Dental jaw recognition and display method and device, oral instrument, storage medium and equipment

Abstract

The invention discloses a tooth jaw recognition and display method and device, an oral instrument, a storage medium and equipment. The tooth jaw recognition method comprises the following steps: respectively obtaining first tooth jaw model data and second tooth jaw model data of a to-be-detected tooth; according to the first dental model data and the second dental model data, obtaining relation identification information of a tooth pair of a to-be-detected tooth; and determining tooth jaw information corresponding to the first tooth jaw model data and the second tooth jaw model data according to the relation identification information. According to the tooth jaw recognition method provided by the invention, the robustness of the tooth jaw recognition process can be stronger, excessive manual intervention and complex preprocessing steps are avoided, the advantages of high efficiency and stability are considered, and the tooth jaw recognition method can adapt to various complex scenes.

Description

Technical Field

[0001] The present invention relates to the technical field of oral digital models, and particularly to a method and device for dental arch recognition and display, an oral instrument, a storage medium, and a device. Background Art

[0002] With the development of society and the improvement of people's requirements for the quality of life, the maintenance of oral health, including oral hygiene and aesthetics, has gradually become the focus of people's attention. When analyzing the organizational structures in the oral cavity, such as dental crowns and gums, it is inevitable to construct a digital model of the oral cavity by means of oral scanning, intraoral photography, digitalization of silicone models, etc., and analyze the digital model to finally determine the treatment plan.

[0003] When performing model analysis, it is necessary to distinguish between the maxilla and the mandible in the model. In the general technical field, mainly manual marking of the dental arch is carried out manually to achieve the effect of distinction, but this method has low efficiency and accuracy, and cannot cope with complex and variable individual differences. There is a method for segmenting the alveolar bone in the prior art, and the model containing only the alveolar bone is sequentially input into the lower alveolar bone segmentation model and the upper alveolar bone segmentation model, but this technical solution has relatively strict requirements for the data preprocessing step, and the error will continue to increase during the iteration process, and the segmentation effect is still not ideal. Summary of the Invention

[0004] One of the purposes of the present invention is to provide a method for dental arch recognition to solve the technical problems in the prior art that rely too much on manual marking and preprocessing steps, and the error increases continuously with the iteration process due to the single operation reference object.

[0005] One of the purposes of the present invention is to provide a display method.

[0006] One of the purposes of the present invention is to provide an oral instrument.

[0007] One of the purposes of the present invention is to provide a computer storage medium.

[0008] One of the purposes of the present invention is to provide a dental arch recognition device.

[0009] One of the purposes of the present invention is to provide an electronic device.

[0010] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides a dental arch recognition method, including: respectively obtaining first dental arch model data and second dental arch model data of a to-be-detected tooth, where the first dental arch model and the second dental arch model correspond to different dental arches; obtaining relationship identification information of tooth pairs of the to-be-detected tooth according to the first dental arch model data and the second dental arch model data, where the tooth pairs include at least one pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs; determining the dental arch information corresponding to the first dental arch model data and the second dental arch model data according to the relationship identification information.

[0011] As a further improvement of an embodiment of the present invention, the dental arch information includes upper jaw determination information or lower jaw determination information for the first dental arch model, and upper jaw determination information or lower jaw determination information for the second dental arch model.

[0012] As a further improvement of an embodiment of the present invention, the dental arch recognition method includes: obtaining tooth pair data of at least one pair of teeth located in different dental arches and corresponding to each other based on the first dental arch model data and the second dental arch model data; parallelly inputting the tooth pair data into a dental arch relationship recognition model to predict corresponding relationship identification information; the dental arch relationship recognition model is a neural network model, and the dental arch relationship recognition model is trained based on preset tooth pair data and relationship labels representing the relative position relationship of the tooth pairs in the preset tooth pair data.

[0013] As a further improvement of an embodiment of the present invention, the dental arch recognition method includes: extracting features from the to-be-detected tooth data according to a first feature extraction module to obtain first tooth data, and extracting features from the to-be-detected tooth data according to a second feature extraction module to obtain second tooth data; obtaining first feature data and second feature data corresponding to the first tooth data and the second tooth data respectively, and generating a tooth pair feature sequence; predicting relationship identification information of the first tooth and the second tooth according to the tooth pair feature sequence.

[0014] As a further improvement of an embodiment of the present invention, the dental arch recognition method includes: obtaining a numerical form of the relationship identification information, and determining upper and lower jaw determination information of the first dental arch model data and the second dental arch model data according to the numerical relationship.

[0015] As a further improvement of an embodiment of the present invention, the jaw recognition method includes: when the teeth to be measured in the first jaw model point to the upper teeth and the teeth to be measured in the second jaw model data point to the lower teeth in the tooth pair data, obtaining relationship identification information in the form of a first tag value in a numerical form; when the teeth to be measured in the first jaw model point to the lower teeth and the teeth to be measured in the second jaw model data point to the upper teeth in the tooth pair data, obtaining relationship recognition information in the form of a second tag value in a numerical form.

[0016] As a further improvement of an embodiment of the present invention, the jaw recognition method includes: based on the tag values in numerical form corresponding to multiple groups of tooth pairs, determining the upper and lower jaw determination information according to the voting scoring result.

[0017] As a further improvement of an embodiment of the present invention, the jaw recognition method includes: segmenting the first jaw model and the second jaw model to obtain a number of first jaw tooth data and a number of second jaw tooth data; matching the first jaw tooth data and the second jaw tooth data according to the tooth position number to obtain tooth pair data.

[0018] As a further improvement of an embodiment of the present invention, the jaw recognition method includes: performing surface simplification on the first jaw model and the second jaw model according to the differentiability of the parts in the jaw model; respectively registering the first jaw model and the second jaw model, and segmenting the first jaw model and the second jaw model based on the tooth segmentation model, where the tooth segmentation model is a neural network model trained based on a preset jaw model and a preset tooth label representing the tooth distribution in the preset jaw model.

[0019] As a further improvement of an embodiment of the present invention, the jaw recognition method includes: determining the patch segmentation method of the jaw model according to the differentiability of adjacent patches in the jaw model, and obtaining the patches of the jaw model accordingly; merging the effective vertices of the obtained patches.

[0020] As a further improvement of an embodiment of the present invention, the jaw recognition method includes: determining the patch segmentation method of the jaw model according to the differentiability between the patch and other intraoral tissues and the cost of adjacent patches belonging to the same intraoral tissue; the differentiability between the patch and other intraoral tissues is determined according to at least one of the following indicators: the degree of the current patch approaching the incisal edge connection line; the degree of the current patch approaching the incisal extreme point of its corresponding tooth; the degree of the current patch being far from the model center of the tooth jaw data to be measured.

[0021] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides a display method, including: displaying a first dental arch model and a second dental arch model according to dental arch information, where the dental arch information is obtained based on the dental arch recognition method described in any of the above technical solutions; the first dental arch model and the second dental arch model are used to design an orthodontic plan.

[0022] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides an oral instrument, which is prepared based on the dental arch information determined by the dental arch recognition method described in any of the above technical solutions.

[0023] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides a computer storage medium, on which an application program is stored. When the application program is executed, the steps of the dental arch recognition method described in any of the above technical solutions are implemented.

[0024] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides a dental arch recognition device, including: a first module for obtaining first dental arch model data and second dental arch model data of a to-be-detected tooth; the first dental arch model and the second dental arch model correspond to different dental arches; a second module for obtaining relationship identification information of tooth pairs of the to-be-detected tooth according to the first dental arch model data and the second dental arch model data, where the tooth pairs include at least a pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs; a third module for determining the dental arch information corresponding to the first dental arch model data and the second dental arch model data according to the relationship identification information.

[0025] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides an electronic device, including a processor, a memory, and a communication bus. The processor and the memory complete mutual communication through the communication bus; the memory is used to store an application program; the processor is used to implement the steps of the dental arch recognition method described in any of the above technical solutions when executing the application program stored on the memory.

[0026] Compared with the prior art, the dental arch recognition method provided by the present invention uses relationship identification information representing the relative relationship of tooth pairs to determine the dental arch information of different dental arches where the tooth pairs are located, which is equivalent to providing tooth pairs as a reference for the dental arch recognition process, making the overall robustness of the recognition process stronger; and, since the tooth recognition scheme is relatively mature, compared with the scheme with excessive manual intervention and complex preprocessing steps, it takes into account the characteristics of high efficiency and stability; due to the discreteness between tooth pairs, it can also avoid the overall result being affected by individual errors and can adapt to the recognition in various complex scenarios. Description of the Drawings

[0027] Figure 1 It is a schematic structural diagram of an electronic device in an embodiment of the present invention.

[0028] Figure 2 It is a schematic structural diagram of a dental arch recognition device in an embodiment of the present invention.

[0029] Figure 3 It is a schematic structural diagram of a dental arch relationship recognition model in an embodiment of the present invention.

[0030] Figure 4 It is a schematic structural diagram of a dental arch model of a tooth to be measured in an embodiment of the present invention.

[0031] Figure 5 It is a partial schematic structural diagram of a dental arch relationship recognition model in another embodiment of the present invention.

[0032] Figure 6 It is a schematic diagram of the steps of a dental arch recognition method in an embodiment of the present invention.

[0033] Figure 7 It is a partial schematic diagram of the steps of the first embodiment of a dental arch recognition method in an embodiment of the present invention.

[0034] Figure 8 It is a partial schematic diagram of the steps of the second embodiment of a dental arch recognition method in an embodiment of the present invention.

[0035] Figure 9 It is a partial schematic diagram of the steps of the third embodiment of a dental arch recognition method in an embodiment of the present invention. Detailed Embodiments

[0036] The present invention will be described in detail below in conjunction with the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, method, or functional transformation made by those of ordinary skill in the art based on these embodiments is included in the protection scope of the present invention.

[0037] It should be noted that the term "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or device. In addition, the terms "first", "second", "third", "fourth", etc. are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0038] The main idea of the present invention is to utilize the relationship identification information that can characterize the relative relationship of teeth pairs to determine the dental arch information of the two dental arches pointed to by the teeth pair. By means of this special local organizational structure in the oral cavity, namely the "teeth pair", the problem of dental arch recognition is transformed into the problem of using the teeth pair relationship. Relying on the characteristics of the teeth pair relationship that are simple and convenient for analysis, the dental arch information is recognized.

[0039] The teeth pair is a combination of at least two corresponding teeth in at least two dental arch models. The construction of the relative relationship between teeth in different dental arch models can be based on the corresponding relationship of numbers in dental position notations such as FDI, or can be constructed according to requirements through algorithms or manual recognition.

[0040] The following will further elaborate on various embodiments, technical principles, and corresponding technical effects of the present invention in conjunction with the accompanying drawings.

[0041] One embodiment of the present invention provides an electronic device, as Figure 1 shown. The electronic device can be a computer, a mobile phone, a tablet computer, etc. The present invention does not limit the specific type of the electronic device.

[0042] The electronic device includes at least one processor, at least one memory, and a communication bus. The at least one processor and the at least one memory communicate with each other through the communication bus.

[0043] The memory is used to store application programs.

[0044] The processor is used to implement the steps of a dental arch recognition method when executing the application program stored in the memory. In one embodiment, the dental arch recognition method may include at least one of the following steps:

[0045] Obtain the first dental arch model data and the second dental arch model data of the teeth to be measured respectively, where the first dental arch model and the second dental arch model correspond to different dental arches;

[0046] According to the first dental arch model data and the second dental arch model data, obtain the relationship identification information of the teeth pair of the teeth to be measured. The teeth pair includes at least a pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the teeth pair;

[0047] According to the relationship identification information, determine the dental arch information corresponding to the first dental arch model data and the second dental arch model data.

[0048] The dental arch recognition method can also be configured based on any of the technical solutions provided later.

[0049] The communication bus may include any number of buses and bridge circuits. In some embodiments, in addition to connecting the processor and the memory, the communication bus may also be used to connect peripheral devices or other peripheral circuits.

[0050] An embodiment of the present invention provides a dental arch recognition device, as Figure 2 shown. The dental arch recognition device includes at least one of the following structures:

[0051] A first module, configured to obtain first dental arch model data and second dental arch model data of a to-be-detected tooth; the first dental arch model and the second dental arch model correspond to different dental arches;

[0052] A second module, configured to obtain relationship identification information of tooth pairs of the to-be-detected tooth according to the first dental arch model data and the second dental arch model data, where the tooth pairs include at least one pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs;

[0053] A third module, configured to determine dental arch information corresponding to the first dental arch model data and the second dental arch model data according to the relationship identification information.

[0054] The use of the dental arch recognition device may also be configured based on the dental arch recognition method in any of the technical solutions provided later. Specifically, based on the association relationship between steps, relevant steps may be implemented in the same or different modules.

[0055] The above device or its modules and units may specifically be implemented by a computer chip or an entity, or be implemented by a product with corresponding functions. When describing the above device, although it is divided into multiple modules for description, in other embodiments, the functions of the modules may also be implemented in the same or multiple software or hardware.

[0056] An embodiment of the present invention provides a storage medium, which may specifically be a computer-readable storage medium. The storage medium may be disposed in a computer and store an application program. At this time, the storage medium may be any available medium that the computer can access data, or may be a storage device such as a server or a data center that includes one or more integrated available media. The available medium may be a magnetic medium such as a floppy disk, a hard disk, or a magnetic tape, or an optical medium such as a DVD (Digital Video Disc), or a semiconductor medium such as an SSD (Solid State Disk).

[0057] When the application program is executed, steps of a dental arch recognition method are implemented. In one embodiment, the dental arch recognition method may include at least one of the following steps:

[0058] Obtain the first dental arch model data and the second dental arch model data of the teeth to be measured respectively, where the first dental arch model and the second dental arch model correspond to different dental arches;

[0059] According to the first dental arch model data and the second dental arch model data, obtain the relationship identification information of the tooth pairs of the teeth to be measured. The tooth pairs include at least one pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs;

[0060] According to the relationship identification information, determine the dental arch information corresponding to the first dental arch model data and the second dental arch model data.

[0061] The stored content of the computer storage medium can also be configured based on the dental arch recognition method in any of the technical solutions provided later.

[0062] An embodiment of the present invention provides an oral appliance.

[0063] The installation environment of the oral appliance can be the actual oral cavity inside the human body, or a simulated oral solid model or a three-dimensional oral model. The oral appliance can be an actual oral appliance, or a three-dimensional model or a solid model of an oral appliance.

[0064] In terms of function and use, the oral appliance can be a tooth orthodontic appliance or a retainer, or an oral appliance for training other intraoral tissues such as orofacial muscle function.

[0065] The oral appliance is prepared based on the dental arch information determined by a dental arch recognition method. The dental arch information can be used to indicate the positional relationship of the dental arch models, for example, to determine the model corresponding to the upper jaw and / or the model corresponding to the lower jaw among several dental arch models. When preparing the oral appliance, an adjustment plan can be determined based on the positional relationship of the dental arch models. The dental arch information can also be used to indicate the characteristics of teeth, gums and other oral tissues.

[0066] In one embodiment, the dental arch recognition method includes at least one of the following steps:

[0067] Obtain the first dental arch model data and the second dental arch model data of the teeth to be measured respectively, where the first dental arch model and the second dental arch model correspond to different dental arches;

[0068] According to the first dental arch model data and the second dental arch model data, obtain the relationship identification information of the tooth pairs of the teeth to be measured. The tooth pairs include at least one pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs;

[0069] Determine the dental arch information corresponding to the first dental arch model data and the second dental arch model data according to the relationship identification information.

[0070] The dental arch recognition method can also be configured based on any of the technical solutions provided later.

[0071] An embodiment of the present invention provides a display method for outputting and displaying to a display unit of at least one device body. The device body can be an electronic device such as a computer or a mobile phone, or can also be a virtual reality (VR) device, an augmented reality (AR) device, a mixed reality (MR) device, etc.

[0072] In one embodiment, the display method includes the steps of: displaying a first dental arch model and a second dental arch model according to dental arch information, where the dental arch information is obtained based on a dental arch recognition method; the first dental arch model and the second dental arch model are used for designing an orthodontic treatment plan.

[0073] The dental arch information can be used to indicate the positional relationship of dental arch models, for example, to determine the model corresponding to the upper jaw and / or the model corresponding to the lower jaw among several dental arch models. When preparing oral appliances, an adjustment plan can be determined based on the positional relationship of the dental arch models. The dental arch information can also be used to indicate the characteristics of teeth, gums, and other oral tissues.

[0074] When the dental arch information can indicate the positional relationship of dental arch models, the first dental arch model and the second dental arch model can be output and displayed at corresponding positions according to the positional relationship. For example, when the first dental arch model is an upper jaw model, the first dental arch model can be output and displayed at a relatively upper position of the display interface; when the second dental arch model is a lower jaw model, the second dental arch model can be output and displayed at a relatively lower position of the display interface.

[0075] In one embodiment, an operator can design an orthodontic treatment plan based on the first dental arch model and the second dental arch model by means of successive manual adjustment or one-time manual adjustment. In another embodiment, other orthodontic treatment plan design methods can be automatically implemented to realize plan design based on the first dental arch model and the second dental arch model.

[0076] In one embodiment, the dental arch recognition method can include at least one of the following steps:

[0077] Obtain the first dental arch model data and the second dental arch model data of the teeth to be measured respectively, where the first dental arch model and the second dental arch model correspond to different dental arches;

[0078] Based on the first dental arch model data and the second dental arch model data, relationship identification information of tooth pairs of the teeth to be measured is obtained. The tooth pairs include at least a pair of teeth located in different dental arches and corresponding to each other. The relationship identification information is used to indicate the relative positional relationship of the tooth pairs;

[0079] Based on the relationship identification information, the dental arch information corresponding to the first dental arch model data and the second dental arch model data is determined.

[0080] The dental arch recognition method can also be configured based on any of the technical solutions provided later.

[0081] The present invention provides a dental arch relationship recognition model. The dental arch relationship recognition model is a neural network model; the dental arch relationship recognition model is trained based on preset tooth pair data and corresponding relationship labels; the relationship labels represent the relative positional relationship of tooth pairs in the preset tooth pair data.

[0082] The preset tooth pair data can be data of standard tooth pairs determined based on rules such as established tooth position notations in a standard dental arch model; it can be data of actual tooth pairs determined after adaptive adjustment in the actual dental arch model of a patient or other user; it can also be data of specific tooth pairs determined by an operator in any dental arch model according to requirements.

[0083] In one implementation, the dental arch relationship recognition model is used to determine the relative positional relationship between dental arches, and uses the corresponding recognition information as the dental arch information. The above dental arch information can also be of other types.

[0084] The dental arch relationship recognition model can be carried in any of the above storage media, memories, modules of devices, or carriers of methods. In one implementation, the dental arch relationship recognition model can be implemented as Figure 3 the structure shown, but any deletion, addition, or other improvement that does not deviate from the inventive concept of the following structure can be regarded as its derivative embodiment.

[0085] The dental arch relationship recognition model includes a first feature extraction module 11. The first feature extraction module 11 is used to extract features from the teeth data to be measured to obtain first tooth data.

[0086] The first feature extraction module 11 can be constructed as a FeatureMap (feature map) to have stronger spatial point cloud analysis capabilities. Specifically, the first feature extraction module 11 at least includes a shared multi-layer perceptron (Shared_MLP) in the FeatureMap.

[0087] The first dental data points to information data of teeth at at least one dental position; the teeth may include only the dental crown, or may include both the dental crown and the gingiva, or may include both the dental crown and the tooth root.

[0088] The dental data to be measured may include spatial point clouds, planar discrete points, grayscale value distributions, or pseudo-color distributions.

[0089] The first feature extraction module 11 may also be used to obtain first feature data corresponding to the first dental data. The first feature data may be obtained by processing the first dental data, and the means of feature extraction may be at least one of SIFT (Scale-Invariant Feature Transform), HOG (Histogram of Oriented Gradient), traditional CNN (Convolutional Neural Networks), SURF (Speeded Up Robust Features), ORB (Oriented FAST and Rotated BRIEF), and LBP (Local Binary Patterns).

[0090] The first feature extraction module 11 may be constructed as a FeatureMap. Specifically, the first feature extraction module 11 at least includes the Max Pool layer in the FeatureMap.

[0091] The dental relationship recognition model includes a second feature extraction module 12. The second feature extraction module 12 is used to extract features from the dental data to be measured to obtain second dental data.

[0092] In terms of definition and configuration, the second dental data and the second feature data may be the same as or approximate to the first dental data and the first feature data respectively; the architecture, algorithm, or included operators of the second feature extraction module 12 may also be the same as or approximate to those of the first feature extraction module 11.

[0093] The teeth pointed to by the first dental data and the teeth pointed to by the second dental data may have an occlusion relationship. There is an occlusion relationship between the first dental data and the second dental data. In other words, the second feature extraction module 12 may be used to receive second dental data having an occlusion relationship with the first dental data; or, the first feature extraction module 11 may be used to receive first dental data having an occlusion relationship with the second dental data.

[0094] In some embodiments, the two tooth data may point to two teeth located in the upper and lower jaws respectively but without an occlusal relationship with each other, or may point to two teeth located in different upper jaws or different lower jaws. For the former, the positional relationship between the corresponding upper and lower jaws can still be determined based on the relationship identification information corresponding to the tooth data; for the latter, the jaw information corresponding to the first jaw model data and the second jaw model data can also be determined based on the corresponding relationship identification information.

[0095] When the jaw information is used to indicate the positional relationship, it may be that the two jaws have an occlusal relationship and one of them is the upper jaw and the other is the lower jaw, it may be that the two jaws are the upper and lower jaws respectively but without an occlusal relationship, it may be that the two jaws are two groups of the same or different upper jaws, or it may be that the two jaws are two groups of the same or different lower jaws.

[0096] The jaw relationship recognition model includes a splicing layer 21. The splicing layer 21 is used to combine the first feature data and the second feature data to generate tooth pair feature data.

[0097] The combination can be a simple splicing. For example, the first feature data in matrix form is arranged on the left, and the second feature data in matrix form is arranged on the right; or the tooth pair feature data can be constructed through weighted fusion or parametric operation. The tooth pair feature data contains the information of both the first tooth data and the second tooth data.

[0098] The jaw relationship recognition model includes a second fully connected layer 22. The second fully connected layer 22 is used to determine the relative positional relationship between the first tooth and the second tooth based on the tooth pair feature data and generate relationship identification information. In a specific embodiment, the relationship identification information may represent at least one of the following information:

[0099] (1) The first tooth is relatively above and the second tooth is relatively below;

[0100] (2) The first tooth is relatively below and the second tooth is relatively above;

[0101] (3) Both the first tooth and the second tooth are relatively above;

[0102] (4) Both the first tooth and the second tooth are relatively below;

[0103] (5) There is an occlusal relationship between the first tooth and the second tooth;

[0104] (6) There is no occlusal relationship between the first tooth and the second tooth.

[0105] Understandably, when applying the FDI tooth position representation method, teeth with an occlusal relationship have the same second "position bit", and the first "limiting bits" correspond to each other (2 corresponds to 3, 1 corresponds to 4). The non-occlusal relationship can include any of the following situations: located in the same dental arch but without an occlusal relationship (such as tooth 11 and tooth 31, or tooth 11 and tooth 12), located in different dental arches (such as tooth 11 in one dental arch and tooth 41 in another dental arch, or tooth 11 in one dental arch and tooth 12 in another dental arch).

[0106] And the occlusal relationship can be as Figure 4 shown. For example, when the first tooth is tooth t12 in the 12th position in the first quadrant and the second tooth is tooth t42 in the 42nd position in the fourth quadrant, the corresponding tooth data of the two form tooth pair data, and at this time, the tooth pair data points to tooth pair Pair12-42. Similarly, the first quadrant can also include Figure 4 at least part of teeth t11 to t18 among the 11th to 18th teeth; the second quadrant can also include Figure 4 at least part of teeth t21 to t28 among the 21st to 28th teeth; the third quadrant corresponding to the second quadrant can also include Figure 4 at least part of teeth t31 to t38 among the 31st to 38th teeth; the fourth quadrant corresponding to the first quadrant can also include Figure 4 at least part of teeth t41 to t48 among the 41st to 48th teeth.

[0107] The relationship identification information is used to determine the relative position relationship between the first dental arch where the first tooth is located and the second dental arch where the second tooth is located. When using the relationship identification information, or when performing a dental arch recognition method using the relationship identification information, at least one of the following information can be generated:

[0108] (1) The first dental arch is the upper jaw and the second dental arch is the lower jaw;

[0109] (2) The first dental arch is the lower jaw and the second dental arch is the upper jaw;

[0110] (3) Both the first dental arch and the second dental arch are the upper jaw;

[0111] (4) Both the first dental arch and the second dental arch are the lower jaw;

[0112] (5) The first dental arch and the second dental arch can be combined to form a complete dental arch;

[0113] (6) The first dental arch and the second dental arch cannot be combined to form a complete dental arch.

[0114] The dental-jaw relationship recognition model includes at least one of a first feature extraction module 11, a second feature extraction module 12, a splicing layer 21, and a quadratic fully-connected layer 22. In a relatively optimal implementation manner, the dental-jaw relationship recognition model includes all of the above structures. In a more optimal implementation manner, based on the above structure, the dental-jaw relationship recognition model constructs a PointNet architecture to better achieve the corresponding technical effects.

[0115] In the above relatively optimal implementation manner, the first feature extraction module 11 and the second feature extraction module 12 can be arranged in parallel. The two respectively send the outputs to the splicing layer 21, and the splicing layer 21 further sends the processed data to the quadratic fully-connected layer 22. In this way, a final prediction output or a final inference output such as relationship identification information is obtained from one side of the quadratic fully-connected layer 22. The second feature extraction module 12 has the same configuration as the first feature extraction module 11. In the above more optimal implementation manner, the parallel first feature extraction module 11 and second feature extraction module 12 constitute a FeatureMap as the front-end network structure of the PointNet architecture. Compared with the traditional PointNet architecture, it can adapt to the paired input of tooth data and perform comprehensive processing to generate an output that can reflect the relationship between two teeth.

[0116] As Figure 5 shown, the quadratic fully-connected layer 22 can include at least one group of multi-layer perceptrons (MLP, Multi-Layer Perceptron). On the one hand, the quadratic fully-connected layer 22 can include several multi-layer perceptrons with layer sizes of 512, 256, and 2 respectively. The three layers of different sizes can be arranged in sequence, which can handle non-linear problems by introducing hidden layers, and can conveniently perform feature extraction, transformation, and information recombination, and is suitable for various classification problems. On the other hand, the multi-layer perceptrons of the quadratic fully-connected layer 22 can be used to form a Shared_MLP (shared multi-layer perceptron), thereby reducing the training parameters of the network and realizing a weight sharing mechanism similar to that of a convolutional neural network. By configuring perceptrons with sizes of 512, 256, and 2, it is possible to gradually extract features from larger input data in sequence, which can increase the operation accuracy.

[0117] In one implementation manner, the first feature extraction module 11 can include two parts for outputting features with different data volumes. For example, the first feature extraction module 11 includes a first fully-connected layer 111 and / or a max pooling layer 112. In one implementation manner, the first feature extraction module 11 simultaneously includes a first fully-connected layer 111 and a max pooling layer 112, and the two are arranged in sequence.

[0118] The first fully-connected layer 111 may include at least one set of multi-layer perceptrons. In one implementation, the first fully-connected layer 111 includes a number of multi-layer perceptrons with layer sizes of 6, 64, 128, and 1024 respectively. When the first input node Input point1 with a size of n×6 forms the input of the first feature extraction module 11, based on the input with a size of 6, the input with a size of 1024 can be gradually constructed to improve the dimension and accuracy of feature processing.

[0119] A number of multi-layer perceptrons in the first fully-connected layer 111 can jointly form a Shared_MLP (shared multi-layer perceptron). After the size of the first input node Input point1 is adjusted, the structure of the first fully-connected layer 111 can also be adjusted correspondingly.

[0120] When the second feature extraction module 12 has the same structure as the first feature extraction module 11, the second feature extraction module 12 can, based on the second input node Input point2 with a size of n×6, construct an output with a size of 1024 through Shared_MLP(6, 64, 128, 1024) and the max pooling layer Max Pool. The output data corresponding to the two tooth data or tooth feature data are concatenated and will form intermediate data with a size of 2048 at the concatenation layer 21, and then are sequentially reduced to 512, 256, and 2 through the second fully-connected layer, and finally an output of relationship identification information with a size of 2 is formed.

[0121] In one implementation, the 6 in the size n×6 of the two input nodes can be 6 coordinate values corresponding to n points in the tooth data to be measured. Among them, three coordinate values are the x coordinate, y coordinate, and z coordinate of the vertex; the other three coordinate values are the coordinate values of the normal vector normal of the plane (which can be a patch) corresponding to the vertex.

[0122] In this way, since the feature extraction, concatenation, and classification layers are respectively configured, and multi-layer perceptrons are used for size expansion or reduction at the feature extraction layer and the classification layer, the model can better adapt to the input data, ensuring the stability of the operation, and this stability is more significant in the scenario of point cloud processing. At the same time, due to the parallel setting of the two feature maps, it can handle the special scenario of tooth pair processing and improve the overall processing efficiency.

[0123] As Figure 6 shown, an embodiment of the present invention provides a dental arch recognition method. The application program or instruction corresponding to this method can be carried on the above-mentioned electronic device, dental arch recognition device, and / or storage medium, or can be carried on the carrier of the display method, or carried on the carrier of implementing oral instrument preparation to achieve the corresponding technical effects. The dental arch recognition method can specifically include the following steps.

[0124] Step S1: Obtain the first dental arch model data and the second dental arch model data of the tooth to be measured respectively. The first dental arch model and the second dental arch model correspond to different dental arches.

[0125] The first dental arch model data and the second dental arch model data can point to the upper and lower jaws with an occlusal relationship in the same maxillofacial region of the same object (patient), or can point to different maxillofacial regions of different objects, or point to two upper jaws or two lower jaws.

[0126] The "obtaining" can represent any means that enables the first dental arch model data and the second dental arch model data to exist at the carrier where the dental arch recognition method is executed. For example, the above-mentioned dental arch model data can be obtained or received from the outside, or the above-mentioned dental arch model data can be called from the internal storage; it can also include obtaining dental arch data after certain data processing. The "obtaining" recorded in this article can be interpreted as above, and will not be elaborated below.

[0127] Step S2: Obtain the relationship identification information of the tooth pairs of the tooth to be measured according to the first dental arch model data and the second dental arch model data.

[0128] The tooth pair includes at least a pair of teeth located in different dental arches and corresponding to each other.

[0129] For example, the first dental arch model data includes the first tooth data, and the second dental arch model data includes the second tooth data. The first tooth data and the second tooth data correspond to each other, and the first tooth and the second tooth pointed to by them are used to form the tooth pair, and the first tooth data and the second tooth data can be the tooth pair data.

[0130] The first tooth data and the second tooth data can be specifically interpreted as follows: the tooth position represented by the first tooth data or the first tooth is located in the first dental arch pointed to by the first dental arch model data, and the tooth position represented by the second tooth data or the second tooth is located in the second dental arch pointed to by the second dental arch model data.

[0131] When implementing this specific example, a corresponding relationship can be set between the first tooth data and the second tooth data, that is, a occlusal relationship is set between the first tooth pointed to by the first tooth data and the second tooth pointed to by the second tooth data. In this way, the error caused by selecting tooth data located in the same upper jaw or the same lower jaw as the input can be naturally avoided. If "both are located in the upper jaw" or "both are located in the lower jaw" is the desired output of the dental arch recognition method, then this occlusal relationship does not need to be set, and there is no so-called error.

[0132] Only one group of tooth pairs can be provided, corresponding to one set of tooth pair data. At this time, the first tooth data and the second tooth data corresponding to a single first tooth and a single second tooth constitute one set of tooth pair data.

[0133] Multiple pairs of teeth can be provided, corresponding to multiple sets of tooth pair data. At this time, several sets of corresponding first teeth and second teeth, corresponding to multiple sets of first tooth data and second tooth data, form multiple sets of tooth pair data.

[0134] The relationship identification information is used to indicate the relative position relationship of the tooth pairs.

[0135] The relationship identification information can be symbols, labels, data, information, etc. that characterize the relative position relationship between the first tooth and the second tooth.

[0136] The relationship identification information can be derived from manual marking or from automated processing of algorithms or program instructions. For the latter, it can specifically be derived from performing point cloud semantic segmentation or from the classification algorithm of a neural network model.

[0137] In the first embodiment provided by the present invention, as Figure 7 shown, the dental arch recognition method further includes the following steps.

[0138] Step S21, based on the first dental arch model data and the second dental arch model data, obtain tooth pair data of at least one pair of teeth located in different dental arches and corresponding to each other.

[0139] The way to determine the tooth pair data can be realized through a preset feature recognition algorithm, especially a point cloud feature extraction algorithm; it can also be obtained by predicting through a trained neural network model. The prediction target can mainly be the crown point cloud data set, which is used as the tooth pair data.

[0140] In the third embodiment provided by the present invention, as Figure 9 shown, the dental arch recognition method further includes the following steps.

[0141] Step S211, segment the first dental arch model and the second dental arch model to obtain a number of first dental arch tooth data and a number of second dental arch tooth data.

[0142] The segmentation can be based on the first dental arch model data and the second dental arch model data to segment several teeth therein, obtaining first dental arch tooth data containing tooth data (first tooth data) corresponding to at least one tooth, and second dental arch tooth data containing tooth data (second tooth data) corresponding to at least one tooth.

[0143] The segmentation is not limited to only including the process of segmenting the dental arch model to obtain teeth, but can also include steps of data processing on the dental arch model data. The data processing of the dental arch model is used to improve the accuracy and efficiency of the segmentation process.

[0144] In an embodiment provided by the present invention, the dental arch recognition method further includes the following steps.

[0145] Step S2111: Perform surface simplification on the first dental arch model and the second dental arch model according to the distinguishability of parts in the dental arch model.

[0146] The distinguishability of the part includes the distinguishability between this part and other intraoral tissues, and / or the distinguishability between adjacent parts.

[0147] The surface simplification can be quadratic surface simplification. The surface simplification is used to reduce the overall data volume of the dental arch model data so as to improve efficiency. During the process of surface simplification, different weights can be set for different regions in the dental arch model data to prevent the loss of important features such as the gingival line and crown information, and prevent key positions from being simplified.

[0148] In an embodiment provided by the present invention, the surface simplification step may specifically include the following steps.

[0149] Step P1: Determine the patch segmentation method of the dental arch model according to the distinguishability of adjacent patches in the dental arch model, and obtain the patches of the dental arch model accordingly.

[0150] The determination of the patch segmentation method may specifically be the segmentation method of simplified patches, which is used to reduce the number of patches obtained by segmentation.

[0151] In an embodiment provided by the present invention, the surface simplification step included in the dental arch recognition method may specifically include the step: determining the patch segmentation method of the dental arch model according to the distinguishability between the patch and other intraoral tissues, and the cost of adjacent patches belonging to the same intraoral tissue.

[0152] The distinguishability between the patch and other intraoral tissues is used to distinguish tooth patches from other intraoral tissue patches.

[0153] The cost of adjacent patches belonging to the same intraoral tissue is used to measure whether the current operation of making adjacent patches belong to the same intraoral tissue will cause patches originally pointing to different intraoral tissues to be wrongly classified as the same intraoral tissue.

[0154] The distinguishability between the patch and other intraoral tissues is determined according to at least one of the following indicators:

[0155] (1) The degree to which the current patch is close to the incisal edge connection line;

[0156] (2) The degree to which the current patch is close to the incisal extreme point of the corresponding tooth;

[0157] (3) The degree to which the current patch is far from the model center of the dental arch model data.

[0158] An embodiment of the present invention provides a solution corresponding to the above-mentioned surface simplification process. Specifically, it includes the following steps.

[0159] Step P11, calculate the Q matrix of the vertices in the dental arch model data.

[0160] The Q matrix can be interpreted as a Quadric matrix, that is, a quadratic surface matrix.

[0161] The technical solution for calculating the Q matrix can be selected according to the needs of those skilled in the art. Here, a technical solution with relatively high efficiency is provided for reference. In this technical solution, step P11 may include the following steps.

[0162] Step P111, determine the vertex and the reference plane where the vertex is located.

[0163] Step P112, determine the plane equation of the reference plane and calculate the quadratic error surface matrix of the vertex.

[0164] Step P113, iteratively obtain the quadratic error surface matrices belonging to the current plane set in the dental arch model data, sum up the quadratic error surface matrices, and obtain the Q matrix corresponding to this vertex.

[0165] The reference plane is a plane jointly formed by the vertex and a preset number of vertices located around it.

[0166] Specifically, the plane equation of the reference plane may have the following form:

[0167] ax + by + cz + d = 0; where a 2 + b 2 + c 2 = 1.

[0168] Then, the plane equation has a coefficient matrix in the form of p = [a b c d] T Based on this, the quadratic error surface matrix corresponding to this vertex and this reference plane may have the following form:

[0169]

[0170] Based on this, the Q matrix corresponding to this quadratic error surface matrix K p may have the following form:

[0171] Q = ∑ p∈planes(v) K p .

[0172] planes(v) represents the current plane set corresponding to this vertex.

[0173] During the execution of the surface simplification process, before calculating the Q matrix, it is also possible to perform a simplification operation on the patches of the dental arch model data, focusing on reducing the data volume in some areas, and improving the operation efficiency without losing important features. That is to say, step P11 can specifically include: determining the patch segmentation method of the dental arch model, and calculating the Q matrix for the vertices in the simplified dental arch model data.

[0174] In one implementation, the patch segmentation strategy can be constrained by the overall energy function, so as to achieve the effect of simplifying the patch segmentation strategy, reducing the number of patches, and even minimizing the number of patches.

[0175] Based on this, the dental arch recognition method can specifically include the following steps.

[0176] Step P1101, fitting the overall energy function of the patches in the dental arch model data.

[0177] Step P1102, minimizing the overall energy function to obtain the simplified dental arch data to be measured.

[0178] The overall energy function is used to evaluate whether the current simplification degree reaches the optimum; the overall energy function includes a patch probability function corresponding to the patch and an adjacent discrimination function corresponding to two adjacent patches. Further, the patch probability function is used to distinguish the oral tissue area, and the adjacent discrimination function represents the cost of adjacent patches belonging to different oral tissue areas.

[0179] In this way, it is possible to evaluate the patch simplification degree from two dimensions: the nature of the oral tissue area itself, especially the importance of the area, and the ability of adjacent patches to be merged within the allowable range, providing the possibility of customization while maintaining the optimal simplification strength.

[0180] The overall energy function can be determined by weighted operation.

[0181] In one implementation, the overall energy function is equal to the weighted sum of the patch probability function and the adjacent discrimination function. For example, define the overall energy function as E(x), the set of patches in the dental arch model data as δ, the patch probability function of the i-th patch as E unary (x i ), the set of adjacent patches in the dental arch model data as the adjacent discrimination function between the i-th patch and the j-th patch as E pairwise (x i , x j ), then the overall energy function E(x) can be configured to at least satisfy the following formula:

[0182]

[0183] λ is the adjacent discrimination function E pairwise (x i , x j ) is the weight relative to the overall energy function E(x).

[0184] In one embodiment, the patch probability function can be set based on the spatial position of the patch in the dental arch model data. Correspondingly, the step P1101 specifically includes: fitting the patch probability function of the patch in the dental arch model data according to the patch spatial position; determining the overall energy function according to the patch probability function.

[0185] Specifically, patches close to the tooth crown, close to the overall center of the model, or close to teeth No. 11, 21, 31, and 41 can be set with a higher patch probability function, which is equivalent to giving them a higher weight or a lower simplification requirement to prevent feature loss caused by excessive simplification.

[0186] The patch probability function can be determined in a weighted manner according to several probability energies. In one embodiment, the patch probability function can be the sum or weighted sum of several probability energies.

[0187] In this way, by setting the probability energies, the patch probability function can be given the ability to consider various spatial positions of the patches. In a preferred embodiment, the patch probability function can include at least one of the first probability energy, the second probability energy, and the third probability energy.

[0188] The first probability energy characterizes the degree to which the current patch is close to the incisal edge connection line. The incisal edge connection line can be interpreted as the connection line fitted according to the incisal edge extension line of the tooth crown, or as the connection line fitted according to the projection line of the tooth crown section on the plane of the labial surface or the lingual surface. Specifically, in the anterior tooth area, the incisal edge connection line can be interpreted as the smile line, and in the posterior tooth area, it can be interpreted as the Spee curve.

[0189] The second probability energy characterizes the degree to which the current patch is close to the incisal extreme point of its corresponding tooth. The incisal extreme point can be interpreted as the point on the incisal edge or section that is farthest from the root of the tooth position. Taking the upper jaw as an example, if the root is relatively above the tooth crown, the incisal extreme point can be interpreted as the lowest point on the z-axis of the incisal edge or section of a certain tooth position; if the root is relatively below the tooth crown, the incisal extreme point can be interpreted as the highest point on the z-axis of the incisal edge or section of a certain tooth position.

[0190] The third probability energy characterizes the degree to which the current patch is far from the model center of the dental arch model. The definition of the model center depends on the importance of the posterior tooth area for subsequent tasks.

[0191] If the importance of the posterior tooth area is similar to that of the anterior tooth area, the center of the model can be set between the upper and lower jaws of the oral cavity; if the importance of the posterior tooth area is significantly less than that of the anterior tooth area, the center of the model can be set at the intersection of the mesial surfaces of teeth 11, 21, 31, and 41. When the center of the model is set between the upper and lower jaws of the oral cavity, based on the adjustment of the patch weight value by the probability energy, the interference of the lingual gingiva can be removed.

[0192] If the patch probability function simultaneously includes the first probability energy, the second probability energy, and the third probability energy, then the patch probability function can be set to be equal to the weighted sum of the first probability energy, the second probability energy, and the third probability energy.

[0193] Set the z-axis as the coordinate axis extending in the vertical direction in the dental arch model, and the coordinate origin is the center of the model. Then:

[0194] (1) The first probability energy can be configured to set the weights of the patches from large to small as the probability energy in the increasing direction of the absolute value of the z-axis. That is, the probability energy of the patches close to the incisal edge connection line is greater than that of the patches far from the incisal edge connection line.

[0195] (2) The second probability energy can be configured to start from the incisal extreme point of the tooth position, search outward and set the weights of the patches from large to small as the probability energy. That is, the probability energy of the patches close to the incisal extreme point is greater than that of the patches far from the incisal extreme point.

[0196] (3) The third probability energy can be configured to calculate the distance between the patch and the center of the model, and set the weights of the patches from small to large as the probability energy according to the distance from near to far. That is, the probability energy of the patches close to the center of the model is less than that of the patches far from the center of the model.

[0197] Based on this, the sum of the absolute values of the weight values of the first probability energy, the second probability energy, and the third probability energy can be configured to be 1. The weight value of the third probability energy has the opposite sign to the weight values of the first probability energy and the second probability energy.

[0198] The patch probability function can be defined as E unary (or define that for a certain patch, its patch probability function value is E unary ; the same applies to the probability energies E1, E2, E3, etc. below), and define the first probability energy as E1, the weight value of the first probability energy as ε1, the second probability energy as E2, the weight value of the second probability energy as ε2, the third probability energy as E3, and the weight value of the third probability energy as ε3. Then the patch probability function is E unary and the weight values at least satisfy:

[0199]

[0200] It can be understood that the weight values ε1, ε2, and ε3 are the weights between three kinds of probability energies with respect to the patch probability function E unary ; while what the probability energies E1, E2, and E3 set is the weight situation between patches on the dental arch model data, and the two can be distinguished and explained.

[0201] The patch probability function E unary The larger it is, the greater the possibility that the patch belongs to a tooth (especially the crown). Further, at least two classification thresholds can be set, and according to the patch probability function E unary , the patches in the dental arch model to be measured are divided into three simplified regions. Thus, the x provided above i can correspond to indicating the i-th patch, based on the patch probability function E unary the meaning of the type to which it belongs.

[0202] In one implementation manner, the adjacent discrimination function can be set based on the angular distance and / or the capacity function value between adjacent patches. Equivalently, the step P1101 specifically includes: fitting the adjacent discrimination function of adjacent patches in the dental arch model data according to the angular distance and / or the capacity function between adjacent patches; determining the overall energy function according to the adjacent discrimination function.

[0203] In one implementation manner, the capacity function can be calculated based on the angular distance value, and then the adjacent discrimination function can be calculated according to the capacity function. Even the capacity function can be directly used as the adjacent discrimination function. Specifically, for example, if the adjacent discrimination function includes a first discrimination function corresponding to a first patch and a second patch, the dental arch recognition method can specifically include the following steps.

[0204] Step P11011, calculate the first angular distance between the first patch and the second patch according to the normal angle between the first patch and the second patch.

[0205] Step P11012, perform maximum flow minimum cut according to the first angular distance and the average angular distance, calculate the first capacity function of the first patch and the second patch, and use the first capacity function as the first discrimination function.

[0206] The first angular distance is used to evaluate the similarity between the first patch and the second patch. The farther the first angular distance is (the larger the value), the worse the similarity between the two patches. On the contrary, the better the similarity. Based on the similarity, the capacity function can be calculated according to the average angular distance of adjacent patches, so as to evaluate the cost generated after fusing the two patches; the better the similarity, the smaller the fusion cost; the lower the similarity, the greater the fusion cost.

[0207] The fusion can be understood as collapsing two surface patches into one surface patch, or merging two surface patches into one surface patch. The average angular distance can be pre-computed or updated in real time during the iteration process.

[0208] Define the i-th surface patch including the first surface patch as f i , and define the j-th surface patch including the second surface patch as f j , define the normal vector of the i-th surface patch f i and the normal vector of the j-th surface patch f j to have an included angle of α ij , and the angular distance Ang i between the i-th surface patch f j and the j-th surface patch f Dist(aij) at least satisfies the following formula:

[0209] Ang Dist(aij) = η(1 - cosα ij ).

[0210] η is the concavo-convex angle parameter.

[0211] In addition to performing the Max-Flow / Min-Cut, the Ford-Fulkerson (FFA) algorithm, the Goldberg-Tarjan algorithm, etc. can also be alternatively implemented to calculate the capacity function or other data with similar properties.

[0212] Taking the Max-Flow / Min-Cut as an example, define the capacity function corresponding to the i-th surface patch f i and the j-th surface patch f j , and including the first capacity function as Cap(i, j), define the adjustment parameter as w, and define the average angular distance as avg(Ang Dist ), then the capacity function Cap(i, j) at least satisfies the following formula:

[0213]

[0214] S is the set of flow source points, and T is the set of flow sink points.

[0215] If the first capacity function is defined as Cap1 and the first discrimination function is E pairwise 1, then there is: Cap1 = E pairwise 1 ∈ {E pairwise (x1, x2), E pairwise (x2, x3),......, E pairwise (x i , x j )}.

[0216] In the process of implementing surface simplification, especially when classifying the internal space of the oral cavity using the patch probability function in the total energy function, the following steps can also be implemented: If the maximum curvature at a vertex falls within a certain range, then the vertex is determined as a seed point. In this way, the position and shape of the segmentation line can be clarified using the seed points, so that the segmentation line can be distributed as much as possible on the gingival line, and avoid being distributed at the interdental spaces or the alveolar lines of the molars.

[0217] Step P2, merge the valid vertices of the obtained patches.

[0218] In one embodiment, specifically, the valid vertices of the obtained patches can be merged according to the merging error.

[0219] Specifically, the dentognathic recognition method may include the following steps.

[0220] Step P21, determine the merging error of the valid vertex combinations.

[0221] Valid vertices can be interpreted as satisfying the following two conditions:

[0222] (1) There is an edge between two vertices;

[0223] (2) The distance between two vertices is less than a certain threshold; specifically, define v1 and v2 as two vertices, then the second condition is |v1 - v1| < t, where t is the set threshold.

[0224] Step P22, select the valid vertex combination with the minimum merging error and update the merging error of the relevant valid vertices.

[0225] To avoid connecting vertices that are far apart from each other, the threshold needs to be set dynamically following the number of vertex combinations. In this process, the algorithm of iterative collapse itself and the merging error Δv between the target vertex and the original vertex need to be considered. In this way, the best iterative collapse point can be found by minimizing the merging error Δv.

[0226] The merging error Δv can at least satisfy the following formula:

[0227] Δv = v T Q(v)v.

[0228] The iterative collapse process can be expressed as (v1, v2) → v, where the first vertex v1 and the second vertex v2 are the original vertices, and v is the target vertex. Q(v) is the Q matrix of the target vertex v. In the process of finding the target vertex v, the arithmetic mean of the first vertex v1 and the second vertex v2 can be used as the initial target vertex, and then the best target vertex can be found through iteration.

[0229] In one embodiment, the optimal target vertex can be quickly determined through operations on the Q matrix. The process of finding the minimum value is abstracted into a linear problem. That is:

[0230]

[0231] Assuming that this matrix is invertible, then the optimal target vertex can be directly calculated according to the following formula

[0232]

[0233] If this matrix is not invertible, then the optimal target vertex can be found at the line segment, endpoints or midpoint formed by connecting the first vertex v1 and the second vertex v2.

[0234] The process of determining the valid vertex combination with the minimum merging error can also be abstracted into a stack problem. Specifically, the valid vertex combination (or, valid vertex pair) can be placed in a heap with the merging error Δv as the key value of the error metric, so that the pair with the minimum cost is at the top of the heap.

[0235] Remove the valid vertex pair with the minimum merging error Δv from the heap, such as (v1, v2), collapse this valid vertex pair, and update the merging error Δv of the valid vertex pairs related to the first vertex v1 (i.e., the current vertex). In this way, the surface simplification is completed, and vertices and patches that do not need to be distinguished from each other are merged from the perspective of vertices.

[0236] Step S2112, respectively register the first dental arch model and the second dental arch model, and segment the first dental arch model and the second dental arch model based on the tooth segmentation model.

[0237] The registration of the dental arch model can include rough registration and / or fine registration.

[0238] In one embodiment, the dental arch recognition method includes performing rough registration and then fine registration on the first dental arch model in sequence. In one embodiment, the dental arch recognition method includes performing rough registration and then fine registration on the second dental arch model in sequence.

[0239] Model registration is used to improve the overall accuracy of the dental arch model data based on a preset standard model, facilitating subsequent prediction.

[0240] Based on this, the dental arch recognition method may include the following steps.

[0241] Step P31, call the dental arch registration model corresponding to the dental arch model data, and perform point cloud sampling on the dental arch model data and the dental arch registration model respectively.

[0242] Define two groups of dental arch models as Mu and M l For the sake of simplicity in description, they can be collectively referred to as the basic model M*; define the two groups of dental arch registration models corresponding to the dental arch models M u and M l as T u and T l respectively. For the sake of simplicity in description, they can be collectively referred to as the registration model T*. The registration model T* can be obtained by matching from a preset model library.

[0243] Based on this, in the Origin data table corresponding to the basic model M* and the registration model T*, which lists data in the form of coordinates X, Y, Z, 5000 data points are sampled respectively and their normal vectors are calculated. In this way, the step of point cloud sampling is completed.

[0244] Step P32, calculate the Fast Point Feature Histogram (FPFH) features based on the sampled points, and perform the Random Sample Consensus (RANSAC) algorithm on the FPFH features to obtain the roughly registered dental arch data.

[0245] Based on the sampled data points and normal vector coordinates (point cloud data), the FPFH (Fast Point Feature Histogram) features can be calculated, and the RANSAC (Random Sample Consensus) algorithm is implemented for the FPFH features to obtain the rough transformation relationship between the two point clouds, so as to achieve the technical effect of roughly registering the sampled point cloud data. It can be understood that the object of the rough registration is the sampled point cloud data, and the roughness in the registration degree is relative to the overall dental arch model data.

[0246] Based on this, the basic model M* or the dental arch model data can be updated based on the rough registration result, so as to obtain the roughly registered dental arch data.

[0247] Step P33, perform the Iterative Closest Point (ICP) algorithm on the roughly registered dental arch data.

[0248] Based on the roughly registered dental arch data, perform fine registration on its point cloud Origin (X, Y, Z) using the ICP (Iterative Closest Point) algorithm, so as to obtain the registered dental arch model data Matched U and Matched L . In one scenario, Matched U points to the registered maxillary data, and Matched L points to the registered mandibular data.

[0249] For coarse registration, since the FPFH combined with RANSAC method is adopted, it can improve the registration accuracy and reduce the sampling pressure while ensuring the operation efficiency. For the combination of coarse registration and fine registration, since the FPFH combined with ICP method is adopted, it can accelerate the registration speed and reduce the error.

[0250] The details in the coarse registration and fine registration processes are difficult to enumerate in this invention. For example, in the coarse registration process, some embodiments also include steps such as performing filtering and denoising on the point cloud data and searching for key points using the ISS (Intrinsic Shape Signatures) algorithm. RANSAC can also be replaced by implementing the SAC-IA (Sampling Consensus Initial Alignment) algorithm to achieve registration.

[0251] In the embodiment of sequentially performing surface simplification and model registration on the dental model data, the input data for performing model registration can specifically be the output data of surface simplification.

[0252] The tooth segmentation model is a neural network model.

[0253] The tooth segmentation model is trained based on a preset dental model and a preset tooth label representing the tooth distribution in the preset dental model.

[0254] The preset dental model can be a standard dental model; it can be the actual dental model of a patient or other users; it can also be a dental model determined by the operator according to requirements. The preset tooth label can be a tooth label obtained by manually or machine-confirming and marking any of the above dental models or other dental models determined by other methods.

[0255] The dental model is segmented to distinguish the crowns of different tooth positions to assist in the formation of tooth pairs, so as to avoid the display of occlusion relationships and the generation of tooth pairs relying too much on manual annotation.

[0256] The process of model segmentation can be achieved by constructing a neural network model.

[0257] On the one hand, the neural network model mentioned below can be a part of the dental relationship recognition model. On the other hand, the neural network model mentioned below can also exist independently, distinct from the dental relationship recognition model.

[0258] In one embodiment, the dental recognition method specifically includes the following steps.

[0259] Step P4, call the tooth segmentation model to perform segmentation on the dental model data in units of tooth positions to obtain tooth position data.

[0260] In one embodiment, the tooth segmentation model includes a deep learning segmentation module and / or a canary module.

[0261] The deep learning segmentation module is used to generate a fine-grained segmentation output.

[0262] The canary module is used to evaluate whether such a segmentation output meets the requirements (especially clinical requirements).

[0263] During the process of model slicing, the tooth segmentation model converts grid data into point cloud data for slicing, and combines the point cloud data and its geometric information (such as patch normal vectors and patch shapes). In this way, it has a better segmentation effect; based on the re-verification process of the canary module, incorrect outputs can be corrected in a timely manner.

[0264] The deep learning segmentation module (Deep Learning Segmentation Module) may include a data preprocessing module, a deep learning segmentation module, and a boundary smoothing module.

[0265] The canary module (Canary Module) may include a confidence evaluation module and an auto-correction module.

[0266] The deep learning segmentation module takes point clouds as input and output, and uses kNN (K-Nearest Neighbor) to map the segmentation result back to the neural network side where the deep learning segmentation module is installed.

[0267] When implementing its functions, the deep learning segmentation module may perform at least one of the following steps:

[0268] Convert the point cloud space;

[0269] Execute EdgeConv (edge convolution, which can also be DGCNN, Dynamic Graph CNN for Learning on Point Clouds) to calculate vertex edge features, and use a 2D convolutional layer to aggregate the vertex edge features;

[0270] Use a mean-pooling layer and a max-pooling layer to connect the output of EdgeConv to form a global feature descriptor;

[0271] Input the one-hot encoded classification vector into the deep learning segmentation module; stack and aggregate several 2D convolutional layers from the outputs of EdgeConv and the global feature descriptor to generate classification labels corresponding to tooth positions.

[0272] In addition to the tooth segmentation model adopting the above special architecture, in some embodiments, a traditional CNN can also be used.

[0273] In the embodiment of successively performing model registration and model segmentation on the dental arch model data, the input data for performing model segmentation can specifically be the output data of the model registration process.

[0274] Step S212, match the first dental arch tooth data and the second dental arch tooth data according to the tooth position numbers to obtain tooth pair data.

[0275] In one embodiment, the dental arch recognition method may further include a data sampling process.

[0276] Data sampling is used to label and process teeth, and based on the labels and quadrant relationships, tooth pairs are constructed and finally tooth pair data is formed. The present invention does not exclude manual marking and combination of tooth pairs.

[0277] Taking the two sets of dental arch model data including the first dental arch model data and the second dental arch model data as an example, the dental arch recognition method specifically includes the following steps.

[0278] Step P51, determine the tooth position numbers, patch center points and patch normal vectors of the teeth in the dental arch model data, and combine them to form several tooth marking data.

[0279] The tooth position numbers can be in the form of Figure 4 the numbers shown, or can be numbered based on FDI.

[0280] The "determination" may include at least one of the following two aspects.

[0281] The first aspect, mask the 6th tooth, 7th tooth and 8th tooth, that is, mask Figure 4 the 16th tooth t16 to the 18th tooth t18 in Figure 4 the 26th tooth t26 to the 28th tooth t28 in Figure 4 the 36th tooth t36 to the 38th tooth t38 in Figure 4 and the 46th tooth t46 to the 48th tooth t48 in

[0282] In a second aspect, the central points of some patches on the tooth positions and the normal vectors of the patches can be randomly sampled, thereby avoiding resource waste caused by sampling the patches on the tooth positions.

[0283] Define the tooth position number as i, then the tooth marking data can be defined as F i , and the patch normal vector can be defined as Normal i .

[0284] Step P52: Perform mean removal processing on the coordinate part in the tooth marking data and update the tooth marking data.

[0285] When the patch normal vector Normal i is also saved in coordinate form, the "coordinate part in the tooth marking data" includes the patch normal vector Normal i and the central point coordinates of the patch; otherwise, the "coordinate part in the tooth marking data" only includes the central point coordinates of the patch. Corresponding to the tooth position number i, define this coordinate part as Coords i . Then the mean removal processing can be interpreted as:

[0286] Coords i ' = Coords i - mean(Coords i ).

[0287] After the above processing, the differences or fluctuations between the elements in the tooth marking data will be more significant, and the data obtained by calculation will have a higher discrimination degree, which is more conducive to subsequent judgment.

[0288] Step P53: According to the tooth position number, perform tooth position pairing on the tooth marking data corresponding to the first dental arch model data and the tooth marking data corresponding to the second dental arch model data to obtain tooth pair data.

[0289] The output tooth pair data can be presented in the form of a sequence, and the sequence can successively include several groups of tooth pair data (for example, Pair12 - 42), and each group of tooth pair data includes at least two items of tooth data.

[0290] The pairing method is determined based on the requirements of dental arch recognition. The same - named teeth at the upper jaw of different maxillofacial regions (for example, both are the eleventh tooth t11) can be combined to form a tooth pair, and the teeth that are centrosymmetric at the upper and lower jaws of the same maxillofacial region (for example, the eleventh tooth t11 and the forty - first tooth t41) can be combined to form a tooth pair. In this embodiment, taking the teeth with an occlusal relationship with each other as a pair as an example, this pairing method can more accurately distinguish which is the upper jaw and which is the lower jaw in the two dental arches.

[0291] The process of constructing tooth marking data can be completed in a four - quadrant coordinate system to lay a foundation for the tooth pairing process required for dental arch recognition. Thus, in one implementation, the dental arch recognition method may include the following steps.

[0292] Step P511: Construct a dental arch coordinate system, set the first dental arch model data in the first quadrant and the second quadrant, and set the second dental arch model data that forms an occlusion relationship with the first dental arch model data in the third quadrant and the fourth quadrant.

[0293] The process of setting the first dental arch model data and the second dental arch model data can be interpreted as the process of placing the models corresponding to the first dental arch model data and the second dental arch model data in the corresponding quadrants respectively. The first quadrant, the second quadrant, the third quadrant, and the fourth quadrant are the same as or similar to the previous quadrant Ⅰ, quadrant Ⅱ, quadrant Ⅲ, and quadrant Ⅳ. When setting the dental arch model corresponding to the first dental arch model data and the dental arch model corresponding to the second dental arch model data, set them as Figure 4 shown without losing their occlusion relationship.

[0294] Step P512: Determine the tooth position numbers, the center points of the patches, and their patch normal vectors of the first tooth, the second tooth, the third tooth, the fourth tooth, and the fifth tooth located in the four quadrants of the dental arch model data respectively, and combine them to form a number of tooth marking data.

[0295] Taking the first quadrant as an example, sample and screen the 11th tooth t11 to the 15th tooth t15, and after determining their tooth position number and other data, combine them to form tooth marking data F 11 、F 12 、F 13 、F 14 、F 15 . It can be seen that in this implementation, ideally, each quadrant can form five groups of tooth marking data, and a total of 20 groups of tooth marking data in the four quadrants. In the case of special situations such as missing teeth, the number of tooth marking data will decrease.

[0296] Step P53 may include the steps of: performing pairing on the tooth data corresponding to the same - numbered teeth located in the first quadrant and the third quadrant respectively, and performing pairing on the teeth corresponding to the same - numbered teeth located in the second quadrant and the fourth quadrant respectively, to obtain tooth pair data

[0297] "Same - numbered teeth" can be interpreted as teeth with the same second "position bit" based on the FDI tooth position representation method. Thus, the same - numbered teeth corresponding to the first quadrant and the fourth quadrant can form five groups of tooth pair data Pair 1j-4j , and the same - numbered teeth corresponding to the second quadrant and the third quadrant can form five groups of tooth pair data Pair 2j-3j . j = 1, 2, 3, 4, 5.

[0298] The connector defining the pairing operation is Then the above operation can be further expanded as follows:

[0299]

[0300]

[0301] In the implementation manner of sequentially performing model segmentation and data sampling on the dental arch model data, the input data for performing data sampling can specifically be the output data of the model segmentation process.

[0302] Among the above several implementation manners, splitting or combination can be performed. The dental arch recognition method may include the steps of: respectively performing at least one of surface simplification, model registration, model segmentation, and data sampling on two groups of dental arch model data.

[0303] Specifically, any combination can be performed among the above four implementation manners, and specifically, it can be a pairwise combination or a three - way combination. And, on the premise of ensuring normal input and output, their sequence can also be adjusted as needed, which will not be elaborated here.

[0304] In one implementation manner, the dental arch recognition method may include the steps of: sequentially performing surface simplification, model registration, model segmentation, and data sampling on one of the two groups of dental arch model data. In this way, it is possible to sequentially remove the edge regions or other unimportant regions that have little influence on the generation of the relationship identification information; it is possible to register and improve the overall accuracy of the model; it is possible to segment into tooth pairs and sample to form data that can characterize the morphology and other features of the tooth pairs; and it is possible to finally generate tooth pair data sufficient to calculate accurate results.

[0305] In one implementation manner, the dental arch recognition method may include the steps of: respectively and sequentially performing surface simplification, model registration, model segmentation, and data sampling on the two groups of dental arch model data. In this way, the same pre - processing is respectively performed on the two groups of dental arch model data, so that there is good consistency between the tooth data used to form the tooth pair data.

[0306] As Figure 7 shown, in the first embodiment provided by the present invention, the dental arch recognition method includes the following steps.

[0307] Step S22, input the tooth pair data in parallel into the dental arch relationship recognition model, and predict the corresponding relationship identification information.

[0308] The dental relationship recognition model is a neural network model, and the dental relationship recognition model is trained based on preset tooth pair data and relationship labels characterizing the relative positional relationship of tooth pairs in the preset tooth pair data.

[0309] In one case, one of the two groups of dental data to be measured points to the upper jaw, and the other points to the lower jaw.

[0310] For example, the first dental model data points to the upper jaw and the second dental model data points to the lower jaw. Thus, the first tooth pointed to by the first tooth data is located in the upper jaw, and the second tooth pointed to by the second tooth data is located in the lower jaw; the relationship identification information correspondingly characterizes that the tooth pointed to by the first tooth data is relatively above the tooth pointed to by the second tooth data.

[0311] The "parallel input" can be understood as inputting the tooth pair data into the dental relationship recognition model respectively, so as to improve the parallel processing ability of several tooth pair data. The "parallel input" can also be understood as taking different tooth data in the tooth pair data as multiple parallel inputs of a single dental relationship recognition model, so as to improve the processing ability of a single group of tooth pair data, facilitate the implementation of configurations such as mutual supervision and weight sharing using this kind of input, and can also improve the overall processing efficiency and facilitate the prediction and generation of relationship identification information.

[0312] In one implementation manner, the dental relationship recognition model includes at least one of the following structures.

[0313] The first feature extraction module 11 and the second feature extraction module 12 are used to receive the tooth data input in parallel;

[0314] The splicing layer 21 is used to combine the tooth pair feature sequences output by the first feature extraction module 11 and the second feature extraction module 12;

[0315] The secondary fully connected layer 22 is used to predict the relationship identification information according to the tooth pair feature sequence.

[0316] The dental relationship recognition model may include all of the above structures. For the specific solutions of the above structures, reference can be made to the technical solutions provided above, especially it can be configured as Figure 5 shown, which will not be elaborated here. It should be noted that when a primary fully connected layer 111 is included in the first feature extraction module 11 or the second feature extraction module 12, the "primary" and the "secondary" only have the connotation of the order, and do not necessarily refer to the number of layers of the fully connected layer; when the primary fully connected layer 111 is not included, the secondary fully connected layer 22 can be expressed as the "fully connected layer".

[0317] In one implementation manner, the dental recognition method includes the following steps.

[0318] Step S221: According to the first feature extraction module, perform feature extraction on the to-be-tested tooth data to obtain the first tooth data. According to the second feature extraction module, perform feature extraction on the to-be-tested tooth data to obtain the second tooth data.

[0319] Step S222: Obtain the first feature data and the second feature data corresponding to the first tooth data and the second tooth data respectively, and generate a tooth pair feature sequence.

[0320] Step S223: According to the tooth pair feature sequence, predict the relationship identification information of the first tooth and the second tooth.

[0321] In this way, by constructing a dental relationship recognition model with parallel input, not only can a larger amount of data be received and processed at one time, but also feature extraction can be performed in parallel to facilitate the generation of relationship identification information by fusion and mutual comparison.

[0322] The dental relationship recognition model can be constructed based on various architectures. In one implementation, the PointNet architecture is adopted to better adapt to the feature extraction scenario of point cloud global features. The PointNet architecture mentioned in this article not only includes the traditional PointNet, but also can include PointNet++ or F-PointNet. Based on this, the dental relationship recognition method includes the steps of: constructing a neural network model based on the PointNet architecture.

[0323] The dental relationship recognition model is based on the neural network model architecture, and is trained with the preset tooth pair data and its preset relationship labels as the training set and the test set, or as the training set, the validation set and the test set. In one embodiment, the dental relationship recognition method further includes the steps of: configuring the optimizer as Adam and the learning rate as 0.001. In this way, the training speed can be stabilized during the iteration process, and the stability and accuracy of the model parameters can be maintained.

[0324] In one embodiment, the dental relationship recognition method can specifically include configuring the loss function as the cross-entropy loss function.

[0325] Before calculating the cross-entropy loss function, label smoothing can also be used to improve the quality of the operation results.

[0326] The number of samples per batch (BatchSize) during training can be configured as 32.

[0327] In one embodiment, the dental relationship recognition method further includes the steps of: according to the preset relationship labels and the preset tooth pair data, configuring the optimizer as Adam, the learning rate as 0.001, and the loss function as the cross-entropy loss function, and performing training on the neural network model.

[0328] The preset relationship label corresponds to the relationship identification information, characterizes the relative position relationship between a number of teeth, and may indicate "upper and lower", "lower and upper", "same above", "same below", "with occlusion", "without occlusion" and the like.

[0329] like Figure 6 As shown, the tooth and jaw identification method provided by the present invention includes the following steps.

[0330] Step S3: determining the dental information corresponding to the first dental model data and the second dental model data according to the relationship identification information.

[0331] In this way, by applying the tooth-jaw recognition method provided by the present invention, the tooth-jaw information corresponding to the teeth can be determined based on the positional relationship between two teeth in the tooth pair determined by the relationship identification information.

[0332] The dental information includes maxillary determination information or mandibular determination information of the first dental model, and maxillary determination information or mandibular determination information of the second dental model.

[0333] In this way, the upper and lower jaws can be identified and distinguished quickly and accurately.

[0334] It can be understood that the technical solution provided by the present invention can also achieve the technical effect of separately identifying and judging the upper jaw, separately identifying and judging the lower jaw, and separately identifying and judging whether the upper and lower jaws belong to the same jaw face. When the first tooth data and the second tooth data point to the data information of the same tooth in different correction states, and the relationship identification information at least records the order of the first tooth data and the second tooth data, the technical solution provided by the present invention can also achieve the effect of determining the order of the dental model data and assisting in the construction of the orthodontic appliance.

[0335] In one embodiment, the tooth and jaw recognition method includes the steps of: obtaining the numerical form of the relationship identification information, and determining the upper and lower jaw determination information of the first tooth and jaw model data and the second tooth and jaw model data according to the numerical relationship.

[0336] The numerical relationship may be the relationship between the numerical relationship identification information and the set threshold, or may be the discreteness between a plurality of numerical relationship identification information. The step of eliminating abnormal values ​​may also be included in the numerical relationship determination.

[0337] The present invention does not limit the relationship identification information to be in the form of a numerical value, such as "yes", "no", etc., as labels that can be converted into numerical values ​​based on preset rules can also be used as the content of the relationship identification information.

[0338] For example, the first tooth being on top and the second tooth being at the bottom can be set as the first type of label, the first tooth being at the bottom and the second tooth being on top as the second type of label, both teeth being on top as the third type of label, and both teeth being at the bottom as the fourth type of label. Compared with setting the first tooth being on top as label A and at the bottom as label B, and the second tooth being on top as label C and at the bottom as label D, with the combinations of the four positional relationships being label AC, label AD, label BC, and label BD, a total of eight labels, it is possible to reduce the number of labels involved in the operation and speed up the training and prediction speed.

[0339] In the second embodiment provided by the present invention, as Figure 8 shown, the dental arch recognition method includes the following steps.

[0340] Step S3A, when in the tooth pair data, the tooth to be measured in the first dental arch model points to the upper teeth and the tooth to be measured in the second dental arch model data points to the lower teeth, obtain the relationship identification information in the form of a first label value.

[0341] Step S3B, when in the tooth pair data, the tooth to be measured in the first dental arch model points to the lower teeth and the tooth to be measured in the second dental arch model data points to the upper teeth, obtain the relationship identification information in the form of a second label value.

[0342] The tooth data pointing to the upper jaw and the tooth data pointing to the lower jaw can be specifically explained as: the tooth position or tooth represented by the tooth data is located in the upper jaw, or the tooth position or tooth represented by the tooth data is located in the lower jaw.

[0343] Steps S3A and S3B list the acquisition situations of two types of relationship identification information, indicating that relationship identification information with different label values is provided in different situations. It should be noted that there is no necessary sequential relationship or dependency relationship between the two steps.

[0344] In one implementation manner, the dental arch recognition method specifically includes the steps of: based on the label values in numerical form corresponding to multiple groups of tooth pairs, determine the upper and lower jaw determination information according to the voting scoring result.

[0345] In one case, different tooth pairs may provide different relationship identification information. When the relationship identification information is used to indicate the relative positional relationship of the tooth pair, or when the dental arch information includes the upper and lower jaw determination information, it may lead to differences in the upper and lower jaw judgments of the dental arch model. At this time, the upper and lower jaw determination information can be determined based on the voting scoring result of "the minority obeys the majority".

[0346] The upper and lower jaw determination information may include the upper jaw determination information or the lower jaw determination information for the first dental arch model, or the upper jaw determination information or the lower jaw determination information for the second dental arch model.

[0347] In one embodiment, the method for identifying dental arches specifically includes the following steps.

[0348] Step S31: Calculate the mean value of the tag values in numerical form corresponding to multiple sets of tooth pair data.

[0349] Step S32: Determine the upper and lower jaw determination information of two sets of dental arch model data according to the numerical relationship between the mean value of the tag values and the prediction threshold.

[0350] The data sample basis for identifying the upper and lower jaw determination information of the dental arch model is expanded to the relationship identification information corresponding to multiple sets of tooth pairs. By calculating the mean value of the numerical tag values for judgment, the overall stability can be increased, and individual errors can be prevented from directly and overly affecting the judgment result.

[0351] The mean value does not necessarily refer to the arithmetic mean. Different weights can also be assigned to tooth pairs in different regions, so as to calculate the weighted mean value. For example, a larger weight can be set for the anterior tooth region with clearer occlusion relationship and more regular arrangement, and a smaller weight can be set for the posterior tooth region.

[0352] In order to compress the data volume to a greater extent, compared with setting tags for each tooth position separately, the present invention sets tags for each combination of positional relationships of tooth pairs as a unit, achieving the effect of simplifying the operation.

[0353] The above embodiments can be combined with the aforementioned steps S3A and S3B. In this way, by setting different tag values for different combinations of tooth pair data pointing to the dental arch, the relative positional relationship between two sets of dental arch model data can be comprehensively and quickly judged.

[0354] It can be to compare the numerical magnitudes between the mean value of the tag values and the prediction threshold. Specifically, it can be determined that the first dental arch model data points to the upper jaw when it is greater, or it can be determined that the first dental arch model data points to the upper jaw when it is less.

[0355] Based on the above idea, set the second tag value to be greater than the first tag value, that is, the tag value when the first tooth is above and the second tooth is below is less than the tag value when the second tooth is above and the first tooth is below. Then the method for identifying dental arches can include the following steps.

[0356] Judge whether the mean value of the tag values is greater than the prediction threshold;

[0357] If it is greater, it is determined that the first dental arch model data points to the lower jaw and the second dental arch model data points to the upper jaw;

[0358] If it is less, it is determined that the first dental arch model data points to the upper jaw and the second dental arch model data points to the lower jaw.

[0359] In this way, based on the combination of the relative positions of several teeth with respect to a tooth and the corresponding relationship between the teeth and the dental arch, the relative position relationship between the corresponding dental arches can be inferred. The determination of the above position relationship is embodied as numerical operations and magnitude judgments, which have higher accuracy and automation.

[0360] The setting of the prediction threshold can be related to whether there is tooth loss in the dental arches indicated by the two sets of dental arch model data. When tooth loss occurs, the prediction threshold can be set relatively high, and when there is no tooth loss, the prediction threshold can be set relatively low. Based on this, the method for identifying a dental arch may include the steps of: determining whether there is tooth loss in the dental arch model data (i.e., the first dental arch model data and the second dental arch model data); if there is tooth loss, setting the prediction threshold as a relatively high first threshold; if there is no tooth loss, setting the prediction threshold as a relatively low second threshold.

[0361] For example, if the second tag value is set to 1 and the first tag value is set to 0. When there is no tooth loss, the predicted relationship identification information may be in the form of a 10-dimensional vector, and at this time, the prediction threshold can be set to 0.25; in one implementation, the mean of the tag values is the arithmetic mean. When tooth loss occurs, the predicted relationship identification information is less than 10 dimensions, and at this time, the prediction threshold can be set to 0.5; in one implementation, the mean of the tag values is the arithmetic mean.

[0362] In summary, the method for identifying a dental arch provided by the present invention uses the relationship identification information representing the relative relationship between tooth pairs to determine the dental arch information of different dental arches where the tooth pairs are located, which is equivalent to providing tooth pairs as a reference for the dental arch identification process, making the overall robustness of the identification process stronger; and, since the tooth identification scheme is relatively mature, compared with the scheme with excessive manual intervention and complex preprocessing steps, it takes into account the characteristics of high efficiency and stability; due to the discreteness between tooth pairs, it can also avoid the influence of individual errors on the overall result and can be adapted to the identification in a variety of complex scenarios.

[0363] It should be understood that although this specification is described according to embodiments, not every embodiment only contains an independent technical solution. This narrative manner of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0364] The series of detailed descriptions listed above are only specific descriptions of the feasible embodiments of the present invention, and they are not intended to limit the protection scope of the present invention. Any equivalent embodiments or modifications made without departing from the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for dental arch recognition, characterized in that, Including: Obtaining first dental arch model data and second dental arch model data of the tooth to be measured respectively, where the first dental arch model and the second dental arch model correspond to different dental arches; Obtaining relationship identification information of tooth pairs of the tooth to be measured according to the first dental arch model data and the second dental arch model data, where the tooth pairs include at least one pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs; Determining the dental arch information corresponding to the first dental arch model data and the second dental arch model data according to the relationship identification information.

2. The dental arch recognition method according to claim 1, wherein The dental arch information includes maxillary determination information or mandibular determination information for the first dental arch model, and maxillary determination information or mandibular determination information for the second dental arch model.

3. The dental arch recognition method according to claim 1 or 2, characterized in that, The dental arch recognition method includes: Based on the first dental arch model data and the second dental arch model data, obtaining tooth pair data of at least one pair of teeth located in different dental arches and corresponding to each other; Parallelly inputting the tooth pair data into a dental arch relationship recognition model to predict corresponding relationship identification information; the dental arch relationship recognition model is a neural network model, and the dental arch relationship recognition model is trained based on preset tooth pair data and relationship labels representing the relative position relationship of tooth pairs in the preset tooth pair data.

4. The method for dental arch recognition according to claim 3, wherein The dental arch recognition method includes: Performing feature extraction on the tooth data to be measured according to a first feature extraction module to obtain first tooth data, and performing feature extraction on the tooth data to be measured according to a second feature extraction module to obtain second tooth data; Obtaining first feature data and second feature data corresponding to the first tooth data and the second tooth data respectively, and generating a tooth pair feature sequence; Predicting relationship identification information of the first tooth and the second tooth according to the tooth pair feature sequence.

5. The dental arch recognition method according to any one of claims 1-4, characterized in that The dental arch recognition method includes: Obtaining a numerical form of the relationship identification information, and determining the maxillary and mandibular determination information of the first dental arch model data and the second dental arch model data according to the numerical relationship.

6. The dental arch recognition method according to any one of claims 1-5, characterized in that, The dental arch recognition method includes: When the tooth to be measured in the first dental arch model points to an upper jaw tooth and the tooth to be measured in the second dental arch model data points to a lower jaw tooth in the tooth pair data, obtaining relationship identification information with a numerical form of a first label value; When the tooth to be measured in the first dental arch model points to a lower jaw tooth and the tooth to be measured in the second dental arch model data points to an upper jaw tooth in the tooth pair data, obtaining relationship recognition information with a numerical form of a second label value.

7. The dental arch recognition method according to claim 5 or 6, characterized in that The dental arch recognition method includes: Based on the label values in numerical form corresponding to multiple groups of tooth pairs, determining the maxillary and mandibular determination information according to the voting scoring result.

8. The dental arch recognition method according to any one of claims 1-7, characterized in that The dental arch recognition method includes: Segmenting the first dental arch model and the second dental arch model to obtain a number of first dental arch tooth data and a number of second dental arch tooth data; Matching the first dental arch tooth data and the second dental arch tooth data according to the tooth position number to obtain tooth pair data.

9. The dental arch recognition method according to claim 8, characterized in that, The dental arch recognition method includes: Performing surface simplification on the first dental arch model and the second dental arch model according to the discrimination degree of parts in the dental arch model; Register the first dental arch model and the second dental arch model separately, and segment the first dental arch model and the second dental arch model based on the tooth segmentation model, where the tooth segmentation model is a neural network model trained based on a preset dental arch model and a preset tooth label representing the tooth distribution in the preset dental arch model.

10. The dental arch recognition method according to claim 8 or 9, characterized in that, The dental arch recognition method includes: Determine the patch segmentation method of the dental arch model according to the distinguishability of adjacent patches in the dental arch model, and obtain the patches of the dental arch model accordingly. Merge the valid vertices of the obtained patches.

11. The method for dental arch recognition according to any one of claims 8-10, characterized in that, The dental arch recognition method includes: Determine the patch segmentation method of the dental arch model according to the distinguishability between the patch and other intraoral tissues and the cost of adjacent patches belonging to the same intraoral tissue; the distinguishability between the patch and other intraoral tissues is determined according to at least one of the following indicators: The degree to which the current patch is close to the incisal edge connection line; The degree to which the current patch is close to the incisal extreme point of the corresponding tooth; The degree to which the current patch is far from the model center of the dental arch model data.

12. A display method, characterized in that, Include: Display the first dental arch model and the second dental arch model according to the dental arch information, where the dental arch information is obtained based on the dental arch recognition method according to any one of claims 1-11; The first dental arch model and the second dental arch model are used to design an orthodontic treatment plan.

13. An oral instrument, characterized in that, The oral appliance is prepared based on the dental arch information determined by the dental arch recognition method according to any one of claims 1-11.

14. A computer storage medium, on which an application program is stored, characterized in that, When the application program is executed, the steps of the dental arch recognition method according to any one of claims 1-11 are implemented.

15. An occlusal recognition device, characterized in that, Include: A first module for obtaining the first dental arch model data and the second dental arch model data of the tooth to be measured; The first dental arch model and the second dental arch model correspond to different dental arches; A second module for obtaining the relationship identification information of the tooth pairs of the tooth to be measured according to the first dental arch model data and the second dental arch model data, where the tooth pairs include at least one pair of teeth located in different dental arches and corresponding to each other, and the relationship identification information is used to indicate the relative position relationship of the tooth pairs; A third module for determining the dental arch information corresponding to the first dental arch model data and the second dental arch model data according to the relationship identification information.

16. An electronic device, characterized in that, It includes a processor, a memory, and a communication bus, and is characterized in that the processor and the memory complete communication with each other through the communication bus; The memory is used to store the application program; The processor is used to implement the steps of the dental arch recognition method according to any one of claims 1-11 when executing the application program stored on the memory.

Images