Support device, support method, and support program

Abstract

This method provides support for appropriate implant procedures using three-dimensional data representing the oral cavity. [Solution] This is a support device for assisting in implant procedures that include multiple components. The support device comprises an acquisition unit that acquires three-dimensional data showing the three-dimensional shape of the oral cavity including the jawbone; a calculation unit that determines the position of each of the multiple components in the oral cavity based on the three-dimensional data acquired by the acquisition unit and adjusts the position of at least one of the determined multiple components based on the positional relationship of the multiple components; and an output unit that outputs support information showing the position of each of the multiple components after adjustment by the calculation unit.

Description

Technical Field

[0001] The present disclosure relates to an assistance device, an assistance method, and an assistance program for assisting a procedure using an implant including a plurality of members.

Background Art

[0002] Conventionally, a technique for acquiring three-dimensional data indicating the three-dimensional shape in the oral cavity has been known, such as using a three-dimensional scanner that scans the surface shape of teeth and gums in the oral cavity or a CT (Computed Tomography) imaging device that performs computed tomography of dental arches and jaws. For example, Patent Document 1 (Japanese Patent Application Laid-Open No. 2020-96691) discloses acquiring three-dimensional data indicating the three-dimensional shape of teeth using a three-dimensional scanner.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] An operator such as a dentist acquires three-dimensional data of teeth using the three-dimensional scanner disclosed in Patent Document 1, and based on the acquired three-dimensional data, produces an artificial tooth (hereinafter also referred to as an "implant") that compensates for the missing tooth. In a procedure (correction) using such an implant, it is common for the attachment position of the implant to be determined based on the experience of the operator etc. Although the situation in the oral cavity where the implant is to be attached varies for each patient, it is not preferable for the attachment position of the implant etc. to vary significantly depending on the experience of the operator. Therefore, there is a need for a technique that enables an operator to optimally perform a procedure using an implant regardless of the operator's experience.

[0005] This disclosure was made to solve the above-mentioned problems and aims to provide a technology that enables optimal implant-based procedures. [Means for solving the problem]

[0006] According to one example of this disclosure, a support device is provided to assist in a procedure using an implant that includes multiple components. The support device includes an acquisition unit that acquires three-dimensional data showing the three-dimensional shape of the oral cavity, including teeth and jawbone that constitute the dentition; a calculation unit that determines the position of each of the multiple components in the oral cavity based on the three-dimensional data acquired by the acquisition unit and adjusts the position of at least one of the determined multiple components based on the positional relationship of the multiple components; and an output unit that outputs support information showing the position of each of the multiple components after adjustment by the calculation unit.

[0007] According to an example of this disclosure, a support method is provided for assisting in a procedure using an implant that includes multiple components, using a computer. The support method includes, as a process performed by the computer, the steps of: acquiring three-dimensional data showing the three-dimensional shape of the oral cavity including teeth and jawbone that constitute the dentition; determining the position of each of the multiple components in the oral cavity based on the acquired three-dimensional data; adjusting the position of at least one of the multiple components determined in the determination step based on the positional relationship of the multiple components; and outputting support information showing the position of each of the adjusted multiple components.

[0008] According to one example of this disclosure, a support program is provided for assisting in a procedure using an implant comprising multiple components, using a computer. The support program causes the computer to perform the following steps: acquire three-dimensional data showing the three-dimensional shape of the oral cavity, including the teeth and jawbone that constitute the dentition; determine the position of each of the multiple components in the oral cavity based on the acquired three-dimensional data; adjust the position of at least one of the multiple components determined in the determination step based on the positional relationship of the multiple components; and output support information showing the position of each of the adjusted multiple components. [Effects of the Invention]

[0009] According to this disclosure, the position of each component of the implant within the oral cavity is determined based on three-dimensional data showing the three-dimensional shape of the oral cavity, and furthermore, the position of at least one component is adjusted based on the positional relationship of each component. Therefore, regardless of the surgeon's experience, the surgeon can perform the implant procedure optimally. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example of the application of the support device according to Embodiment 1. [Figure 2] This is a block diagram showing the hardware configuration of the support device according to Embodiment 1. [Figure 3] This is a diagram illustrating the procedure using implants. [Figure 4] This diagram illustrates the estimation of the implant's placement position and orientation. [Figure 5] This figure illustrates an example of training an estimation model for estimating the optimal sub-positions of the superstructure. [Figure 6] This figure illustrates an example of training an estimation model for estimating the optimal sub-position of an abutment. [Figure 7] This figure illustrates an example of training an estimation model that estimates the suboptimal position of a fixture. [Figure 8]This figure illustrates an example of learning using an estimation model that estimates the overall optimal position. [Figure 9] This is a flowchart of the support processes performed by the support device. [Modes for carrying out the invention]

[0011] <Embodiment 1> Embodiment 1 of this disclosure will be described in detail with reference to the drawings. Note that identical or corresponding parts in the drawings are denoted by the same reference numerals, and their descriptions will not be repeated.

[0012] [Examples of application] An example of the application of the support device 1 according to Embodiment 1 will be explained with reference to Figure 1. Figure 1 is a diagram showing an example of the application of the support device 1 according to Embodiment 1.

[0013] As a treatment for tooth loss, implants 70 are commonly used. For example, an implant 70 includes a superstructure (artificial tooth) 71, an abutment (connecting part) 72, and a fixture 73 (artificial tooth root). The general procedure for implant treatment involves embedding a fixture 73, made of a biocompatible material, into the jawbone where the tooth is missing, and then attaching the superstructure 71 to the top of the fixture 73 via an abutment 72. The fixture 73 and the superstructure 71 can be connected by the abutment 72. Note that the implant 70 may include only the superstructure 71 and fixture 73 without the abutment 72.

[0014] The placement and type of implant 70 must be appropriately determined based on the patient's oral condition. However, variations may occur in the placement and type of implant 70 depending on factors such as the surgeon's priorities, skill level, and experience.

[0015] Specifically, among operators, there may be differences in the position and orientation of the fixture 73 with respect to the jawbone, the length of the fixture 73 embedded in the jawbone, the position and orientation of the abutment 72, the position and orientation of the superstructure 71, and the like. Also, the type of each of the superstructure 71, the abutment 72, and the fixture 73 selected by the operator may differ among operators. For example, generally, in order to emphasize pain reduction of the patient and the resistance of the implant 70, the operator determines the position and orientation of each member included in the implant 70 in consideration of the balance of the entire implant 70. On the other hand, an operator who emphasizes the aesthetics of teeth, such as in cosmetic dentistry, may first determine the position and orientation of the superstructure 71 and then determine the position and orientation of the fixture 73 based on the position and orientation of the superstructure 71. Thus, due to various factors such as the operator's experience, the viewpoint and purpose of the implant surgery, the position, orientation, type, etc. of each member of the implant 70 may differ.

[0016] Therefore, the support device 1 according to the first embodiment is configured to estimate (derive) support information for assisting the surgery using the implant 70 by using AI (Artificial Intelligence) technology.

[0017] Specifically, the user of the support device 1 acquires three-dimensional data showing a living tissue including the jawbone in the oral cavity by performing computed tomography of the dental arch and the jawbone using a CT imaging device (not shown).

[0018] Note that the "user" includes operators (such as doctors) or assistants (such as dental assistants, dental technicians, nurses, etc.) in various fields such as dentistry, oral surgery, orthopedic surgery, plastic surgery, and cosmetic surgery. Also, the "patient" includes patients in dentistry, oral surgery, orthopedic surgery, plastic surgery, and cosmetic surgery. The user obtains three-dimensional volume (voxel) data of the hard tissue parts (such as bone and teeth) including soft tissue parts (such as skin and gums) around the upper and lower jaws of the patient by using a CT imaging device (not shown) to photograph the upper and lower jaws of the patient. The CT imaging device is an X-ray imaging device that performs computed tomography of the upper and lower jaws of the patient by rotating an X-ray transmitter and receiver, which are a type of radiation, around the patient's face. The user generates three-dimensional data of the living tissue by using the volume data of the living tissue to be imaged obtained by the CT imaging device. Note that soft tissue parts are less detectable by X-rays and have lower data acquisition accuracy compared to hard tissue parts. Therefore, when soft tissue parts are displayed in an image, there is a possibility that the soft tissue parts can be faintly recognized or that parts that cannot be recognized may occur.

[0019] Also, the user can generate a rendering image (tomographic image or appearance image) showing the three-dimensional shape of the living tissue by using the volume data. A "rendering image" is an image generated by processing or editing certain data. Hereinafter, the three-dimensional volume data obtained by the CT imaging device is also referred to as "CT data", and the rendering image generated based on the CT data is also referred to as "CT image". For example, the user can generate a rendering image showing a two-dimensional living tissue (a part of the living tissue that can be shown by the CT data) viewed from a predetermined viewpoint by processing or editing the volume data of the living tissue obtained by the CT imaging device, and further, by changing the predetermined viewpoint in multiple directions, a plurality of rendering images showing the two-dimensional living tissue (a part of the living tissue that can be shown by the CT data) viewed from multiple directions can be generated.

[0020] Furthermore, the user may obtain optical scanner data, including positional information for each point in a point cloud (multiple points) representing the surface of biological tissue, including teeth and gums, within the oral cavity, by scanning the patient's oral cavity using a three-dimensional scanner (optical scanner) not shown. The optical scanner data includes, as positional information, the coordinates (X, Y, Z) of each point representing the surface of the biological tissue in predetermined horizontal (X-axis), vertical (Y-axis), and height (Z-axis) directions. Furthermore, the optical scanner data may include color information indicating the actual color of the portion (surface portion of the biological tissue) corresponding to each point in the point cloud (multiple points) representing the surface of biological tissue, including teeth and gums, within the oral cavity.

[0021] A three-dimensional scanner is a so-called intra-oral scanner (IOS) that can optically image the inside of a patient's mouth using methods such as confocal or triangulation. It can acquire positional information for each point in the point cloud that constitutes the surface of the biological tissue being scanned (for example, teeth and gums in the oral cavity) placed in a certain coordinate space. By using the optical scanner data acquired by the three-dimensional scanner, the user can generate a rendering image (external image) that shows the three-dimensional shape of the biological tissue. Hereinafter, the optical scanner data containing the positional information of each point in the point cloud representing the surface of the biological tissue acquired by the three-dimensional scanner will also be referred to as "IOS data," and the rendering image generated based on the IOS data will also be referred to as an "IOS image." For example, by processing or editing the optical scanner data of the biological tissue acquired by the three-dimensional scanner, the user can generate a rendering image that shows the two-dimensional biological tissue (the part of the biological tissue that can be shown by the IOS data) as seen from a predetermined viewpoint. Furthermore, by changing a predetermined viewpoint across multiple directions, the user can generate multiple rendering images showing two-dimensional biological tissue (parts of the biological tissue that can be shown by IOS data) viewed from multiple directions. Alternatively, the three-dimensional scanner may be optical scanner data acquired with a desktop scanner.

[0022] IOS images can show the surface shape of the scanned biological tissue in great detail, but they cannot show the internal structure that is not visible on the surface of the biological tissue (such as the alveolar bone and root apex). On the other hand, CT images can show the hard tissue portion (bones, teeth, etc.) of the scanned object in relatively detail, but they cannot show the soft tissue portion (skin, gums, etc.) in as much detail as the hard tissue portion.

[0023] The user may combine IOS data and CT data obtained for the same patient. By combining IOS data and CT data, composite data is generated. Here, IOS data and CT data have different data formats. Therefore, both data are made common. For example, the user may convert the data format of IOS data to the data format of CT data, and then use the converted data to perform pattern matching on the three-dimensional shape of the surface of the biological tissue to generate composite data by combining IOS data and CT data. Alternatively, the user may generate composite data by combining IOS data and CT data using the position of the tooth crown in IOS data and the position of the tooth crown in CT data as a reference. The user may also convert the data format of CT data to the data format of IOS data, and then use the converted data to perform pattern matching on the three-dimensional shape of the surface of some biological tissue, and then use the matched position as a reference to recalculate the three-dimensional position of the other biological tissue to generate composite data. Pattern matching is performed on the three-dimensional shape of the tooth crown portion of teeth where the degree of agreement is relatively large between IOS data and CT data. Since IOS data and CT data are acquired using different methods, and although the matching level is high, 100% matching may not be possible, the matching level can be considered to be above a predetermined level (e.g., 95% or higher). Alternatively, the user may convert the data format of the CT data and IOS data to a common data format and generate composite data by pattern matching the three-dimensional shape of the surface of the biological tissue using both converted data. By processing or editing the composite data, the user can generate a rendering image (for example, the composite image shown in Figure 1) that shows the two-dimensional biological tissue (the part of the biological tissue that can be shown by both IOS data and CT data) as viewed from a predetermined viewpoint. Alternatively, instead of unifying the data, the user may generate data at the corresponding coordinate positions of the IOS data and CT data in a certain three-dimensional space to generate composite data of both data. Such a method of generating data is also called overlay, fusion, or superimpose, but in this application, it is defined as synthesis in a broad sense.As shown in Figure 1, the composite image can display in three dimensions the surface shape of biological tissue shown by IOS data and the tomographic structure or appearance of hard tissue parts (bones, teeth, etc.) shown by CT data. The user may adjust the brightness, contrast, and transmittance of both the IOS and CT data as needed when generating the composite data. Furthermore, the user may pre-segment each of the multiple teeth, the jawbone, and the alveolar bone in both the IOS and CT data before generating the composite data.

[0024] The support device 1 may acquire CT data from the CT imaging device and IOS data from the three-dimensional scanner, and generate composite data based on the acquired IOS data and CT data according to user input. Alternatively, the support device 1 may acquire composite data generated by the user using other devices from other devices without acquiring IOS data and CT data.

[0025] In composite images generated based on composite data, the three-dimensional shape of hard tissues such as the alveolar bone and root apex is shown by CT data, while the three-dimensional shape of soft tissues such as the gingiva, which cannot be shown by CT data, can be shown by IOS data. As a result, composite images can show soft tissues such as the gingiva, which cannot be represented by CT data alone, in detail along with hard tissues such as the alveolar bone and root apex, by supplementing them with IOS data.

[0026] As will be explained in more detail later, the support device 1 estimates the appropriate position for attaching each component of the implant 70 and at least one of the types of each component of the implant 70, based on synthetic data of biological tissue including at least the jawbone. The support device 1 derives support information, including information on at least one of the attachment position and type of the implant 70, using estimation models 30A to 30D, which will be described later. Below, an example of estimating the appropriate position for attaching the implant 70 will be explained.

[0027] The support information may, for example, be composite data that shows the user and patient an image of the area after the implant 70 procedure is completed, and may be composite data in which three-dimensional data showing the implant 70 is superimposed on the missing tooth area. Figure 1 shows three-dimensional data showing the implant 70 superimposed on the composite data as support information. The support device 1 may also provide the user with numerical values ​​indicating the position and orientation of each component constituting the implant 70 (superstructure 71, abutment 72, fixture 73) as support information. For example, the support information may include oral cavity position information (X, Y, Z) of each component, or information indicating the oral cavity angle of each component. That is, the support device 1 may provide the user with an image representing at least one of the position and orientation of each component as support information, or it may provide the user with numerical values ​​indicating at least one of the position and orientation of each component constituting the implant 70. This allows the user to easily determine at which position in the actual patient's oral cavity it is appropriate to attach each component of the implant 70.

[0028] [Hardware configuration of the support device] The hardware configuration of the support device 1 according to Embodiment 1 will be described with reference to Figure 2. Figure 2 is a block diagram showing the hardware configuration of the support device 1 according to Embodiment 1. The support device 1 may be implemented, for example, by a general-purpose computer, or by a computer dedicated to a system for estimating support information.

[0029] As shown in Figure 2, the support device 1 comprises, as its main hardware elements, an arithmetic unit 11, a memory 12, a storage device 13, an input interface 14, a display interface 15, a peripheral device interface 16, a media reader 17, and a communication device 18.

[0030] The arithmetic unit 11 is an arithmetic entity (computer) that performs various processes by executing various programs, and is an example of an "arithmetic unit". The arithmetic unit 11 is composed of processors such as a CPU (central processing unit), MPU (Micro-processing unit), NPU (Neural network Processing Unit), TPU (Tensor Processing Unit), or GPU (Graphics Processing Unit). Although the processor, which is an example of the arithmetic unit 11, has the function of performing various processes by executing programs, some or all of these functions may be implemented using dedicated hardware circuits such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field-Programmable Gate Array). The term "processor" is not limited to processors in the narrow sense that execute processing in a stored-program manner, such as a CPU, MPU, NPU, TPU, or GPU, but may also include hardwired circuits such as ASICs or FPGAs. For this reason, the "processor," which is an example of the arithmetic unit 11, can also be read as processing circuitry in which processing is predefined by computer-readable code and / or hardwired circuits. The arithmetic unit 11 may consist of one chip or multiple chips. Furthermore, the processor and associated processing circuits may consist of multiple computers interconnected by wired or wireless connections via a local area network or wireless network. The processor and associated processing circuits may also consist of a cloud computer that remotely performs calculations based on input data and outputs the calculation results to other devices located at a distance.

[0031] Memory 12 includes a volatile storage area (for example, a working area) for temporarily storing program code and work memory when the arithmetic unit 11 executes various programs. Examples of memory 12 include volatile memory such as DRAM (dynamic random access memory) and SRAM (static random access memory), or non-volatile memory such as ROM (Read Only Memory) and flash memory.

[0032] The storage device 13 stores various programs or data executed by the arithmetic unit 11. The storage device 13 may be one or more non-transitory computer-readable media or one or more computer-readable storage media. Examples of storage devices 13 include HDDs (Hard Disk Drives) and SSDs (Solid State Drives).

[0033] The storage device 13 stores the support program 20 and the estimation model 30. The support program 20 describes the content of the support processing that the computing device 11 uses the estimation model 30 to estimate support information based on image data (for example, synthesized data) showing the three-dimensional shape of biological tissue.

[0034] The estimation model 30 includes a neural network 31 and a dataset 32 ​​used by the neural network 31. The estimation model 30 is trained to estimate support information based on synthetic data showing a state in which at least one tooth is missing, using machine learning with training data that includes image data showing the three-dimensional shape of biological tissue in the oral cavity (e.g., synthetic data) and ground truth data associated with said image data (synthetic data after the implant 70 has actually been attached).

[0035] The neural network 31 may be any algorithm applicable to the neural network 31 of Embodiment 1, such as an autoencoder, convolutional neural network (CNN), recurrent neural network (RNN), transformer, state-space model, or generative adversarial network (GAN). The estimation model 30 is not limited to the neural network 31; it may also include other known algorithms such as Bayesian estimation or support vector machines (SVM).

[0036] Dataset 32 ​​includes weight coefficients used in calculations performed by the neural network 31 and decision thresholds used for making decisions during calculations.

[0037] The input interface 14 is an example of an "acquisition unit". The input interface 14 acquires, for example, composite data of the oral cavity. The composite data input from the input interface 14 is stored in the memory 12 or storage device 13 and used by the arithmetic unit 11 when estimating support information. The input interface 14 may also acquire CT data and IOS data before synthesis. For example, the input interface 14 may be connected to a three-dimensional scanner (not shown) in a communicative manner and acquire IOS data from the three-dimensional scanner. Alternatively, the input interface 14 may be connected to a CT imaging device (not shown) in a communicative manner and acquire CT data from the CT imaging device. In this case, the arithmetic unit 11 generates the composite data described above by synthesizing the IOS data and CT data input from the input interface 14, and estimates support information based on the generated composite data. The data used for estimating support information is not limited to composite data; for example, it may be CT data before synthesis. That is, the support device 1 may estimate support information using only CT data without using IOS data.

[0038] The display interface 15 is an example of an "output unit" and is an interface for connecting the display 40. The display interface 15 enables data input and output between the support device 1 and the display 40. The display interface 15 outputs image data to the display 40 to show the support information estimated using the estimation model 30, and causes the display 40 to display the support information.

[0039] The peripheral device interface 16 is an interface for connecting peripheral devices such as a keyboard 51 and a mouse 52. The peripheral device interface 16 enables data input and output between the support device 1 and the peripheral devices.

[0040] The media reader 17 reads various types of data stored in the storage medium 60 and writes various types of data to the storage medium 60. For example, the media reader 17 may acquire a support program 20 from the storage medium 60, or it may write support information estimated by the arithmetic unit 11 to the storage medium 60. The storage medium 60 may be one or more non-transitory computer-readable media or one or more computer-readable storage media. When the arithmetic unit 11 acquires image data (for example, composite data) from the storage medium 60 via the media reader 17, the media reader 17 may be an example of an "acquisition unit".

[0041] The communication device 18 transmits and receives data to and from an external device (not shown) via wired or wireless communication. For example, the communication device 18 may transmit support information estimated by the arithmetic unit 11 to an external device. When the arithmetic unit 11 acquires image data (e.g., composite data) from an external device via the communication device 18, the communication device 18 can be an example of an "acquisition unit". Also, when the support information estimated using the estimation model 30 is output to an external device different from the display 40, the communication device 18 can be an example of an "output unit".

[0042] [Procedures using implants] Figure 3 is a diagram illustrating a procedure using implant 70. As shown in Figure 3, implant 70 includes a superstructure 71, an abutment 72, and a fixture 73, each of which is an independent component. Implant 70 is installed, for example, in a dentition where a tooth is missing and the gum is exposed. First, the fixture 73 is embedded and fixed in the jawbone covered by the gum, and the superstructure 71 is attached to the fixture 73 via the abutment 72. Implant 70 is installed in the oral cavity to maintain an optimal position and posture, taking into account the position and posture of the adjacent teeth adjacent to the treatment site, and the position, posture and jawbone condition of the opposing teeth facing the treatment site.

[0043] [Estimation of implant placement position and orientation] Figure 4 is a diagram illustrating the estimation of the mounting position and orientation of the implant 70. The computing device 11 of the support device 1 determines the optimal position of each of the multiple components included in the implant 70 based on three-dimensional data of the tooth missing state. As shown in Figure 4, the computing device 11 estimates the optimal positions of the superstructure 71, abutment 72, and fixture 73, respectively, using estimation models 30A, 30B, and 30C. Estimation models 30A, 30B, and 30C are examples of the "first estimation models" of this disclosure. That is, the computing device 11 independently estimates the optimal position of each component using estimation models 30A to 30C. In addition to the process of estimating the optimal position of the components, estimation models 30A to 30C can also perform a process of adjusting the estimated optimal position of the components. Hereinafter, the optimal position of each component estimated using estimation models 30A, 30B, and 30C will be referred to as the "partially optimal position". Note that estimated models 30A, 30B, and 30C are included in estimated model 30 shown in Figure 2.

[0044] The partially optimal position is the position of each component (superstructure 71, abutment 72, fixture 73) estimated based on three-dimensional data of the tooth missing state, without considering its relationship to other components. That is, the partially optimal position of the superstructure 71 is estimated by estimation model 30A without considering the positions of the abutment 72 and fixture 73. Similarly, the partially optimal position of the abutment 72 is estimated by estimation model 30B without considering the positions of the superstructure 71 and fixture 73. Furthermore, the partially optimal position of the fixture 73 is estimated by estimation model 30C without considering the positions of the superstructure 71 and abutment 72. The partially optimal position of each component is estimated considering predetermined viewpoints corresponding to each component. These predetermined viewpoints include at least one of the following: the patient's pain during the procedure, the tolerance of each component, the patient's occlusion, and the patient's oral aesthetics. For example, the partially optimal position of the superstructure 71 is estimated so that its position is optimal from the viewpoint of aesthetics and occlusion. Similarly, the partially optimal positions of the abutment 72 and fixture 73 are estimated so that their positions are optimal from the viewpoint of pain associated with the patient and the tolerance of each component. In other words, the superstructure 71 corresponds to the viewpoint of aesthetics and occlusion, while the abutment 72 and fixture 73 correspond to the viewpoint of pain and tolerance. Note that the predetermined viewpoints to which each component corresponds are not limited to these, and other viewpoints or different combinations may be used. For example, the superstructure 71 may correspond to the viewpoint of tolerance, and the fixture 73 may correspond to the viewpoint of occlusion.

[0045] The partially optimal position is, for example, a position relative to the jawbone, and may be 3D data including the distance between the jawbone and each component, and the orientation (angle) of each component relative to the jawbone, or it may be a numerical value. Estimation models 30A to 30C independently estimate the partially optimal position of each component based on a predetermined viewpoint corresponding to each component. Therefore, even if each component is adopted at the position of the output result of each estimation model 30A to 30C, it may not be possible to assemble each component whose partially optimal position is adopted as a single implant 70. For example, the distance between the partially optimal positions of each component may be too far apart, the partially optimal positions of each component may overlap on the three-dimensional coordinate system, or the axes of each component may not lie on the same axis.

[0046] Therefore, the computing device 11 of Embodiment 1 uses estimation model 30D to adjust the partially optimal positions independently estimated by each estimation model 30A to 30C so that they can be assembled as a single implant 70. Note that estimation model 30D is an example of the "second estimation model" of this disclosure. Specifically, the computing device 11 uses estimation model 30D to estimate the optimal position considering the entire implant 70, by setting the three components that constitute the implant 70: the superstructure 71, the abutment 72, and the fixture 73. Hereinafter, the optimal position of the entire implant 70 estimated using estimation model 30D will also be referred to as the "overall optimal position". The overall optimal position includes the positional information of each of the superstructure 71, the abutment 72, and the fixture 73, and is the position in which the positions of each of these components can be assembled as a single implant 70. The overall optimal position estimated using estimation model 30D is a position in which the positions of the superstructure 71, the abutment 72, and the fixture 73 are considered in relation to each other.

[0047] Estimation Model 30D is a model that learns from three-dimensional data of the oral cavity of a specific patient after a specific surgeon has actually performed an implant procedure on that patient, using that data as ground truth. Therefore, Estimation Model 30D is configured to output results that reflect the actual treatment results of the surgeon who performed the procedure on the three-dimensional data that serves as the ground truth. In the following, Estimation Models 30A, 30B, 30C, and 30D may be collectively referred to as "Estimation Model 30," or each of Estimation Models 30A, 30B, 30C, and 30D may be referred to as "Estimation Model 30."

[0048] The computing unit 11 estimates the overall optimal position using estimation model 30D, then segments each component, compares the position of each component at the overall optimal position with the partially optimal positions of each component estimated independently by estimation models 30A to 30C, and adjusts the partially optimal positions of each component estimated by estimation models 30A to 30C. Specifically, each of estimation models 30A to 30C queries whether the partially optimal position it estimated is the optimal position compared to the overall optimal position estimated by estimation model 30D. In other words, each of estimation models 30A to 30C requests an answer from estimation model 30D indicating whether the partially optimal position it estimated is the optimal position when considering the overall balance of the implant 70. In response, estimation model 30D compares the overall optimal position with the partially optimal positions obtained from each of estimation models 30A to 30C and generates an answer indicating whether the partially optimal position is the optimal position when considering the overall balance of the implant 70. The response includes the result of whether the partially optimal position from the estimated models 30A to 30C is optimal, and information to adjust the partially optimal position if it is not optimal. The information to adjust the partially optimal position includes information on the adjusted position of the partially optimal position so that it is the optimal position when considering the balance of the entire implant 70. The computing device 11 adjusts each partially optimal position to a position where it can be assembled as a single implant 70 by repeatedly exchanging requests and responses between each of the estimated models 30A to 30C and estimated model 30D. The training method for each estimated model 30 will be described in detail below for each estimated model 30A to 30D.

[0049] [Training the estimation model] The training of estimation models 30A to 30D using machine learning will be explained, referring to Figures 5 to 8.

[0050] Figure 5 illustrates an example of training using the estimation model 30A, which estimates the partially optimal position of the superstructure 71. As shown in Figure 5, in Embodiment 1, the training data includes three-dimensional data showing the three-dimensional shape of the oral cavity, including the teeth and jawbone that constitute the dentition. For example, during training of the estimation model 30A, machine learning is performed using the three-dimensional data and ground truth data including the position and orientation of the superstructure 71. The ground truth data including the position and orientation of the superstructure 71 is, for example, three-dimensional data showing the position of the superstructure 71 after it has actually been treated by a practitioner who prioritizes aesthetics, in contrast to three-dimensional data showing a state where teeth are missing. The ground truth data only needs to show the position of the superstructure 71, and may be numerical data showing positional information and orientation, rather than three-dimensional data. The three-dimensional data indicating the position of the superstructure 71, with an emphasis on aesthetics, includes, for example, three-dimensional data indicating the position that forms a beautiful U-shaped dental arch, the position where the height matches that of adjacent teeth, the position where the interdental space is not widened and the space is optimal, and the position where only the crown is exposed and the root is not exposed from the gum.

[0051] The perspectives reflected in the output results may differ depending on the source of the ground truth data. Generally, surgeons in the field of cosmetic surgery tend to place more emphasis on aesthetics than surgeons in other fields. For example, if three-dimensional data including the position and posture of a superstructure 71 actually performed by a surgeon who prioritizes aesthetic appearance is used as the ground truth data when training the estimation model 30A, the estimation model 30A can estimate the position and posture of the superstructure 71 that prioritizes aesthetic appearance. Also, if three-dimensional data including the position and posture of a superstructure 71 actually performed by a surgeon who prioritizes bite alignment is used as the ground truth data when training the estimation model 30A, the estimation model 30A can estimate the position and posture of the superstructure 71 that prioritizes bite alignment.

[0052] When estimation model 30A acquires three-dimensional data of a missing tooth, the neural network 31A estimates the partially optimal position of the superstructure 71 based on the input three-dimensional data. At this time, estimation model 30A estimates the optimal position and orientation for the superstructure 71 without considering the position and orientation of other components (abutment 72, fixture 73) that constitute the implant 70. The optimal position and orientation for the superstructure 71 is, for example, a position and orientation that takes into account the position and orientation of teeth adjacent to the treatment site and the position and orientation of opposing teeth facing the treatment site. Note that the neural network 31A is included in the neural network 31 shown in Figure 2. Estimation model 30A determines whether the estimated partially optimal position of the superstructure 71 matches the position information of the superstructure 71, which is the ground truth data associated with the input three-dimensional data. If the two match, estimation model 30A does not update the dataset 32A, but if they do not match, it updates the dataset 32A to optimize the dataset 32A. Note that dataset 32A is included in dataset 32 ​​shown in Figure 2. Optimization of dataset 32A involves setting parameters such as thresholds and weighting coefficients corresponding to aesthetics, which are included in dataset 32A, so that the position of the superstructure 71 is in a position that prioritizes aesthetics. The neural network 31A sets the priority of predetermined viewpoints by optimizing dataset 32A.

[0053] In this way, the estimation model 30A is trained to accurately estimate the partially optimal position of the superstructure 71 based on the input data by optimizing the dataset 32A using training data that includes three-dimensional data of the tooth in a missing state before implant treatment and ground truth data of the position information of the superstructure 71.

[0054] Figure 6 illustrates an example of training using the estimation model 30B to estimate the partially optimal position of the abutment 72. As shown in Figure 6, during the training of the estimation model 30B, machine learning is performed using three-dimensional data showing the three-dimensional shape of the oral cavity, including the teeth and jawbone that constitute the dentition, and training data including the position information of the abutment 72 after the implant procedure has actually been performed as ground truth data.

[0055] If three-dimensional data including the position and posture of abutments 72 actually performed by surgeons who prioritize aesthetics and preventing patient pain is used as ground truth data when training the estimation model 30B, the estimation model 30B can estimate the position and posture of abutments 72 that prioritizes aesthetics and preventing patient pain. If three-dimensional data including the position and posture of abutments 72 actually performed by surgeons who prioritize aesthetics when the superstructure 71 is attached and strengthening the durability of the abutments 72 is used as ground truth data when training the estimation model 30B, the estimation model 30B can estimate the position and posture of abutments 72 that prioritizes aesthetics when the superstructure 71 is attached and strengthening the durability of the abutments 72 is As shown in Figure 3, when the abutment 72 is implanted, a portion of it is located within the superstructure 71, so aesthetic considerations are taken into account when the superstructure 71 is attached. Furthermore, because a portion of it is located within the alveolar bone and jawbone, tolerance and pain based on the thickness of the alveolar bone and the bone volume and density of the jawbone are taken into consideration. However, when comparing aesthetics with tolerance and pain, tolerance and pain tend to be given more importance than aesthetics.

[0056] When estimation model 30B acquires three-dimensional data of a missing tooth, the neural network 31B estimates the partially optimal position of the abutment 72 based on the acquired three-dimensional data. At this time, estimation model 30B estimates the optimal position and orientation for the abutment 72 without considering the position and orientation of other components constituting the implant 70 (superstructure 71, fixture 73). The optimal position and orientation for the abutment 72 is, for example, a position and orientation that takes into account the tolerance of the abutment 72 after it is attached and the pain experienced by the patient after the abutment 72 is attached. Note that the neural network 31B is also included in the neural network 31 shown in Figure 2.

[0057] In this way, the estimation model 30B also optimizes the dataset 32B by determining whether the estimated suboptimal position of the abutment 72 matches the ground truth data, which is the position information of the abutment 72. This trains the estimation model 30B to accurately estimate the suboptimal position of the abutment 72. Optimizing the dataset 32B involves setting parameters such as thresholds and weighting coefficients corresponding to aesthetics, tolerance, and pain, which are included in the dataset 32B, so that the position prioritizes aesthetics and tolerance / pain. The neural network 31B sets priorities for predetermined viewpoints by optimizing the dataset 32B.

[0058] Figure 7 illustrates an example of training using the estimation model 30C to estimate the partial optimal position of fixture 73. As shown in Figure 7, during training of the estimation model 30C, machine learning is performed using three-dimensional data showing the three-dimensional shape of the oral cavity, including the teeth and jawbone that constitute the dentition, and training data including the position information of fixture 73 after the implant procedure has actually been performed as ground truth data.

[0059] If three-dimensional data including the position and posture of fixture 73 actually performed by a practitioner who prioritizes preventing patient pain and maximizing tolerance is used as ground truth data during training of estimation model 30C, estimation model 30C can estimate the position and posture of fixture 73 that prioritizes preventing patient pain and maximizing tolerance. If three-dimensional data including the position and posture of fixture 73 actually performed by a practitioner who prioritizes preventing patient pain and maximizing tolerance is used as ground truth data during training of estimation model 30C, estimation model 30C can estimate the position and posture of fixture 73 that prioritizes maximizing tolerance. Patient pain is influenced by factors such as the superstructure 71 contacting adjacent teeth too much, fixture 73 being too deep, thin alveolar bone, contact with the mandibular canal, and low surrounding bone volume and density. Tolerance is influenced by factors such as low surrounding bone volume and density.

[0060] When the estimation model 30C acquires three-dimensional data of a missing tooth, the neural network 31C estimates the partially optimal position of the fixture 73 based on the acquired three-dimensional data. At this time, the estimation model 30C estimates the optimal position and orientation for the fixture 73 without considering the position and orientation of other components constituting the implant 70 (superstructure 71, abutment 72). The optimal position and orientation for the fixture 73 is included in the neural network 31 shown in Figure 2, and includes factors such as the durability of the fixture 73 after it is attached and the pain experienced by the patient after the fixture 73 is attached.

[0061] In this way, the estimation model 30C also optimizes the dataset 32C by determining whether the estimated suboptimal position of fixture 73 matches the positional information of fixture 73, which is the ground truth data. This trains the estimation model 30C to accurately estimate the suboptimal position of fixture 73. Optimizing the dataset 32C involves setting parameters such as thresholds and weighting coefficients corresponding to tolerance and pain, which are included in the dataset 32C, so that the positions prioritize tolerance and pain. The neural network 31C sets priorities for predetermined viewpoints by optimizing the dataset 32C.

[0062] Next, we will describe the estimation model 30D that estimates the overall optimal position. Figure 8 is a diagram illustrating an example of machine learning in the learning phase of the estimation model 30D that estimates the overall optimal position. As shown in Figure 8, in Embodiment 1, the training data includes three-dimensional data of a state in which at least one tooth is missing, and as ground truth data, the position information of the implant 70 after the implant procedure is actually performed and assembled (position information of the superstructure 71, position information of the abutment 72, and position information of the fixture 73).

[0063] When three-dimensional data representing a missing tooth is input to the estimation model 30D, the neural network 31D estimates the overall optimal position of the implant 70 based on the input three-dimensional data. The estimation model 30D also optimizes the dataset 32D by determining whether the estimated overall optimal position of the implant 70 matches the ground truth data, which includes the position information of the superstructure 71, the abutment 72, and the fixture 73. In this way, the estimation model 30D is trained to estimate the overall optimal position of the implant 70 with high accuracy. Below, an example of adjusting the partial optimal position will be explained using the flowchart shown in Figure 9.

[0064] [Support Processing] Figure 9 is a flowchart of the support process performed by the support device. The flowchart shown in Figure 9 is realized when the support program 20 is executed by the arithmetic unit 11.

[0065] The calculation device 11 acquires three-dimensional data showing the three-dimensional shape of the oral cavity, including the teeth and jawbone that constitute the dentition (step S1). For example, the three-dimensional data shows the oral cavity of a patient who is missing teeth and is the target of implant treatment. The calculation device 11 inputs the three-dimensional data acquired in step S1 into estimation model 30A to determine the partially optimal position of the superstructure 71 (step S2). The calculation device 11 inputs the three-dimensional data acquired in step S1 into estimation model 30B to determine the partially optimal position of the abutment 72 (step S3). The calculation device 11 inputs the three-dimensional data acquired in step S1 into estimation model 30C to determine the partially optimal position of the fixture 73 (step S4). Note that the calculation device 11 may execute steps S2 to S4 simultaneously, or in a different order than the example in Figure 9. The calculation device 11 acquires the positional relationships of each component determined in steps S2, S3, and S4 from the respective estimation models 30A to 30C (step S5). In other words, in step S5, the computing device 11 acquires information indicating the position and orientation of the partially optimal position of each member, which was obtained in steps S2 to S4.

[0066] The computing device 11 determines whether the partially optimal position of each component determined in steps S2 to S4 is the optimal position when viewed as a whole (step S6). Specifically, the computing device 11 compares the partially optimal position of each component determined in steps S2 to S4 with the overall optimal position estimated by the estimation model 30D to determine whether the partially optimal position is the optimal position considering the balance of the entire implant 70. For example, the computing device 11 may determine whether the partially optimal position of each component is the optimal position when viewed as a whole based on the condition that each component can be assembled as a single implant. A position that can be assembled as a single implant means a position where each component can be connected to each other in order to function as an implant 70. A position that can be assembled as a single implant can be uniquely determined depending on the type of each component.

[0067] If at least one of the partially optimal positions of each component determined in steps S2 to S4 is not the optimal position (NO in step S6), each of the estimated models 30A to 30C requests estimated model 30D to adjust the partially optimal positions of the superstructure 71, abutment 72, and fixture 73, respectively, and the adjusted partially optimal positions are determined again in the S2 to S4 process. Adjustment means moving and rotating the partially optimal position of each component in three dimensions.

[0068] For example, the computing unit 11 inputs the three-dimensional data acquired in step S1 into the estimation model 30D to determine the overall optimal position of the implant 70. The estimation model 30D acquires the partially optimal positions output by estimation models 30A to 30C. If the estimation model 30D determines that each partially optimal position is not the optimal position, it provides adjusted position information to estimation models 30A to 30C, for example, so that each partially optimal position approaches the positions of the superstructure 71, abutment 72, and fixture 73 in the overall optimal position. In this way, the estimation model 30D adjusts the partially optimal positions of the superstructure 71, abutment 72, and fixture 73 to a position where they can be assembled as a single implant, based on the current positional relationship of each partially optimal position.

[0069] For example, if the estimation model 30D determines that the distance between the partially optimal position of the superstructure 71 and the position of the superstructure 71 in the overall optimal position is 50.0 μm, it generates adjusted position information that brings the partially optimal position of the superstructure 71 closer to the position of the superstructure 71 in the overall optimal position by 5.0 μm. In other words, estimation model 30D causes estimation model 30A to adjust the partially optimal position of the superstructure 71 closer to the position of the superstructure 71 in the overall optimal position by 1%. Estimation model 30D responds by bringing the partially optimal position of each component closer to the overall optimal position according to a predetermined percentage for each component.

[0070] Each of the estimated models 30A to 30C determines, based on the limit position, whether the partially optimal position of each component can be moved and rotated as indicated by the answer from estimated model 30D. The limit position is a predetermined position for each superstructure 71, abutment 72, and fixture 73, where, when each component is viewed independently, it is not possible to attach each component based on a predetermined viewpoint. The predetermined limit position is, for example, a position that is a predetermined percentage (e.g., 1%) away from the partially optimal position.

[0071] For example, the limit position of the superstructure 71 can be determined from the perspective of occlusion and aesthetics. When considering the limit position of the superstructure 71 from the perspective of occlusion, the calculation device 11 focuses only on the superstructure 71 and determines that the position of the superstructure 71 has exceeded the limit position when the relative distance between the tooth axis of the superstructure 71 and the tooth axis of the opposing tooth, and the relative positional relationship of their respective inclinations, exceeds the limit range from the perspective of aesthetics. In other words, the limit position of the superstructure 71 is determined based on the relative positional relationship between the superstructure 71 and the opposing tooth. For example, the calculation device 11 may determine that the partially optimal position of the superstructure 71 has exceeded the limit position if the area of ​​the superstructure 71 exposed from the gingiva is larger than a predetermined area.

[0072] Furthermore, the limit positions of the abutment 72 and fixture 73 can be determined from the perspective of pain and tolerance. The calculation unit 11 determines the limit positions of the abutment 72 and fixture 73 from the perspective of pain based on the relative positional relationship between the positions of the abutment 72 and fixture 73 and the nerves. If the limit positions of the abutment 72 and fixture 73 are set from the perspective of tolerance, the calculation unit 11 determines that the partially optimal position of the abutment 72 and fixture 73 has exceeded the limit position if, when viewing the jawbone in plan, the position of the jawbone through which the extension of the tooth axis of the abutment 72 and fixture 73 passes is thinner than a predetermined thickness.

[0073] Each of the estimation models 30A to 30C, if it determines that moving or rotating the partially optimal position of each component would cause it to exceed its limit position, will reduce the amount of movement or rotation. Using the example of the superstructure 71 described above, if estimation model 30A determines that moving the partially optimal position of the superstructure 71 by 50.0 μm based on the adjusted position information would cause it to exceed its limit position, it will, for example, move the partially optimal position of the superstructure 71 closer to its position in the overall optimal position by 5.0 μm instead of 50.0 μm. In other words, estimation model 30A moves the partially optimal position of the superstructure 71 closer to its position in the overall optimal position by 0.01% in response to the answer from estimation model 30D. Subsequently, each of the estimation models 30A to 30C again requests estimation model 30D whether the partially optimal position it estimated is the optimal position when considering the overall balance of the implant 70, and estimation model 30D responds to this request, again considering the overall optimal position. This causes steps S2 to S6 to be repeated.

[0074] In other words, the estimation model 30D determines that if the calculation unit 11 determines that the partially optimal position of each component updated in steps S2 to S4 is not the optimal position when viewing each component as a whole (NO in step S6), it will execute steps S2 to S5 again. In this way, the calculation unit 11 brings the partially optimal position of each component as close as possible to the overall optimal position. By repeating the process in steps S2 to S4, each partially optimal position can be updated to a position close to the overall optimal position, starting from each predetermined viewpoint. If, even after repeating the request and response a predetermined number of times, each partially optimal position does not become a position where it can be assembled as a single implant, the calculation unit 11 adjusts the current partially optimal positions so that they become positions where they can be assembled as a single implant, without considering the overall optimal position. Specifically, the calculation unit 11 uses one of the partially optimal positions as a reference and updates the other partially optimal positions so that other components can be connected to the component at the reference partially optimal position. In the end, the calculation unit 11 can position each partially optimal position so that it can be assembled as a single implant.

[0075] If the calculation unit 11 determines that the partially optimal position of each component updated in steps S2 to S4 is the optimal position when viewing each component as a whole (YES in step S6), it generates support information (step S7) and outputs it (step S8). The estimation model 30D may not perform the process of bringing each partially optimal position closer to the overall optimal position, but may only provide a response that moves and rotates the partially optimal positions of the other components so that they can be connected to one of the partially optimal positions as a reference component. In this case, a component corresponding to a predetermined viewpoint having a higher priority than other priorities may be determined as the reference component. For example, if the user wants to prioritize the aesthetics of the superstructure 71, the estimation model 30D may move and rotate the partially optimal positions of the abutment 72 and fixture 73 as a reference to the position of the superstructure 71 estimated by the estimation model 30A.

[0076] Thus, in this embodiment, the optimal position of each component, which is output independently, is used as a reference to bring the components closer to their optimal position when viewed as a whole. This makes it possible to derive the optimal position of each component, which can be assembled as a single implant, as support information, while reflecting the optimal position of each component, which is output independently. The surgeon can grasp the appropriate mounting position of the implant 70 simply by looking at this support information, and the support device 1 of Embodiment 1 determines the position of each component of the implant in the oral cavity, and further, the position of at least one component is adjusted based on the positional relationship of each component, so that the surgeon can perform the implant procedure optimally regardless of the surgeon's experience.

[0077] <Embodiment 2> Embodiment 1 describes an example in which support information is output using the partially optimal position estimated by estimation models 30A to 30C. However, the support device 1 may output support information without using estimation models 30A to 30C, and Embodiment 2 describes an example in which support information is generated using rule-based processing that does not use so-called artificial intelligence, such as a neural network.

[0078] In Embodiment 2, the support device 1 also executes the flowchart shown in Figure 9 and outputs support information, similar to Embodiment 1. However, the processing in steps S2 to S6 uses rule-based processing instead of the estimation model 30. The rule-based processing described in Embodiment 2 is written within the support program 20. Below, the rule-based processing described in Embodiment 2 will be explained using Figure 9. Note that in Embodiment 2, the description of configurations that overlap with the support device 1 in Embodiment 1 will not be repeated.

[0079] The computing device 11 is configured to determine the optimal partial position of each component for each predetermined viewpoint based on the three-dimensional data acquired in step S1. These predetermined viewpoints may include, for example, aesthetics, bite, durability, and patient-related pain. The optimal partial position of each component can be determined using a rule-based method with the support program 20.

[0080] The calculation device 11 independently determines the partially optimal position of each component based on the three-dimensional data acquired in step S1 and predetermined rules. For example, from the viewpoint of aesthetics, the calculation device 11 determines the partially optimal position of the superstructure 71 using the gingival margin line of the composite data as a reference. More specifically, when focusing only on the superstructure 71 in the three-dimensional data of a missing tooth, the calculation device 11 determines the position where the aesthetics of the superstructure 71 are highest as the partially optimal position for aesthetics of the superstructure 71. For example, the calculation device 11 determines the partially optimal position for aesthetics of the superstructure 71 according to the gingival margin line of the composite data, the relative position with adjacent teeth, the balance of shape, etc. That is, the partially optimal position is the position when a predetermined component is considered independently from a predetermined viewpoint, without considering other components and other viewpoints, similar to the partially optimal position. In Embodiment 2, the partially optimal position may be determined using the mean and median based on data obtained from statistical results of past patient three-dimensional data. Furthermore, the partially optimal position in Embodiment 2 may be manually entered by the user, and the entered data may be used to train the estimation models 30A to 30C that estimate the partially optimal position as described in Embodiment 1.

[0081] Similarly, the computing device 11 determines the partially optimal occlusal position of the superstructure 71 based on the positions and tooth axes of the opposing teeth in three-dimensional data from the perspective of occlusion. For example, the computing device 11 determines the position of the superstructure 71 when the tooth axes of the opposing teeth and the tooth axes of the superstructure 71 are aligned as the partially optimal occlusal position of the superstructure 71. The partially optimal occlusal position of the superstructure 71 is determined without considering the aesthetics of the superstructure 71.

[0082] Furthermore, the calculation unit 11 determines the partially optimal position of the abutment 72 from the standpoint of tolerance. The calculation unit 11 determines the partially optimal position of the abutment 72 for tolerance, provided that the abutment 72 does not come into contact with the surface of the gingiva when the superstructure 71 and the fixture 73 are connected. In addition, from the standpoint of pain and tolerance, the calculation unit 11 determines the position of the fixture 73 when the extension of the tooth axis of the fixture 73 passes through the thickest part of the jawbone as the partially optimal position for pain and tolerance of the fixture 73. The calculation unit 11 may determine these partially optimal positions based on three-dimensional data by executing the support program 20. The conditions for determining the partially optimal positions may be predetermined as described above, or they may be set by the user.

[0083] In Embodiment 2, the calculation device 11 sets priorities for these predetermined viewpoints. For example, the calculation device 11 sets the priority of these viewpoints in the order of pain, tolerance, bite, and aesthetics. Alternatively, the calculation device 11 may set the priority of these viewpoints in the order of tolerance, pain, bite, and aesthetics.

[0084] Furthermore, the priority of predetermined viewpoints set by the calculation device 11 may be set by the user. For example, the user sets numerical values ​​representing the priority of each viewpoint so that the sum of the priorities of each viewpoint is "100". Specifically, the user may set "40" for the pain viewpoint, "30" for the tolerance viewpoint, "20" for the bite viewpoint, and "10" for the aesthetic viewpoint. The calculation device 11 determines that a viewpoint has a high priority if the numerical value set for that viewpoint is high, and determines that a viewpoint has a low priority if the numerical value set for that viewpoint is low. In this embodiment, pain, tolerance, bite, and aesthetics are adopted as predetermined viewpoints, but other viewpoints may be adopted, and these other viewpoints may be set by the user.

[0085] Furthermore, the numerical values ​​representing the priority of each perspective may be automatically determined by the calculation unit 11 based on the input data corresponding to each perspective. For example, the user inputs the disease state of the dentition as information corresponding to pain into the support device 1. The disease state of the dentition is information related to pain that may occur after the implant 70 is installed, such as premature contact, gingival condition, and the presence or absence of alveolar bone condition. The calculation unit 11 automatically determines a numerical value representing the priority of pain based on the input disease state of the dentition. Specifically, if the input disease state of the dentition is not good, the calculation unit 11 increases the numerical value representing the priority of pain, and if the disease state of the dentition is good, it decreases the numerical value representing the priority of pain. The user may also input information indicating the state of periodontal disease as information corresponding to pain. The state of periodontal disease may be determined using other estimation models that estimate the state of periodontal disease from images showing the oral cavity, for example.

[0086] The user inputs the member retention status as information corresponding to resistance into the support device 1. Member retention status refers to information indicating the state in which each member is held after it has been attached, and may include the direction of the tooth axes of opposing teeth and adjacent teeth, and the state of the alveolar bone, such as the state of the bone trabeculae and bone density. The calculation device 11 automatically determines a numerical value representing the priority of resistance based on the input member retention status. If information indicating that the member retention status is not good is input, the calculation device 11 increases the numerical value representing the priority of resistance, and if information indicating that the member retention status is good is input, it decreases the numerical value representing the priority of resistance.

[0087] The user inputs the occlusal state of opposing teeth as information corresponding to the bite into the support device 1. The occlusal state of the teeth is information indicating whether or not occlusion with opposing teeth is easy, and may include the tooth axis and shape of the opposing teeth. The calculation device 11 automatically determines a numerical value representing the priority of the bite based on the input occlusal state of the teeth. The calculation device 11 increases the numerical value representing the priority of the bite when information indicating that occlusion with opposing teeth is not easy is input, and decreases the numerical value representing the priority of the bite when information indicating that occlusion with opposing teeth is easy is input. In addition, the user may input jaw movement data (not shown) or the output results of occlusal simulation software as information corresponding to the bite.

[0088] The user inputs information corresponding to aesthetics into the support device 1, such as the positional relationships of adjacent teeth, the positional relationships of opposing teeth, or the condition of the gums. The calculation device 11 automatically determines a numerical value representing the priority of aesthetics based on the input positional relationships of adjacent teeth, the positional relationships of opposing teeth, or the condition of the gums. The calculation device 11 increases the numerical value representing the priority of aesthetics if it determines from the input data that aesthetics should be given high importance, and decreases the numerical value representing the priority of aesthetics if it determines from the input data that aesthetics should not be given high importance. The user may also input data showing the shape of the user's face scanned by a face scanning device (not shown) as information corresponding to aesthetics. In this case, the calculation device 11 may determine the priority of aesthetics based on the position of the user's cheeks. The support device 1 may also determine the priority of each viewpoint using an estimation model for determining the priority of each viewpoint.

[0089] In step S2, the calculation device 11 determines the partially optimal position of the superstructure 71 from the viewpoint of aesthetics and from the viewpoint of occlusion. In step S3, the calculation device 11 determines the partially optimal position of the abutment 72 from the viewpoint of durability. Furthermore, in step S4, the calculation device 11 determines the partially optimal position of the fixture 73 from the viewpoint of pain and durability. In step S5, the calculation device 11 obtains the positional relationship of the partially optimal positions of each component from each viewpoint.

[0090] In step S6, the computing unit 11 determines whether each partially optimal position is the optimal position when viewing each component as a whole. In Embodiment 2, since partially optimal positions are determined for the superstructure 71 from two viewpoints, the computing unit 11 can determine in step S6 that if either of the partially optimal positions of the superstructure 71 is connectable to the partially optimal position of the abutment 72 and the partially optimal position of the fixture 73, then it is the optimal position when viewing each component as a whole.

[0091] In Embodiment 2, if it is determined that each partially optimal position cannot be attached as a single implant (NO in step S6), the computing device 11 determines the overall optimal position which is the center of each partially optimal position, without using the overall optimal position determined by the estimation model 30D, based on, for example, the positional relationship of each partially optimal position of each member from each viewpoint. In Embodiment 1, the overall optimal position was estimated using the estimation model 30D, but in Embodiment 2, the computing device 11 determines the center position of each partially optimal position output in a rule-based manner, and determines the position which can be assembled as a single implant at that center position as the overall optimal position.

[0092] Subsequently, the calculation unit 11 of Embodiment 2 repeats steps S2 to S6, similar to Embodiment 1, to gradually adjust each of the current partial optimal positions so that they approach the overall optimal position which is the center of each partial optimal position. Similar to Embodiment 1, the calculation unit 11 approaches each of the partial optimal positions to the overall optimal position which is the center of each partial optimal position by a predetermined ratio.

[0093] In Embodiment 2, the calculation unit 11 may determine a predetermined ratio for bringing each partial optimal position closer together, according to the priority set for the viewpoint described above. The calculation unit 11 may reduce the amount of movement and rotation of the partial optimal position corresponding to the viewpoint with a high priority, and increase the amount of movement and rotation of the partial optimal position corresponding to the viewpoint with a low priority.

[0094] For example, if the priority of a predetermined viewpoint is set in the order of pain, tolerance, occlusion, and aesthetics, the calculation unit 11 moves and rotates the partially optimal position of the fixture 73 in terms of pain and tolerance so that it is 0.1% closer to the overall optimal position in Embodiment 2. The calculation unit 11 also moves and rotates the partially optimal position of the abutment 72 in terms of tolerance so that it is 0.2% closer to the overall optimal position in Embodiment 2. Furthermore, the calculation unit 11 moves and rotates the partially optimal position of the superstructure 71 in terms of occlusion so that it is 0.5% closer to the overall optimal position in Embodiment 2. Furthermore, the calculation unit 11 moves and rotates the partially optimal position of the superstructure 71 in terms of aesthetics so that it is 0.8% closer to the overall optimal position in Embodiment 2. In this way, the calculation unit 11 of Embodiment 2 updates each partially optimal position. Based on the updated partially optimal positions, the calculation unit 11 again calculates the overall optimal position which is the center of each partially optimal position and updates the overall optimal position which is the center of each partially optimal position. Thus, in Embodiment 2, when bringing each partially optimal position closer to the overall optimal position, each partially optimal position is moved and rotated by a ratio corresponding to a numerical value set as the priority of a predetermined viewpoint. In Embodiment 1 as well, each partially optimal position may be moved and rotated according to a numerical value set as the priority of a predetermined viewpoint.

[0095] If the calculation device 11 cannot assemble the implant as a single unit even after repeating steps S2 to S6 a predetermined number of times, it sets one of the individual optimal positions as the reference optimal position, similar to Embodiment 1, and moves and rotates the other optimal positions to match the reference optimal position. When each optimal position can be assembled as a single implant (YES in step S6), the calculation device 11 generates the position of the implant 70 based on the finally updated optimal positions of each component as support information (step S7) and outputs it (step S8). As a result, even in the support device 1 of Embodiment 2, the operator can grasp the appropriate mounting position of the implant 70 simply by looking at the support information, and the support device 1 of Embodiment 2 can support appropriate implant procedures.

[0096] [Differentiation] In the support device 1 according to Embodiments 1 and 2, the estimation models 30A to 30D were models that estimated the position of each component. However, the estimation models 30A to 30D may also be models that can estimate, in addition to the position of each component, the base material of each component, the model of each component, the manufacturer, and the combination of each component determined by the manufacturer, based on three-dimensional data of a missing tooth. This makes it possible for the support device 1 to output not only the position of the implant 70, but also the optimal type of implant 70 for each component as support information.

[0097] In step S1 described above, an example was explained in which three-dimensional data with at least one missing tooth is input. However, the data input in step S1 may also be three-dimensional data with no missing teeth. In this case, the user performs segmentation on the teeth on the dental arch and generates new three-dimensional data of a state in which a specific tooth specified by the user is missing. The calculation unit 11 may estimate the position of each component, assuming that the implant 70 will be attached to the location where the specific tooth is missing. Furthermore, the calculation unit 11 may automatically detect the position where the implant 70 should be inserted, or it may estimate the position of each component, assuming that it will be inserted at a position specified by the user. For example, if three-dimensional data of a state in which multiple teeth are missing is input, the user may input information to the support device 1 indicating which location the support device 1 should estimate the position of the implant 70 to be inserted at.

[0098] As described above, Embodiment 1 describes an example in which the partially optimal position and the overall optimal position are estimated using estimation models 30A to 30D, while Embodiment 2 describes an example in which the partially optimal position and the overall optimal position are determined by the support program 20 using a rule-based approach. Whether to use estimation model 30 or determine the position using a rule-based approach may be decided for each partially optimal position and overall optimal position. For example, the partially optimal position of the superstructure 71 may be estimated using estimation model 30A, while the partially optimal positions of the abutment 72 and fixture 73 may be determined using a rule-based approach. Alternatively, each partially optimal position may be determined using a rule-based approach, while only the overall optimal position is estimated using estimation model 30D.

[0099] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of this disclosure is indicated by the claims rather than the foregoing description, and all modifications within the meaning and scope of the claims are intended to be included. The configurations illustrated in these embodiments and those illustrated in the variations may be combined as appropriate. [Explanation of Symbols]

[0100] 1. Support device, 11. Processing unit, 12. Memory, 13. Storage device, 14. Input interface, 15. Display interface, 16. Peripheral device interface, 17. Media reader, 18. Communication device, 20. Support program, 30, 30A-30D. Estimation model, 31, 31A-31D. Neural network, 32, 32A-32D. Dataset, 40. Display, 51. Keyboard, 52. Mouse, 60. Storage medium, 70. Implant, 71. Superstructure, 72. Abutment, 73. Fixture.

Claims

1. A support device for assisting in procedures using implants that include multiple components, An acquisition unit that acquires three-dimensional data showing the three-dimensional shape of the oral cavity, including the teeth and jawbone that make up the dentition, A calculation unit determines the position of each of the plurality of members within the oral cavity based on the three-dimensional data acquired by the acquisition unit, and adjusts the position of at least one of the determined plurality of members based on the positional relationship of the plurality of members. A support device comprising: an output unit that outputs support information indicating the position of each of the plurality of members after adjustment by the calculation unit.

2. The support device according to claim 1, wherein the positional relationship follows the priority of a predetermined viewpoint.

3. The support device according to claim 1, wherein the three-dimensional data includes CT data obtained by computed tomography of the dentition and the jawbone.

4. The support device according to claim 3, wherein the three-dimensional data includes composite data obtained by combining optical scanner data, which includes positional information of each point in a point cloud representing the surface of the dentition, and the CT data, based on the position of the tooth crown.

5. The support device according to any one of claims 1 to 4, wherein the calculation unit determines the position of the specific member based on the three-dimensional data using a first estimation model trained to determine the partially optimal position of a specific member included in the plurality of members based on the three-dimensional data.

6. The support device according to any one of claims 1 to 4, wherein the calculation unit adjusts the position of at least one of the members using a second estimation model trained to estimate the position of each of the plurality of members such that the overall position of the plurality of members becomes the overall optimal position, based on the three-dimensional data.

7. The support device according to any one of claims 1 to 4, wherein the calculation unit adjusts the position of at least one of the plurality of members based on a setting value set by the user to determine the position of each of the plurality of members, or a predetermined setting value.

8. The aforementioned predetermined perspective includes at least one of the following: the pain of the person being treated, the tolerance of the plurality of components, the occlusion of the dentition, and the aesthetics of the oral cavity. The support device according to claim 2, wherein a numerical value indicating priority is set for each of the aforementioned pain, tolerance, bite, and aesthetics.

9. The pain includes the diseased condition of the dentition, The resistance includes the holding state of the plurality of members, The aforementioned bite includes the occlusal state of opposing teeth, The aforementioned aesthetics include the positional relationship of multiple adjacent teeth, the positional relationship of multiple opposing teeth, or the condition of the gums. The support device according to claim 8, wherein the numerical values ​​set for each of the pain, tolerance, bite, and aesthetics can be changed by the user according to the state or positional relationship of each of the pain, tolerance, bite, and aesthetics.

10. The support device according to claim 9, wherein the aesthetics further include the shape of the user's face scanned by a face scanning device.

11. The support device according to claim 8, wherein the priority of the predetermined viewpoint is higher in the order of pain, tolerance, bite, and aesthetics, or in the order of pain, bite, tolerance, and aesthetics.

12. The support device according to claim 8, wherein the aforementioned predetermined viewpoint can be set by the user.

13. The first estimation model is further trained to determine the type of the specific member based on the three-dimensional data, The support device according to claim 5, wherein the calculation unit determines the type of the specific member based on the three-dimensional data using the first estimation model.

14. The support device according to claim 13, wherein the type of the specific member includes at least one of the manufacturers of the specific member and the combination of the plurality of members determined by the manufacturer.

15. The support device according to any one of claims 1 to 4, wherein the plurality of members include a superstructure and a fixture.

16. The support device according to claim 15, wherein the plurality of members further include abutments.

17. A computer-assisted method for assisting in implant procedures involving multiple components, The process to be performed by the aforementioned computer is: The steps include: acquiring three-dimensional data showing the three-dimensional shape of the oral cavity, including the jawbone; A step of determining the position of each of the plurality of members within the oral cavity based on the acquired three-dimensional data, A step of adjusting the position of at least one of the plurality of members determined in the determination step, based on the positional relationship of the plurality of members, A support method comprising the step of outputting support information indicating the position of each of the adjusted members.

18. A support program for assisting procedures using implants containing multiple components, which is powered by a computer. To the aforementioned computer, The steps include: acquiring three-dimensional data showing the three-dimensional shape of the oral cavity, including the jawbone; A step of determining the position of each of the plurality of members within the oral cavity based on the acquired three-dimensional data, A step of adjusting the position of at least one of the plurality of members determined in the determination step, based on the positional relationship of the plurality of members, A support program that performs the step of outputting support information indicating the position of each of the multiple components after adjustment.

Images