network and a Transformer model. The above-mentioned machine learning model or neural network model can be trained by the point cloud data of the sample node and the type of the sample node. By inputting the point cloud data of the node into the above-mentioned machine learning model or neural network model, the classification information of the node output by the machine learning model or neural network model is obtained.

[0049] Optionally, the point cloud data of each node is classified based on a pre-trained point cloud recognition model to obtain classification information of each node, wherein the classification information of the node is probability information that the node belongs to the abutment type.

[0050] The point cloud recognition model can be a neural network model, which can be used to classify the point cloud data of each node and output the probability information that the point cloud data of the node belongs to the abutment type. The point cloud recognition model can be a PointTransformer model. The model structure of the point cloud recognition model is not limited here and can be created according to user needs.

[0051] Optionally, the point cloud recognition model may include a feature extraction module and a semantic prediction module. The semantic prediction module is used to predict the classification information of the node, and the feature extraction module is used to extract the input features of the point cloud recognition model. The feature extraction module may include a point conversion module and a downward transition module. The point conversion module may include a point conversion layer, and a linear layer before and after, and a residual structure is adopted. The point conversion layer is responsible for performing self-attention operations and feature conversions on the point cloud data to capture the complex geometric relationships and features in the point cloud data. Before the point conversion layer, a front linear layer is provided for mapping the features of the input point cloud data to a new feature space to provide preprocessing for the self-attention operation. After the point conversion layer, a post linear layer is provided for further mapping and converting the output of the layer to meet the needs of subsequent network layers or tasks. For example, see Figure 2 , Figure 2 is a schematic diagram of the structure of a point conversion module provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the structure of the downward transition module provided in an embodiment of the present invention.

[0052] The processing flow of the point conversion layer is as follows: By adding the position code δ = θ (p i -p j ) to obtain spatial information, where p i， p j are the three-dimensional coordinates corresponding to the two nodes, θ(·) is the multi-layer perceptron layer; using linear transformation ψ(·) calculates the query vector q and key vector k of point i respectively. Then, the features are aggregated by the γ(·) mapping function to generate the attention vector. The attention weight is obtained by the SoftMax function and multiplied by the numerical vector v to obtain the output feature y. i The calculation formula is as follows:

[0053]

[0054] Among them, ρ(·) represents the SoftMax function, γ(·) represents the multi-layer perceptron, ψ(·), α(·) denote linear layers, and ⊙ denotes dot products.

[0055] The downward transition module aims to aggregate features and reduce the resolution of the point cloud through the downsampling process. The processing flow is as follows: the input point cloud data is sampled at the farthest point with a sampling rate of 4 to obtain a sub-point set of the input point set; for each point in the sub-point set, the k points in its neighborhood are obtained by the K nearest neighbor algorithm, where the value of k is set to 16; for the k points in the neighborhood of each point, a multi-layer perceptron is used to extract features, which includes a linear layer and a batch normalization layer, as well as a ReLU activation function; the output features are subjected to local maximum pooling, that is, the feature vectors of the k points in the neighborhood are replaced by the maximum value of each dimension in the output feature vector, thereby completing the process of downsampling and feature aggregation.

[0056] The semantic prediction module can be a stacked upward transition to gradually upsample the features of the point cloud data to restore to the same number of points N as the original input point cloud. The processing flow is as follows: after a multi-layer perceptron, including a linear layer and Batch Normalization, the activation function is the ReLU activation function; perform trilinear interpolation to obtain a super point set of the input point set, thereby improving the resolution of the point set.

[0057] S130, constructing an undirected graph corresponding to the oral scan data, wherein the undirected graph includes multiple nodes in the oral scan data and edges between the nodes, and weights corresponding to the edges between the nodes are determined based on point cloud data of the nodes and / or classification information of the nodes.

[0058] In this embodiment, the undirected graph can be a graph composed of multiple nodes in the oral scan data and edges between the nodes. A weight is set on the edge between any two nodes, which can be used to characterize the association relationship between the nodes. The larger the weight of the edge between the two nodes, the higher the similarity between the two nodes, the stronger the association between the two nodes, and the greater the probability of belonging to the same type; the smaller the weight of the edge between the two nodes, the lower the similarity between the two nodes, the weaker the association between the two nodes, and the smaller the probability of belonging to the same type.

[0059] By using multiple nodes in the oral scan data as nodes in an undirected graph, by determining the weights between any two nodes, and establishing edges between the two nodes, an undirected graph is formed. In some embodiments, the undirected graph can be represented as G = (V, E), where V represents a node set of multiple nodes in the oral scan data, including the three-dimensional coordinates P and feature information F of the point cloud data, and E represents an edge set of the node, each node v∈V corresponds to a point in the oral scan data, and each edge (v i ,v j )∈E connects the node v i and v j .

[0060] Optionally, the weight corresponding to the edge between nodes can be determined based on the point cloud data of the nodes. For example, the similarity data of the point cloud data of two nodes can be calculated and the weight can be determined based on the similarity data. The weight can be positively correlated with the similarity data. For example, the similarity data of the point cloud data of two nodes can be determined as the weight corresponding to the edge between the two nodes.

[0061] Optionally, the weight corresponding to the edge between nodes can be determined based on the classification information of the nodes. For example, the weight corresponding to the edge between nodes can be obtained by calculating the difference between the classification information of two nodes. When the difference between the classification information of two nodes is larger, it means that the probability that the two points belong to the same type is smaller, and the weight corresponding to the edge between the two nodes is smaller. Accordingly, the weight corresponding to the edge between two nodes is negatively correlated with the difference between the classification information of the two nodes.

[0062] Optionally, the weight corresponding to the edge between nodes can be determined based on the classification information of the nodes and the point cloud data of the nodes. Specifically, the method for determining the weight corresponding to the edge between any two nodes in an undirected graph is: determining the initial weight of the edge between the two nodes based on the similarity data of the point cloud data of the two nodes; determining the weight adjustment information of the edge based on the classification information of the two nodes, adjusting the initial weight of the edge based on the weight adjustment information of the edge, and determining the target weight of the edge.

[0063] Specifically, the similarity data can represent the similarity between the point cloud data of two nodes. For example, the similarity data of one or more of the position information, color information and normal feature information of the two nodes can be calculated. For example, the initial weight of the edge is calculated as follows:

[0064]

[0065] Among them, I(v i ) and I(v j ) is the node v i and node v j Point cloud data.

[0066] The weight adjustment information is information used to optimize the initial weight, and can be obtained by converting the classification information of two nodes based on the arc cosine similarity mapping function. The edge weight adjustment information is calculated as follows:

[0067]

[0068] Among them, p(v i ) and p(v j ) are nodes v i and node v j The probability of belonging to abutment tooth.

[0069] The target weight is obtained by calculating the product of the edge weight adjustment information and the initial weight of the edge. The target weight of the edge is calculated as follows:

[0070] w ij ′=w ij ·P(e ij )

[0071] Among them, w ij ′ is the target weight of the edge.

[0072] S140, segmenting the undirected graph based on the weights of the edges in the undirected graph to obtain abutment subgraph and background subgraph, determining abutment edge nodes based on the abutment subgraph and the background subgraph, and generating abutment edge lines based on the abutment edge nodes.

[0073] In this embodiment, the abutment subgraph may be the part of the abutment in the segmented undirected graph, and the background subgraph may be the part of the background in the segmented undirected graph. The undirected graph is segmented to update the type of the node. The type of the node in the abutment subgraph is the abutment type, and the type of the node in the background subgraph is the background type. It can be understood that in the process of classifying the point cloud data of the node based on the point cloud recognition model, due to the processing accuracy of the point cloud recognition model, the classification information of the node may have a certain error. The node type is updated through the graph segmentation processing of the undirected graph to improve the accuracy of the node type.

[0074] Optionally, the undirected graph can be segmented using a neural network model. For example, the undirected graph is input into a graph segmentation model (ie, a neural network model) to obtain a base sub-graph and a background sub-graph output by the graph segmentation model.

[0075] Optionally, the graph segmentation of the undirected graph can also be achieved by threshold judgment. For example, a threshold is set to split the edges in the undirected graph that are smaller than the threshold to obtain two sub-graphs. It can be understood that discrete nodes can also be included in addition to the two sub-graphs. For discrete nodes, the weights of the discrete nodes and the nodes in the base sub-graph and the background sub-graph can be calculated respectively. The larger the weight, the greater the probability that the node belongs to the corresponding base sub-graph or background sub-graph. The node can be connected to the corresponding sub-graph to obtain the base sub-graph and the background sub-graph.

[0076] Optionally, the undirected graph is segmented to obtain abutment subgraph and background subgraph, including: segmenting the undirected graph based on the minimum cut principle to obtain abutment subgraph and background subgraph. Specifically, the minimum cut principle can be to find the minimum edge weight and cut set that divides the graph into two subsets, for example, the undirected graph can be segmented by minimizing the total weight of the cut edges. The undirected graph can be segmented into abutment subgraph and background subgraph by the minimum cut principle, which is helpful for the subsequent determination of abutment edge nodes.

[0077] The undirected graph is segmented by the weight of the edges in the undirected graph to obtain abutment subgraph and background subgraph. The nodes between the abutment subgraph and the background subgraph are determined as abutment edge nodes, and the abutment edge nodes are fitted into abutment edge lines, thereby improving the accuracy of abutment edge line generation.

[0078] The abutment edge node is a node located at the edge of the abutment area. A plurality of abutment edge nodes may form an abutment edge line. The abutment edge line may be understood as a boundary line between the abutment area and the background area in the oral scan data.

[0079] Optionally, the abutment edge nodes are determined in the abutment subgraph, for example, in a graph partitioning process of the undirected graph, a node located in the abutment subgraph among two nodes corresponding to the partitioned edge is determined as the abutment edge node.

[0080] Generate an abutment edge line based on the abutment edge nodes in the oral scan data. For example, the abutment edge nodes in the oral scan data are connected to obtain the abutment edge line. For example, the abutment edge line can be obtained by fitting the abutment edge nodes in the oral scan data. For example, the tooth edge nodes can be fitted into a continuous and smooth curve by the least square method, and the curve is the abutment edge line.

[0081] The technical solution of the present embodiment obtains oral scanning data including point cloud data of multiple nodes, and performs one or more of data preprocessing such as alignment processing, centroid centering processing, scale normalization processing and resampling processing on the point cloud data of multiple nodes in the oral scanning data with a predefined tooth coordinate system, classifies each node based on the point cloud data of each node, obtains classification information of each node, constructs an undirected graph corresponding to the oral scanning data, segments the undirected graph to minimize the total weight of the cut edges, obtains abutment subgraph and background subgraph, determines the nodes between the abutment subgraph and the background subgraph as abutment edge nodes, forms abutment edge lines based on the abutment edge nodes, thereby improving the accuracy of abutment edge node generation, and further improving the accuracy of abutment edge line generation.

[0082] Embodiment 2

[0083] Figure 4 FIG. 1 is a flow chart of another method for generating abutment edge lines provided by Embodiment 2 of the present invention. This embodiment is an optimization of the above embodiment. Figure 4 As shown, the method includes:

[0084] S210 , obtaining oral scan data, where the oral scan data includes point cloud data of multiple nodes.

[0085] S220 , classifying each node based on the point cloud data of each node to obtain classification information of each node, where the node type includes abutment type and background type.

[0086] S230, constructing an undirected graph corresponding to the oral scan data, wherein the undirected graph includes multiple nodes in the oral scan data and edges between the nodes, and weights corresponding to the edges between the nodes are determined based on point cloud data of the nodes and / or classification information of the nodes.

[0087] S240, segmenting the undirected graph based on the weights of the edges in the undirected graph to obtain abutment subgraph and background subgraph; setting abutment labels for nodes in the abutment subgraph; setting background labels for nodes in the background subgraph; when the label of any node is the abutment label and the labels of the adjacent nodes of the node are background labels, determining the node as an abutment edge node; fitting the abutment edge node to an abutment edge line.

[0088] In this embodiment, the abutment label can indicate that the node is an abutment, and the background label can indicate that the node is a background. For example, the abutment label can be 1, and the corresponding background label can be 0. By setting abutment labels for nodes in the abutment subgraph and background labels for nodes in the background subgraph, when the label of any node is an abutment label and the label of its adjacent node is a background label, it indicates that the type of the node is an abutment type and the type of its adjacent node is a background type. Therefore, the node can be determined as an abutment edge node, and all abutment edge nodes can be fitted to obtain an abutment edge line.

[0089] There is a certain interval between nodes. The resampling process of point cloud data of multiple nodes in oral scanning data may cause large intervals between nodes and errors in fitting the abutment edge line. In order to make the abutment edge line smoother, connection points can be set to further fit the abutment edge line.

[0090] Optionally, fitting the abutment edge nodes into the abutment edge lines includes: determining connection points between adjacent abutment edge nodes based on adjacent abutment edge nodes; and fitting the abutment edge nodes and the connection points to obtain the abutment edge lines.

[0091] Specifically, the connection point can be a discrete point set used to form the edge of the abutment tooth, for example, it can be obtained by calculating the midpoint between two adjacent abutment tooth edge nodes. The connection point is obtained by calculating the midpoint between two adjacent abutment tooth edge nodes, and the abutment tooth edge nodes and the connection point are sequentially connected. The connection point can be fitted into a continuous edge curve by the least square method, and the continuous edge curve is the abutment tooth edge line, which improves the fitting effect of the abutment tooth edge line, thereby improving the accuracy of the abutment tooth edge line generation.

[0092] After obtaining the fitted abutment tooth edge line, in order to reduce the error of the abutment tooth edge line and reduce abnormal points, the abutment tooth edge line can be corrected based on the curvature.

[0093] Optionally, after fitting the abutment edge nodes as the abutment edge line, the method further includes: determining the curvature distribution on the abutment edge line; determining an abnormal point set on the abutment edge line based on the curvature distribution on the abutment edge line, and correcting the abnormal point set to obtain a corrected abutment edge line.

[0094] Specifically, the curvature distribution can be the curvature of each connection point on the edge line of the abutment tooth, and the curvature calculation formula is as follows:

[0095]

[0096] Where κ is the curvature, r′ and r″ are the first-order derivative vector and second-order derivative vector of the abutment edge line at the connection point, respectively.

[0097] The curvature distribution of the abutment tooth edge line is calculated, and the curvature average value is calculated, the curvature of all connection points is compared with the curvature average value, the abnormal point set with excessive curvature is identified, and the abnormal point set is corrected, for example, by removing the abnormal point set with excessive curvature or by smoothing the abnormal point set, so as to optimize the quality of the abutment tooth edge line, improve the fitting effect of the abutment tooth edge line, make the fitted abutment tooth edge line closer to the real abutment tooth edge line, and thus improve the accuracy of the abutment tooth edge line generation.

[0098] The technical solution of the present embodiment obtains oral scan data including point cloud data of multiple nodes, and performs one or more of data preprocessing such as alignment processing, centroid centering processing, scale normalization processing and resampling processing on the point cloud data of multiple nodes in the oral scan data with a predefined tooth coordinate system, classifies each node based on the point cloud data of each node, obtains classification information of each node, constructs an undirected graph corresponding to the oral scan data, sets abutment labels for nodes located in the abutment subgraph, sets background labels for nodes located in the background subgraph, and when the label of any node is an abutment label and the label of an adjacent node of the node is a background label, the node is determined to be an abutment edge node, and the abutment edge node is fitted as an abutment edge line, thereby improving the fitting effect of the abutment edge line and improving the accuracy of abutment edge line recognition and extraction.

[0099] Embodiment 3

[0100] Figure 5 : is a flow chart of a method for training a point cloud recognition model provided by Embodiment 3 of the present invention. Figure 5 As shown, the training method of the point cloud recognition model includes:

[0101] S310, constructing an initial model, the initial model includes a feature extraction module, a semantic prediction module and an offset prediction module, wherein the semantic prediction module and the offset prediction module are respectively connected to the feature extraction module, the semantic prediction module is used to predict the classification information of the node, and the offset prediction module is used to predict the center of mass offset of the node.

[0102] In this embodiment, the initial model is an untrained point cloud recognition model, and the feature extraction module can be a module for extracting model input features, such as a convolutional neural network model. The offset prediction module can be a module that calculates the offset of each node O = {o1, ..., o N}∈R N×3 , each offset is a vector from the node coordinates of the input point cloud to its corresponding tooth centroid.

[0103] S320 , obtaining a sample set, where the sample set includes sample nodes, target types of the sample nodes, and centroid information corresponding to the sample nodes.

[0104] In this embodiment, the sample set may be point cloud data of the sample, which may be obtained by scanning with a scanner. The sample node may be a node in the point cloud data, and the target type may be the type of the sample node that the point cloud recognition model needs to predict, such as abutment type and background type. The centroid information may be the three-dimensional coordinate information of the centroid. By scanning the sample with a scanner to obtain the point cloud data of the sample, relevant information of the sample node may be obtained, providing data support for the point cloud recognition model.

[0105] S330, performing prediction processing on the sample nodes based on the initial model to obtain predicted classification information and predicted centroid offset; generating a loss function based on the target type of the sample nodes, the centroid information corresponding to the sample nodes, the predicted classification information and the predicted centroid offset.

[0106] In this embodiment, the predicted classification information can be the type of the sample node obtained by the initial point cloud recognition model predicting the sample node, for example, it can be a base tooth type or a background type. The predicted center of mass offset can be the offset of the center of mass of the sample node relative to the predefined tooth coordinate system obtained by the initial model predicting the sample node. The loss function can be a standard for measuring the training effect of the point cloud recognition model. The smaller the loss function, the better the training effect of the point cloud recognition model, and vice versa. The loss function can be obtained by calculating the difference between the target type of the sample node and the predicted classification information, and the difference between the center of mass information corresponding to the sample node and the predicted center of mass offset.

[0107] S340, with the goal of reducing the loss function, the initial model is trained, and when the training end conditions are met, a point cloud recognition model is formed based on the feature extraction module and the semantic prediction module in the trained initial model.

[0108] In some embodiments, the initial model is trained with the goal of reducing the loss function, and when the training termination condition is met, a point cloud recognition model is formed based on the feature extraction module, semantic prediction module, and offset prediction module in the trained initial model. The termination condition may include, but is not limited to, one or more of the number of iterative training times reaching a preset number, the accuracy of the model output result reaching a preset accuracy, and the loss function of the model reaching a convergence state.

[0109] During the model training process, corresponding loss functions are generated based on the output results of the semantic prediction module and the offset prediction module, respectively. Based on the loss functions corresponding to the semantic prediction module and the offset prediction module, respectively, the model parameters of the initial model are adjusted. For example, the loss function corresponding to the semantic prediction module is reversely input from the output end of the semantic prediction module, and the loss function corresponding to the offset prediction module is reversely input from the output end of the offset prediction module, and the parameter adjustment of the initial model can be achieved based on the gradient descent method. For example, the sum of the loss functions corresponding to the semantic prediction module and the offset prediction module can be used as the target loss function, and the target loss function is reversely input from the output end of the offset prediction module and the output end of the semantic prediction module, respectively, to achieve parameter adjustment of the initial model.

[0110] The loss function corresponding to the semantic prediction module can represent the difference between the target type of the sample node and the predicted classification information, for example, it can be a cross entropy loss function, which can be expressed by L sem The loss function corresponding to the offset prediction module can represent the difference between the centroid information corresponding to the sample node and the predicted centroid offset. For example, it can be an L1 regression loss function, which can be expressed by L o_reg The loss function L of the semantic prediction module is represented by o_reg The calculation is as follows:

[0111]

[0112] Among them, i is the offset information of node i, and the output of the offset prediction module is O = {o1, ..., o N}∈R N×3 ; N represents the number of points; m i Represented as a binary mask, if node i is in the abutment, then m i =1, otherwise 0; is the centroid of the abutment to which node i belongs. The calculation formula is as follows:

[0113]

[0114] Among them, g(i) represents the index of node i mapped to the truth graph, that is, the abutment containing node i; Abutment I g(i) 's points.

[0115] The technical solution of this embodiment obtains a sample set including sample nodes, target types of sample nodes, and centroid information corresponding to the sample nodes, performs prediction processing on the sample nodes based on the initial model, obtains predicted classification information and predicted centroid offset, generates a loss function based on the target types of sample nodes, centroid information corresponding to the sample nodes, predicted classification information, and predicted centroid offset, trains the initial model with the goal of reducing the loss function, and forms a point cloud recognition model based on the feature extraction module and semantic prediction module in the trained initial model when the training end conditions are met, thereby accelerating the convergence speed of the point cloud recognition model and improving the accuracy of the point cloud recognition model prediction.

[0116] Embodiment 4

[0117] Figure 6 Schematic diagram of the structure of a device for generating abutment edge line provided by the fourth embodiment of the present invention. Figure 6 As shown, the device comprises:

[0118] An oral scan data acquisition module 410 is used to acquire oral scan data, where the oral scan data includes point cloud data of multiple nodes;

[0119] A classification module 420, for classifying each node based on the point cloud data of each node to obtain classification information of each node, where the type of the node includes abutment type and background type;

[0120] An undirected graph construction module 430 is used to construct an undirected graph corresponding to the oral scan data, wherein the undirected graph includes a plurality of nodes in the oral scan data and edges between the nodes, and weights corresponding to the edges between the nodes are determined based on point cloud data of the nodes and / or classification information of the nodes;

[0121] The abutment edge line generation module 440 is used to segment the undirected graph based on the weights of the edges in the undirected graph to obtain an abutment subgraph and a background subgraph, determine abutment edge nodes based on the abutment subgraph and the background subgraph, and generate abutment edge lines based on the abutment edge nodes.

[0122] On the basis of the above embodiment, optionally, the classification module 420 is also used to classify the point cloud data of each node based on the point cloud recognition model to obtain classification information of each node, wherein the classification information of the node is the probability information that the node belongs to the abutment type.

[0123] Optionally, the classification module 420 also includes a point cloud recognition model training module, which is used to construct an initial model. The initial model includes a feature extraction module, a semantic prediction module and an offset prediction module, wherein the semantic prediction module and the offset prediction module are respectively connected to the feature extraction module; the semantic prediction module is used to predict the classification information of the node, and the offset prediction module is used to predict the center of mass offset of the node.

[0124] Optionally, the point cloud recognition model training module of the classification module 420 is also used to obtain a sample set, which includes sample nodes, target types of sample nodes, and centroid information corresponding to the sample nodes; predict the sample nodes based on the initial model to obtain predicted classification information and predicted centroid offset; generate a loss function based on the target type of the sample nodes, the centroid information corresponding to the sample nodes, the predicted classification information, and the predicted centroid offset; train the initial model with the goal of reducing the loss function, and when the training end conditions are met, form a point cloud recognition model based on the feature extraction module and semantic prediction module in the trained initial model.

[0125] Optionally, the undirected graph construction module 430 is also used to determine the initial weight of the edge between two nodes based on the similarity data of the point cloud data of the two nodes; determine the weight adjustment information of the edge based on the classification information of the two nodes, adjust the initial weight of the edge based on the weight adjustment information of the edge, and determine the target weight of the edge.

[0126] Optionally, the abutment edge line generating module 440 is further configured to perform graph segmentation on the undirected graph based on a minimum cut principle to obtain an abutment subgraph and a background subgraph.

[0127] Optionally, the abutment edge line generation module 440 is also used to set abutment labels for nodes located in the abutment subgraph; set background labels for nodes located in the background subgraph; when the label of any node is an abutment label and the label of the node's adjacent node is a background label, determine the node as an abutment edge node; and fit the abutment edge node as an abutment edge line.

[0128] Optionally, the abutment edge line generating module 440 is further used to determine the connection points between adjacent abutment edge nodes based on adjacent abutment edge nodes; and to fit the abutment edge nodes and the connection points to obtain the abutment edge line.

[0129] Optionally, the abutment edge line generating module 440 is further used to determine the curvature distribution on the abutment edge line; determine the abnormal point set on the abutment edge line based on the curvature distribution on the abutment edge line, correct the abnormal point set, and obtain a corrected abutment edge line.

[0130] Optionally, the device further includes a data preprocessing module, which is configured to perform one or more of the following steps after acquiring the oral scan data:

[0131] The point cloud data of multiple nodes in the oral scan data are aligned with the predefined tooth coordinate system; the point cloud data of multiple nodes in the oral scan data are centroided; the point cloud data of multiple nodes in the oral scan data are scale-normalized; and the point cloud data of multiple nodes in the oral scan data are resampled.

[0132] The abutment edge line generating device provided in the embodiment of the present invention can execute the abutment edge line generating method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the executing method.

[0133] Embodiment 5

[0134] Figure 7 1 is a schematic diagram of the structure of an electronic device provided in Embodiment 5 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0135] like Figure 7 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0136] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0137] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the abutment edge line generation method.

[0138] In some embodiments, the abutment edge line generation method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the abutment edge line generation method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the abutment edge line generation method in any other appropriate manner (e.g., by means of firmware).

[0139] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0140] The computer program for implementing the method for generating abutment edge lines of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program can be executed completely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or completely on a remote machine or server.

[0141] Embodiment 6

[0142] Embodiment 6 of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to execute a method for generating an abutment edge line, the method comprising:

[0143] Acquire oral scan data, the oral scan data including point cloud data of multiple nodes; classify each node based on the point cloud data of each node to obtain classification information of each node, the type of node including abutment type and background type; construct an undirected graph corresponding to the oral scan data, the undirected graph including multiple nodes in the oral scan data and edges between the nodes, the weights corresponding to the edges between the nodes are determined based on the point cloud data of the nodes and / or the classification information of the nodes; segment the undirected graph based on the weights of the edges in the undirected graph to obtain abutment subgraph and background subgraph, determine abutment edge nodes based on the abutment subgraph and the background subgraph, and generate abutment edge lines based on the abutment edge nodes.

[0144] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0145] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0146] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0147] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0148] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0149] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating an abutment edge line, characterized in that: include: Acquire oral scan data, wherein the oral scan data includes point cloud data of multiple nodes; Classifying each of the nodes based on the point cloud data of each of the nodes to obtain classification information of each of the nodes, wherein the types of the nodes include abutment types and background types; Constructing an undirected graph corresponding to the oral scan data, wherein the undirected graph includes the plurality of nodes in the oral scan data and edges between the nodes, and weights corresponding to the edges between the nodes are determined based on point cloud data of the nodes and classification information of the nodes; The undirected graph is segmented based on the weights of the edges in the undirected graph to obtain abutment subgraph and background subgraph, abutment edge nodes are determined based on the abutment subgraph and the background subgraph, and abutment edge lines are generated based on the abutment edge nodes.

2. The method according to claim 1, characterized in that The classifying process of each node based on the point cloud data of each node to obtain the classification information of each node includes: The point cloud data of each of the nodes is classified based on the point cloud recognition model to obtain classification information of each of the nodes, wherein the classification information of the nodes is probability information that the nodes belong to the abutment type.

3. The method according to claim 2, characterized in that The training method of the point cloud recognition model includes: Constructing an initial model, the initial model comprising a feature extraction module, a semantic prediction module and an offset prediction module, wherein the semantic prediction module and the offset prediction module are respectively connected to the feature extraction module; the semantic prediction module is used to predict the classification information of the node, and the offset prediction module is used to predict the centroid offset of the node; Acquire a sample set, the sample set including sample nodes, target types of the sample nodes, and centroid information corresponding to the sample nodes; Based on the initial model, the sample node is predicted and processed to obtain predicted classification information and predicted centroid offset; based on the target type of the sample node, the centroid information corresponding to the sample node, the predicted classification information and the predicted centroid offset, a loss function is generated; The initial model is trained with the goal of reducing the loss function, and when the training end condition is met, the point cloud recognition model is formed based on the feature extraction module and the semantic prediction module in the trained initial model.

4. The method according to claim 1, characterized in that: The method for determining the weight corresponding to the edge between any two nodes in the undirected graph is: Determining an initial weight of an edge between two nodes based on similarity data of the point cloud data of the two nodes; The weight adjustment information of the edge is determined based on the classification information of the two nodes, and the initial weight of the edge is adjusted based on the weight adjustment information of the edge to determine the target weight of the edge.

5. The method according to claim 1, characterized in that The step of segmenting the undirected graph to obtain a base subgraph and a background subgraph includes: The undirected graph is segmented based on the minimum cut principle to obtain abutment subgraphs and background subgraphs.

6. The method according to claim 1, characterized in that The step of determining abutment edge nodes based on the abutment sub-graph and the background sub-graph, and generating abutment edge lines based on the abutment edge nodes comprises: Setting abutment labels for nodes in the abutment subgraph; setting background labels for nodes in the background subgraph; When the label of any of the nodes is the abutment label, and the label of the node adjacent to the node is the background label, determining the node as an abutment edge node; The abutment edge nodes are fitted as abutment edge lines.

7. The method according to claim 6, characterized in that The step of fitting the abutment edge nodes into abutment edge lines comprises: Based on adjacent abutment edge nodes, determining a connection point between adjacent abutment edge nodes; The abutment edge nodes and the connection points are fitted to obtain the abutment edge line.

8. The method according to claim 6, characterized in that After fitting the abutment edge nodes as abutment edge lines, the method further comprises: determining a curvature distribution on the edge line of the abutment tooth; An abnormal point set on the abutment tooth edge line is determined based on the curvature distribution on the abutment tooth edge line, and the abnormal point set is corrected to obtain a corrected abutment tooth edge line.

9. The method according to claim 1, characterized in that: After obtaining the oral scan data, one or more of the following may also be included: Aligning the point cloud data of multiple nodes in the oral scan data with a predefined tooth coordinate system; Performing centroid centering processing on point cloud data of multiple nodes in the oral scan data; Performing scale normalization processing on point cloud data of multiple nodes in the oral scan data; Resampling is performed on the point cloud data of multiple nodes in the oral scan data.

10. A device for generating an abutment edge line, characterized in that: include: An oral scan data acquisition module, used to acquire oral scan data, wherein the oral scan data includes point cloud data of multiple nodes; A classification module, used for classifying each of the nodes based on the point cloud data of each of the nodes to obtain classification information of each of the nodes, wherein the types of the nodes include abutment types and background types; An undirected graph construction module, used to construct an undirected graph corresponding to the oral scan data, wherein the undirected graph includes the multiple nodes in the oral scan data and the edges between the nodes, and the weights corresponding to the edges between the nodes are determined based on the point cloud data of the nodes and the classification information of the nodes; The abutment edge line generation module is used to segment the undirected graph based on the weights of the edges in the undirected graph to obtain an abutment subgraph and a background subgraph, determine abutment edge nodes based on the abutment subgraph and the background subgraph, and generate an abutment edge line based on the abutment edge nodes.

11. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the abutment edge line generating method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the abutment edge line generation method according to any one of claims 1 to 9 when executed.

Images