A monitoring imaging system for oral care

Abstract

The present invention relates to the field of image recognition and processing technology, and in particular to a monitoring imaging system for oral care. The system includes the following functions: establishing an oral health baseline feature model based on historical oral health image data; performing block processing on the patient's oral structure three-dimensional model, and extracting the oral structure model blocks with high incidence of abnormalities and transmitting them to the oral health baseline feature model for oral structure abnormality area analysis processing, generating oral structure abnormality area data; analyzing oral structure abnormality quantitative data based on oral structure abnormality area data; analyzing oral care time series state feature data based on oral structure abnormality quantitative data; performing dynamic optimization and adjustment processing on the patient's oral care decision through the oral care time series state feature data, and generating the patient's oral care optimization decision. The present invention realizes efficient and accurate oral care monitoring by performing efficient analysis of the care status while ensuring the monitoring accuracy through oral care monitoring imaging.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition and processing, and in particular to a monitoring imaging system for oral care. Background Art

[0002] With people's increasing attention to oral health, oral care has become an important part of modern medical care and health management. The health of teeth, gums and other oral tissues is directly related to the health of the whole body. Regular oral examination and maintenance are effective means to prevent oral diseases. In recent years, with the rapid development of imaging technology and artificial intelligence technology, digital imaging and automated diagnosis technology in oral medicine have been widely used. Through high-resolution oral endoscopes, digital X-rays, laser fluorescence imaging, optical coherence tomography (OCT), etc., teeth, gums and surrounding tissues are carefully imaged, and artificial intelligence technology is used to capture subtle lesions under oral care monitoring imaging, such as caries, periodontal disease, gingivitis, oral precancerous lesions, etc., to provide doctors with accurate information on the location, size and development of lesions. However, the existing oral care monitoring imaging system requires detailed inspection through high-resolution monitoring imaging, which greatly increases the pressure of data processing, or the limitation of resolution and accuracy. It is difficult to comprehensively and finely detect minor lesions or early abnormalities of oral tissues such as teeth and periodontal tissues; and it provides static image data, lacks dynamic tracking and analysis methods for oral health status, and cannot provide effective support for long-term oral health management. Summary of the invention

[0003] Based on this, the present invention provides a monitoring imaging system for oral care to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a monitoring imaging system for oral care includes the following modules:

[0005] The oral health image feature analysis module is used to obtain historical oral health image data in the oral care database; perform multi-scale fusion feature mining and processing of oral health images based on the historical oral health image data to generate oral health image feature data;

[0006] A health baseline characteristic model construction module is used to design a preliminary baseline characteristic mapping relationship of oral health based on oral health image characteristic data to generate a preliminary oral health baseline characteristic model; bias weight adjustment processing of model parameters of the preliminary oral health baseline characteristic model is performed to generate an oral health baseline characteristic model;

[0007] The oral structure abnormal area analysis module is used to perform oral structure imaging scanning and processing on the patient through an oral scanner to establish a three-dimensional model of the oral structure; the oral structure abnormal area analysis and processing is performed on the three-dimensional model of the oral structure using an oral health baseline feature model to generate oral structure abnormal area data;

[0008] The oral structure abnormality quantification module is used to mine the spatial diffusion influence relationship of oral abnormalities based on the oral structure abnormality area data to generate oral structure abnormality spatial diffusion influence relationship data; mine the oral structure abnormality pattern characteristics based on the oral structure abnormality area data to generate oral structure abnormality pattern characteristic data; establish an oral structure abnormality quantitative analysis model based on the oral structure abnormality pattern characteristic data and the oral structure abnormality spatial diffusion influence relationship data; transmit the oral structure abnormality area data to the oral structure abnormality quantitative analysis model to perform oral structure abnormality quantitative analysis to generate oral structure abnormality quantitative data;

[0009] An oral structure care status analysis module is used to perform oral structure care status time series characteristic analysis based on oral structure abnormality quantitative data, and generate oral structure care status time series characteristic data;

[0010] The oral care decision intelligent adjustment module is used to obtain the patient's oral care decision; dynamically optimize and adjust the patient's oral care decision through the oral structure care status time series characteristic data to generate the patient's oral care optimization decision; transmit the patient's oral care optimization decision to the terminal to execute the oral care intelligent adjustment operation.

[0011] Furthermore, the oral health image feature analysis module includes the following functions:

[0012] Access historical oral health imaging data from the oral care database;

[0013] Perform image edge analysis processing based on historical oral health image data to generate historical oral health image edge data; perform historical oral area division processing on historical oral health image data based on historical oral health image edge data to generate historical partitioned oral image data;

[0014] Perform multi-scale feature analysis of partitioned oral images on historical partitioned oral image data to generate partitioned multi-scale feature data;

[0015] Performing partition multi-scale feature weighted fusion processing according to the partition multi-scale feature data to generate partition multi-scale fusion feature data;

[0016] Based on the partitioned multi-scale fusion feature data, the multi-scale fusion feature mining of oral health images is performed on the historical partitioned oral image data to generate oral health image feature data.

[0017] Furthermore, the health baseline characteristic model building module includes the following functions:

[0018] Design a preliminary oral health baseline feature mapping relationship based on oral health image feature data to generate a preliminary oral health baseline feature model;

[0019] Using the preset baseline characteristic model test threshold to screen the test samples of historical oral health image data, and obtain oral health test image data;

[0020] Perform model fusion feature bias anomaly analysis on the preliminary oral health baseline feature model through oral health examination image data to generate model fusion feature bias anomaly data;

[0021] The model bias adjustment weight analysis is performed based on the model fusion feature bias abnormal data to generate the model bias adjustment weight data; the model bias adjustment weight data is used to adjust the model parameter weights of the preliminary oral health baseline feature model to generate the oral health baseline feature model.

[0022] Furthermore, the oral structure abnormal area analysis module includes the following functions:

[0023] Perform multi-angle oral structure imaging scans on patients through an oral scanner to generate multi-angle oral structure image data, and establish a three-dimensional oral structure model based on the multi-angle oral structure image data;

[0024] The three-dimensional model of the oral structure is divided into blocks according to a preset model division threshold value to generate an oral structure model block; the internal grayscale gradient analysis of the oral structure model block is performed to generate internal grayscale gradient data of the model block;

[0025] When the extreme value difference in the grayscale gradient data inside the model block is greater than a preset grayscale gradient extreme value difference threshold, the oral structure model block corresponding to the grayscale gradient data inside the model block is identified as a target model block to generate target oral structure model block data;

[0026] The target oral structure model block is transmitted to the oral health baseline feature model to perform oral health abnormality analysis and processing on the model block, including: when the oral health baseline feature model identifies that the target oral structure model block has a local abnormality, generating local oral abnormality data of the model block; when the oral health baseline feature model identifies that the target oral structure model block has a global abnormality, generating global oral abnormality data of the model block; when the oral health baseline feature model does not identify that the target oral structure model block has an abnormality, generating oral health data of the model block;

[0027] The oral structure abnormality area data is analyzed based on the local oral abnormality data of the model block and the global oral abnormality data of the model block to generate the oral structure abnormality area data.

[0028] Furthermore, performing oral structure abnormality area analysis based on the local oral abnormality data of the model block and the global oral abnormality data of the model block includes:

[0029] When local oral abnormality data of the model block is identified, a first abnormal area of ​​the oral structure is located based on the local oral abnormality data of the model block to generate positioning data of the first abnormal area of ​​the oral structure;

[0030] When the global oral abnormality data of the model block is identified, a second abnormal area positioning process of the oral structure is performed based on the global oral abnormality data of the model block to generate the second abnormal area positioning data of the oral structure;

[0031] Based on the first oral structure abnormal area positioning data and the second oral structure abnormal area positioning data, the oral structure abnormal area data is integrated to generate the oral structure abnormal area data.

[0032] Furthermore, the second abnormal area positioning process of the oral structure based on the global oral abnormality data of the model block includes the following functions:

[0033] Perform local abnormality descriptor analysis on the local oral abnormality data of the model block to generate a local abnormality descriptor of the model segmentation block; perform local abnormality direction traversal analysis and processing based on the local abnormality descriptor of the model segmentation block to generate local abnormality extension direction traversal data; identify oral abnormality feature data of the local abnormality extension direction traversal based on the local abnormality extension direction traversal data; locate the first abnormal area of ​​the oral structure based on the oral abnormality feature data of the local abnormality extension direction traversal and the local oral abnormality data of the model block to generate oral structure first abnormal area positioning data.

[0034] Furthermore, the second abnormal area positioning process of the oral structure based on the global oral abnormality data of the model block includes:

[0035] Perform global abnormality direction expansion traversal on the global oral abnormality data of the model block to generate global abnormality direction expansion traversal data; identify oral abnormality feature data of the global abnormality expansion direction traversal based on the global abnormality analysis expansion traversal data; perform second abnormal area positioning processing on the oral structure based on the oral abnormality feature data of the global abnormality expansion direction traversal and the global oral abnormality data of the model block to generate oral structure second abnormal area positioning data.

[0036] Furthermore, the oral structure abnormality quantification module includes the following functions:

[0037] Performing spatial distribution characteristic analysis of the abnormal area based on the data of the abnormal area of ​​the oral structure to generate spatial distribution characteristic data of the abnormal area;

[0038] Mining and processing the spatial diffusion influence relationship of oral abnormalities based on the spatial distribution characteristic data of abnormal areas to generate spatial diffusion influence relationship data of oral abnormalities;

[0039] Perform oral structure abnormality data analysis based on oral structure abnormality area data to generate oral structure abnormality data;

[0040] Performing cluster analysis on the oral structure abnormality data to generate clustered oral structure abnormality data; performing abnormal pattern recognition on the clustered oral structure abnormality data using a preset abnormal pattern rule library to generate oral structure abnormality pattern data;

[0041] Performing oral structure abnormality pattern feature mining processing according to the oral structure abnormality pattern data to generate oral structure abnormality pattern feature data;

[0042] Based on the oral structure abnormality pattern feature data and the oral abnormality spatial diffusion influence relationship data as oral abnormality risk association factors, a mapping relationship between oral abnormality characteristics and abnormality quantification degree is established through a preset logistic regression model and abnormal pattern risk association factors to generate an oral abnormality quantitative analysis model;

[0043] The data of the oral structure abnormality area is transmitted to the oral structure abnormality quantitative analysis model for oral structure abnormality quantitative analysis to generate oral structure abnormality quantitative data.

[0044] Furthermore, the oral structure care status analysis module includes the following functions:

[0045] Analyze the oral structure care status based on the quantitative data of oral structure abnormalities and generate oral structure care status data;

[0046] Perform oral structure care status analysis of the care cycle according to the oral structure care status data to generate periodic oral structure care status data;

[0047] Performing short-term trend analysis of the oral structure care status based on the oral structure care status data to generate short-term trend oral structure care status data;

[0048] Based on the periodic oral structure care status data and the short-term trend oral structure care status data, the oral structure care status time series feature mining is performed to generate the oral structure care status time series feature data.

[0049] Furthermore, the oral care decision intelligent adjustment module includes the following functions:

[0050] Obtaining patient oral care decisions;

[0051] Conduct oral care decision gradient benefit analysis on patients’ oral care decisions based on oral structure care status time series characteristic data, and generate oral care decision gradient benefit data;

[0052] Based on the oral care decision gradient benefit data, the patient's oral care decision is dynamically optimized and adjusted to generate the patient's oral care optimization decision;

[0053] Transmit the patient's oral care optimization decision to the terminal to execute intelligent oral care adjustment operations.

[0054] The beneficial effect of the present application is that the oral health image feature analysis module can accurately extract and analyze the multi-scale image features of oral health by acquiring historical image data in the oral care database and using multi-scale fusion feature mining technology, which significantly improves the analysis accuracy of oral health images. The edge analysis and partition processing of historical oral image data can refine the identification and division of oral areas, help identify subtle abnormalities that are difficult to detect, and are particularly suitable for the detection of early oral lesions. Through multi-scale feature weighted fusion processing, it is possible to comprehensively integrate features at all levels and provide a more detailed description of oral health features, thereby providing high-quality basic data for subsequent abnormal area analysis and quantification. In the health baseline feature model construction module, a preliminary baseline feature model of oral health is constructed based on historical image feature data, and through the adjustment of bias weights, it can adapt to the personalized needs of different patients and provide an accurate health baseline reference model. The historical image data is tested and the model bias abnormality analysis is performed, and the model parameters can be adaptively adjusted to ensure the accuracy of the baseline feature model and improve the judgment accuracy of the oral health status. The analysis and adjustment of the model bias weights realize the dynamic optimization of the baseline feature model to ensure that the model can adapt to the historical health characteristics of different patients. In the oral structure abnormal area analysis module, the system uses multi-angle oral structure imaging technology to generate an accurate three-dimensional oral model, which can carefully reflect the internal structure of the oral cavity, realize abnormal area analysis under the panoramic perspective of the oral cavity, and ensure comprehensive detection of local and global abnormalities. Through grayscale gradient analysis and model block technology, the model block is refined and analyzed to reduce the amount of data processing. Under the condition of ensuring the accuracy of oral abnormality detection, the abnormal area is accurately located, and the local and global oral abnormal areas are distinguished to improve the processing ability of oral abnormality identification. Not only can static abnormalities be identified, but also dynamic monitoring of abnormal areas can be performed according to the changing trend of oral structure, which improves the predictability of oral health management. In the oral structure abnormality quantification module, the spatial diffusion influence relationship of abnormal areas is analyzed, which can quantify the spatial expansion of abnormal areas, help doctors understand the diffusion trend and potential risks of abnormalities, and provide reference for formulating treatment plans. Based on the cluster analysis of abnormal areas, the diffusion pattern and risk characteristics of abnormal areas are identified, and different types of abnormalities are accurately distinguished, which improves the abnormality identification ability and processing accuracy of the system. Through the oral abnormality quantitative analysis model, the system can quantify the severity of the abnormal area and compare it with the oral health baseline characteristic model to provide scientific basis and data support for subsequent nursing decisions. Based on the oral structure abnormality quantitative data, the oral structure care status analysis module can perform time series analysis on the patient's care status, identify health fluctuation trends in the care cycle, analyze the short-term trend of the care status, predict the patient's oral health development direction, identify potential health risks in advance, help patients maintain the stability of oral health, and reduce the sudden or worsening of oral problems.The oral care decision intelligent adjustment module uses time series feature data to dynamically adjust the care decision, which can provide personalized care optimization solutions according to the patient's real-time care status and health trends, ensuring the efficiency and adaptability of the care plan. The optimized care decision is automatically transmitted to the terminal device, supporting intelligent oral care operations, reducing the complexity of manual intervention, and improving the automation level and execution efficiency of nursing operations. Through the care decision gradient benefit analysis, different care can be evaluated for dynamic optimization and adjustment, greatly improving the automation, intelligence and precision level of oral care monitoring. Therefore, the oral care monitoring imaging system of the present invention performs abnormality detection on the model blocks corresponding to the high-incidence abnormal areas in the oral care monitoring images, ensuring the data processing capability and abnormality detection accuracy for high-resolution oral images, thereby improving the analysis capability and efficiency of oral care monitoring imaging, and can comprehensively and finely detect microlesions or early abnormalities of oral tissues such as teeth and periodontal tissues. And provide a dynamic analysis technology for oral care monitoring imaging, realize dynamic tracking and analysis means of oral health status, and provide effective support for long-term oral health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of the module flow of a monitoring imaging system for oral care of the present invention;

[0056] Figure 2 for Figure 1 Detailed implementation functional flow chart of the oral structure abnormal area analysis module;

[0057] Figure 3 for Figure 1 Detailed implementation functional flow diagram of the oral structure abnormality quantification module;

[0058] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0059] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0060] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0061] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0062] To achieve this, please refer to Figures 1 to 3 The present invention provides a monitoring imaging system for oral care, including the following modules. In the embodiments of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of a module flow of a monitoring imaging system for oral care of the present invention, wherein the monitoring imaging system for oral care comprises the following steps:

[0063] S1: Oral health image feature analysis module, used to obtain historical oral health image data in the oral care database; perform multi-scale fusion feature mining and processing of oral health images based on the historical oral health image data to generate oral health image feature data;

[0064] In an embodiment of the present invention, the patient's historical oral health image data is retrieved from the oral care database. The image data in the database includes two-dimensional X-ray images, three-dimensional CT images, and optical scanning data. These image data reflect the patient's oral status at different time periods. During the data retrieval process, the system first pre-processes the image data, unifies the image resolution to 300 DPI, ensures that the format of all images is consistent (such as DICOM format), and eliminates noise interference in the image through a denoising algorithm (such as Gaussian filtering). For example, the historical image data of a patient includes three X-ray images and one 3D CT scan image. The system unifies all these images into a standard format and processes them. After obtaining the processed image data, the image is analyzed using a multi-scale feature extraction method. Specifically, the image data is decomposed using wavelet transform, and the image is decomposed into details and contour information of multiple scales. In the first level of wavelet decomposition, the system decomposes high-frequency details (such as tooth texture, edges, etc.) and low-frequency components (such as the overall shape of the teeth). The system continues to perform secondary and tertiary decomposition on each image partition to extract more subtle structural information and global features. For example, local cracks in tooth structure are more obvious in high-frequency components, while global features such as the curvature and thickness of teeth are reflected in low-frequency components. The system normalizes the features of each level of decomposition to ensure that the data range of features at different scales is consistent. After completing the multi-scale feature extraction, the system performs weighted fusion of features at different scales. The system assigns weights to features at each level according to their importance in the image, and performs weighted fusion processing on features at each level according to the weights to generate comprehensive image data containing features at different levels. Through this multi-scale feature fusion, the system can simultaneously reflect the overall morphology of teeth and local micro-lesions. For example, in the fused image, the global features of gingival recession and the details of local caries are presented with high precision. The fused feature data is formatted and key information is extracted to generate "oral health image feature data". The feature data includes, for example, the morphology, curvature, occlusal angle, tooth thickness, etc. of the teeth, as well as the texture characteristics of different areas in the image. The contrast, homogeneity, etc. are calculated through the grayscale co-occurrence matrix.

[0065] S2: Health baseline characteristic model construction module, used to design the preliminary baseline characteristic mapping relationship of oral health based on the oral health image characteristic data, and generate a preliminary oral health baseline characteristic model; adjust the bias weight of the model parameters of the preliminary oral health baseline characteristic model to generate an oral health baseline characteristic model;

[0066] In an embodiment of the present invention, oral health image feature data are first obtained from the oral health image feature analysis module, and these data include multiple oral health features such as tooth morphology, gum thickness, boundary features between teeth and gums, and tooth surface. These features are used to construct a baseline feature mapping relationship. These data are analyzed by statistical methods to extract key oral health features and establish a preliminary baseline mapping relationship. For example, the system determines that the standard range of tooth thickness is 2.0 mm to 2.5 mm and the standard range of gum thickness is 0.7 mm to 0.9 mm by statistically analyzing the tooth thickness distribution of healthy people. The mapping relationship design adopts multivariate regression analysis, and a mathematical model is established through the association between multiple features such as tooth morphology and thickness to form a preliminary baseline feature model of oral health status. For example, the system establishes the following linear mapping formula: Y=β0+β1X1+β2X2+...+βnXn. Among them, Y represents the characteristic value of the health state, Xi represents different oral health image features (such as tooth thickness, caries area, etc.), and βi is the weight coefficient in the mapping relationship. Through this model, the system can determine the ideal baseline value for each oral health feature and generate a preliminary oral health baseline feature model. After generating the preliminary baseline feature model, the system uses oral health image feature data from real patients to verify the model, and adjusts the model parameters by bias, and analyzes the deviation of each feature value by comparing the actual health data and the preliminary model calculation results. The bias weight adjustment process uses an adaptive optimization algorithm (such as gradient descent) to adjust the parameters of the model. The system adjusts the weight βi of each feature according to the size and frequency of the deviation value. For example, if the deviation of a feature is large, the system increases the weight of the feature for more refined adjustments. Assume that the initial weight of tooth thickness is 0.3, and it is increased to 0.4 after adjustment to make the feature more consistent with the actual health status. Through multiple iterative adjustments, the system gradually reduces the bias of the model parameters to ensure that the output of the model is more consistent with the actual health status characteristics. Finally, an optimized oral health baseline feature model is generated, which can accurately reflect the distribution of features of different individuals in a healthy state.

[0067] S3: Oral structure abnormal area analysis module, used to perform oral structure imaging scanning and processing on patients through oral scanners to establish an oral structure three-dimensional model; use the oral health baseline feature model to perform oral structure abnormal area analysis and processing on the oral structure three-dimensional model to generate oral structure abnormal area data;

[0068] In an embodiment of the present invention, a high-precision oral scanner is used to perform multi-angle imaging scans on the patient's oral structure. The resolution of the scanner is set to 100 microns, which can capture the fine structure inside the oral cavity. The scanning process covers multiple angles (such as 0°, 30°, 60°, 90°, etc.) to ensure complete imaging of teeth, gums and other oral tissues. The multi-angle image data obtained by scanning is transmitted to the processing unit, and the images at different angles are aligned and spliced ​​using a registration algorithm (such as an ICP algorithm, an iterative closest point algorithm). Through triangulation processing (such as Delaunay triangulation), the system converts these two-dimensional image data into a three-dimensional model. A three-dimensional oral structure model containing structural details such as tooth morphology, gum contour, and tooth gap is established, and the resolution reaches sub-millimeter level (such as 0.1 mm), ensuring that the accuracy is sufficient to detect tiny abnormalities. After the three-dimensional model is established, the three-dimensional model of the oral structure is divided into blocks to screen model blocks in high-incidence areas, such as model blocks with large grayscale value differences, and the model is divided into several model blocks with a volume of less than 5 cubic millimeters according to spatial coordinates (such as each block is 0.5 mm × 0.5 mm × 0.5 mm). The threshold setting of the block ensures that the model block is small enough to capture the slight changes of the local structure while ensuring the computational efficiency. After the block processing, the 3D oral structure model is decomposed into multiple model blocks, each of which contains the local structure information of the teeth, gums or jaws. Grayscale gradient analysis is performed on each oral structure model block after the block. Grayscale gradient analysis uses the Sobel operator to process the image data of the model block to extract the intensity and direction of its grayscale change. The grayscale value range is set to 0 to 255, where 0 represents the darkest part and 255 represents the brightest part, which usually reflects the density or tissue difference of the structure. The grayscale gradient inside each model block is calculated to generate a gradient vector. Assuming that the grayscale gradient extreme value difference of a model block is 30, the system records and stores the difference value as the grayscale gradient data inside the model block. Areas with obvious gradient changes are often the concentration points of potential abnormalities, so these data are crucial for subsequent abnormality analysis. The preset grayscale gradient extreme value difference threshold is 20. The system checks the extreme value difference of the grayscale gradient of the model blocks one by one. When it is found that the extreme value difference of the grayscale gradient of a model block is greater than the threshold (for example, the grayscale difference is 30), the model block is identified as the target model block. The data of the target model block is extracted and marked to generate the target oral structure model block data. These target model blocks correspond to high-incidence areas of tooth surface damage or caries such as lesions, structural abnormalities, and tooth gaps inside the oral cavity, and reduce the amount of data processed. It is compared and analyzed with the previously generated oral health baseline feature model. The health baseline feature model contains feature data under normal oral conditions (such as tooth thickness, gum morphology, etc.). Through these baseline data, the system can identify areas that deviate from the standard health state. The system uses a feature matching algorithm to compare each area in the three-dimensional model point by point.Assuming that the baseline range of tooth thickness is 2.0 mm to 2.5 mm, the system detects that the thickness of a certain area is 2.7 mm, which exceeds the baseline range, and the system marks the area as abnormal. Similarly, the system analyzes the morphology of the gums (such as the degree of recession) and the texture changes on the tooth surface to detect whether there are local or global abnormalities in the model block. And by performing an abnormal expansion analysis on the model blocks with local or global abnormalities, the detection adjacent areas of the abnormal model blocks are expanded, and the three-dimensional model is divided into regions according to abnormal features (such as thickness deviation, grayscale difference, surface irregularity, etc.), and the areas that meet the abnormal conditions are marked as abnormal areas. For example, the caries area on the surface of a tooth shows a significant decrease in grayscale value (such as a decrease of 30%), and the area is marked as abnormal. After completing the abnormal area analysis, the data of the abnormal area of ​​the oral structure is generated. The data includes key information such as the three-dimensional coordinates, abnormality type, volume and morphology of each abnormal area. For example, the volume of an abnormal area is 1.2 cubic millimeters, and the main feature is gingival recession, which is manifested as a reduction in gingival thickness to 0.5 mm. The data of each abnormal area is stored in the system database for subsequent treatment planning and analysis.

[0069] S4: Oral structure abnormality quantification module, used for mining and processing the spatial diffusion influence relationship of oral abnormalities based on the oral structure abnormality area data, and generating oral structure abnormality spatial diffusion influence relationship data; mining and processing the oral structure abnormality pattern characteristics based on the oral structure abnormality area data, and generating oral structure abnormality pattern characteristic data; establishing an oral abnormality quantitative analysis model based on the oral structure abnormality pattern characteristic data and the oral abnormality spatial diffusion influence relationship data; transmitting the oral structure abnormality area data to the oral abnormality quantitative analysis model to perform oral structure abnormality quantitative analysis, and generating oral structure abnormality quantitative data;

[0070] In the embodiment of the present invention, the three-dimensional coordinates, volume, grayscale gradient, surface morphology and other characteristics of the abnormal area are obtained from the data of the abnormal area of ​​the oral structure. Based on these data, the system uses a spatial diffusion model to analyze the diffusion influence relationship of the abnormal area. Using the concepts of nodes and edges in graph theory, each abnormal area is regarded as a node, and the spatial relationship between adjacent abnormal areas is represented as an edge. Using the Dijkstra algorithm or Search algorithm, the system calculates the diffusion path and diffusion influence range between each abnormal area. For example, the diffusion rates of two adjacent abnormal areas are 0.3 mm / week and 0.5 mm / week respectively. The system simulates the diffusion impact and analyzes the influence relationship coefficient of the two adjacent abnormal areas, such as analyzing whether the abnormal area with a faster diffusion rate causes the adjacent area to cause abnormal lesions through time series characteristics. Generate oral abnormal spatial diffusion influence relationship data, which includes the diffusion rate, diffusion direction and diffusion influence range between each abnormal area. Perform pattern feature mining on the data of oral structure abnormal areas, extract the key features of the abnormal areas, and use pattern recognition algorithms to classify the abnormal types. For example, the system identifies different types of abnormalities such as caries and gingival recession based on the volume, morphology, grayscale gradient changes and other features of the abnormal area. Use clustering algorithms (such as K-means or DBSCAN) to cluster the features of the abnormal area, and the system classifies the abnormal areas into patterns. For example, the system finds that a certain type of abnormal area has similar features (such as a 30% decrease in grayscale value and a 15% increase in surface roughness) through feature analysis, and clusters these areas into one category and identifies them as caries patterns. The system also analyzes the temporal characteristics of these patterns, such as whether the expansion speed of the abnormal area changes in the short term and whether the evolution trend of the abnormality is stable. Through this mining, the system generates oral structure abnormality pattern feature data, describing the pattern category, feature value and development trend of the abnormal area. The oral abnormality quantitative analysis model is established using the oral structure abnormality pattern feature data and the oral abnormality spatial diffusion influence relationship data. The model is established based on machine learning algorithms such as logistic regression or random forest. By training these data, the system establishes a mapping relationship between abnormal characteristics and abnormal quantitative degree. The system sets a risk assessment model, takes abnormal pattern characteristics (such as the volume growth rate of caries pattern) and spatial diffusion influence (such as the fusion speed of two abnormal areas) as input factors, and generates a comprehensive risk score through a weighted algorithm. For example, the system sets the diffusion rate weight of the caries pattern to 0.4 and the grayscale change weight of gingival recession to 0.6, and generates an abnormal quantitative analysis model that can predict the future expansion trend and severity of the abnormal area. The results output by the model include a quantitative score for each abnormal area, ranging from 0 to 1, where a score close to 1 indicates that the area has a higher risk of diffusion and deterioration. After the quantitative analysis model is established, the system inputs the data of the oral structure abnormal area into the model for quantitative analysis. The system calculates the quantitative score and diffusion prediction of each abnormal area through the model. Combined with these calculation results, the quantitative data of oral structure abnormalities is generated, which contains key information such as the score of each abnormal area, diffusion speed, and future expansion range prediction. It helps to identify high-risk areas and carry out timely intervention and treatment.

[0071] S5: Oral structure care status analysis module, used to perform oral structure care status time series feature analysis based on oral structure abnormality quantitative data, and generate oral structure care status time series feature data;

[0072] In an embodiment of the present invention, a time series analysis algorithm (such as an ARIMA model or an LSTM neural network) is used to analyze the time series data of the abnormal area. The system analyzes the historical data of these abnormal areas, identifies their changing rules, and predicts the future development trend of the nursing status. For example, the expansion rate of a certain abnormal area shows a linear growth trend. The system predicts that the expansion rate of the area in the next four weeks may reach 0.45 mm / week through a linear regression model. Trend analysis is also performed on the changes in abnormal areas within multiple nursing cycles. For example, the volume of a certain abnormal area has gradually expanded from 2.0 cubic millimeters to 2.8 cubic millimeters in the past three nursing cycles (each cycle is one month). Through this time series analysis, the system identifies that the expansion mode of the area is accelerated expansion, indicating that the area needs special attention. After completing the time series analysis, the time series change characteristics and prediction results of each abnormal area are used to generate oral structure nursing status time series feature data. Describe the key change information of the abnormal area at different time points. For example, the expansion rate of a certain abnormal area increased from 0.25 mm / week last month to 0.35 mm / week this month, and the volume increased by 20%. Predict changes in abnormal areas during future care cycles based on time series data. For example, the system predicts that an abnormal area may continue to expand by 1.5 mm in the next month, with a 15% increase in volume. The generated oral structure care status time series feature data is stored in the database for subsequent care decisions and treatment planning. Doctors can use this data to understand the evolution trend of patients' oral abnormalities and adjust care plans based on future expansion predictions.

[0073] S6: Oral care decision intelligent adjustment module, used to obtain the patient's oral care decision; dynamically optimize and adjust the patient's oral care decision through the oral structure care status time series characteristic data to generate the patient's oral care optimization decision; transmit the patient's oral care optimization decision to the terminal to execute the oral care intelligent adjustment operation.

[0074] In an embodiment of the present invention, historical oral care decision data is obtained from the patient's electronic health record. The patient's care decision includes the doctor's advice on the patient's oral health status, the implementation record of the historical care plan and the patient's compliance. The care decision includes a monthly regular oral examination and a treatment plan for specific problems, such as caries restoration or gum care. The patient's existing care decision is optimized by a dynamic optimization algorithm (such as a genetic algorithm or a particle swarm optimization algorithm), and the algorithm comprehensively considers the oral structure care state time series characteristic data and multiple factors of the patient's oral care decision, including: the expansion rate of the abnormal area, the impact of the care frequency on the abnormal state, the limitation of medical resources, etc., and readjusts the care frequency and treatment plan based on these data. Calculate the optimized care plan and generate the patient's oral care optimization decision, which includes a new care frequency and recommended care measures. After the optimized care decision is generated, the patient's oral care optimization decision is transmitted to the intelligent terminal device to perform intelligent adjustment operations. Connect to the intelligent oral care device used by the patient, and synchronize the optimized care plan to the device through wireless transmission technology (such as Bluetooth or Wi-Fi). The terminal device adjusts the care mode according to the optimization decision. For example, after receiving the optimized cleaning frequency and duration instructions, the smart toothbrush will automatically remind the patient at the specified time and adjust the brushing duration and mode according to the adjusted care plan. The optimized care decisions are also synchronized to the doctor's terminal device so that the doctor can view the adjusted care plan and provide guidance for subsequent examinations and treatments. After optimizing the care decision, real-time feedback is also obtained through the patient's smart device. The system will regularly obtain data from the care equipment (such as cleaning frequency, cleaning time, etc.), and compare these data with the "oral structure care status time series feature data" to track the execution effect of the care and automatically adjust the care strategy.

[0075] Furthermore, the oral health image feature analysis module includes the following functions:

[0076] Access historical oral health imaging data from the oral care database;

[0077] In an embodiment of the present invention, oral health image data within the past year are obtained from an oral care database containing 500 patients. These image data include dental X-ray images, oral endoscope images, CT scan images, etc., with a resolution of 300 to 600 DPI (dots per inch) to ensure sufficient clarity of image details. Each patient's data set contains an average of 30 oral images, and the total number of images in the database is about 15,000.

[0078] Perform image edge analysis processing based on historical oral health image data to generate historical oral health image edge data; perform historical oral area division processing on historical oral health image data based on historical oral health image edge data to generate historical partitioned oral image data;

[0079] In an embodiment of the present invention, the acquired oral health image is processed by the Canny edge detection algorithm, and the low threshold is set to 100 and the high threshold is set to 200 to generate clear edge images of structures such as teeth and gums. After edge detection, an oral X-ray image with a resolution of 512x512 pixels generates an edge data map, and the number of edge pixels is approximately 2,000, which clearly presents the outline of the teeth. According to the anatomical characteristics of the oral cavity, each oral image is divided into multiple functional areas, and seed points are selected. The seed points are selected according to the anatomical characteristics of the oral image, such as the center point of the tooth and the intersection point of the gum boundary. Then, according to the grayscale similarity between pixels (the difference is less than 10), adjacent pixels are classified into the same area, thereby separating teeth, gums and other oral areas. These divided data are stored as historical partitioned oral image data for subsequent multi-scale feature analysis.

[0080] Perform multi-scale feature analysis of partitioned oral images on historical partitioned oral image data to generate partitioned multi-scale feature data;

[0081] In an embodiment of the present invention, multi-scale feature analysis is performed on the data of each partition, and the image data is decomposed using wavelet transform. The image is decomposed into detail images of different scales, including low-frequency components and high-frequency components, through three-level wavelet decomposition. The low-frequency component is used to extract large-scale features, such as the overall shape and boundary contour of the teeth; the high-frequency component is used to identify local fine features, such as the texture and grayscale features of the tooth surface. For example, the feature data of each scale includes analyzing the statistical texture features (such as grayscale co-occurrence matrix, energy, contrast, homogeneity, etc.) to obtain texture features, and collecting the grayscale value quantization degree to obtain grayscale quantization information, thereby generating the partition multi-scale feature data of each partition.

[0082] Performing partition multi-scale feature weighted fusion processing according to the partition multi-scale feature data to generate partition multi-scale fusion feature data;

[0083] In the embodiment of the present invention, for multi-scale feature data of different partitions, the system performs weighted fusion processing on the features of each scale. The weight setting is configured according to the importance of the feature, for example, the weight of the tooth shape feature is set to 0.7, and the weight of the local texture feature is set to 0.3. Through the fusion processing of these weights, comprehensive partitioned multi-scale fusion feature data is generated, which accurately reflects the overall characteristics and local details of each oral partition at different scales, ensuring that the analysis results can reflect the integrity of the macroscopic structure and capture microscopic abnormalities.

[0084] Based on the partitioned multi-scale fusion feature data, the multi-scale fusion feature mining of oral health images is performed on the historical partitioned oral image data to generate oral health image feature data.

[0085] In the embodiment of the present invention, after obtaining the multi-scale fusion features, they are further mined to extract deep-level feature data. Principal component analysis (PCA) is performed on the fused feature data to extract the most discriminative feature vectors. The feature data is subjected to dimensionality reduction processing to remove redundant information and retain key features that are helpful for anomaly detection. At the same time, combined with the time dimension of historical images, trend analysis is performed on these features to identify the patterns of oral health features changing over time and generate comprehensive oral health image feature data, such as the weight value of oral health texture 0.4, texture complexity, etc.

[0086] Furthermore, the health baseline characteristic model building module includes the following functions:

[0087] Design a preliminary oral health baseline feature mapping relationship based on oral health image feature data to generate a preliminary oral health baseline feature model;

[0088] In an embodiment of the present invention, multiple core feature data points are extracted from historical oral health image feature data, including tooth height, width, thickness, surface curvature, gingival margin position, and gaps between teeth. Taking tooth thickness as an example, its feature value can be obtained by calculating the pixel distance in the image data. The system measures the thickness from the outer edge of the tooth to the root part through the cross-sectional image of the tooth in the image, assuming that the thickness of the tooth in the image is 2 mm. At the same time, the gingival margin feature extracts the contour curve at the junction of the tooth and the gingiva, and measures the vertical distance between the curve and the tooth surface, assuming that the value is 0.8 mm. When establishing a mapping relationship, these features are associated with the standard values ​​of a healthy oral state. For example, if the tooth thickness of a healthy individual is in the range of 1.8 mm to 2.2 mm, 2 mm is used as the reference value. Similarly, the healthy range of the gingival margin is set to 0.7 mm to 0.9 mm, and the reference value is 0.8 mm. These key feature values ​​are used as input feature vectors, and a mapping relationship is established using a supervised learning algorithm, such as a support vector machine, to form a preliminary oral health baseline feature mapping relationship. By comparing these mapping relationships with the data of multiple healthy samples, a preliminary oral health baseline feature model is generated. This model contains the mapping relationship between each feature value and its health standard, ensuring that subsequent oral health abnormality detection has sufficient reference basis.

[0089] Using the preset baseline characteristic model test threshold to screen the test samples of historical oral health image data, and obtain oral health test image data;

[0090] In an embodiment of the present invention, historical oral health image data is screened by a preset baseline feature model test threshold to obtain test samples that meet health standards. The test threshold is set according to the image feature distribution of multiple healthy oral cavity. For example, the threshold range of tooth thickness is set to 1.8 mm to 2.2 mm, and the threshold range of gum thickness is set to 0.7 mm to 0.9 mm. The system compares the features of each image data in turn. If the feature values ​​in a sample are all within the threshold range, the sample is screened as qualified oral health test image data, and samples that do not meet the threshold range are excluded.

[0091] Perform model fusion feature bias anomaly analysis on the preliminary oral health baseline feature model through oral health examination image data to generate model fusion feature bias anomaly data;

[0092] In an embodiment of the present invention, the characteristic value of each test sample is compared with the standard characteristic value in the preliminary baseline characteristic model. The system will calculate the deviation value of each feature. For example, the average tooth thickness of a sample is 2.1 mm, while the standard value in the baseline model is 1.8-2 mm, which generates a positive bias of 0.1-0.3 mm. By comparing the features of all samples, a bias data set containing each feature is generated, and the model fusion feature bias abnormality data is generated, which can accurately reflect the difference between the preliminary baseline characteristic model and the actual health test sample. This is done through a variety of methods, such as error analysis methods such as mean square error (MSE), to ensure the accuracy and rationality of the data.

[0093] The model bias adjustment weight analysis is performed based on the model fusion feature bias abnormal data to generate the model bias adjustment weight data; the model bias adjustment weight data is used to adjust the model parameter weights of the preliminary oral health baseline feature model to generate the oral health baseline feature model.

[0094] In an embodiment of the present invention, the bias information of each feature is extracted from the model fusion feature bias anomaly data. The model fusion feature bias anomaly data contains the deviation values ​​of each key feature (such as tooth thickness, gum height, tooth gap, etc.) relative to the baseline feature model. Weight analysis is performed according to the size and frequency of the deviation value. The bias degree of each feature is normalized so that the bias values ​​of different features can be compared on the same scale. Assuming that the deviation of tooth thickness is the largest and the impact is the most significant, after normalization, the weight of the tooth thickness feature is set to 0.6, the weight of the gum texture feature is set to 0.3, and the weight of the tooth gap is 0.1. The corresponding adjustment weight is assigned according to the importance of each feature in the health assessment. The weight assignment depends not only on the size of the deviation value, but also on the degree of influence of the feature on the overall oral health. For example, tooth thickness and gum texture are more important in judging the health status of teeth, so their bias adjustment weight is larger. Although tooth gap is an auxiliary feature, its adjustment weight is lower due to its small deviation. Through this analysis process, the model bias adjustment weight data of each feature is generated. The generated adjustment weight is used to adjust the preliminary baseline feature model. Taking tooth thickness as an example, the tooth thickness characteristic parameters in the preliminary model are adjusted, and the initial standard value of 2.2 mm is corrected to 2.15 mm through offset. After adjustment, the value is closer to the actual measurement result of the test sample. Similarly, the parameter value of gum thickness is adjusted, and the initial standard value of 0.8 mm is adjusted to 0.78 mm. After all characteristic parameters are adjusted, the system generates the final oral health baseline characteristic model, which focuses on the significant features that affect oral health. This model can more accurately reflect the oral structure characteristics in a healthy state.

[0095] Further, as an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed functional flow chart of the oral structure abnormal area analysis module. In this embodiment, the oral structure abnormal area analysis module includes the following functions:

[0096] S31: performing a multi-angle oral structure imaging scan on the patient through an oral scanner to generate multi-angle oral structure image data, and establishing an oral structure three-dimensional model according to the multi-angle oral structure image data;

[0097] In an embodiment of the present invention, a high-precision oral scanner is used to perform multi-angle imaging scans on the patient's oral cavity, and the scanning angle is set to obtain complete image data of the internal structure of the oral cavity from different orientations (such as 0°, 30°, 60°, 90°, 120°, etc.). The resolution of the scanner is 100 microns, ensuring that the complex details and minor abnormalities in the oral cavity can be captured. During the scanning process, the image data collected by the device each time will be automatically transmitted to the processing unit. The system will align and splice the images at multiple angles, and use the Delaunay triangulation method to splice the two-dimensional image data into a complete three-dimensional oral structure model based on the tooth protrusion node or tooth gap node as the matching node. The three-dimensional model includes major structures such as teeth, gums and jaws, and the accuracy can reach sub-millimeter level (0.1 mm), thereby ensuring the accuracy of the model.

[0098] S32: performing model block processing on the oral structure three-dimensional model according to a preset model block threshold to generate an oral structure model block; performing internal grayscale gradient analysis on the oral structure model block to generate internal grayscale gradient data of the model block;

[0099] In the embodiment of the present invention, the generated three-dimensional oral structure model is processed by block according to the set block threshold, and the system divides the model into a number of model blocks with a volume less than 5 cubic millimeters according to the spatial coordinates (for example, each block is 0.5 mm × 0.5 mm × 0.5 mm in size). The block threshold setting ensures that the model block is small enough to capture the slight changes in the local structure while ensuring the calculation efficiency. After the block processing, the three-dimensional oral structure model is decomposed into multiple model blocks, each of which contains the local structure information of the teeth, gums or jaws.

[0100] S33: when the extreme value difference in the grayscale gradient data inside the model block is greater than a preset grayscale gradient extreme value difference threshold, the oral structure model block corresponding to the grayscale gradient data inside the model block is identified as a target model block to generate target oral structure model block data;

[0101] In an embodiment of the present invention, a grayscale gradient analysis is performed on each oral structure model block after block division. The grayscale gradient analysis uses the Sobel operator to process the image data of the model block and extract the intensity and direction of its grayscale change. The grayscale value range is set to 0 to 255, where 0 represents the darkest part and 255 represents the brightest part, which usually reflects the density or tissue difference of the structure. The grayscale gradient inside each model block is calculated to generate a gradient vector. Assuming that the grayscale gradient extreme value difference of a certain model block is 30, the system records and stores the difference value as the grayscale gradient data inside the model block. Areas with obvious gradient changes are often the concentration points of potential abnormalities, so these data are crucial for subsequent abnormality analysis. The preset grayscale gradient extreme value difference threshold is 20. The system checks the grayscale gradient extreme value difference of the model blocks one by one. When it is found that the grayscale gradient extreme value difference of a certain model block is greater than the threshold (for example, the grayscale difference is 30), the model block is identified as the target model block. The data of the target model block is extracted and marked to generate the target oral structure model block data. These target model blocks correspond to high-incidence areas of tooth surface damage or caries such as lesions inside the oral cavity, structural abnormalities, and tooth gaps. Therefore, their labeling process provides key data for subsequent abnormality analysis.

[0102] S34: transmitting the target oral structure model block to the oral health baseline feature model to perform oral health abnormality analysis and processing on the model block, including: when the oral health baseline feature model identifies that the target oral structure model block has a local abnormality, generating local oral abnormality data of the model block; when the oral health baseline feature model identifies that the target oral structure model block has a global abnormality, generating global oral abnormality data of the model block; when the oral health baseline feature model does not identify that the target oral structure model block has an abnormality, generating oral health data of the model block;

[0103] In an embodiment of the present invention, the target oral structure model block is transmitted to the oral health baseline feature model to perform oral health abnormality analysis and processing on the model block. The oral health baseline feature model is matched with the target oral structure model block for detection, such as matching detection of features such as morphology and texture. When a part of the target oral structure model block does not meet the health baseline requirements of the oral health baseline feature model, it is marked as local oral abnormality data, such as if the detection model block has surface damage on all parts; when a global part of the target oral structure model block does not meet the health baseline requirements of the oral health baseline feature model, it is marked as global oral abnormality data, such as if the detection model block has surface damage on the whole; when the oral health baseline feature model does not identify the existence of an abnormality in the target oral structure model block, it is marked as model block oral health data, such as if the detection model block has no surface damage on the whole.

[0104] S35: Perform oral structure abnormality area analysis based on the local oral abnormality data of the model block and the global oral abnormality data of the model block to generate oral structure abnormality area data.

[0105] In the embodiment of the present invention, the local oral abnormality data of the model block and the global oral abnormality data of the model block generated after analysis are obtained, and each data records the coordinates, size, gray gradient, and abnormality degree of the abnormal model block. The system adopts a three-dimensional spatial clustering algorithm to cluster adjacent abnormal model blocks. Specifically, the DBSCAN algorithm (Density-Based Spatial Clustering of Applications with Noise) is adopted. The algorithm performs clustering operations on abnormal model blocks adjacent to each other in space by setting a distance threshold Eps=1 mm and a minimum number of points MinPts=5. If there are 5 or more abnormal model blocks in a certain area, and their center distance is less than 1 mm, the system clusters these model blocks into an abnormal area. After completing the clustering operation, the clustered abnormal area is subjected to diffusion analysis. For local abnormalities, abnormal diffusion analysis is performed in the direction of local abnormalities, and for global abnormalities, abnormal diffusion analysis is performed in the direction of global abnormalities to accurately locate the shape characteristics and volume of each abnormal area, generate feature data of the abnormal area, and normalize the data such as the center position, volume and shape of the abnormal area to form oral structure abnormal area data. The center position is obtained by calculating the geometric center of all abnormal model block coordinates, the volume is calculated by the cumulative volume of all model blocks, and the shape of the abnormal area is inferred by the arrangement relationship of each model block, such as whether the area is spherical or ellipsoidal. This data structure fully describes the abnormal area inside the oral cavity and is stored in the system database for subsequent analysis.

[0106] Furthermore, performing oral structure abnormality area analysis based on the local oral abnormality data of the model block and the global oral abnormality data of the model block includes:

[0107] When local oral abnormality data of the model block is identified, a first abnormal area of ​​the oral structure is located based on the local oral abnormality data of the model block to generate positioning data of the first abnormal area of ​​the oral structure;

[0108] In an embodiment of the present invention, local oral abnormality data of the model block is extracted from the analysis data of the model block. These data record the model block coordinates, abnormality type, and grayscale gradient characteristics of the local abnormality. The system uses a spatial positioning algorithm to cluster these local abnormal points through the three-dimensional coordinate data of each model block. Specifically, the K-means clustering algorithm is used, and the clustering radius is set to 1 mm, and the abnormal model blocks within the radius are clustered into a group. For each group of local abnormal model blocks, the system calculates its geometric center coordinates and boundary range, and generates the first abnormal area positioning data of the oral structure. For example, if a certain area contains 8 local abnormal model blocks, the system calculates its center position coordinates as (X=2.5 mm, Y=1.8 mm, Z=3.2 mm), and the space volume it occupies is 0.8 cubic millimeters. At this time, the system marks the area as the first abnormal area positioning data of the oral structure, and generates positioning data containing the area position and volume.

[0109] When the model block global oral abnormality data is identified, a second abnormal area positioning process of the oral structure is performed based on the model block global oral abnormality data to generate the second abnormal area positioning data of the oral structure;

[0110] In the embodiment of the present invention, the system adopts different processing methods for the global abnormal data of the model block. First, the system extracts the information of the global abnormal model block from the global oral abnormal data of the model block, which includes the grayscale distribution, volume, and position of the abnormal model block in the oral structure. The system determines the global abnormal area by analyzing the grayscale distribution characteristics of these model blocks. The definition of the global abnormal area is: the grayscale gradient change amplitude exceeds the local threshold of 20%, and the abnormality of the model block occupies the whole world, involving multiple oral structures (such as teeth and gums are affected at the same time). In order to spatially locate these global abnormalities, the system uses a three-dimensional space traversal algorithm to check the spatial correlation between the model blocks one by one, and confirm whether the interconnected model blocks belong to the same abnormal area. For example, in the concentrated area of ​​the global abnormal model blocks, if the grayscale change amplitude exceeds 30%, the system merges these model blocks and marks them as the second abnormal area positioning data of the oral structure. Then, the system calculates the total volume (such as 2.3 cubic millimeters) and the geometric center position (such as X=1.5 mm, Y=3.2 mm, Z=4.8 mm) of the area.

[0111] Based on the first oral structure abnormal area positioning data and the second oral structure abnormal area positioning data, the oral structure abnormal area data is integrated to generate the oral structure abnormal area data.

[0112] In an embodiment of the present invention, after completing the positioning of the local and global abnormal areas, the system integrates the positioning data of the first abnormal area of ​​the oral structure and the positioning data of the second abnormal area of ​​the oral structure. The integration processing step adopts a spatial clustering algorithm. The system determines whether there are overlapping or adjacent abnormal areas by analyzing the relative position and shape characteristics between these abnormal areas. If the distance between the local abnormal area and the global abnormal area is less than a set threshold (for example, 1 mm), the system merges them into a comprehensive abnormal area. For example, the boundary distance between the local abnormal area 1 and the global abnormal area 1 is 0.8 mm. The system merges the two areas to generate an abnormal area with a total area of ​​3.1 cubic millimeters, and records its new geometric center position and volume. If the local abnormal area is not associated with the global abnormal area, the abnormal area data is generated independently. Finally, the system generates complete oral structure abnormal area data, which includes the spatial coordinates, volume, shape, and distribution characteristics of the comprehensive abnormal area.

[0113] Furthermore, the second abnormal area positioning process of the oral structure based on the global oral abnormality data of the model block includes the following functions:

[0114] Perform local abnormality descriptor analysis on the local oral abnormality data of the model block to generate a local abnormality descriptor of the model segmentation block; perform local abnormality direction traversal analysis and processing based on the local abnormality descriptor of the model segmentation block to generate local abnormality extension direction traversal data; identify oral abnormality feature data of the local abnormality extension direction traversal based on the local abnormality extension direction traversal data; locate the first abnormal area of ​​the oral structure based on the oral abnormality feature data of the local abnormality extension direction traversal and the local oral abnormality data of the model block to generate oral structure first abnormal area positioning data.

[0115] In an embodiment of the present invention, the local abnormal data of the model block is analyzed to extract the local features of each abnormal model block. The generation process of the local abnormal descriptor is implemented by the SIFT (Scale-Invariant Feature Transform) algorithm, which can extract the key feature points of the model block and generate a descriptor. The grayscale gradient, surface contour features, and abnormality degree of each model block are described in the form of a feature vector to generate a local descriptor of the abnormal feature. The dimension of the descriptor is set to 128 dimensions according to the complexity of the abnormal feature. The system stores the descriptor of each local abnormal model block as a local abnormal descriptor of the model segmentation block for subsequent analysis. For example, if the grayscale gradient of a model block is 25, the contour change rate is 10%, and the abnormality degree is 15%, its descriptor will contain these key features. After generating the local abnormal descriptor, the system traverses and analyzes the extension direction of the abnormality based on the descriptor data. Specifically, the direction traversal algorithm (DirectionSearch Algorithm) is used to traverse the directional expansion trend of each local abnormality. The system sets 8 directions (such as 0°, 45°, 90°, etc.), traverses the adjacent model blocks one by one in each direction, and calculates the grayscale gradient difference, morphological change rate, etc. between the adjacent blocks and the current model block. Through the traversal process, the system generates abnormal expansion data in each direction. For example, the grayscale difference of a local abnormality in the 0° direction is 20%, and the grayscale difference in the 45° direction is 35%. This information will be recorded as the local abnormal expansion direction traversal data. Based on the local abnormal expansion direction traversal data, the expansion trend of the abnormal area is identified. By comparing the traversal data in different directions, the system determines the main expansion direction of the abnormal area. If the grayscale difference or morphological change rate in a certain direction (such as the 45° direction) exceeds the set threshold (for example, 30%), the system marks this direction as the main expansion direction of the abnormality. The system also calculates the expansion rate and range of the abnormal area in this direction. For example, the expansion rate in the 45° direction is 0.5 mm / week, and the expansion range is 2 mm. The system summarizes these data into oral abnormality feature data traversed in the local abnormal expansion direction. These data are used to describe the expansion morphology and trend of the abnormal area. The first abnormal area is located by combining the oral abnormality feature data traversed in the direction of local abnormal expansion and the local oral abnormality data of the model block. By analyzing the expansion direction, rate, and range of the abnormal area, the system determines the spatial position and boundary of the abnormal area, calculates the geometric center of the abnormal area, and uses the coordinate points to accurately locate the abnormal area. For example, if the center coordinates of the local abnormal area are (X=2.0 mm, Y=3.5 mm, Z=1.8 mm), and the expansion range is a circular area with a diameter of 1.5 mm, the system will store this data as the first abnormal area positioning data of the oral structure. This positioning data will provide the precise spatial position of the abnormal area and its expansion, providing a reference for subsequent analysis and diagnosis.

[0116] Furthermore, the second abnormal area positioning process of the oral structure based on the global oral abnormality data of the model block includes:

[0117] Perform global abnormality direction expansion traversal on the global oral abnormality data of the model block to generate global abnormality direction expansion traversal data; identify oral abnormality feature data of the global abnormality expansion direction traversal based on the global abnormality analysis expansion traversal data; perform second abnormal area positioning processing on the oral structure based on the oral abnormality feature data of the global abnormality expansion direction traversal and the global oral abnormality data of the model block to generate oral structure second abnormal area positioning data.

[0118] In an embodiment of the present invention, global abnormal features are extracted from the global oral abnormality data of the model block, and a traversal process of global abnormal direction expansion is performed. The system sets multiple expansion directions, specifically the X, Y, Z axes and the oblique directions of each axis (such as X+Y, Y+Z, etc.), and a total of 12 traversal directions are set. The step size in each traversal direction is set to 0.5 mm, and the adjacent global abnormal model blocks are gradually analyzed to see whether they are associated with the current abnormal block, such as whether the grayscale gradient difference exceeds the set threshold (20%), and whether the morphological change rate reaches the set standard (10%). Through this directional traversal analysis, the system generates extended data of the abnormal area in each direction, marking the expansion amplitude and range of the abnormal area in each direction. Taking a global abnormal model block as an example, the expansion distance in the X direction is 2 mm, and the expansion distance in the Y direction is 1.5 mm. The system records these expansion information and generates global abnormal direction expansion traversal data. After completing the global abnormal direction expansion traversal, the system identifies the global abnormal expansion features based on the traversal data, determines the main expansion direction and features of the abnormal area by analyzing the expansion data in each direction, and performs weighted processing on the expansion distance and grayscale change data in each direction, setting the weight value, for example, the weight of the expansion distance is 0.7, and the weight of the grayscale change is 0.3. The system calculates the comprehensive expansion features in each direction and screens out the main expansion direction. Assuming that the expansion feature weight value in the X+Y direction is the highest (such as more than 60%), the system marks this direction as the main expansion direction of the global abnormality. In addition, the system also identifies the expansion rate of the abnormal area. Assuming that the expansion rate in the X+Y direction is 0.2 mm / week and the expansion range is 2 mm, these data are recorded as the oral abnormality feature data of the global abnormal expansion direction traversal, which describes the spatial expansion trend and characteristics of the abnormal area. After identifying the global abnormal expansion feature data, the system combines the global oral abnormality data of the model block to perform the second abnormal area positioning processing of the oral structure. According to the global abnormal expansion direction and range, a spatial clustering algorithm (such as the DBSCAN algorithm) is used to spatially integrate the global abnormal area, aggregate adjacent abnormal model blocks with similar expansion directions into an abnormal area, and calculate the geometric center coordinates and volume of the area. For example, the main expansion range of an abnormal area in the X+Y direction is 3 mm, and the total volume after expansion is 5 cubic millimeters. The system calculates the center position of the area as (X=2.0 mm, Y=3.0 mm, Z=1.5 mm). Through this positioning processing, the system generates the positioning data of the second abnormal area of ​​the oral structure, which accurately describes the spatial position, volume, expansion direction and morphological characteristics of the global abnormal area.

[0119] Further, as an embodiment of the present invention, refer to Figure 3 As shown, Figure 1Detailed functional flow chart of the oral structure abnormality quantification module. In this embodiment, the oral structure abnormality quantification module includes the following functions:

[0120] S41: performing spatial distribution characteristic analysis of the abnormal region according to the oral structure abnormal region data to generate spatial distribution characteristic data of the abnormal region;

[0121] In an embodiment of the present invention, the spatial distribution analysis of the abnormal area data of the oral structure is performed using a spatial feature analysis algorithm. The abnormal area is spatially divided using a Voronoi diagram algorithm, and the distribution density, shape, volume and other characteristics of each abnormal area in three-dimensional space are calculated. The system sets a threshold for the spatial distribution density. For example, when the density is greater than 5 abnormal model blocks / cubic millimeter, the area is marked as a "high-density area" to generate abnormal area spatial distribution feature data, including the spatial position, density, volume, shape and other information of each abnormal area. For example, the volume of an abnormal area is 2.5 cubic millimeters, and the distribution density is 6 model blocks / cubic millimeter.

[0122] S42: performing spatial diffusion influence relationship mining processing on oral abnormalities according to the spatial distribution characteristic data of the abnormal area, and generating spatial diffusion influence relationship data on oral abnormalities;

[0123] In an embodiment of the present invention, a graph-based diffusion model is used to mine the spatial diffusion influence of abnormal regions on the spatial distribution feature data of abnormal regions, and a graph model with abnormal regions as nodes is established. Each edge in the graph represents the adjacent relationship and diffusion path between different abnormal regions. The system calculates the diffusion influence factor on each edge, which includes the size of the abnormal region, the expansion direction, and the diffusion rate. For example, the system detects that there is a diffusion path between two abnormal regions, and the diffusion rate is 0.3 mm / week. The system generates oral abnormal spatial diffusion influence relationship data based on the diffusion influence factor, describing the diffusion pattern and influence degree between abnormal regions.

[0124] S43: performing oral structure abnormality data analysis according to the oral structure abnormality area data to generate oral structure abnormality data;

[0125] In the embodiment of the present invention, the data of the abnormal area of ​​the oral structure is used to perform a deep analysis of each abnormal area through a morphological analysis method to identify the characteristic attributes of the abnormal area, including the boundary of the abnormal area, the thickness change rate, and the surface morphological characteristics. The system evaluates the degree of abnormality by calculating the convex hull volume of the abnormal area. For example, the convex hull volume of a certain abnormal area is 4.5 cubic millimeters and the surface roughness is 15%. These data are analyzed to obtain the oral structure abnormality data, which provides basic data for subsequent clustering and pattern recognition.

[0126] S44: performing cluster analysis on the oral structure abnormality data to generate clustered oral structure abnormality data; performing abnormal pattern recognition on the clustered oral structure abnormality data using a preset abnormal pattern rule library to generate oral structure abnormality pattern data;

[0127] In an embodiment of the present invention, the DBSCAN clustering algorithm (Density-Based Spatial Clustering of Applications with Noise) is used to perform clustering analysis of oral structure abnormality data. The algorithm uses two key parameters: the minimum number of points MinPts=5, and the distance threshold Eps=1 mm. Adjacent abnormal regions with similar features are clustered into one category according to the spatial distance and feature similarity between abnormal regions. For example, a clustering result shows that 10 abnormal gingival regions are included, with an average volume of 3.2 cubic millimeters and a surface roughness of 12%; a clustering result shows that 6 abnormal interdental regions are included, with an average volume of 1.2 cubic millimeters and a grayscale difference of 50%. The system records the clustering result as clustered oral structure abnormality data. The pattern type of the abnormal region is identified by matching the clustered oral structure abnormality data with a preset abnormal pattern rule library. The abnormal pattern rule library includes typical oral lesion patterns, such as caries, periodontal lesions, etc. The system uses a pattern matching algorithm to identify the pattern type corresponding to the abnormal region by feature comparison (such as volume, morphology, diffusion rate, etc.). For example, if a cluster of abnormal data highly matches the caries extension pattern in the rule base (matching degree is 90%), the system will mark the cluster as a caries pattern and generate oral structure abnormality pattern data, including pattern type, matching degree and other information.

[0128] S45: performing oral structure abnormality pattern feature mining processing according to the oral structure abnormality pattern data to generate oral structure abnormality pattern feature data;

[0129] In the embodiment of the present invention, the identified oral structure abnormal pattern data is further subjected to feature mining. Through the pattern mining algorithm, the key features of each abnormal pattern are extracted, including the evolution path, expansion trend, and abnormal degree of each pattern. For example, it is found that the expansion trend of a certain caries pattern is to gradually spread toward the root direction, with a diffusion rate of 0.4 mm / week. The system generates oral structure abnormal pattern feature data based on these data, describing the key features and development trends of the abnormal pattern.

[0130] S46: Based on the oral structure abnormality pattern feature data and the oral abnormality spatial diffusion influence relationship data as oral abnormality risk association factors, a mapping relationship between oral abnormality characteristics and abnormality quantification degree is established through a preset logistic regression model and abnormal pattern risk association factors to generate an oral abnormality quantitative analysis model;

[0131] In the embodiment of the present invention, the oral structure abnormality pattern feature data includes geometric features, morphological changes, grayscale gradients, volume changes, abnormal evolution trends, etc. of the abnormal area. These feature data can reflect the development trend of the abnormal area, for example, the volume change rate of a certain abnormal area is 12%, and the grayscale change is 25%. The oral abnormality spatial diffusion influence relationship data reflects the spatial diffusion influence between abnormal areas, including the diffusion rate, diffusion direction, and diffusion path between adjacent abnormal areas. For example, the diffusion rate of two adjacent abnormal areas is 0.3 mm / week, and the diffusion direction is along the X-axis. Based on the oral structure abnormality pattern feature data and the oral abnormality spatial diffusion influence relationship data as oral abnormality risk association factors, key information extracted from the oral structure abnormality pattern feature data, such as the volume of the area, the volume change rate, the grayscale gradient change rate, etc. The weight value of each feature is specifically set. For example, the weight of the volume change rate is 0.4, the weight of the grayscale change rate is 0.3, and the weight of the abnormal expansion speed is 0.3. Key information extracted from the oral abnormality spatial diffusion influence relationship data, such as the diffusion rate and diffusion range between adjacent abnormal areas. The weight of the diffusion rate is set to 0.6, and the weight of the diffusion direction is set to 0.4. These factors will be used as inputs to the logistic regression model to quantify the risk. The logistic regression model is used to quantify the risk of abnormal areas. The characteristic data of abnormal patterns and spatial diffusion effects are used as model parameters, and the weights and bias terms of the model parameters are set. The model parameters are trained with historical data of oral structure abnormality pattern characteristic data and oral abnormality spatial diffusion influence relationship data. The gradient descent algorithm is used to optimize the model parameters to minimize the loss function (such as the cross entropy loss function). The preprocessed oral structure abnormality pattern characteristic data and oral abnormality spatial diffusion influence relationship data are input into the trained logistic regression model. The model calculates the characteristics of each abnormal area and outputs a risk score. The risk scores of each abnormal area calculated by the logistic regression model are summarized to form a quantitative analysis model for oral abnormalities. The model not only contains the risk scores of each abnormal area, but also includes its related diffusion prediction information (such as diffusion rate, direction, and impact range).

[0132] S47: Transmitting the oral structure abnormality area data to the oral structure abnormality quantitative analysis model to perform oral structure abnormality quantitative analysis to generate oral structure abnormality quantitative data.

[0133] In an embodiment of the present invention, after generating a quantitative analysis model, the data of the oral structure abnormal area is input into the model for quantitative analysis. The input data includes the spatial position, volume, morphology, grayscale gradient and other characteristics of each abnormal area. The logistic regression model performs quantitative analysis on the abnormal areas item by item according to the preset weights and input factors, and calculates the quantitative score and risk level of each abnormal area. For example, the grayscale gradient change of an abnormal area is 20%, and its quantitative score is 0.75, indicating a medium risk. The system calculates the possibility and range of future spread based on the score, for example, predicting that the area will expand to the adjacent gingival area within the next two weeks. The system outputs all calculation results as quantitative data of oral structure abnormalities, including risk scores, diffusion rates, predicted expansion ranges and other information of abnormal areas, providing accurate quantitative assessments, which facilitates the further formulation of personalized oral care plans.

[0134] Furthermore, the oral structure care status analysis module includes the following functions:

[0135] Analyze the oral structure care status based on the quantitative data of oral structure abnormalities and generate oral structure care status data;

[0136] In an embodiment of the present invention, quantitative data of oral structure abnormalities are obtained from the oral abnormality quantitative analysis model. These data include key information such as the risk score, diffusion rate, and abnormal volume of each abnormal area. Based on these data, the system uses a state analysis algorithm to evaluate the overall care status of the abnormal area. The system uses a multidimensional feature weighting method (for example, the risk score weight is 0.6 and the diffusion rate weight is 0.4) to perform weighted calculations on each abnormal area to generate oral structure care status data. For example, the risk score of an abnormal area is 0.85, and the diffusion rate is 0.3 mm / week. The system calculates the comprehensive care status score of 0.78 based on the weight calculation, indicating that the area requires high attention. Finally, the system generates a care status score for each abnormal area and records the oral structure care status data.

[0137] Perform oral structure care status analysis of the care cycle according to the oral structure care status data to generate periodic oral structure care status data;

[0138] In an embodiment of the present invention, the oral structure care status data is used to analyze the care cycle of each abnormal area. The care cycle analysis is based on the patient's historical data, and the system summarizes the abnormal changes in each care cycle (such as once a month) of the patient. For example, the volume changes of a patient's dental abnormal area in the past three months of care cycles are 2.0, 2.1, and 2.5 cubic millimeters. The system calculates its volume change rate based on these data and generates periodic oral structure care status data. The periodic data will reflect the changing trend of the abnormal area in different cycles, helping to determine the effectiveness of the care plan.

[0139] Performing short-term trend analysis of the oral structure care status based on the oral structure care status data to generate short-term trend oral structure care status data;

[0140] In an embodiment of the present invention, the short-term trend of changes in abnormal areas is evaluated through a short-term trend analysis algorithm. The evaluation is based on the recent care status data (e.g., the volume change and diffusion rate of the abnormal area in the last 4 weeks), and a linear regression model is used to predict the development trend of the abnormal area in the short term. Assuming that the diffusion rates of an abnormal area in the last four weeks are 0.25, 0.28, 0.30, and 0.35 mm / week, respectively, the diffusion rate of the area in the next week is predicted to be 0.38 mm / week through regression analysis, and short-term trend oral structure care status data is generated to provide doctors with short-term care recommendations.

[0141] Based on the periodic oral structure care status data and the short-term trend oral structure care status data, the oral structure care status time series feature mining is performed to generate the oral structure care status time series feature data.

[0142] In an embodiment of the present invention, a complete time series feature is generated by integrating periodic oral structure care status data and short-term trend oral structure care status data through a time series feature mining algorithm. For example, the system performs time series fusion of periodic data and short-term trend data, performs time series prediction through an ARIMA (autoregressive integrated moving average) model, and analyzes the future development trajectory of abnormal areas. The system smoothes the status data of each abnormal area to identify the relationship between its long-term care effect and short-term abnormal fluctuations. For example, the periodic expansion rate of an abnormal area is stable at 1.5%, while the short-term expansion rate fluctuates greatly. After analyzing these change data through an ARIMA model, the system predicts that the area may remain stable or expand slightly in the next month. Generate oral structure care status time series feature data, which not only includes the historical development trajectory of the abnormal area, but also provides prediction results for future care effects.

[0143] Furthermore, the oral care decision intelligent adjustment module includes the following functions:

[0144] Obtaining patient oral care decisions;

[0145] In an embodiment of the present invention, historical nursing decision data of the patient is extracted from the database, which generally includes the patient's previous nursing plan, the doctor's nursing advice, and the patient's execution record. For example, a patient's nursing decision includes weekly teeth cleaning, monthly oral examination, and a treatment plan to repair tooth caries. The system uses these decision data as basic input to prepare for subsequent gradient benefit analysis and optimization processing.

[0146] Conduct oral care decision gradient benefit analysis on patients’ oral care decisions based on oral structure care status time series characteristic data, and generate oral care decision gradient benefit data;

[0147] In the embodiment of the present invention, the patient's current nursing decision is evaluated by a gradient benefit analysis algorithm, and the benefit change rate of each nursing decision is calculated. For example, the system evaluates the effect of weekly tooth cleaning on controlling diffusion based on the diffusion rate of the abnormal area (such as 0.3 mm / week), and quantifies the effect as a gradient benefit value (for example, the benefit value is 0.75). The benefit of each nursing measure is analyzed one by one, and the oral care decision gradient benefit data is generated, which includes the degree of influence of each nursing decision on controlling abnormal expansion and reducing risk scores.

[0148] Based on the oral care decision gradient benefit data, the patient's oral care decision is dynamically optimized and adjusted to generate the patient's oral care optimization decision;

[0149] In an embodiment of the present invention, the patient's nursing decision is dynamically optimized and adjusted based on the generated oral care decision gradient benefit data. The system uses an optimization algorithm based on benefit maximization, such as a linear programming model, to find the nursing plan combination with the highest benefit. The system inputs the benefit value of each nursing decision into the optimization model and sets nursing resources (such as time, frequency, etc.) as constraints. For example, the system sets that two nursing operations can be performed per week, and analyzes the improvement effect of each nursing operation on the patient's oral health status. The patient's oral care optimization decision is designed based on the improvement effect of the operation on the patient's oral health status. The optimization decision may recommend adding an additional check once a month and timely repair of caries.

[0150] Transmit the patient's oral care optimization decision to the terminal to execute intelligent oral care adjustment operations.

[0151] In an embodiment of the present invention, after the optimization decision is generated, the patient's oral care optimization decision is transmitted to the intelligent nursing terminal. The nursing terminal performs specific operations according to the optimization decision through intelligent devices (such as oral cleaners, treatment equipment, etc.). For example, the system sends the monthly nursing decision for repairing caries to the terminal of the medical institution to arrange the patient's treatment plan. At the same time, the intelligent device adjusts the cleaning frequency according to the optimization decision and reminds the patient to perform daily care operations according to the new nursing decision. The terminal device of the system records and executes these optimization decisions to ensure that the patient's nursing work is intelligently executed.

[0152] The beneficial effect of the present application is that the oral health image feature analysis module can accurately extract and analyze the multi-scale image features of oral health by acquiring historical image data in the oral care database and using multi-scale fusion feature mining technology, which significantly improves the analysis accuracy of oral health images. The edge analysis and partition processing of historical oral image data can refine the identification and division of oral areas, help identify subtle abnormalities that are difficult to detect, and are particularly suitable for the detection of early oral lesions. Through multi-scale feature weighted fusion processing, it is possible to comprehensively integrate features at all levels and provide a more detailed description of oral health features, thereby providing high-quality basic data for subsequent abnormal area analysis and quantification. In the health baseline feature model construction module, a preliminary baseline feature model of oral health is constructed based on historical image feature data, and through the adjustment of bias weights, it can adapt to the personalized needs of different patients and provide an accurate health baseline reference model. The historical image data is tested and the model bias abnormality analysis is performed, and the model parameters can be adaptively adjusted to ensure the accuracy of the baseline feature model and improve the judgment accuracy of the oral health status. The analysis and adjustment of the model bias weights realize the dynamic optimization of the baseline feature model to ensure that the model can adapt to the historical health characteristics of different patients. In the oral structure abnormal area analysis module, the system uses multi-angle oral structure imaging technology to generate an accurate three-dimensional oral model, which can carefully reflect the internal structure of the oral cavity, realize abnormal area analysis under the panoramic perspective of the oral cavity, and ensure comprehensive detection of local and global abnormalities. Through grayscale gradient analysis and model block technology, the model block is refined and analyzed to reduce the amount of data processing. Under the condition of ensuring the accuracy of oral abnormality detection, the abnormal area is accurately located, and the local and global oral abnormal areas are distinguished to improve the processing ability of oral abnormality identification. Not only can static abnormalities be identified, but also dynamic monitoring of abnormal areas can be performed according to the changing trend of oral structure, which improves the predictability of oral health management. In the oral structure abnormality quantification module, the spatial diffusion influence relationship of abnormal areas is analyzed, which can quantify the spatial expansion of abnormal areas, help doctors understand the diffusion trend and potential risks of abnormalities, and provide reference for formulating treatment plans. Based on the cluster analysis of abnormal areas, the diffusion pattern and risk characteristics of abnormal areas are identified, and different types of abnormalities are accurately distinguished, which improves the abnormality identification ability and processing accuracy of the system. Through the oral abnormality quantitative analysis model, the system can quantify the severity of the abnormal area and compare it with the oral health baseline characteristic model to provide scientific basis and data support for subsequent nursing decisions. Based on the oral structure abnormality quantitative data, the oral structure care status analysis module can perform time series analysis on the patient's care status, identify health fluctuation trends in the care cycle, analyze the short-term trend of the care status, predict the patient's oral health development direction, identify potential health risks in advance, help patients maintain the stability of oral health, and reduce the sudden or worsening of oral problems.The oral care decision intelligent adjustment module uses time series feature data to dynamically adjust the care decision, which can provide personalized care optimization solutions according to the patient's real-time care status and health trends, ensuring the efficiency and adaptability of the care plan. The optimized care decision is automatically transmitted to the terminal device, supporting intelligent oral care operations, reducing the complexity of manual intervention, and improving the automation level and execution efficiency of nursing operations. Through the care decision gradient benefit analysis, different care can be evaluated for dynamic optimization and adjustment, greatly improving the automation, intelligence and precision level of oral care monitoring. Therefore, the oral care monitoring imaging system of the present invention performs abnormality detection on the model blocks corresponding to the high-incidence abnormal areas in the oral care monitoring images, ensuring the data processing capability and abnormality detection accuracy for high-resolution oral images, thereby improving the analysis capability and efficiency of oral care monitoring imaging, and can comprehensively and finely detect microlesions or early abnormalities of oral tissues such as teeth and periodontal tissues. And provide a dynamic analysis technology for oral care monitoring imaging, realize dynamic tracking and analysis means of oral health status, and provide effective support for long-term oral health management.

[0153] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0154] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An oral care monitoring imaging system, characterized in that: Includes the following modules: The oral health image feature analysis module is used to obtain historical oral health image data in the oral care database; perform multi-scale fusion feature mining and processing of oral health images based on the historical oral health image data to generate oral health image feature data; A health baseline characteristic model building module is used to design a preliminary oral health baseline characteristic mapping relationship based on oral health image characteristic data and generate a preliminary oral health baseline characteristic model; Adjusting the bias weights of model parameters of the preliminary oral health baseline characteristic model to generate an oral health baseline characteristic model; The oral structure abnormal area analysis module is used to perform oral structure imaging scanning and processing on the patient through an oral scanner to establish a three-dimensional model of the oral structure; The oral health baseline characteristic model is used to analyze and process the oral structure abnormality area of ​​the three-dimensional model of the oral structure to generate data on the oral structure abnormality area; The oral structure abnormality quantification module is used to mine and process the spatial diffusion influence relationship of oral abnormalities based on the oral structure abnormality area data, and generate oral abnormality spatial diffusion influence relationship data; Performing oral structure abnormality pattern feature mining based on oral structure abnormality area data to generate oral structure abnormality pattern feature data; Establish an oral abnormality quantitative analysis model based on oral structure abnormality pattern feature data and oral abnormality spatial diffusion influence relationship data; transmit oral structure abnormality area data to the oral abnormality quantitative analysis model to perform oral structure abnormality quantitative analysis and generate oral structure abnormality quantitative data; An oral structure care status analysis module is used to perform oral structure care status time series characteristic analysis based on oral structure abnormality quantitative data, and generate oral structure care status time series characteristic data; Oral care decision intelligent adjustment module, used to obtain patients' oral care decisions; Through the oral structure care status time series characteristic data, the patient's oral care decision is dynamically optimized and adjusted to generate the patient's oral care optimization decision; the patient's oral care optimization decision is transmitted to the terminal to execute the oral care intelligent adjustment operation.

2. The oral care monitoring imaging system according to claim 1, characterized in that: The oral health image feature analysis module includes the following functions: Access historical oral health imaging data from the oral care database; Perform image edge analysis processing based on historical oral health image data to generate historical oral health image edge data; perform historical oral area division processing on historical oral health image data based on historical oral health image edge data to generate historical partitioned oral image data; Perform multi-scale feature analysis of partitioned oral images on historical partitioned oral image data to generate partitioned multi-scale feature data; Performing partition multi-scale feature weighted fusion processing according to the partition multi-scale feature data to generate partition multi-scale fusion feature data; Based on the partitioned multi-scale fusion feature data, the multi-scale fusion feature mining of oral health images is performed on the historical partitioned oral image data to generate oral health image feature data.

3. The oral care monitoring imaging system according to claim 1, characterized in that: The health baseline characteristic model building module includes the following functions: Design a preliminary oral health baseline feature mapping relationship based on oral health image feature data to generate a preliminary oral health baseline feature model; Using the preset baseline characteristic model test threshold to screen the test samples of historical oral health image data, and obtain oral health test image data; Perform model fusion feature bias anomaly analysis on the preliminary oral health baseline feature model through oral health examination image data to generate model fusion feature bias anomaly data; The model bias adjustment weight analysis is performed based on the model fusion feature bias abnormal data to generate the model bias adjustment weight data; the model bias adjustment weight data is used to adjust the model parameter weights of the preliminary oral health baseline feature model to generate the oral health baseline feature model.

4. The oral care monitoring imaging system according to claim 1, characterized in that: The oral structure abnormal area analysis module includes the following functions: Perform multi-angle oral structure imaging scans on patients through an oral scanner to generate multi-angle oral structure image data, and establish a three-dimensional oral structure model based on the multi-angle oral structure image data; The three-dimensional model of the oral structure is divided into blocks according to a preset model division threshold value to generate an oral structure model block; Performing grayscale gradient analysis on the oral structure model block to generate grayscale gradient data inside the model block; When the extreme value difference in the grayscale gradient data inside the model block is greater than a preset grayscale gradient extreme value difference threshold, the oral structure model block corresponding to the grayscale gradient data inside the model block is identified as a target model block to generate target oral structure model block data; The target oral structure model block is transmitted to the oral health baseline characteristic model to perform oral health abnormality analysis and processing on the model block, including: when the oral health baseline characteristic model identifies that the target oral structure model block has a local abnormality, generating local oral abnormality data of the model block; When the oral health baseline feature model identifies that the target oral structure model block has global abnormalities, the model block global oral abnormality data is generated; when the oral health baseline feature model does not identify that the target oral structure model block has abnormalities, the model block oral health data is generated; The oral structure abnormality area data is analyzed based on the local oral abnormality data of the model block and the global oral abnormality data of the model block to generate the oral structure abnormality area data.

5. The oral care monitoring imaging system according to claim 4, characterized in that: The oral structure abnormality area analysis is performed according to the local oral abnormality data of the model block and the global oral abnormality data of the model block to generate the oral structure abnormality area data, including: When the local oral abnormality data of the model block is identified, a first abnormal area positioning process of the oral structure is performed based on the local oral abnormality data of the model block to generate the first abnormal area positioning data of the oral structure; When the model block global oral abnormality data is identified, a second abnormal area positioning process of the oral structure is performed based on the model block global oral abnormality data to generate the second abnormal area positioning data of the oral structure; Based on the first oral structure abnormal area positioning data and the second oral structure abnormal area positioning data, the oral structure abnormal area data is integrated to generate the oral structure abnormal area data.

6. The oral care monitoring imaging system according to claim 5, characterized in that: The first abnormal area positioning process of the oral structure based on the local oral abnormality data of the model block includes: Perform local abnormality descriptor analysis on the local oral abnormality data of the model block to generate a local abnormality descriptor of the model segmentation block; perform local abnormality direction traversal analysis and processing based on the local abnormality descriptor of the model segmentation block to generate local abnormality extension direction traversal data; identify oral abnormality feature data of the local abnormality extension direction traversal based on the local abnormality extension direction traversal data; locate the first abnormal area of ​​the oral structure based on the oral abnormality feature data of the local abnormality extension direction traversal and the local oral abnormality data of the model block to generate oral structure first abnormal area positioning data.

7. The oral care monitoring imaging system according to claim 5, characterized in that: The second abnormal area positioning process of the oral structure based on the global oral abnormality data of the model block includes: Perform global abnormality direction expansion traversal on the global oral abnormality data of the model block to generate global abnormality direction expansion traversal data; identify oral abnormality feature data of the global abnormality expansion direction traversal based on the global abnormality direction expansion traversal data; perform second abnormal area positioning processing on the oral structure based on the oral abnormality feature data of the global abnormality expansion direction traversal and the global oral abnormality data of the model block to generate oral structure second abnormal area positioning data.

8. The oral care monitoring imaging system according to claim 1, characterized in that: The oral structure abnormality quantification module includes the following functions: Performing spatial distribution characteristic analysis of the abnormal area based on the data of the abnormal area of ​​the oral structure to generate spatial distribution characteristic data of the abnormal area; Mining and processing the spatial diffusion influence relationship of oral abnormalities based on the spatial distribution characteristic data of abnormal areas to generate spatial diffusion influence relationship data of oral abnormalities; Perform oral structure abnormality data analysis based on oral structure abnormality area data to generate oral structure abnormality data; Performing cluster analysis on the oral structure abnormality data to generate clustered oral structure abnormality data; Using a preset abnormal pattern rule library to perform abnormal pattern recognition on clustered oral structure abnormal data, and generate oral structure abnormal pattern data; Performing oral structure abnormality pattern feature mining processing according to the oral structure abnormality pattern data to generate oral structure abnormality pattern feature data; Based on the oral structure abnormality pattern feature data and the oral abnormality spatial diffusion influence relationship data as oral abnormality risk association factors, a mapping relationship between oral abnormality characteristics and abnormality quantification degree is established through a preset logistic regression model and oral abnormality risk association factors to generate an oral abnormality quantitative analysis model; The data of the oral structure abnormality area is transmitted to the oral structure abnormality quantitative analysis model for oral structure abnormality quantitative analysis to generate oral structure abnormality quantitative data.

9. The oral care monitoring imaging system according to claim 1, characterized in that: The oral structure care status analysis module includes the following functions: Analyze the oral structure care status based on the quantitative data of oral structure abnormalities and generate oral structure care status data; Perform oral structure care status analysis of the care cycle according to the oral structure care status data to generate periodic oral structure care status data; Performing short-term trend analysis of the oral structure care status based on the oral structure care status data to generate short-term trend oral structure care status data; Based on the periodic oral structure care status data and the short-term trend oral structure care status data, the oral structure care status time series feature mining is performed to generate the oral structure care status time series feature data.

10. The oral care monitoring imaging system according to claim 1, characterized in that: The oral care decision intelligent adjustment module includes the following functions: Obtaining patient oral care decisions; Conduct oral care decision gradient benefit analysis on patients’ oral care decisions based on oral structure care status time series characteristic data, and generate oral care decision gradient benefit data; Based on the oral care decision gradient benefit data, the patient's oral care decision is dynamically optimized and adjusted to generate the patient's oral care optimization decision; Transmit the patient's oral care optimization decision to the terminal to execute intelligent oral care adjustment operations.

Images