Remote diagnosis method and system for pathogenesis of tooth fixation

Abstract

The embodiment of the invention discloses a remote diagnosis method and system for a tooth fixation disease cause, belongs to the field of tooth fixation disease cause diagnosis, and solves the problem that an existing remote diagnosis method for tooth fixation is relatively low in accuracy. Remote diagnosis login information of the user is received, tooth fixation connection disease cause feature extraction is carried out on the illness state information and historical diagnosis information of the user, and needed historical case data are screened out based on weight values corresponding to the tooth fixation connection disease cause features respectively; key point matching is carried out on the tooth fixation disease condition area, and a user tooth fixation three-dimensional model is constructed based on a matching result; inputting the required historical case data and the three-dimensional model of the fixed tooth connection of the user into a preset fixed tooth connection pathogenesis prediction model to obtain predicted fixed tooth connection pathogenesis data; matching a corresponding tooth fixation connection treatment scheme based on the required historical case data and the predicted tooth fixation connection pathogenesis data; and performing encrypted storage on the predicted tooth fixation disease cause data and the tooth fixation treatment scheme, and calling encrypted data under the condition of receiving remote diagnosis login information of the user next time.

Description

Technical Field

[0001] The present application relates to the field of diagnosis of the cause of tooth fixation, and in particular to a remote diagnosis method and system for the cause of tooth fixation. Background Art

[0002] In the field of stomatology, tooth adhesion is a common clinical disease with complex and diverse causes, involving the interaction of multiple factors such as genetics, environment, local inflammation, trauma, physiology and metabolism. Accurately predicting the cause of tooth adhesion is crucial for developing a reasonable treatment plan and improving the treatment effect of patients.

[0003] The traditional diagnosis of the cause of tooth fixation mainly relies on the doctor's clinical experience and some basic examination methods, and requires the patient to go to the hospital for diagnosis. Therefore, it not only has great subjectivity and limitations, but also causes a waste of time.

[0004] With the development of information technology and computer-aided diagnosis technology, remote diagnosis technology for dental fixation has gradually been recognized by patients. However, in the existing technology, when conducting remote diagnosis of the cause of dental fixation, it is difficult to fully explore the characteristics of the cause of dental fixation hidden in the patient information and a large amount of historical case information. For the large amount of information provided by users, a lot of useful information is often ignored or not fully utilized, making it difficult to specifically screen out high-value case information and treatment plans for different users as references, resulting in a low accuracy rate of remote diagnosis of dental fixation. Summary of the invention

[0005] The embodiments of the present application provide a method and system for remote diagnosis of the cause of dental consolidation, which are used to solve the following technical problems: the existing remote diagnosis technology for dental consolidation makes it difficult to specifically screen out high-value case information and treatment plans for different users as references, resulting in a low accuracy rate of remote diagnosis of dental consolidation.

[0006] The present application embodiment adopts the following technical solutions:

[0007] The embodiment of the present application provides a method for remote diagnosis of the cause of tooth consolidation. The method comprises: receiving user remote diagnosis login information, obtaining user condition information and historical diagnosis information according to the login information; extracting tooth consolidation cause features from the user condition information and historical diagnosis information, and filtering out required historical case data based on the weight values ​​corresponding to each tooth consolidation cause feature; determining the tooth consolidation condition area based on the user condition information and historical diagnosis information, matching key points of the tooth consolidation condition area, and building a user tooth consolidation three-dimensional model based on the matching result; inputting the required historical case data and the user tooth consolidation three-dimensional model into a preset tooth consolidation cause prediction model to obtain predicted tooth consolidation cause data; matching the corresponding tooth consolidation treatment plan based on the required historical case data and the predicted tooth consolidation cause data; encrypting and storing the predicted tooth consolidation cause data and the tooth consolidation treatment plan, so that the encrypted data can be called when the user remote diagnosis login information is received next time.

[0008] The embodiment of the present application extracts the characteristics of the cause of tooth fixation from the user's medical condition information and historical diagnosis information, and refines complex medical information into key information related to the cause of tooth fixation, thereby improving the efficiency and accuracy of diagnosis. By determining the area of ​​tooth fixation, it helps doctors to clarify the specific location and scope of the patient's tooth fixation problem, so as to more accurately assess the severity and spatial distribution of the disease. By presetting the tooth fixation cause prediction model, the cause of tooth fixation of the current patient is predicted, avoiding the limitations of relying on the doctor's personal experience, and improving the scientificity and accuracy of the cause prediction. According to the required historical case data and the predicted tooth fixation cause data, the corresponding tooth fixation treatment plan is matched, so that the formulation of the treatment plan is more personalized, and the most suitable treatment means are provided for patients with different causes and conditions, thereby improving the effectiveness and success rate of the treatment. The predicted tooth fixation cause data and the tooth fixation treatment plan are encrypted and stored to protect the patient's privacy and sensitive medical information, prevent information leakage and illegal access, and ensure the security and confidentiality of patient information.

[0009] In one implementation of the present application, tooth fixation etiology features are extracted from user medical condition information and historical diagnosis information, and required historical case data are screened out based on the weight values ​​corresponding to each tooth fixation etiology feature, specifically including: determining multiple tooth fixation etiology features based on the user medical condition information; sorting multiple tooth fixation etiology features based on a preset tooth fixation etiology feature importance sequence, and based on the sorting order, sequentially screening out historical case data that meet the feature similarity threshold in the database to construct multiple historical case sets; sorting the case data in the multiple historical case sets according to the feature similarity, and assigning weights to each case data according to the sorting order; determining the feature scores corresponding to each case data based on the weights corresponding to the multiple historical case sets and the weights corresponding to each case data; and screening out the required historical case data based on the feature scores.

[0010] In one implementation of the present application, based on the weights corresponding to the multiple historical case sets and the weights corresponding to the case data, the feature scores corresponding to the case data are determined, specifically including: according to the function:

[0011]

[0012] Determine the characteristic scores corresponding to each case data; among them, is the feature score; H is the set of historical case sets; h ij For the historical case set H i The jth data point in ; wH is the weight set of the historical case set, where wH={wH1,wH2,...,wH n}; For each historical case set H i , the weight set of its data points is {wh i1 ,wh i2 ,...,wh im}; F = {f1,f2,...,f k}, where F is the feature set; is the data point h ij In the feature f l α is the adjustment factor, which is used to adjust the degree to which the feature score is affected by the weight of the historical case set and the weight of the data point, and its value range is [0,1]; n is the number of historical case sets; m p For the historical case set H p The number of data points in .

[0013] In one implementation of the present application, based on the user's medical condition information and historical diagnosis information, the tooth joint disease area is determined, and key points of the tooth joint disease area are matched to build a user's tooth joint three-dimensional model based on the matching result, specifically including: based on the user's medical condition information, building an initial tooth joint three-dimensional model corresponding to the user; extracting the disease area of ​​the initial tooth joint three-dimensional model to obtain a disease area image, matching the disease area image with a disease image database, and determining the disease type corresponding to the disease area based on the matching result; based on the edge detection algorithm, pixel point identification is performed on the disease area corresponding to each disease type, and the area of ​​the disease area corresponding to each disease type is determined based on the identified pixel set; based on the user information, a reference tooth three-dimensional model corresponding to the user is built; key points of the reference tooth three-dimensional model and the initial tooth joint three-dimensional model are matched, and based on the matching result, the area of ​​the disease area is adjusted to determine the user's tooth joint three-dimensional model based on the adjusted area of ​​the disease area.

[0014] In one implementation of the present application, key points of a reference tooth 3D model and an initial tooth-attached 3D model are matched, and based on the matching results, the area of ​​the diseased region is adjusted, specifically including: extracting key points of the initial tooth-attached model and the reference tooth 3D model respectively through a preset key point extraction model; generating descriptors from the extracted key points, and matching the key points based on the similarity between the descriptors; based on the matched key points, determining the deformation field from the reference tooth 3D model to the initial tooth-attached 3D model through thin plate spline interpolation; based on the deformation field, mapping the diseased region in the reference tooth 3D model to the initial tooth-attached 3D model, and adjusting the area of ​​the diseased region according to the deformation field.

[0015] In one implementation of the present application, the required historical case data and the user's three-dimensional model of tooth fixation are input into a preset tooth fixation etiology prediction model to obtain the predicted tooth fixation etiology data. The method also includes: determining a first information entropy based on a preset tooth fixation etiology prediction data set; dividing the preset tooth fixation etiology prediction data set based on each candidate tooth fixation etiology feature, and determining a second information entropy and information gain corresponding to the divided preset tooth fixation etiology prediction data set; randomly extracting data from the preset tooth fixation etiology prediction data set, and determining a Gini coefficient based on the extraction result; determining the optimal tooth fixation etiology feature according to the first information entropy, the second information entropy, the information gain and the Gini coefficient; starting from the root node, dividing the preset training set data according to the preset division criteria and the optimal tooth fixation etiology feature to obtain multiple subsets; wherein each subset corresponds to a child node; dividing the features of each child node by the optimal tooth fixation etiology feature corresponding to each child node until the stop condition is met, so as to construct a preset tooth fixation etiology prediction model.

[0016] In one implementation of the present application, before determining the first information entropy based on a preset tooth fixation etiology prediction data set, the method also includes: determining tooth fixation feature pairs based on the user's personal physical information and the required historical case data, and dividing the tooth fixation feature pairs to obtain feature pairs with interactive relationships and feature pairs without interactive relationships; performing linear combination processing on the feature pairs with interactive relationships, and performing mathematical transformation processing on the feature pairs without interactive relationships; performing feature fusion on the high-order features obtained after the processing to obtain a first high-order feature set; and performing dimensionality reduction processing on the first high-order feature set through principal component analysis to obtain a second high-order feature set.

[0017] In one implementation of the present application, matching corresponding tooth consolidation treatment plans based on required historical case data and predicted tooth consolidation etiology data specifically includes: building a case library based on the required historical case data and the treatment data corresponding to the required historical case data; performing a similarity match between the predicted tooth consolidation etiology data and the case library to obtain a first tooth consolidation treatment plan set; dynamically adjusting the quantity threshold based on the number of treatment plans in the first tooth consolidation treatment plan set; obtaining the user's historical case data, and performing a secondary similarity match in the first tooth consolidation treatment plan set based on the historical case data and the quantity threshold to obtain a second tooth consolidation treatment plan set; obtaining the user's basic information, building a personalized similarity model based on the user's basic information, inputting the treatment plans in the second tooth consolidation treatment plan set into the personalized similarity model, and matching the corresponding tooth consolidation treatment plans based on the output personalized similarity value.

[0018] In one implementation of the present application, the predicted cause of tooth consolidation data and the tooth consolidation treatment plan are encrypted and stored so that the encrypted data can be called when the user's remote diagnosis login information is received next time, specifically including: encrypting the predicted cause of tooth consolidation data and the tooth consolidation treatment plan through the SM3 encryption algorithm to obtain the first summary information; encrypting the first summary information through the SM2 algorithm and the private key to generate a digital signature ciphertext; generating an encrypted medical information package based on the encrypted predicted cause of tooth consolidation data, the tooth consolidation treatment plan and the digital signature ciphertext; re-determining the usage frequency and importance of the encrypted medical information package after a preset interval, and dividing the encrypted medical information package into different levels based on the determination result; storing the encrypted medical information packages of different levels in different storage media, and adjusting the encryption strength of the encrypted medical information package based on the difference in storage media; when receiving a user system login request, verifying the user's identity and authority, and calling the corresponding encrypted medical information package if the verification passes.

[0019] The embodiment of the present application provides a remote diagnosis system for the cause of tooth fixation, including: an information acquisition unit, which receives user remote diagnosis login information and acquires user condition information and historical diagnosis information according to the login information; a required historical case data screening unit, which extracts tooth fixation cause features from the user condition information and historical diagnosis information, and screens out required historical case data based on weight values ​​corresponding to each tooth fixation cause feature; a user tooth fixation three-dimensional model construction unit, which determines a tooth fixation condition area based on the user condition information and historical diagnosis information, matches key points of the tooth fixation condition area, and constructs a user tooth fixation three-dimensional model based on the matching result; a predicted tooth fixation cause data acquisition unit, which inputs the required historical case data and the user tooth fixation three-dimensional model into a preset tooth fixation cause prediction model to obtain predicted tooth fixation cause data; a treatment plan matching unit, which matches a corresponding tooth fixation treatment plan based on the required historical case data and the predicted tooth fixation cause data; and a data calling unit, which encrypts and stores the predicted tooth fixation cause data and the tooth fixation treatment plan, so that the encrypted data can be called when the user remote diagnosis login information is received next time.

[0020] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: The embodiments of the present application extract the characteristics of the cause of tooth fixation from the user's condition information and historical diagnosis information, and refine the complex medical information into key information related to the cause of tooth fixation, thereby improving the efficiency and accuracy of diagnosis. By determining the condition area of ​​tooth fixation, it helps doctors to clarify the specific location and scope of the patient's tooth fixation problem, so as to more accurately assess the severity and spatial distribution of the condition. By presetting the tooth fixation cause prediction model, the cause of tooth fixation of the current patient is predicted, avoiding the limitations of relying on the doctor's personal experience, and improving the scientificity and accuracy of the cause prediction. According to the required historical case data and the predicted tooth fixation cause data, the corresponding tooth fixation treatment plan is matched, so that the formulation of the treatment plan is more personalized, and the most suitable treatment means are provided for patients with different causes and conditions, thereby improving the effectiveness and success rate of the treatment. The predicted tooth fixation cause data and the tooth fixation treatment plan are encrypted and stored to protect the privacy and sensitive medical information of the patient, prevent information leakage and illegal access, and ensure the security and confidentiality of the patient's information. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. In the drawings:

[0022] Figure 1A flow chart of a remote diagnosis method for the cause of tooth fixation provided in an embodiment of the present application;

[0023] Figure 2 A schematic diagram of the structure of a remote diagnosis system for the cause of tooth fixation provided in an embodiment of the present application.

[0024] Reference numerals:

[0025] 200 remote diagnosis system for etiology of tooth fixation, 210 information acquisition unit, 220 required historical case data screening unit, 230 user tooth fixation three-dimensional model building unit, 240 predicted tooth fixation cause data acquisition unit, 250 treatment plan matching unit, 260 data calling unit. DETAILED DESCRIPTION

[0026] The embodiments of the present application provide a method and system for remote diagnosis of the cause of tooth fixation.

[0027] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0028] The technical solution proposed in the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0029] Figure 1 A flow chart of a remote diagnosis method for the cause of tooth fixation provided in an embodiment of the present application is as follows: Figure 1 As shown, the remote diagnosis method for the cause of tooth fixation includes the following steps:

[0030] Step 101: Receive user remote diagnosis login information, and obtain user condition information and historical diagnosis information based on the login information.

[0031] In one embodiment of the present application, a user logs in through a specific network platform, such as a medical APP, an online medical service website, etc. If the user logs in successfully, the system will retrieve the corresponding data from the database storing the user's medical information according to the user's ID. These data are stored in a special medical information database, which contains a large number of patients' medical records.

[0032] Furthermore, the user's medical condition information in the embodiment of the present application includes a description of the symptoms related to tooth fixation that the user is currently suffering from, such as whether the teeth are loose, the nature of the pain: such as dull pain, tingling, dull pain, the duration of the pain, the frequency of pain attacks, whether it affects the chewing or biting function, changes in the appearance of the teeth, etc. In addition, it may also include some clinical examination results, such as X-ray imaging manifestations and laboratory test results.

[0033] Furthermore, the historical diagnosis information in the embodiment of the present application covers all the diagnosis records related to oral health that the user has previously undergone in medical institutions, including oral diseases that have been suffered, such as caries, periodontitis, pulpitis, etc., and their treatment processes and results, whether there has been tooth fixation before and the diagnosis and treatment methods at that time, and also includes the history of systemic diseases that may affect oral health, such as diabetes, osteoporosis, etc.

[0034] Step 102: extract the tooth fixation etiology features from the user's condition information and historical diagnosis information, and filter out the required historical case data based on the weight values ​​corresponding to each tooth fixation etiology feature.

[0035] In one embodiment of the present application, a plurality of tooth fixation etiology features are determined based on the user's medical condition information. The plurality of tooth fixation etiology features are sorted based on a preset tooth fixation etiology feature importance sequence, so as to screen out historical case data that meet the feature similarity threshold in the database in turn based on the sorting order, and construct a plurality of historical case sets. According to the feature similarity, the case data in the plurality of historical case sets are sorted, and weights are assigned to each case data according to the sorting order. Based on the weights corresponding to the plurality of historical case sets and the weights corresponding to each case data, the feature scores corresponding to each case data are determined. According to the feature scores, the required historical case data are screened out.

[0036] Specifically, the embodiments of the present application extract various features related to the cause of tooth fixation from the user's medical information. These features may include the user's current symptoms, clinical examination results, living habits, and past medical history. For example, the characteristics of the cause of tooth fixation extracted from the user's medical information may include: the degree of alveolar bone absorption, tooth mobility, inflammatory index level, whether there is occlusal trauma, whether there is a history of orthodontic or endodontic treatment, the patient's age, whether there is a systemic disease such as diabetes, osteoporosis, etc. These features reflect various potential factors that may lead to tooth fixation and provide basic information for subsequent analysis.

[0037] Furthermore, the preset tooth fixation etiology feature importance sequence in the embodiment of the present application is a pre-set list, which ranks the importance of various tooth fixation etiology features. According to the preset tooth fixation etiology feature importance sequence, the tooth fixation etiology features extracted from the user's medical condition information are ranked. Then, according to the ranked feature order, the database storing the historical case data is screened. For the most important tooth fixation etiology feature, first find the historical cases with a high similarity. In each historical case set, the case data is sorted again according to the similarity between the case data and the corresponding features in the user's medical condition information. Weights are assigned to the sorted case data, and the higher the similarity, the higher the weight is assigned to the case. For the case data in each historical case set, its feature score is obtained by comprehensively considering the weight of the historical case set to which it belongs and its own weight in the set. The embodiment of the present application sets the screening criteria for the feature score in advance, and only selects the historical case data with a feature score exceeding a certain value as the required historical case data.

[0038] Specifically, in the embodiment of the present application, the characteristic score is calculated according to the function:

[0039]

[0040] Determine the characteristic scores corresponding to each case data; among them, is the feature score; H is the set of historical case sets; h ij For the historical case set H i The jth data point in ; wH is the weight set of the historical case set, where wH={wH1,wH2,...,wH n}; For each historical case set H i , the weight set of its data points is {wh i1 ,wh i2 ,...,wh im}; F = {f1,f2,...,f k}, where F is the feature set; is the data point h ij In the feature f l α is the adjustment factor, which is used to adjust the degree to which the feature score is affected by the weight of the historical case set and the weight of the data point, and its value range is [0,1]; n is the number of historical case sets; m p For the historical case set H p The number of data points in .

[0041] Step 103: Based on the user's condition information and historical diagnosis information, determine the tooth fixation condition area, perform key point matching on the tooth fixation condition area, and construct a user's tooth fixation three-dimensional model based on the matching result.

[0042] In one embodiment of the present application, an initial tooth-joint three-dimensional model corresponding to the user is constructed based on the user's medical condition information. The disease area of ​​the initial tooth-joint three-dimensional model is extracted to obtain a disease area image, and the disease area image is matched with a disease image database, and the disease type corresponding to the disease area is determined based on the matching result. Based on the edge detection algorithm, pixel points of the disease areas corresponding to each disease type are identified, and the area of ​​the disease area corresponding to each disease type is determined based on the identified pixel set. Based on the user information, a reference tooth three-dimensional model corresponding to the user is constructed. Key point matching is performed between the reference tooth three-dimensional model and the initial tooth-joint three-dimensional model, and based on the matching result, the area of ​​the disease area is adjusted to determine the user's tooth-joint three-dimensional model based on the adjusted area of ​​the disease area.

[0043] Specifically, based on the user's medical condition information, such as clinical examination data and other diagnostic information, the user's initial tooth fixation 3D model is constructed using computer graphics and image processing technology. This information includes the position, shape, density, state of surrounding tissues, etc. of the teeth. For example, through CT scan data, the 3D coordinate information of the teeth and surrounding tissues can be extracted, and the 3D reconstruction software can be used to convert this data into a 3D visualization model.

[0044] Furthermore, image processing technology is used to separate areas where lesions may exist from the constructed initial tooth fixation three-dimensional model. For example, by setting a certain density threshold or texture feature, areas that are significantly different from normal tissues are identified as disease areas. These areas may appear as alveolar bone absorption areas, root and alveolar bone adhesion areas, etc. The extracted disease area is displayed in the form of an image to obtain a disease area image. The disease image database in the embodiment of the present application stores a large number of standard images of different types of tooth fixation diseases. These images are marked with various disease types, such as tooth fixation areas caused by inflammation, tooth fixation areas caused by trauma, and tooth fixation areas caused by developmental abnormalities. The user's disease area image is compared and matched with the image in the disease image database. By comparing the image features and texture information, the standard image that is most similar to the user's disease area image is found, thereby determining the disease type corresponding to the disease area.

[0045] Furthermore, the edge detection algorithm is used to calculate the gradient of the pixels in the image and find the location where the gray value changes dramatically, so as to determine the edge of the diseased area. Once the edge is determined, the pixels inside the edge can be regarded as pixels belonging to the diseased area. By counting the number of pixels belonging to the diseased area and combining it with the resolution of the image, that is, knowing the actual physical area represented by each pixel, the area of ​​the diseased area can be calculated.

[0046] Furthermore, the user information in the embodiment of the present application may include the user's age, gender, normal tooth development data, and oral examination data when not sick. A reference tooth three-dimensional model is constructed using this information, which represents the ideal state of the user's teeth and surrounding tissues under normal conditions. Corresponding key feature points are found in the reference tooth three-dimensional model and the initial tooth-fixed three-dimensional model, such as the apex of the tooth, the root tip, the alveolar ridge top, etc. By calculating the transformation relationship between these key points, such as translation, rotation, and scaling, a mapping relationship between the two models is established. According to the above transformation relationship, the area of ​​the diseased area calculated previously is adjusted. For example, if it is found that the user's tooth-fixed three-dimensional model has been scaled or rotated relative to the reference model, the area of ​​the diseased area also needs to be adjusted accordingly. By mapping the area of ​​the diseased area to the space of the reference model, the true size and range of the diseased area can be more accurately evaluated, avoiding errors in the model construction process or deviations caused by different perspectives, and finally determining a more accurate user tooth-fixed three-dimensional model.

[0047] In one embodiment of the present application, by presetting a key point extraction model, key points are extracted from the tooth-attached model and the reference tooth three-dimensional model respectively. Descriptors are generated from the extracted key points, and the key points are matched based on the similarity between the descriptors. Based on the matched key points, the deformation field from the reference tooth three-dimensional model to the initial tooth-attached three-dimensional model is determined by thin plate spline interpolation. Based on the deformation field, the diseased area in the reference tooth three-dimensional model is mapped to the initial tooth-attached three-dimensional model, and the area of ​​the diseased area is adjusted according to the deformation field.

[0048] Specifically, the preset key point extraction model in the embodiment of the present application is used to extract key points from the tooth fixation model and the reference tooth three-dimensional model respectively. The training process of the preset key point extraction model is to use the preset tooth fixation model sample and the preset reference tooth three-dimensional model sample as input samples, and the model with key points marked corresponding to the input sample as the output sample, and train the preset neural network model to obtain the preset key point extraction model. The key points in the embodiment of the present application are usually located at the key positions of the teeth or surrounding tissues, which are very important for describing the shape, structure and positional relationship of the teeth. Common key points may include the cusps of the teeth, such as the cusps, root tips, bifurcation points, apex of the alveolar ridge, edge points where the teeth and alveolar bones contact, etc.

[0049] Furthermore, the preset key point extraction model is applied to the initial tooth fixation model and the reference tooth three-dimensional model respectively to extract the corresponding key points. In order to more accurately describe and compare the key points, the embodiment of the present application generates a descriptor for each extracted key point. The descriptor may contain local geometric features of the key point, such as local curvature, normal direction, shape description of the surrounding neighborhood, and other information. By calculating the similarity between different key point descriptors, the key points in the initial tooth fixation model are matched with the key points in the reference tooth three-dimensional model.

[0050] Furthermore, in the three-dimensional space, the reference tooth three-dimensional model is regarded as a "thin plate". According to the displacement of the matched key points, a deformation field from the reference tooth three-dimensional model space to the tooth-fixed three-dimensional model space can be obtained through thin plate spline interpolation. For any point in the reference tooth three-dimensional model, the deformation field can give the corresponding position of the point in the tooth-fixed three-dimensional model. For example, for a point in the reference tooth three-dimensional model, its new position in the tooth-fixed three-dimensional model can be calculated according to the deformation field, thereby realizing the spatial transformation from the normal state to the pathological state.

[0051] Furthermore, the diseased area in the reference tooth three-dimensional model in the embodiment of the present application is a possible abnormal area obtained based on the previous analysis. Through the deformation field, each point in this diseased area can be mapped to the corresponding position in the initial tooth-fixed three-dimensional model to show the actual position of the diseased area in the tooth-fixed state. For example, if an area that may be affected by inflammation is marked in the reference tooth three-dimensional model, the points of this area are mapped through the deformation field, and the exact position of the area in the pathological state can be found in the initial tooth-fixed three-dimensional model. Since there may be deformations in shape and position between the reference tooth three-dimensional model and the initial tooth-fixed three-dimensional model, the change in area can be calculated through the deformation field. For example, a circular diseased area originally marked in the reference tooth three-dimensional model may become an ellipse due to deformation after being mapped to the initial tooth-fixed three-dimensional model. By calculating the geometric transformation before and after the deformation, the new area of ​​the diseased area can be accurately calculated.

[0052] Step 104: input the required historical case data and the user's three-dimensional model of tooth fixation into a preset tooth fixation etiology prediction model to obtain predicted tooth fixation etiology data.

[0053] In one embodiment of the present application, based on the user's personal physical information and the required historical case data, the tooth fixation feature pairs are determined, and the tooth fixation feature pairs are divided to obtain feature pairs with interactive relationships and feature pairs without interactive relationships. Linear combination processing is performed on the feature pairs with interactive relationships, and mathematical transformation processing is performed on the feature pairs without interactive relationships. Feature fusion is performed on the high-order features obtained after processing to obtain a first high-order feature set. The first high-order feature set is subjected to dimensionality reduction processing through principal component analysis to obtain a second high-order feature set.

[0054] Specifically, first, a series of features that may be related to tooth fixation are extracted from the user's personal physical information, such as age, gender, general health status, etc., and the required historical case data, such as previous oral disease history, treatment history, examination indicators, etc. Then these features are combined in pairs to form tooth fixation feature pairs. For example, possible feature pairs include: age and alveolar bone density, periodontitis history and tooth mobility, diabetes history and gingival inflammation indicators, etc.

[0055] It should be noted that the feature pair with an interactive relationship in the embodiments of the present application refers to the mutual influence between the two features, which have a joint effect on the occurrence or development of tooth fixation. For example, (history of periodontitis, tooth mobility) may be a feature pair with an interactive relationship, because periodontitis usually leads to loose teeth, and they may influence each other in the formation process of tooth fixation. Feature pairs that do not have an interactive relationship are relatively independent and may not have a direct correlation with the impact on tooth fixation, such as (gender, alveolar bone density). In some cases, the impact of gender on alveolar bone density does not directly lead to tooth fixation, and the relationship between them is relatively weak.

[0056] Furthermore, for feature pairs with interactive relationships, linear combination operations are performed to capture the joint effect between features. For example, for the feature pair (history of periodontitis, tooth mobility), a new feature F = history of periodontitis a + tooth mobility b can be created, where a and b are weight coefficients obtained based on experience or data learning. For feature pairs that do not have interactive relationships, mathematical transformations are performed to explore their potential nonlinear relationships or enhance their expressive power. The features processed by linear combination and mathematical transformation are combined to form a first high-order feature set. These first high-order features contain new features after the original feature pairs are processed. They contain not only the original information, but also the processed joint effects or potential relationship information.

[0057] Furthermore, the first high-order feature set is used as input to calculate its covariance matrix, solve the eigenvalues ​​and eigenvectors, select the most important principal components according to the size of the eigenvalues, project the original high-order feature set onto these principal components, and obtain the second high-order feature set after dimensionality reduction.

[0058] In one embodiment of the present application, a first information entropy is determined based on a preset tooth fixation etiology prediction data set. Based on each candidate tooth fixation etiology feature, the preset tooth fixation etiology prediction data set is divided, and the second information entropy and information gain corresponding to the divided preset tooth fixation etiology prediction data set are determined. Data is randomly extracted from the preset tooth fixation etiology prediction data set, and the Gini coefficient is determined based on the extraction result. The optimal tooth fixation etiology feature is determined according to the first information entropy, the second information entropy, the information gain and the Gini coefficient. Starting from the root node, the preset training set data is divided according to the preset division criteria and the optimal tooth fixation etiology feature to obtain multiple subsets; wherein each subset corresponds to a child node. The features of each child node are divided by the optimal tooth fixation etiology feature corresponding to each child node until the stop condition is met to construct a preset tooth fixation etiology prediction model.

[0059] Specifically, the information entropy in the embodiment of the present application is an indicator to measure the purity of the data set. The higher the purity, the lower the information entropy. The preset tooth fixation etiology prediction data set is a data set containing multiple tooth fixation etiology samples. The first information entropy is the information entropy of the entire data set in the initial state, reflecting the initial purity of the data set. Assuming that there is a tooth fixation etiology data set containing 100 samples, 50 of which are etiology A, 30 are etiology B, and 20 are etiology C, then the calculated first information entropy can reflect the distribution of these three categories in the data set. The candidate tooth fixation etiology features in the embodiment of the present application are used to distinguish the features of different etiologies, such as age, gender, tooth position, etc. For each candidate feature, the data set is divided into multiple subsets, each subset corresponds to a value of the feature, and the second information entropy is the information entropy of each subset. Information gain is the amount of reduction in information entropy before and after division, reflecting the degree to which the division improves the purity of the data set. Assuming there is a candidate feature "age", the data set can be divided into three subsets: "less than 30 years old", "30-50 years old" and "greater than 50 years old". Then, the information entropy of these three subsets is calculated respectively and compared with the information entropy before division to calculate the information gain.

[0060] Furthermore, the Gini coefficient is another indicator to measure the purity of a data set. After randomly extracting samples, the proportion of samples of different categories in the extraction results is calculated, and then the Gini coefficient is calculated based on these proportions. Compare the information gain and Gini coefficient of different candidate features. For each candidate feature, there is a corresponding second information entropy, information gain and Gini coefficient. In the embodiment of the present application, features with large information gain and small Gini coefficient after division are used as better division features. Starting from the root node, the preset training set data is divided using preset division criteria such as maximum information gain, minimum Gini coefficient, etc. and optimal tooth fixation etiology features to obtain multiple subsets. Each subset corresponds to a child node, and the child node continues to be divided using the optimal feature until the stopping condition is met, such as the node purity reaches a threshold, the number of samples contained in the node is less than a certain value, etc. The final decision tree is the preset tooth fixation etiology prediction model, which can be used to predict new tooth fixation etiology samples.

[0061] Step 105: Match a corresponding treatment plan for tooth fixation based on the required historical case data and the predicted etiology data of tooth fixation.

[0062] In one embodiment of the present application, a case library is constructed based on required historical case data and treatment data corresponding to the required historical case data. The predicted etiology data of dental consolidation is matched with the case library once for similarity to obtain a first set of dental consolidation treatment plans. The quantity threshold is dynamically adjusted based on the number of treatment plans in the first set of dental consolidation treatment plans. The user's historical case data is obtained, and a secondary similarity match is performed in the first set of dental consolidation treatment plans based on the historical case data and the quantity threshold to obtain a second set of dental consolidation treatment plans. The user's basic information is obtained, and a personalized similarity model is constructed based on the user's basic information. The treatment plans in the second set of dental consolidation treatment plans are input into the personalized similarity model, and the corresponding dental consolidation treatment plans are matched based on the output personalized similarity value.

[0063] Specifically, the collected historical case data and their corresponding treatment data are sorted and stored to construct a case library. Each case is a record, which contains the characteristic information of the case and the corresponding treatment plan information. The similarity between the predicted tooth fixation etiology data and each case in the case library is calculated. For example, if cosine similarity is used, for each case, the predicted data and the characteristic vector of the case are regarded as two vectors in space, and the cosine value of the angle between them is calculated as the similarity. An initial similarity threshold is set, and the treatment plans corresponding to the cases whose similarity exceeds the threshold are extracted to form a first tooth fixation treatment plan set. When the number of treatment plans in the first tooth fixation treatment plan set is large, the similarity threshold is appropriately increased to narrow the range of the set; when the number is small, the similarity threshold is appropriately lowered to expand the range of the set.

[0064] Furthermore, the user's historical case data reflects the user's unique oral health trajectory and treatment experience. Based on key information in the historical case data, such as previous treatment effects, disease recurrence, disease development trends, etc., the applicability of each treatment plan is re-evaluated. The historical case data is re-calculated for similarity with the cases in the first set of tooth-fixed treatment plans, and the treatment plans that are more in line with the user's historical situation are screened out to form the second set of tooth-fixed treatment plans. For example, for a user who has experienced multiple recurrences of periodontitis, during the secondary similarity matching, he or she will be more inclined to choose a treatment plan that has a better treatment effect on recurrent periodontitis in the case.

[0065] Furthermore, the basic information of the user in the embodiment of the present application includes the user's gender, age, occupation, living habits, etc. Based on the basic information of the user, the weight values ​​corresponding to the treatment factors in the treatment plan are determined, and the personalized similarity model is constructed through the weight values ​​and the treatment factors corresponding to the weight values. The treatment plans in the second set of tooth fixation treatment plans are input into the personalized similarity model to obtain a personalized similarity value, so as to determine the tooth fixation treatment plan based on the personalized similarity value.

[0066] Step 106: Encrypt and store the predicted cause of tooth fixation and the tooth fixation treatment plan, so that the encrypted data can be called next time when the user's remote diagnosis login information is received.

[0067] In one embodiment of the present application, the predicted tooth fixation etiology data and the tooth fixation treatment plan are encrypted by the SM3 encryption algorithm to obtain the first summary information. The first summary information is encrypted by the SM2 algorithm and the private key to generate a digital signature ciphertext. Based on the encrypted predicted tooth fixation etiology data, the tooth fixation treatment plan and the digital signature ciphertext, an encrypted medical information package is generated. After a preset time interval, the usage frequency and importance of the encrypted medical information package are re-determined, and the encrypted medical information package is divided into different levels based on the determination result. The encrypted medical information packages of different levels are stored in different storage media, and the encryption strength of the encrypted medical information package is adjusted based on the different storage media. When a user system login request is received, the user's identity and authority are verified, and if the verification is passed, the corresponding encrypted medical information package is called.

[0068] Specifically, the data containing detailed tooth fixation etiology prediction results, such as alveolar bone density, inflammatory indicators, possible etiological factors, etc. and corresponding treatment plans, such as drug treatment details, surgical plans, rehabilitation suggestions, etc., are input into the SM3 algorithm to generate the first summary information. The first summary information is encrypted using the SM2 algorithm and the private key, and the generated digital signature ciphertext can not only protect the confidentiality of the summary information, but also serve as a digital signature to prove the source and integrity of the information. The predicted tooth fixation etiology data and tooth fixation treatment plan obtained through SM3 encryption processing and the generated digital signature ciphertext are combined together to form a complete encrypted medical information package.

[0069] Furthermore, over time, different encrypted medical information packages may have different usage frequencies and importance. For example, some patients' recent case information may be frequently used, while some historical case information is less frequently used; at the same time, information about some serious or complex cases may have a higher importance. At preset time intervals (such as monthly or quarterly), the system will re-evaluate these encrypted medical information packages. Based on the evaluation results, the encrypted medical information packages are divided into different levels. For example, they can be divided into high frequency and high importance, high frequency and low importance, low frequency and high importance, low frequency and low importance, and other levels.

[0070] Furthermore, different storage media are selected for encrypted medical information packages at different levels. High-frequency information may be stored in storage media with better performance, such as solid-state drives, for quick access; low-frequency information can be stored in lower-cost storage media, such as magnetic tapes or large-capacity hard drives. Also, the strength of encryption is adjusted according to the storage medium. For encrypted medical information packages stored in storage media with a high security level, a stronger encryption algorithm or a longer key length can be used to improve security; for information stored in relatively secure storage media, relatively weak encryption can be used to balance security and performance costs.

[0071] Furthermore, when a user initiates a system login request, the system will verify the user's identity information. Only after the user's identity and authority are verified, the user can access the encrypted medical information package for which he has permission. According to the user's role and authority scope, the system will retrieve the corresponding encrypted medical information package, and decrypt the encrypted information based on the stored information using the corresponding decryption key and algorithm to provide the user with the required medical information.

[0072] Figure 2 This is a schematic diagram of the structure of a remote diagnosis system for the cause of tooth fixation provided in an embodiment of the present application. Figure 2As shown, the remote diagnosis system 200 for the cause of tooth fixation includes: an information acquisition unit 210, which receives user remote diagnosis login information and acquires user condition information and historical diagnosis information according to the login information. A required historical case data screening unit 220 extracts tooth fixation cause features from the user condition information and historical diagnosis information, and screens out required historical case data based on the weight values ​​corresponding to each tooth fixation cause feature. A user tooth fixation three-dimensional model construction unit 230 determines the tooth fixation condition area based on the user condition information and historical diagnosis information, and performs key point matching on the tooth fixation condition area to construct the user tooth fixation three-dimensional model based on the matching result. A predicted tooth fixation cause data acquisition unit 240 inputs the required historical case data and the user tooth fixation three-dimensional model into a preset tooth fixation cause prediction model to obtain predicted tooth fixation cause data. A treatment plan matching unit 250 matches the corresponding tooth fixation treatment plan based on the required historical case data and the predicted tooth fixation cause data. The data calling unit 260 encrypts and stores the predicted tooth fixation etiology data and the tooth fixation treatment plan, so as to call the encrypted data when receiving the user's remote diagnosis login information next time.

[0073] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0074] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the embodiments of the present application may have various changes and variations. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A remote diagnosis method for the cause of tooth fixation, characterized in that: The method comprises: Receive user remote diagnosis login information, and obtain user condition information and historical diagnosis information according to the login information; Extracting the tooth fixation etiology features from the user's condition information and the historical diagnosis information, and filtering out the required historical case data based on the weight values ​​corresponding to the tooth fixation etiology features; Based on the user's condition information and the historical diagnosis information, determine the tooth fixation condition area, perform key point matching on the tooth fixation condition area, and construct a user's tooth fixation three-dimensional model based on the matching result; Input the required historical case data and the user's three-dimensional model of tooth fixation into a preset tooth fixation etiology prediction model to obtain predicted tooth fixation etiology data; Matching a corresponding treatment plan for tooth fixation based on the required historical case data and the predicted etiology data of tooth fixation; The predicted tooth fixation etiology data and the tooth fixation treatment plan are encrypted and stored, so that the encrypted data can be called when the user's remote diagnosis login information is received next time.

2. A remote diagnosis method for the cause of tooth fixation according to claim 1, characterized in that: The step of extracting the tooth fixation etiology features from the user's condition information and the historical diagnosis information, and filtering out the required historical case data based on the weight values ​​corresponding to the tooth fixation etiology features, specifically includes: Based on the user's medical condition information, determining a plurality of tooth fixation etiology characteristics; Based on a preset tooth fixation etiology feature importance sequence, a plurality of tooth fixation etiology features are sorted, so as to sequentially select historical case data that meet the feature similarity threshold in the database based on the sorting order, and construct a plurality of historical case sets; Sorting the case data in the plurality of historical case sets according to feature similarity, and assigning a weight to each case data according to the sorting order; Based on the weights corresponding to the plurality of historical case sets and the weights corresponding to the respective case data, determining the feature scores corresponding to the respective case data; According to the characteristic scores, the required historical case data are screened out.

3. A remote diagnosis method for the cause of tooth fixation according to claim 2, characterized in that: The determining of the characteristic scores corresponding to each of the case data based on the weights corresponding to the plurality of historical case sets and the weights corresponding to each of the case data specifically includes: According to the function: Determine the characteristic scores corresponding to each of the case data; wherein, is the feature score; H is the set of historical case sets; h ij For the historical case set H i The jth data point in ; wH is the weight set of the historical case set, where wH={wH1,wH2,...,wH n }; For each historical case set H i , the weight set of its data points is {wh i1 ,wh i2 ,...,wh im }; F = {f1,f2,...,f k }, where F is the feature set; is the data point h ij In the feature f l α is the adjustment factor, which is used to adjust the degree to which the feature score is affected by the weight of the historical case set and the weight of the data point, and its value range is [0,1]; n is the number of historical case sets; m p For the historical case set H p The number of data points in .

4. A remote diagnosis method for the cause of tooth fixation according to claim 1, characterized in that: The determining of a tooth fixation condition region based on the user's condition information and the historical diagnosis information, and matching key points of the tooth fixation condition region to construct a user's tooth fixation three-dimensional model based on the matching result specifically includes: Based on the user's medical condition information, construct the user's corresponding initial tooth-fixed connection three-dimensional model; Extracting the diseased area from the initial three-dimensional model of tooth fixation to obtain a diseased area image, matching the diseased area image with a diseased image database, and determining the disease type corresponding to the diseased area based on the matching result; Based on the edge detection algorithm, pixel points of the disease area corresponding to each disease type are identified, and the area of ​​the disease area corresponding to each disease type is determined based on the identified pixel point set; Based on the user information, a reference tooth three-dimensional model corresponding to the user is constructed; The reference tooth three-dimensional model is matched with the initial tooth-attachment three-dimensional model at key points, and the area of ​​the diseased region is adjusted based on the matching result, so as to determine the user tooth-attachment three-dimensional model based on the adjusted area of ​​the diseased region.

5. A remote diagnosis method for the cause of tooth fixation according to claim 4, characterized in that: The step of matching the reference tooth three-dimensional model with the initial tooth fixed connection three-dimensional model at key points and adjusting the area of ​​the diseased region based on the matching result specifically includes: By presetting a key point extraction model, key points are extracted from the initial tooth fixed connection model and the reference tooth three-dimensional model respectively; Generating descriptors from the extracted key points, and matching the key points based on the similarity between the descriptors; Based on the matched key points, determining the deformation field from the reference tooth three-dimensional model to the initial tooth fixed connection three-dimensional model by thin plate spline interpolation; Based on the deformation field, the diseased area in the reference tooth three-dimensional model is mapped onto the initial tooth fixed three-dimensional model, and the area of ​​the diseased area is adjusted according to the deformation field.

6. A remote diagnosis method for the cause of tooth fixation according to claim 1, characterized in that: Before inputting the required historical case data and the user's three-dimensional model of tooth fixation into a preset tooth fixation etiology prediction model to obtain predicted tooth fixation etiology data, the method further includes: Based on a preset tooth fixation etiology prediction data set, a first information entropy is determined; Based on the characteristics of each candidate tooth fixation cause, the preset tooth fixation cause prediction data set is divided, and the second information entropy and information gain corresponding to the divided preset tooth fixation cause prediction data set are determined; Randomly extracting data from the preset tooth fixation etiology prediction data set, and determining a Gini coefficient based on the extraction result; Determine the optimal tooth fixation etiology feature according to the first information entropy, the second information entropy, the information gain and the Gini coefficient; Starting from the root node, the preset training set data is divided according to the preset division criteria and the optimal tooth fixation etiology characteristics to obtain multiple subsets; wherein each of the subsets corresponds to a child node; The characteristics of each sub-node are divided according to the optimal tooth fixation etiology characteristics corresponding to each sub-node, until a stop condition is met, so as to construct the preset tooth fixation etiology prediction model.

7. A remote diagnosis method for the cause of tooth fixation according to claim 6, characterized in that: Before determining the first information entropy based on the preset tooth fixation etiology prediction data set, the method further includes: Based on the user's personal physical information and the required historical case data, determining tooth fixation feature pairs, and dividing the tooth fixation feature pairs to obtain feature pairs with interactive relationships and feature pairs without interactive relationships; Performing linear combination processing on the feature pairs with interactive relationship, and performing mathematical transformation processing on the feature pairs without interactive relationship; Performing feature fusion on the high-order features obtained after processing to obtain a first high-order feature set; The first high-order feature set is subjected to dimensionality reduction processing by principal component analysis to obtain a second high-order feature set.

8. A remote diagnosis method for the cause of tooth fixation according to claim 1, characterized in that: The matching of a corresponding treatment plan for tooth fixation based on the required historical case data and the predicted tooth fixation etiology data specifically includes: constructing a case library based on the required historical case data and treatment data corresponding to the required historical case data; Performing a similarity match between the predicted tooth fixation etiology data and the case library to obtain a first tooth fixation treatment plan set; Dynamically adjusting the quantity threshold based on the quantity of treatment plans in the first set of tooth fixation treatment plans; Acquire historical case data of the user, and perform secondary similarity matching in the first set of tooth fixation treatment plans based on the historical case data and the quantity threshold to obtain a second set of tooth fixation treatment plans; Obtain user basic information, build a personalized similarity model based on the user basic information, input the treatment plan in the second tooth fixation treatment plan set into the personalized similarity model, and match the corresponding tooth fixation treatment plan based on the output personalized similarity value.

9. A remote diagnosis method for the cause of tooth fixation according to claim 1, characterized in that: The step of encrypting and storing the predicted tooth fixation etiology data and the tooth fixation treatment plan so as to call the encrypted data when receiving user remote diagnosis login information next time specifically includes: The predicted tooth fixation etiology data and the tooth fixation treatment plan are encrypted by using an SM3 encryption algorithm to obtain first summary information; Encrypt the first summary information using the SM2 algorithm and a private key to generate a digital signature ciphertext; Generate an encrypted medical information package based on the encrypted predicted tooth fixation etiology data, the tooth fixation treatment plan and the digital signature ciphertext; After a preset time interval, the usage frequency and importance of the encrypted medical information package are re-determined, and the encrypted medical information package is divided into different levels based on the determination result; Storing the encrypted medical information packages at different levels in different storage media, and adjusting the encryption strength of the encrypted medical information packages based on the difference in storage media; When receiving a user system login request, the user's identity and authority are verified, and if the verification is passed, the corresponding encrypted medical information package is called.

10. A remote diagnosis system for the cause of tooth fixation, characterized in that: The system comprises: An information acquisition unit receives user remote diagnosis login information, and acquires user condition information and historical diagnosis information according to the login information; The required historical case data screening unit extracts the tooth fixation etiology features from the user's condition information and the historical diagnosis information, and screens out the required historical case data based on the weight values ​​corresponding to the tooth fixation etiology features respectively; A user-tooth joint three-dimensional model building unit, which determines a tooth joint condition area based on the user condition information and the historical diagnosis information, and performs key point matching on the tooth joint condition area to build a user-tooth joint three-dimensional model based on the matching result; The unit for obtaining the data for predicting the etiology of tooth fixation is configured to input the required historical case data and the three-dimensional model of tooth fixation of the user into a preset tooth fixation etiology prediction model to obtain the data for predicting the etiology of tooth fixation; A treatment plan matching unit, which matches a corresponding treatment plan for tooth fixation based on the required historical case data and the predicted tooth fixation etiology data; A data calling unit is used to encrypt and store the predicted tooth fixation etiology data and the tooth fixation treatment plan, so as to call the encrypted data when receiving the user's remote diagnosis login information next time.

Images