Method and device for determining CBCT image distance between maxillary sinus and maxillary posterior tooth, medium and product

Abstract

The invention discloses a CBCT (Cone Beam Computed Tomography) image distance determination method and device for maxillary sinus and maxillary posterior teeth, a medium and a product, and relates to the field of image processing, the method comprises the following steps: acquiring CBCT images of maxillary sinus and maxillary posterior teeth; according to the CBCT images of the maxillary sinus and the maxillary posterior teeth, performing three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth by adopting a structure segmentation model; the structure segmentation model is constructed based on a deep learning network; according to a three-dimensional segmentation result, generating a surface mesh model based on a marching cube algorithm; determining point cloud data based on a point cloud generation algorithm according to the surface mesh model; according to the point cloud data, the nearest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding nearest apical point of the maxillary posterior teeth are determined; and determining the position relationship between the maxillary sinus and the maxillary posterior teeth according to the nearest distance, and classifying the position relationship. The distance between the maxillary sinus and the maxillary posterior tooth can be efficiently and accurately measured, and the position relation can be judged.

Description

Technical Field

[0001] The present application relates to the field of image processing, and in particular to a method, device, medium and product for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth. Background Art

[0002] The maxillary sinus (MS) is one of the important paranasal sinuses located in the midface. Anatomically, the maxillary sinus is adjacent to the roots of the maxillary posterior teeth (MPT). Studies have reported that approximately 10% of maxillary sinusitis is odontogenic, and in cases of unilateral maxillary sinusitis, this proportion is as high as 75%. Among the causes of odontogenic maxillary sinusitis, iatrogenic damage caused by dental procedures such as tooth extraction, root canal treatment, and implant treatment accounts for 65%. Therefore, before performing oral treatment in the maxillary posterior area, it is necessary to accurately assess the positional relationship between the maxillary sinus and the roots of the maxillary posterior teeth to minimize the occurrence of iatrogenic maxillary sinus damage.

[0003] In the existing technology, imaging examination is an objective indicator for evaluating the positional relationship between the maxillary sinus and the maxillary posterior teeth. In clinical practice, two-dimensional periapical films and curved tomographic films are the preferred examination methods. However, when the positional relationship between the two is close, two-dimensional imaging examinations have the problem of being unable to accurately distinguish between anatomical structure protrusion and image overlap. Three-dimensional imaging cone beam computed tomography (CBCT) can provide high-resolution three-dimensional reconstruction images of any plane, which can accurately evaluate the morphology and adjacent relationships of anatomical structures such as the jaw, teeth, and maxillary sinus, and thus provide a basis for positional relationship judgment and treatment plan formulation. However, in existing studies, the judgment of the positional relationship between the maxillary sinus floor and the maxillary posterior tooth roots is performed by physicians with professional training backgrounds. After making appropriate angle adjustments in the CBCT multi-planar reconstruction view, the nearest distance layer is observed layer by layer, the distance is manually measured on the two-dimensional image of this layer, and the relationship between the two is judged. This process is not only time-consuming, but also highly technically sensitive, and the accuracy of interpretation is affected by the experience of the physician reading the film. Against the backdrop of a massive increase in data volume due to the widespread clinical application of CBCT and a shortage of professional physicians, new intelligent medical technologies are being developed to assist professional physicians in CBCT imaging diagnosis, enabling efficient and accurate measurement of the distance between the maxillary sinus and the maxillary posterior teeth and determination of their positional relationship. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, medium and product for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth, which can efficiently and accurately measure the distance between the maxillary sinus and the maxillary posterior teeth and determine the positional relationship.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a method for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images, the method comprising:

[0007] Obtain CBCT images of the maxillary sinus and maxillary posterior teeth;

[0008] Based on the CBCT images of the maxillary sinus and the maxillary posterior teeth, a structure segmentation model is used to perform three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth; the structure segmentation model is constructed based on a deep learning network;

[0009] According to the 3D segmentation results, a surface mesh model is generated based on the marching cube algorithm;

[0010] According to the surface mesh model, point cloud data is determined based on the point cloud generation algorithm;

[0011] Based on the point cloud data, the closest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding distance between the maxillary posterior teeth and the nearest root apex are determined;

[0012] The positional relationship between the maxillary sinus and the maxillary posterior teeth was determined based on the closest distance, and the positional relationship was classified.

[0013] Optionally, the training process of the structure segmentation model specifically includes:

[0014] Constructing a data set; the data set includes: CBCT images and corresponding annotation results; the annotation results include: maxillary sinus, maxillary first premolar, maxillary second premolar, maxillary first molar and maxillary second molar;

[0015] Sampling the CBCT images in the data set to a set voxel spacing and performing preprocessing; the preprocessing includes normalization and data enhancement;

[0016] According to the processed data set, the deep learning network is trained using a three-fold cross-validation training strategy to obtain a structural segmentation model.

[0017] Optionally, the deep learning network includes: a U-type network of a convolutional block attention module, a SegNet model, a DeepLab model, and a Transformer model.

[0018] Optionally, determining point cloud data based on a point cloud generation algorithm according to the surface mesh model specifically includes:

[0019] The surface mesh model was scaled using the voxel spacing of the CBCT image;

[0020] Save the scaled surface mesh model as an STL format file;

[0021] According to the saved file, the point cloud generation algorithm is used to obtain point cloud data.

[0022] Optionally, the point cloud generation algorithm includes Poisson disk sampling.

[0023] Optionally, the positional relationship between the maxillary sinus and the maxillary posterior teeth is determined based on the closest distance, and the positional relationship is classified, specifically including:

[0024] Determine the positional relationship between the maxillary sinus and maxillary posterior teeth based on the closest distance;

[0025] Get the position relationship demarcation point;

[0026] Determine whether the positional relationship between the maxillary sinus and the maxillary posterior teeth is greater than the positional relationship cutoff point, and obtain a comparison result;

[0027] The position relationship classification result is determined based on the comparison result.

[0028] In a second aspect, the present application provides a device for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images, the device comprising:

[0029] An image acquisition unit, used for acquiring CBCT images of the maxillary sinus and maxillary posterior teeth;

[0030] a three-dimensional segmentation unit for performing three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth based on CBCT images of the maxillary sinus and the maxillary posterior teeth using a structural segmentation model constructed based on a deep learning network;

[0031] A surface mesh model generating unit, configured to generate a surface mesh model based on a marching cubes algorithm according to the three-dimensional segmentation result;

[0032] A point cloud data determination unit, configured to determine point cloud data based on a surface mesh model and a point cloud generation algorithm;

[0033] The closest distance measurement unit is used to determine the closest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding distance between the maxillary posterior teeth and the nearest root apex point based on the point cloud data;

[0034] The position relationship determination and classification unit is used to determine the position relationship between the maxillary sinus and the maxillary posterior teeth based on the closest distance and to classify the position relationship.

[0035] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth.

[0036] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth.

[0037] In a fifth aspect, the present application provides a computer program product, comprising a computer program, characterized in that when the computer program is executed by a processor, it implements the method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth.

[0038] According to the specific embodiments provided in this application, this application has the following technical effects:

[0039] The present application provides a method, device, medium and product for determining the distance between the maxillary sinus and the maxillary posterior teeth on CBCT images. The method automatically segments the CBCT images of the maxillary sinus and the maxillary posterior teeth using a structure segmentation model built based on a deep learning network (DL). The method generates point cloud data based on the three-dimensional segmentation results using a marching cube algorithm and a point cloud generation algorithm. The method then determines the closest distance between the maxillary sinus and the maxillary posterior teeth in three-dimensional space based on the point cloud data, identifies the root apex closest to the maxillary sinus floor, determines the positional relationship between the maxillary sinus and the maxillary posterior teeth based on the distance between the maxillary sinus and the maxillary posterior teeth, and classifies the positional relationship based on clinical operation risks. The method utilizes artificial intelligence (AI) and three-dimensional point cloud technology to automatically measure the distance between the maxillary sinus and the maxillary posterior teeth on CBCT images, and can efficiently and accurately measure the distance between the maxillary sinus and the maxillary posterior teeth and determine the positional relationship. This method significantly improves the efficiency of clinical diagnosis, reduces the possibility of misjudgment, and reduces treatment risks. The method provides clinicians with a more accurate decision-making basis based on the positional relationship classification results, promoting the development of intelligent medicine. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0041] Figure 1 Schematic diagram of a flow chart of a method for determining the distance between the maxillary sinus and the maxillary posterior teeth in CBCT images according to one embodiment of the present application;

[0042] Figure 2 This is a schematic diagram of the structural segmentation process of the structural segmentation model;

[0043] Figure 3A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0045] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0046] In an exemplary embodiment, Figure 1 As shown, a method for determining the distance between the maxillary sinus and the maxillary posterior teeth in CBCT images is provided, the method comprising the following S101 to S106.

[0047] S101, obtain CBCT images of the maxillary sinus and maxillary posterior teeth;

[0048] S102, based on the CBCT images of the maxillary sinus and the maxillary posterior teeth, use a structural segmentation model to perform three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth; the structural segmentation model is constructed based on a deep learning network; the deep learning network includes but is not limited to: a U-Network with Convolutional Block Attention Module (U-Net CBAM), a SegNet model, a DeepLab model, and a Transformer model.

[0049] The training process of the structure segmentation model is:

[0050] S21, constructing a data set; the data set includes: CBCT images and corresponding annotation results; the annotation results include: maxillary sinus, maxillary first premolar, maxillary second premolar, maxillary first molar, and maxillary second molar;

[0051] As a specific embodiment, the process of constructing the data set specifically includes the following steps:

[0052] 1) Case selection

[0053] Methods: 44 CBCT images of patients who were treated at Peking University School of Stomatology between April 1, 2022, and April 1, 2024 were collected and screened. The CBCT brands and models were NewTom Vgi and NewTomVG (Quantitative Radiology, Italy), and the scan field of view was 12 cm × 8 cm. Inclusion criteria included: (1) clear images without motion artifacts or sclerosis artifacts that affect observation; (2) permanent dentition with complete root development; the maxillary dentition and maxillary sinus floor were scanned completely; (3) no history of maxillary trauma, surgery, tumors, or systemic diseases. Exclusion criteria included: (1) apical lesions or previous apical surgery in the maxillary posterior teeth area; (2) congenital malformations, osteomyelitis, and maxillary sinus diseases in the segmented area.

[0054] 2) Image annotation

[0055] (1) Annotation of the maxillary sinus and maxillary posterior teeth segmentation training samples: An endodontist, A, with 3 years of clinical experience, delineated the maxillary sinus, maxillary first premolar, maxillary second premolar, maxillary first molar, and maxillary second molar in the CBCT images using ITK-SNAP 4.2 (www.itksnap.org). The delineated images were then modified and confirmed by another dental radiologist, B, with more than 10 years of experience.

[0056] (2) Manual measurement of the distance between the maxillary sinus and maxillary posterior teeth and marking of the nearest root apex: An endodontist (A) with 3 years of clinical experience manually measured the distance between the two in CBCT images. This was then reviewed and corrected by another dental imaging specialist (B) with more than 10 years of experience. All CBCT data were recorded twice by evaluator A, one month apart, to verify consistency. ITK-SNAP 4.2 (www.itksnap.org) was used as the software.

[0057] The specific measurement process is as follows: Each maxillary posterior tooth is measured individually, and the CBCT multiplanar reconstruction image is aligned so that the coronal plane is parallel to the long axis of the tooth and the sagittal plane is parallel to the maxillary posterior dentition. Layer by layer in the sagittal and coronal planes is observed, and the closest layer is selected. The closest distance between the maxillary tooth root apex and the corresponding inferior wall of the maxillary sinus is measured and recorded. For multi-rooted teeth, only the root apex closest to the sinus floor is calibrated. If the root apex contacts the sinus floor, the distance is considered zero.

[0058] (3) Classification of the positional relationship between the maxillary sinus and the maxillary posterior teeth: Based on clinical needs, 1 mm was selected as the positional relationship cutoff. A distance between the two of 0-1 mm was considered a Type I relationship (indicating a high risk of clinical operation); a distance >1 mm was considered a Type II relationship (indicating a relatively low risk of clinical operation).

[0059] S22, the CBCT images in the data set are sampled to a set voxel spacing and preprocessed; the preprocessing includes normalization and data enhancement; the data enhancement includes spatial enhancement and voxel enhancement, thereby increasing the diversity of the data. The voxel is the resolution of the three-dimensional CBCT image (0.3*0.3*0.3mm 3 );

[0060] Prior to S22, the following steps were also included: converting the CBCT image data format to NIfTI format and resizing the CBCT images to a consistent size (256*256*256); using SimpleITK to read the CBCT images and labels; SimpleITK is a medical image segmentation and registration toolkit, an open source tool library for medical image processing;

[0061] S23. Based on the processed data set, a three-fold cross-validation training strategy is used to train the deep learning network to obtain a structure segmentation model.

[0062] As a specific embodiment, the deep learning network adopts U-Net CBAM, such as Figure 2 As shown in the figure, the mask of the result of manual segmentation is used as the spatial template, and the minimum bounding box (Bounding Box) of all the annotated data in three-dimensional space is calculated. The boundary is expanded according to the maximum expansion principle (axial ±5mm, coronal / sagittal ±3mm, not exceeding the image boundary) to form a region of interest.

[0063] U-Net CBAM consists of an encoder (downsampling) and a decoder (upsampling), connected by skip connections. The encoder extracts features through multiple convolutions (copy and crop) and max pooling (2x2 maxpooling) to reduce spatial dimensionality. The decoder restores spatial dimensionality through up-convolutions (2x2 up-conv) and combines the encoder's feature map (via skip connections) to improve segmentation accuracy. The CBAM module is introduced within the downsampling convolutional layer of U-Net. This module consists of two submodules: channel attention and spatial attention. By first performing channel attention and then spatial attention, the attention map is added to the input feature map to achieve feature recalibration. This allows the model to more effectively focus on important regions of interest while suppressing less important information. Post-processing in U-Net CBAM includes morphological operations, thresholding, and connected region analysis. Morphological operations include: ① Erosion: removing isolated noise points, smoothing the contours of the segmented structure, and eliminating burrs; ② Dilation: filling small gaps along the segmentation edges and internal holes. Thresholding converts the model's output probability map into a binary mask. A threshold of 0.5 is set to distinguish foreground from background, yielding the segmentation result. Connected region analysis is performed to identify adjacent teeth or structures, as they often appear connected in the segmentation results. This requires performing connected region analysis to obtain independent regions for label assignment, facilitating the output of segmentation results for multiple structures.

[0064] The structure segmentation model adopts a three-fold cross-validation training strategy. During the training process, a combination of Dice loss (Dice Loss) and Categorical Cross-Entropy loss (CCE) is used. The total loss function is Total Loss = α·Dice Loss + (1-α)·CCE Loss; where α = 0.5;

[0065] Dice loss function Dice Loss is:

[0066]

[0067] In the above formula, y true ∈{0,1} is the true label (binarized mask, 1 represents the target area); y pred ∈[0,1] is the probability value predicted by the model; N is the total number of pixels / voxels; ∈ is a smoothing term (usually 1e-5) to prevent the denominator from being zero.

[0068] The cross entropy loss function CCE Loss is:

[0069]

[0070] In the above formula, C is the total number of categories; y true,c ∈{0,1} is the true label of category c; y pred,c ∈[0,1] is the model's predicted probability for the category.

[0071] The structural segmentation model can optimize the classification performance of different labels while maintaining segmentation accuracy. The Adam optimizer is used for parameter update, with an initial learning rate set to 0.0001. The cosine annealing strategy is used to fine-tune the parameters and obtain the structural segmentation model.

[0072] The accuracy of the structural segmentation model was evaluated using the Dice Similarity Coefficient (DSC) and Jaccard Index, while the accuracy of the classification diagnosis was evaluated using the confusion matrix, accuracy, precision, and sensitivity.

[0073] S103, generating a surface mesh model based on the 3D segmentation results using the Marching Cubes algorithm; the 3D segmentation results include multi-category annotation information for the teeth and maxillary sinus;

[0074] S104 , determining point cloud data according to the surface mesh model and based on a point cloud generation algorithm; the point cloud generation algorithm includes Poisson disk sampling.

[0075] S104 specifically includes:

[0076] S41, scaling the surface mesh model using the voxel spacing of the CBCT image;

[0077] S42, saving the scaled surface mesh model as a Standard Tessellation Language (STL) format file;

[0078] S43, using a point cloud generation algorithm according to the saved file to obtain point cloud data;

[0079] S105, determining the closest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding distance between the maxillary posterior teeth and the nearest root apex point based on the point cloud data;

[0080] Calculate the closest distance from each point in one point cloud to another point cloud based on the point cloud data, store it in a distance array, sort the distance storage array to obtain the minimum distance value, and output the corresponding tooth point cloud coordinates as the nearest root apex point;

[0081] S106, determining the positional relationship between the maxillary sinus and the maxillary posterior teeth based on the closest distance, and classifying the positional relationship.

[0082] S106 specifically includes:

[0083] Determine the positional relationship between the maxillary sinus and maxillary posterior teeth based on the closest distance;

[0084] Get the position relationship demarcation point;

[0085] Determine whether the positional relationship between the maxillary sinus and the maxillary posterior teeth is greater than the positional relationship cutoff point, and obtain a comparison result;

[0086] The position relationship classification result is determined based on the comparison result.

[0087] In a specific embodiment, based on actual clinical needs, the closest distance of 1mm is set as the positional relationship demarcation point. When the closest distance is less than or equal to 1mm, the positional relationship between the maxillary sinus and the maxillary posterior teeth is classified as Class I, indicating that the distance is close and the risk of maxillary sinus-related complications during oral surgery in the maxillary posterior teeth area is high, and clinicians need to pay attention. When the closest distance between the two is greater than 1mm, the positional relationship is classified as Class II, indicating that the distance is relatively far and the risk of maxillary sinus-related complications during treatment is relatively low.

[0088] Based on the method provided in this application, in a specific embodiment, the steps of automatic segmentation, distance measurement, and automatic classification of CBCT images are integrated to establish a complete set of AI-assisted automatic measurement software for the CBCT image distance between the maxillary sinus and the maxillary posterior teeth.

[0089] This application uses artificial intelligence technology combined with a three-dimensional point cloud algorithm to automatically measure the CBCT impact distance between the maxillary sinus and the maxillary posterior teeth. The deep learning-based U-Net CBAM algorithm automatically segmented the multi-structure of the maxillary sinus and the maxillary posterior teeth. The point cloud generation algorithm automatically measured the closest distance between the maxillary sinus and the maxillary posterior teeth and automatically identified the closest root apex. The positional relationship between the maxillary sinus and the maxillary posterior teeth was automatically classified based on the minimum distance. This application has the following technical effects:

[0090] (1) Improve measurement accuracy: Through a single-architecture multi-structure segmentation model, the maxillary sinus and maxillary posterior teeth can be accurately segmented, thereby improving the accuracy of subsequent distance measurement and relationship classification.

[0091] (2) Improve work efficiency: Utilizing 3D point cloud reconstruction technology, high-precision 3D models can be automatically generated, reducing the need for manual intervention by doctors during image analysis. This not only improves work efficiency but also reduces errors in judgment caused by human factors.

[0092] (3) Optimize the data processing process: In the distance measurement step, the closest distance between the root apex and the maxillary sinus floor is automatically calculated through a software algorithm, reducing the number of messages transmitted between the server and the client, allowing the system to process large amounts of image data more quickly.

[0093] (4) Rapid classification and risk assessment: Through automatic classification of measurement results, treatment risks can be quickly assessed, allowing doctors to obtain key information in a short period of time, thereby improving the efficiency of clinical decision-making.

[0094] Based on the same inventive concept, embodiments of the present application also provide a device for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images, for implementing the aforementioned method for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of the device for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images provided below can be found in the aforementioned method for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images, and will not be further elaborated upon here.

[0095] In an exemplary embodiment, a device for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images is provided, comprising:

[0096] An image acquisition unit, used for acquiring CBCT images of the maxillary sinus and maxillary posterior teeth;

[0097] a three-dimensional segmentation unit for performing three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth based on CBCT images of the maxillary sinus and the maxillary posterior teeth using a structural segmentation model constructed based on a deep learning network;

[0098] A surface mesh model generating unit, configured to generate a surface mesh model based on a marching cubes algorithm according to the three-dimensional segmentation result;

[0099] A point cloud data determination unit, configured to determine point cloud data based on a surface mesh model and a point cloud generation algorithm;

[0100] The closest distance measurement unit is used to determine the closest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding distance between the maxillary posterior teeth and the nearest root apex point based on the point cloud data;

[0101] The position relationship determination and classification unit is used to determine the position relationship between the maxillary sinus and the maxillary posterior teeth based on the closest distance and to classify the position relationship.

[0102] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 3As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth is implemented.

[0103] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application and does not constitute a limitation on the computer device to which the solution of the present application is applied. A specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the above-mentioned method embodiments when executing the computer program.

[0104] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0105] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0106] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0107] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0108] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0109] In this application, all actions to obtain signals, information or data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0110] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for determining the distance between the maxillary sinus and the maxillary posterior teeth in CBCT images, characterized in that: The method for determining the distance between the maxillary sinus and the maxillary posterior teeth based on CBCT images comprises: Obtain CBCT images of the maxillary sinus and maxillary posterior teeth; Based on the CBCT images of the maxillary sinus and the maxillary posterior teeth, a structure segmentation model is used to perform three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth; the structure segmentation model is constructed based on a deep learning network; According to the 3D segmentation results, a surface mesh model is generated based on the marching cube algorithm; According to the surface mesh model, point cloud data is determined based on the point cloud generation algorithm; Based on the point cloud data, the closest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding distance between the maxillary posterior teeth and the nearest root apex are determined; The positional relationship between the maxillary sinus and the maxillary posterior teeth was determined based on the closest distance, and the positional relationship was classified.

2. The method for determining the distance between the maxillary sinus and the maxillary posterior teeth according to claim 1, characterized in that: The training process of the structure segmentation model specifically includes: Constructing a data set; the data set includes: CBCT images and corresponding annotation results; the annotation results include: maxillary sinus, maxillary first premolar, maxillary second premolar, maxillary first molar and maxillary second molar; Sampling the CBCT images in the data set to a set voxel spacing and performing preprocessing; the preprocessing includes normalization and data enhancement; According to the processed data set, the deep learning network is trained using a three-fold cross-validation training strategy to obtain a structural segmentation model.

3. The method for determining the distance between the maxillary sinus and the maxillary posterior teeth based on CBCT images according to claim 1 or claim 2, wherein: The deep learning network includes: a U-shaped network of a convolutional block attention module, a SegNet model, a DeepLab model, and a Transformer model.

4. The method for determining the distance between the maxillary sinus and the maxillary posterior teeth in CBCT images according to claim 1, characterized in that: Determining point cloud data based on the surface mesh model and the point cloud generation algorithm specifically includes: The surface mesh model was scaled using the voxel spacing of the CBCT image; Save the scaled surface mesh model as an STL format file; According to the saved file, the point cloud generation algorithm is used to obtain point cloud data.

5. The method for determining the distance between the maxillary sinus and the maxillary posterior teeth based on CBCT images according to claim 1 or claim 4, wherein: The point cloud generation algorithm includes Poisson disk sampling.

6. The method for determining the distance between the maxillary sinus and the maxillary posterior teeth in CBCT images according to claim 1, characterized in that: The positional relationship between the maxillary sinus and the maxillary posterior teeth is determined based on the closest distance, and the positional relationship is classified, including: Determine the positional relationship between the maxillary sinus and maxillary posterior teeth based on the closest distance; Get the position relationship demarcation point; Determine whether the positional relationship between the maxillary sinus and the maxillary posterior teeth is greater than the positional relationship cutoff point, and obtain a comparison result; The position relationship classification result is determined based on the comparison result.

7. A device for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images, characterized in that: The device for determining the distance between the maxillary sinus and the maxillary posterior teeth using CBCT images comprises: An image acquisition unit, used for acquiring CBCT images of the maxillary sinus and maxillary posterior teeth; a three-dimensional segmentation unit for performing three-dimensional segmentation of the maxillary sinus and the maxillary posterior teeth based on CBCT images of the maxillary sinus and the maxillary posterior teeth using a structural segmentation model constructed based on a deep learning network; A surface mesh model generating unit, configured to generate a surface mesh model based on a marching cubes algorithm according to the three-dimensional segmentation result; A point cloud data determination unit, configured to determine point cloud data based on a surface mesh model and a point cloud generation algorithm; The closest distance measurement unit is used to determine the closest distance between the maxillary sinus and the maxillary posterior teeth and the corresponding distance between the maxillary posterior teeth and the nearest root apex point based on the point cloud data; The position relationship determination and classification unit is used to determine the position relationship between the maxillary sinus and the maxillary posterior teeth based on the closest distance and to classify the position relationship.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for determining the CBCT image distance between the maxillary sinus and the maxillary posterior teeth according to any one of claims 1 to 6 is implemented.

Images