Precise quantitative evaluation method of periodontal disease based on ios image and cbct image

Abstract

The present application relates to the technical field of dental diagnosis, and particularly relates to a periodontal disease precise quantitative evaluation method based on IOS images and CBCT images, comprising the following steps: performing image segmentation on an intraoral scanning IOS image to obtain an IOS segmentation result of each tooth; performing image segmentation on a CBCT image to obtain a CBCT segmentation result of each tooth; performing preprocessing and registration on the IOS segmentation result and the CBCT segmentation result to obtain a data fusion model; obtaining a gum contour and a tooth long axis according to the data fusion model, obtaining a gum contour point according to the gum contour, obtaining a alveolar bone contour on the alveolar bone along the direction of the tooth long axis and obtaining a GBD distance according to the gum contour point; and evaluating periodontal disease according to the GBD distance. According to the present application, the data fusion model is obtained according to the segmented IOS result and the CBCT result, the gum contour and the tooth long axis are obtained through the data fusion model, the GBD distance of the gum contour and the alveolar bone contour is measured along the direction of the tooth long axis, and the periodontal disease health condition is evaluated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of dental diagnosis, and particularly relates to a periodontal disease precise quantitative evaluation method, system and device based on IOS images and CBCT images and a storage medium. BACKGROUND

[0002] Periodontal disease is a global health concern, characterized by progressive and largely irreversible loss of gingiva and alveolar bone, ultimately leading to tooth loss. The Global Oral Health Report shows that about 19% of individuals over the age of 15 have severe periodontal disease, with a total of more than 1 billion cases worldwide, and the number is increasing year by year. Therefore, timely and accurate assessment is crucial for effective treatment intervention. Clinically, periodontal pocket depth measured by periodontal probe is the standard method for quantitative assessment of disease severity. Shallow periodontal pocket depth usually indicates healthier gums, while deeper periodontal pocket usually indicates inflammation and potential periodontal disease. However, this probing technique, although standard, is subjective, prone to human error, invasive, and only measures the values of six key points to measure the degree of tooth health. Therefore, developing an accurate, efficient and non-invasive assessment system is crucial for effective management of periodontal disease.

[0003] Recently, with the development of medical imaging technology, some digital dental technologies (such as two-dimensional panoramic X-ray imaging, cone beam computed tomography (CBCT) imaging and intraoral scanning (IOS) technology) have become important tools to assist in periodontal assessment. Among the above technologies, X-ray imaging is designed based on the principle of tomography, and the obtained X-ray image can only provide two-dimensional projection information of teeth and bones; CBCT imaging scans the target area in all directions, and the obtained CBCT image can provide three-dimensional information of teeth and alveolar bone; IOS technology is to superimpose the data collected at different positions in the mouth, and finally form a complete three-dimensional IOS image, providing high-resolution crown (the part of the tooth exposed outside the gum) and gum information.

[0004] With the development of computer technology, many existing researches use deep learning technology based on dental images (e.g., two-dimensional panoramic X-ray images and cone beam computed tomography (CBCT) images) to evaluate periodontal health. Among them, some researchers detect radiological bone level and cement-enamel junction (CEJ) level from X-ray images respectively, then measure the distance of key points in the two levels along the long axis of the tooth, thereby performing percentage analysis of radiological bone loss (RBL) and staging. In addition, some researchers discuss the role of three-dimensional CBCT images in periodontal disease evaluation. They use commercial software to semi-automatically export 3D models of teeth and bones, and use manual measurement method to measure the distance of bone level to determine the degree of periodontitis. Although these methods provide a feasible strategy for periodontal disease evaluation, they still have the following limitations. First, the information provided by the images used by these methods is insufficient, and the usefulness of gingival information for early periodontal disease evaluation is not considered. Among them, X-ray images reflect the overlapping information of tooth bone and other structures, which is distorted and difficult to accurately determine the three-dimensional position and shape information of the lesion; CBCT images can provide complete three-dimensional bone information, but cannot reflect the health status of the gums. Second, these methods evaluate the degree of alveolar bone loss, and only locate a single tooth and calculate the distance of a few points, which cannot fully show the degree of periodontal disease of each tooth, especially the degree of gingival loss in the early stage. Finally, the measurement steps of these methods require additional image key point labeling, and quantitative measurement still heavily relies on the professional knowledge of dentists, and cannot automatically provide quantitative results for periodontal evaluation. SUMMARY

[0005] The purpose of the present application is to solve the problems existing in the prior art, and to provide a precise quantitative evaluation method for periodontal disease based on IOS images and CBCT images, comprising the following steps:

[0006] Image segmentation is performed on the intraoral scanning IOS image to obtain IOS segmentation results for each tooth;

[0007] Image segmentation is performed on the CBCT image to obtain CBCT segmentation results for each tooth;

[0008] The IOS segmentation results and the CBCT segmentation results are preprocessed and registered to obtain a data fusion model;

[0009] The gingival contour and the tooth long axis are obtained according to the data fusion model, the gingival contour point is obtained according to the gingival contour, the alveolar bone contour is obtained on the alveolar bone along the direction of the tooth long axis, and the GBD distance is obtained;

[0010] The periodontal disease is evaluated according to the GBD distance.

[0011] Preferably, the intraoral scan IOS image is image segmented to obtain an IOS segmentation result of each tooth, comprising:

[0012] The intraoral scan IOS image is down-sampled to obtain sampling points, and the intraoral scan IOS images in different directions are rotated to a unified position to obtain a unified IOS image;

[0013] According to the unified IOS image, tooth center data is obtained from the sampling points, and according to the tooth center data and the unified IOS image, a tooth image block of each tooth is generated;

[0014] According to the unified IOS image, gingival label data is obtained, and according to the gingival label data, gingival average curvature information of the tooth image block is calculated to obtain tooth and gingival boundary information;

[0015] The tooth image block and the tooth and gingival boundary information are segmented into IOS segmented teeth and gingiva according to a coronal segmentation network, and the IOS segmented teeth are corrected to obtain a corrected tooth result.

[0016] Preferably, the CBCT image is image segmented to obtain a CBCT segmentation result of each tooth, further comprising:

[0017] A second label classification network outputs a predicted tooth label and a first predicted background label according to the CBCT image;

[0018] According to the CBCT image, a tooth region of interest ROI is obtained, and a third label classification network outputs a predicted tooth contour, a predicted tooth content, and a second predicted background label from the tooth region of interest ROI;

[0019] Each classified tooth is classified according to the tooth label, the first predicted background label, the predicted tooth contour, the predicted tooth content, and the second predicted background label;

[0020] A tooth label classification network obtains an FD I label of each tooth from the tooth region of interest ROI;

[0021] According to the FD I label, each classified tooth is given a correct tooth label.

[0022] Preferably, the IOS result and the CBCT result are preprocessed and registered to obtain a data fusion model, further comprising:

[0023] According to a moving cube algorithm, a tooth and alveolar bone mesh is reconstructed from the CBCT segmentation result to obtain the reconstructed CBCT segmentation result;

[0024] extracting a principal component direction of the reconstructed CBCT segmentation result and the IOS segmentation result respectively by using principal component analysis, and integrating CBCT center point data, calculating transformation parameters between the center point data of the reconstructed CBCT segmentation result and the IOS segmentation result by using singular value decomposition, so as to complete coarse registration by aligning the positions and directions of the dental arches;

[0025] On the basis of the coarse registration, point-to-surface registration is performed on the reconstructed CBCT segmentation result and the IOS segmentation result for each tooth by using an iterative closest point matching registration algorithm, so as to complete fine registration, thereby realizing the unification of the IOS segmentation result and the reconstructed CBCT segmentation result to the same position and obtaining a data fusion model.

[0026] Preferably, the direction of the tooth long axis obtained along the data fusion model according to the gingival contour point on the data fusion model obtains a alveolar bone contour on the alveolar bone and acquires a GBD distance, further comprising:

[0027] The gingival contour is obtained according to the vertex on the edge of the crown mesh of the data fusion model, and the number of gingival contour points is specified according to the gingival contour, and the gingival contour points are obtained by interpolation method;

[0028] The tooth long axis is generated from the tooth vertex data of the data fusion model by the principal component analysis;

[0029] The alveolar bone contour is obtained on the alveolar bone according to the direction of the tooth long axis along the gingival contour point, so as to obtain the GBD distance.

[0030] Preferably, the periodontal disease is evaluated according to the GBD distance, further comprising:

[0031] The degree of tooth health is judged according to the set threshold and the GBD distance, wherein the degree of tooth health is divided into grades including healthy, mild periodontal disease, moderate periodontal disease and severe periodontal disease.

[0032] Preferably, judging the degree of tooth health according to the set threshold and the GBD distance further comprises:

[0033] If the GBD distance is within the first threshold, it is healthy;

[0034] If the GBD distance is between the first threshold and the second threshold, it is mild periodontal disease;

[0035] If the GBD distance is between the second threshold and the third threshold, it is moderate periodontal disease;

[0036] If the GBD distance is greater than the third threshold, it is severe periodontal disease.

[0037] The application also provides a periodontal disease precise quantitative evaluation device based on IOS images and CBCT images, comprising:

[0038] An IOS image segmentation module is configured to perform image segmentation on the IOS image to obtain IOS segmentation results of each tooth.

[0039] A CBCT image segmentation module is configured to perform image segmentation on the CBCT image to obtain CBCT segmentation results of each tooth.

[0040] A multi-modal data fusion module is configured to perform format matching and registration on the IOS results and the CBCT results to obtain a data fusion model.

[0041] A GBD measurement and quantitative evaluation module is configured to obtain a gum profile and a tooth long axis according to the data fusion model, obtain a gum profile point according to the gum profile, obtain a GBD distance by obtaining a alveolar bone profile on the alveolar bone along a direction of the tooth long axis and obtaining the GBD distance, and evaluate periodontal disease according to the GBD distance.

[0042] The application also provides a computer device comprising a memory and a processor, wherein the memory stores computer readable instructions, and the computer readable instructions are executed by the processor to enable the processor to perform steps of the periodontal disease precise quantitative evaluation method based on IOS images and CBCT images as described in the embodiments.

[0043] The application also provides a storage medium storing computer readable instructions, wherein the computer readable instructions are executed by one or more processors to enable the one or more processors to perform steps of the periodontal disease precise quantitative evaluation method based on IOS images and CBCT images as described in the embodiments.

[0044] Compared with the prior art, the application has the following beneficial effects:

[0045] 1) The application provides complete and high-precision crown, tooth and alveolar bone data by combining three-dimensional IOS images and CBCT images to obtain a data fusion model comprising aligned crowns, teeth and alveolar bones.

[0046] 2) The application measures the GBD distance of the gum profile and the alveolar bone profile along the tooth long axis direction in a three-dimensional digital space to evaluate the periodontal disease health status, and reduces the error of the evaluation method of measuring a small number of points in the clinic and the prior art by customizing the number of measured points. BRIEF DESCRIPTION OF DRAWINGS

[0047] Various other advantages and benefits will become apparent to those of ordinary skill in the art, upon reading the following detailed description of the preferred embodiment. The accompanying drawings are included to provide a description of preferred embodiments and are not meant to limit the present application.

[0048] Figure 1 Flow chart of the periodontal disease precise quantitative evaluation method based on IOS image and CBCT image of the present application;

[0049] Figure 2 Flow chart of the periodontal disease precise quantitative evaluation method based on IOS image and CBCT image of the present application;

[0050] Figure 3 Structural schematic diagram of IOS image segmentation of the present application;

[0051] Figure 4 Structural schematic diagram of CBCT image segmentation of the present application;

[0052] Figure 5 Structural schematic diagram of GBD measurement and quantitative evaluation of the present application;

[0053] Figure 6 Structural schematic diagram of GBD measurement and quantitative evaluation of the present application; DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0055] Those skilled in the art can understand that, unless specifically stated otherwise, the singular form "a", "an", "the" used herein can also include the plural form. It should be further understood that the use of the word "comprise" in the specification of the present application means that the features, integers, steps, operations, elements and / or components exist, but do not exclude the existence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0056] Embodiment 1

[0057] Referring to Figure 1 and Figure 2 The periodontal disease precise quantitative evaluation method based on IOS image and CBCT image provided by the present embodiment includes the following steps:

[0058] The IOS image is image segmented to obtain IOS segmentation results of each tooth, and specifically, in this embodiment, the points of the input IOS image containing teeth are down-sampled, and after processing, the segmented tooth crown and gingiva are obtained.

[0059] Please refer to Figure 3 The IOS image is image segmented to obtain IOS segmentation results of each tooth, and specifically, in this embodiment, the points of the input IOS image containing teeth are down-sampled, and after processing, the segmented tooth crown and gingiva are obtained.

[0060] The IOS image is down-sampled to obtain sampling points, and IOS images in different directions are rotated to a unified position to obtain a unified IOS image, and specifically, in this embodiment, the model rotation network (Rotation Network) in the Pointnet++ network is used to rotate the IOS image to the unified position, and in other embodiments, other networks can be used to rotate the IOS image to the unified position.

[0061] According to the tooth center data obtained from the sampling points and the unified IOS image, the tooth image block of each tooth is generated, and specifically, in this embodiment, the centroid prediction network (Centroid Prediction Network) in the Pointnet++ network is used to obtain the tooth center point data of the IOS image rotated to the unified position, and in other embodiments, other networks can be used to obtain the tooth center point data of the IOS image rotated to the unified position.

[0062] According to the gingiva label data obtained from the unified IOS image, the gingiva average curvature information of the tooth image block is calculated to obtain the tooth and gingiva boundary information, and specifically, in this embodiment, the gingiva separation network (Gingival Separation Network) in the Pointnet++ network is used to obtain the gingiva label data of the IOS image rotated to the unified position, and in other embodiments, other networks can be used to obtain the gingiva label data of the IOS image rotated to the unified position.

[0063] The tooth image block and the tooth and gingiva boundary information are segmented by the crown segmentation network to obtain the IOS segmented tooth and gingiva, and the IOS segmented tooth is corrected to obtain the corrected tooth result, and specifically, in this embodiment, the crown segmentation network in the Pointnet++ network is used to segment 32 kinds of teeth and 1 kind of gingiva, and in other embodiments, other networks can be used to segment 32 kinds of teeth and 1 kind of gingiva, and the IOS segmented tooth label is output.

[0064] More preferably, the IOS segmented teeth are corrected to obtain a corrected tooth result, further comprising:

[0065] To handle the case that there may be FDI label errors in the correction result, a label correction submodule (Label Correction) is used to correct the segmentation result of each tooth crown (Crown) according to a cost ranking algorithm and a set of dental arch curves, so as to minimize the problems of label reversal and the same label.

[0066] The present embodiment can accurately segment all tooth crowns of a patient by segmenting an IOS image, and number and classify them according to the FDI system proposed by the International Dental Federation, so as to obtain a unique identification of each tooth crown.

[0067] Referring to Figure 4 The CBCT image is segmented to obtain a CBCT segmentation result of each tooth.

[0068] Preferably, the CBCT image is segmented to obtain a CBCT segmentation result of each tooth, further comprising:

[0069] A second label classification network (Tooth Region Segmentation Network) outputs a predicted tooth label and a first predicted background label according to the CBCT image. Specifically, in the present embodiment, the second label classification network uses an encoder and a decoder in the nnUNet network.

[0070] A tooth region of interest (ROI) is obtained from the CBCT image, and a third label classification network (Tooth Boundary and Interior Segmentation Network) outputs a predicted tooth contour, a predicted tooth content, and a second predicted background label from the tooth region of interest (ROI). Specifically, in the present embodiment, the third label classification network uses an encoder and a decoder in the nnUNet network to locate the tooth region of interest (ROI) and segment the predicted tooth contour, the predicted tooth content, and the second predicted background label based on the tooth region of interest (ROI). Because the calculation amount of tooth segmentation directly on the CBCT three-dimensional image is large, the tooth region of interest (ROI) is used to exclude irrelevant regions and only keep the tooth region of interest, so as to reduce the calculation amount of subsequent segmentation.

[0071] According to the tooth label, the first predicted background label, the predicted tooth contour, the predicted tooth content and the second predicted background label, each classified tooth is classified, specifically, in the embodiment, the tooth label, the first predicted background label, the predicted tooth contour, the predicted tooth content and the second predicted background label are decoded by using a decoder in the nnUNet network, and the results of the tooth two-label classification and the tooth three-label classification are integrated by using a watershed algorithm to obtain the complete shape of each tooth, and the alveolar bone is segmented at the same time;

[0072] The tooth labeling network obtains the FDI label of each tooth from the tooth region of interest (ROI);

[0073] According to the FDI label, each classified tooth is given a correct tooth label, specifically, in the embodiment, the probability of the label of the overlapping part of each tooth in the result of the watershed algorithm and the result of the tooth labeling network is calculated by using a voting strategy, and the final tooth label is obtained based on the probability voting, but the implementation method is not limited to this voting mechanism.

[0074] The embodiment can accurately segment all teeth of a patient by segmenting the CBCT image, and obtain a standard FDI number, and the alveolar bone can be segmented at the same time.

[0075] The IOS segmentation result and the CBCT segmentation result are preprocessed and registered to obtain a data fusion model, specifically, in the embodiment, the aligned tooth crown, tooth and alveolar bone are obtained according to the IOS segmentation result and the CBCT segmentation result.

[0076] Preferably, the IOS result and the CBCT result are preprocessed and registered to obtain a data fusion model, further comprising:

[0077] According to the marching cubes algorithm, the tooth and alveolar bone grid are reconstructed from the CBCT segmentation result to obtain a reconstructed CBCT segmentation result;

[0078] The principal component directions of the reconstructed CBCT segmentation result and the IOS segmentation result are extracted by using principal component analysis respectively, the center point data of the reconstructed CBCT segmentation result and the IOS segmentation result are integrated, the transformation parameters between the center point data of the reconstructed CBCT segmentation result and the IOS segmentation result are calculated by using singular value decomposition (SVD), and the coarse registration is completed by aligning the position and direction of the dental arch, specifically, in the embodiment, the principal component directions of the reconstructed CBCT segmentation result and the IOS segmentation result are extracted by using principal component analysis (PCA) respectively, and in the embodiment, three principal component directions are extracted.

[0079] On the basis of the coarse registration, the point-to-surface registration algorithm of the iterative closest point (ICP) is used to register the reconstructed CBCT segmentation result and the IOS segmentation result for each tooth to complete the fine registration, so as to realize the unification of the IOS segmentation result and the reconstructed CBCT segmentation result to the same position, obtain the data fusion model, and further obtain the dental arch curve according to the data fusion model.

[0080] Referring to Figure 5 As shown in the figure, the gingival contour and the tooth’s longtudinal axis are obtained according to the data fusion model, the gingival contour points are obtained according to the gingival contour, the alveolar bone contour is obtained on the alveolar bone along the direction of the tooth’s longtudinal axis, and the GBD (gingival-bone) distance is obtained. Specifically, in this embodiment, the gingival contour is obtained according to the crown of the data fusion model, and the tooth’s longtudinal axis and the alveolar bone contour are obtained according to the tooth mesh data and the alveolar bone mesh data of the data fusion model.

[0081] Preferably, the alveolar bone contour is obtained on the alveolar bone along the direction of the tooth’s longtudinal axis obtained according to the gingival contour points on the data fusion model, and the GBD distance is obtained, further comprising:

[0082] The gingival contour is obtained according to the vertexes on the edge of the crown mesh of the data fusion model, and the number of gingival contour points is specified according to the gingival contour, and the gingival contour points are obtained by interpolation method;

[0083] The tooth’s longtudinal axis is generated from the tooth vertex data of the data fusion model by principal component analysis, specifically, in this embodiment, the principal component analysis (PCA) is used to identify the main features with the largest amount of information from the high-dimensional tooth vertex data, and the tooth’s longtudinal axis is generated to describe the geometric shape thereof, and in other embodiments, other methods can be used to obtain the tooth’s longtudinal axis;

[0084] The alveolar bone contour is obtained on the alveolar bone along the direction of the tooth’s longtudinal axis according to the gingival contour points, so as to obtain the GBD distance, specifically, in this embodiment, the alveolar bone contour points are obtained by measuring the gingival contour points along the direction of the tooth’s longtudinal axis, and the GBD distance between the corresponding points of the gingival contour points and the alveolar bone contour points is obtained.

[0085] According to the obtained gingival contour and alveolar bone contour, the gingival-bone GBD distance is measured along the tooth’s longtudinal axis representing the tooth direction information, and the accurate quantitative evaluation result of periodontal disease is obtained according to the GBD value and the set threshold value.

[0086] The periodontal disease is evaluated according to the GBD distance, and specifically, in the present embodiment, the average GBD distance of all points of the gingival contour points and the alveolar bone contour points is compared with a set threshold value.

[0087] The periodontal disease is evaluated according to the GBD distance, and further comprising:

[0088] The degree of tooth health is determined according to the set threshold value and the GBD distance, wherein the degree of tooth health is classified into levels including healthy, mild periodontal disease, moderate periodontal disease, and severe periodontal disease.

[0089] Referring to Figure 6 As shown, the determination of the degree of tooth health according to the set threshold value (settable classification threshold) and the GBD distance further comprises:

[0090] If the GBD distance is within the first threshold value, it is healthy (Hea lthy), and specifically, in the present embodiment, the first threshold value is 2 mm, and when the GBD distance (GBD Results) is less than 2 mm, the tooth is in a healthy state;

[0091] If the GBD distance is between the first threshold value and the second threshold value, it is mild periodontal disease (M i l d Per i odontal D i sease), and specifically, in the present embodiment, the second threshold value is 4 mm, and when the GBD distance is between 2 mm and 4 mm, the tooth is in a mild periodontal disease, i.e., when the tooth label is 47 and the GBD distance is 3.14 mm, the degree of tooth health is mild periodontal disease;

[0092] If the GBD distance is between the second threshold value and the third threshold value, it is moderate periodontal disease (Moderate Per i odontal D i sease), and specifically, in the present embodiment, the third threshold value is 6 mm, and when the GBD distance is between 4 mm and 6 mm, the tooth is in a moderate periodontal disease, i.e., when the tooth label is 11 and the GBD distance is 5.14 mm, the degree of tooth health is moderate periodontal disease, and when the tooth label is 12 and the GBD distance is 5.40 mm, the degree of tooth health is moderate periodontal disease;

[0093] If the GBD distance is greater than the third threshold value, it is severe periodontal disease (Server Per i odontal D i sease), and specifically, in the present embodiment, when the GBD distance is greater than 6 mm, the tooth is in a severe periodontal disease, i.e., when the tooth label is 13 and the GBD distance is 6.40 mm, the degree of tooth health is severe periodontal disease.

[0094] Embodiment 2

[0095] The embodiment provides a periodontal disease precision quantitative evaluation system based on IOS images and CBCT images, which is used for executing the method in the above embodiment 1, and comprises the following steps of:

[0096] An IOS image segmentation module is configured to perform image segmentation on the IOS image of the intraoral scan to obtain an IOS segmentation result of each tooth.

[0097] A CBCT image segmentation module is configured to perform image segmentation on the CBCT image to obtain a CBCT segmentation result of each tooth.

[0098] A multi-modal data fusion module is configured to perform format matching and registration on the IOS result and the CBCT result to obtain a data fusion model.

[0099] A GBD measurement and quantitative evaluation module is configured to obtain a gum contour and a tooth long axis according to the data fusion model, obtain a gum contour point according to the gum contour, obtain a jawbone contour on the jawbone along the direction of the tooth long axis and obtain the GBD distance according to the gum contour point, and evaluate periodontal disease according to the GBD distance.

[0100] Embodiment 3

[0101] In some embodiments of the present application, a computer device is also provided, comprising a memory and a processor, wherein the memory stores computer readable instructions, and the computer readable instructions are executed by the processor to enable the processor to perform the steps of the periodontal disease precision quantitative evaluation method based on IOS images and CBCT images according to any one of the above embodiments.

[0102] The present application also provides a storage medium storing computer readable instructions, wherein the computer readable instructions are executed by one or more processors to enable the one or more processors to perform the steps of the periodontal disease precision quantitative evaluation method based on IOS images and CBCT images according to any one of the above embodiments.

[0103] It can be understood that, for the aforementioned periodontal disease precision quantitative evaluation method based on IOS images and CBCT images, if they are all implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer server, a network device, etc.) to execute all or part of the steps of the various embodiment methods of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0104] The computer readable storage medium can include a data signal carried in the baseband or as a part of a carrier wave propagating through the program code. Such a propagating data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium, which can send, propagate or transmit programs for use by or in conjunction with an instruction execution system, device or apparatus. The program code contained on the readable storage medium can be transmitted by any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0105] The above is only the preferred embodiment of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solution falling within the scope of the present application should be considered as falling within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application should also be considered as falling within the protection scope of the present application.

Claims

1. A precise quantitative assessment method for periodontal disease based on ISO images and CBCT images, characterized in that, Includes the following steps: Image segmentation was performed on the intraoral scan IOS images to obtain the IOS segmentation results for each tooth; Image segmentation was performed on the CBCT images to obtain the CBCT segmentation results for each tooth; The IOS segmentation results and the CBCT segmentation results are preprocessed and registered to obtain a data fusion model; The gingival contour and tooth long axis are obtained according to the data fusion model. Gingival contour points are obtained according to the gingival contour. The gingival contour points are used to obtain the alveolar bone contour along the direction of the tooth long axis and the GBD distance is obtained. Periodontal disease is assessed based on the GBD distance; The process of preprocessing and registering the IOS results and the CBCT results to obtain a data fusion model further includes: The reconstructed CBCT segmentation results are obtained by reconstructing the tooth and alveolar bone mesh from the CBCT segmentation results using the moving cube algorithm. Principal component analysis is used to separate the principal component directions of the reconstructed CBCT segmentation results and the IOS segmentation results, and the CBCT center point data is integrated. Singular value decomposition is used to calculate the transformation parameters between the center point data of the reconstructed CBCT segmentation results and the IOS segmentation results, so as to complete the coarse registration by aligning the dental arch position and dental arch direction. Based on the coarse registration, an iterative nearest point matching registration algorithm is used for each tooth to perform point-to-surface registration on the reconstructed CBCT segmentation results and the IOS segmentation results to complete the fine registration, so as to unify the IOS segmentation results and the reconstructed CBCT segmentation results to the same position and obtain the data fusion model. Based on the gingival contour points on the data fusion model, the alveolar bone contour is obtained along the direction of the tooth long axis obtained from the data fusion model, and the GBD distance is acquired. Further steps include: The gingival contour is obtained from the vertices on the edge of the crown mesh of the data fusion model, and the number of gingival contour points is specified according to the gingival contour. The gingival contour points are then obtained by interpolation. The tooth long axis is generated from the tooth vertex data of the data fusion model through the principal component analysis. The alveolar bone contour is obtained on the alveolar bone based on the gingival contour point along the direction of the long axis of the tooth, so as to obtain the GBD distance.

2. The method for precise quantitative assessment of periodontal disease based on ISO images and CBCT images according to claim 1, characterized in that, Image segmentation was performed on the intraoral scan IOS images to obtain the IOS segmentation results for each tooth, including: The intraoral scanning IOS image is downsampled to obtain sampling points, and the intraoral scanning IOS images in different directions are rotated to a unified position to obtain a unified IOS image; Based on the unified IOS image, tooth center data is obtained from the sampling points, and tooth image blocks for each tooth are generated based on the tooth center data and the unified IOS image. Based on the unified ISO image, gingival label data is obtained, and based on the gingival label data, the average curvature information of the gingival margin of the tooth image block is calculated to obtain the tooth-gingival boundary information. The tooth image block and the tooth-gingival boundary information are segmented into IOS segmented teeth and gingiva using a coronal segmentation network, and the IOS segmented teeth are corrected to obtain corrected tooth results.

3. The method for precise quantitative assessment of periodontal disease based on ISO images and CBCT images according to claim 1, characterized in that, Image segmentation is performed on the CBCT images to obtain the CBCT segmentation results for each tooth, further including: The second label classification network outputs predicted tooth labels and a first predicted background label based on the CBCT image; Based on the CBCT image, the region of interest (ROI) of the tooth is obtained, and the third label classification network outputs a predicted tooth outline, a predicted tooth contents, and a second predicted background label from the ROI. Each tooth is classified according to the tooth label, the first predicted background label, the predicted tooth outline, the predicted tooth contents, and the second predicted background label; The tooth label classification network obtains the FDI label for each tooth from the region of interest (ROI). Each classified tooth is assigned the correct tooth label according to the FDI label.

4. The method for precise quantitative assessment of periodontal disease based on ISO images and CBCT images according to claim 1, characterized in that, Assessing periodontal disease based on the GBD distance further includes: The degree of dental health is determined based on a set threshold and the GBD distance, wherein the degree of dental health is divided into levels including healthy, mild periodontitis, moderate periodontitis and severe periodontitis.

5. The method for precise quantitative assessment of periodontal disease based on ISO images and CBCT images according to claim 4, characterized in that, Determining the degree of dental health based on a set threshold and the GBD distance further includes: If the GBD distance is within the first threshold, then it is considered healthy; If the GBD distance is between the first threshold and the second threshold, it indicates mild periodontitis. If the GBD distance is between the second threshold and the third threshold, it is considered moderate periodontitis. If the GBD distance is greater than the third threshold, it is considered severe periodontitis.

6. A precise quantitative assessment system for periodontal disease based on ISO images and CBCT images, characterized in that, include: The IOS image segmentation module is used to segment intraoral scan IOS images to obtain IOS segmentation results for each tooth. The CBCT image segmentation module is used to segment CBCT images to obtain the CBCT segmentation results for each tooth. A multimodal data fusion module is used to perform format matching and registration of the IOS results and the CBCT results to obtain a data fusion model; The GBD measurement and quantitative assessment module is used to obtain the gingival contour and tooth long axis according to the data fusion model, obtain gingival contour points according to the gingival contour, obtain alveolar bone contour along the direction of the tooth long axis and obtain the GBD distance, and assess periodontal disease according to the GBD distance. The process of preprocessing and registering the IOS results and the CBCT results to obtain a data fusion model further includes: The reconstructed CBCT segmentation results are obtained by reconstructing the tooth and alveolar bone mesh from the CBCT segmentation results using the moving cube algorithm. Principal component analysis is used to separate the principal component directions of the reconstructed CBCT segmentation results and the IOS segmentation results, and the CBCT center point data is integrated. Singular value decomposition is used to calculate the transformation parameters between the center point data of the reconstructed CBCT segmentation results and the IOS segmentation results, so as to complete the coarse registration by aligning the dental arch position and dental arch direction. Based on the coarse registration, an iterative nearest point matching registration algorithm is used for each tooth to perform point-to-surface registration on the reconstructed CBCT segmentation results and the IOS segmentation results to complete the fine registration, so as to unify the IOS segmentation results and the reconstructed CBCT segmentation results to the same position and obtain the data fusion model. Based on the gingival contour points on the data fusion model, the alveolar bone contour is obtained along the direction of the tooth long axis obtained from the data fusion model, and the GBD distance is acquired. Further steps include: The gingival contour is obtained from the vertices on the edge of the crown mesh of the data fusion model, and the number of gingival contour points is specified according to the gingival contour. The gingival contour points are then obtained by interpolation. The tooth long axis is generated from the tooth vertex data of the data fusion model through the principal component analysis. The alveolar bone contour is obtained on the alveolar bone based on the gingival contour point along the direction of the long axis of the tooth, so as to obtain the GBD distance.

7. A computer device, characterized in that, The device includes a memory and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, cause the processor to perform the steps of the method for precise quantitative assessment of periodontal disease based on IOS images and CBCT images as described in any one of claims 1 to 5.

8. A storage medium storing computer-readable instructions, characterized in that, When the computer-readable instructions are executed by one or more processors, the one or more processors perform the steps of the method for precise quantitative assessment of periodontal disease based on IOS images and CBCT images as described in any one of claims 1 to 5.

Images