A method and system for scoring children's oral hygiene

Abstract

The present invention discloses a method and system for evaluating children's oral cleaning, which relates to the technical field of oral cleaning. The method includes the following steps: obtaining the basic identity information and cleaning behavior information of children participating in oral cleaning evaluation; through information recognition and behavior analysis, marking the current oral development stage of children, and preliminarily evaluating the cleaning behavior score of children according to the behavior habits in the oral development stage; through data processing and feature extraction of oral examination data, generating oral evaluation indicators, inputting them into a constructed multi-instance scoring model to evaluate the oral health score; integrating the cleaning behavior score and the oral health score, and comprehensively outputting the oral cleaning score. The present invention can evaluate cleaning behavior personalized according to the oral development stage of children, avoid the insufficient applicability of traditional scoring methods to children in different development stages, more comprehensively reflect the overall oral cleaning level of children, and greatly improve the practicability and pertinence of the evaluation report.

Description

Technical Field

[0001] The present invention relates to the technical field of oral cavity cleaning, and particularly relates to a method and system for scoring children's oral cavity cleaning. Background Art

[0002] Oral health is an important part of overall health. Good oral health not only concerns the state of teeth and gums, but also affects digestion, nutrient absorption and mental health. The key to maintaining oral health lies in regular oral examinations, cleaning and developing good oral hygiene habits. Oral cavity cleaning mainly includes brushing teeth, using dental floss, etc., aiming to remove food residues and dental plaque on the tooth surface and between tooth gaps, and prevent dental caries and gum diseases. Daily cleaning should include brushing teeth at least twice a day, using fluoride toothpaste, and regularly replacing toothbrushes to maintain the cleaning effect.

[0003] Oral cavity cleaning for children is particularly important because their teeth and gums are more vulnerable and susceptible to bacterial invasion. When children are about two years old, parents should start to help them with oral cavity cleaning and teach them the correct toothbrushing method. In children's oral cavity cleaning, using a soft toothbrush and low-fluoride toothpaste specifically for children is a recommended choice. Since children's toothbrushing skills are not yet mature, parents should accompany and guide them to ensure that every part can be adequately cleaned. In addition, regular dental examinations and professional cleaning cannot be ignored, which can timely detect potential problems and intervene.

[0004] Current oral cavity cleaning detection and scoring technologies mainly focus on imaging analysis, manual inspection, sensor technology, etc. Common oral cavity cleaning detection means include dental plaque staining detection, oral endoscope imaging analysis, periodontal probe detection, optical detection, and using image recognition technology to evaluate the cleanliness and health status of teeth. In addition, some intelligent toothbrushes and cleaning devices are equipped with sensors that can real-time feedback data such as brushing strength and duration, and give cleaning scores through supporting software. These technologies have high accuracy and operability when applied to the adult population.

[0005] However, there are certain particularities and challenges in children's oral cavity cleaning detection and scoring. First of all, children are in the mixed dentition stage, with deciduous teeth and permanent teeth coexisting, and the teeth are not yet fully developed, so the difficulty of tooth cleaning is relatively large. Secondly, children's toothbrushing habits are often not standardized, and it is easy to miss cleaning areas, making the traditional scoring method ineffective for children. And although the dental plaque staining detection for children is intuitive, concerns about the impact of the staining agent on tooth health and psychology limit the application of this technology. Summary of the Invention

[0006] Based on this, it is necessary to provide a method and system for scoring children's oral cavity cleaning in view of the above technical problems.

[0007] In a first aspect, the present invention provides a method for scoring children's oral cleaning, which includes the following steps:

[0008] S1. Obtain the basic identity information and cleaning behavior information of children participating in oral cleaning scoring, and detect the oral examination data of the children through an oral health detection device.

[0009] S2. Mark the current oral development stage of the child through information recognition and behavior analysis, and preliminarily evaluate the cleaning behavior score of the child according to the behavior habits in the current oral development stage.

[0010] S3. Generate oral evaluation indicators through data processing and feature extraction of oral examination data, input them into a constructed multi-instance scoring model, and evaluate the oral health score.

[0011] S4. Integrate the cleaning behavior score and the oral health score, comprehensively output the oral cleaning score, and generate a scoring evaluation report according to the scoring result. Display the oral cleaning score of the child through a data visualization tool and provide personalized cleaning improvement suggestions.

[0012] Generating oral evaluation indicators through data processing and feature extraction of oral examination data, inputting them into a constructed multi-instance scoring model, and evaluating the oral health score includes the following steps:

[0013] S31. Divide the oral examination data into image data and numerical data according to the detection method of the oral examination data, and perform data preprocessing by adopting corresponding data processing methods according to the data type.

[0014] S32. Extract the effective features of each type of oral examination data to generate oral evaluation indicators.

[0015] S33. Set the index value range and convert the oral evaluation indicators into multiple groups of binary inputs.

[0016] S34. Establish a multi-instance scoring model by defining bags and examples, calculate the individual scores of each example respectively, and then integrate and calculate the overall score of the bag as the oral health score.

[0017] Furthermore, marking the current oral development stage of the child through information recognition and behavior analysis, and preliminarily evaluating the cleaning behavior score of the child according to the behavior habits in the current oral development stage includes the following steps:

[0018] S21. Mark the current oral development stage of the child by extracting and identifying the age information, gender and tooth growth record in the basic identity information, and the oral development stage includes the primary tooth eruption stage, the primary tooth occlusion completion stage, the mixed dentition stage and the permanent dentition stage.

[0019] S22. By identifying the cleaning behavior information, extract the cleaning frequency, cleaning duration, cleaning tools and eating habits of the child, and match the corresponding cleaning scoring criteria according to the current oral development stage.

[0020] S23. According to the cleaning scoring criteria, evaluate the actual scores of the data in the cleaning behavior information, and use the cumulative summation method to take the calculation result as the cleaning behavior score of the preliminary evaluation.

[0021] Further, to extract the effective features of various types of oral examination data and generate oral evaluation indicators, the following steps are included:

[0022] S321. Extract the image data after data preprocessing, use a deep learning model to automatically segment the teeth, identify the position and shape of each tooth, and use a classification model for identification according to the current oral development stage, and mark the type and number of each tooth.

[0023] S322. Detect the health status of each tooth, extract the cleaning effect features by analyzing the tooth surface image, and use the image analysis method to quantify the cleaning effect to form feature indicators.

[0024] S323. Extract the numerical data after data preprocessing, match each type of numerical data with the tooth position in the image data, and map it to the corresponding tooth number.

[0025] S324. According to the tooth position distribution, integrate the feature indicators extracted from the image data and numerical data, and combine with the tooth number to generate oral evaluation indicators.

[0026] Further, to set the index value range and convert the oral evaluation indicators into multiple groups of binary inputs, the following steps are included:

[0027] S331. Obtain the quantified oral evaluation indicators. According to the nine-point system, divide the value range of each oral evaluation indicator into (1, C n - 1), and set the classification threshold. For each classification threshold, generate a binary input, and C n represents the total score, and the classification threshold ∈ (1, C n - 1).

[0028] S332. For the divided classification thresholds, construct binary classifiers. Each type of oral evaluation indicator corresponds to a binary classifier. Take the oral evaluation indicators corresponding to each tooth as the input of the binary classifier and output the corresponding binary results.

[0029] S333. Integrate the binary results corresponding to each tooth to form multiple groups of binary inputs.

[0030] Further, by defining bags and examples, a multi-instance scoring model is established. The individual scores of each example are calculated respectively, and then the overall score of the bag is calculated through integration as the oral health score, including the following steps:

[0031] S341. Define the entire oral cavity of a child as a bag, which contains all oral evaluation indicators, and then define each tooth as an example, and each example contains the oral evaluation indicators under the corresponding number.

[0032] S342. Integrate the binary classifier and the Gaussian model, adjust the weight values of different binary classifiers to form a combined classifier, and calculate the individual scores of each example.

[0033] S343. Integrate the individual scores of all examples, combine the distribution of pathological individuals, construct a multi-instance scoring model, and output the oral health score of the entire oral cavity through model calculation.

[0034] Further, integrating the binary classifier and the Gaussian model, adjusting the weight values of different binary classifiers to form a combined classifier, and calculating the individual scores of each example includes the following steps:

[0035] S3421. For each example, use the oral evaluation indicators of each tooth as input features.

[0036] S3422. According to the deviation degree of the position of the binary classifier and the middle score, use the Gaussian model to assign weight values to each binary classifier, adjust the contribution degree of each binary classifier to the individual score, and form a combined classifier by integrating all binary classifiers.

[0037] S3423. Input the input features into the combined classifier to calculate the individual scores of each example.

[0038] Further, the calculation formula for the individual score of each example is:

[0039] ;

[0040] In the formula, S i represents the individual score of the i th example; j represents the binary classifier serial number; M represents the number of binary classifiers; g (·) represents the weight function; C i represents the historical average score value of the i th example; C n represents the total score value; f ij represents thei the score of the j second binary classifier in one example.

[0041] Furthermore, by integrating the individual scores of all examples and combining the distribution of pathological individuals, a multi-instance scoring model is constructed. Through model calculation, the oral health score of the entire oral cavity is output, including the following steps:

[0042] S3431. Obtain the individual scores of all examples in the oral cavity. At the same time, sort out the types of teeth corresponding to each example and assign individual weights to the examples.

[0043] S3432. According to the examination results of the oral cavity and dental health status, count the types and quantities of pathological teeth in the oral cavity, and set a penalty factor to represent the non-linear negative impact.

[0044] S3433. Integrate the scores of all examples and the penalties for pathological individuals to construct a multi-instance scoring model. Through model calculation, obtain the oral health score of the child's oral cavity.

[0045] Furthermore, the expression of the multi-instance scoring model is:

[0046] ;

[0047] In the formula, Q represents the oral health score; S i represents the i individual score of the α i i-th example; β represents the penalty factor; N represents the number of examples; U represents the number of pathological teeth in the oral cavity.

[0048] On the second aspect, the present invention also provides a child oral cavity cleaning scoring system, which includes:

[0049] A data acquisition module, which is used to obtain the basic identity information and cleaning behavior information of children participating in oral cavity cleaning scoring, and detect the oral examination data of the child through an oral health detection device.

[0050] An identification and analysis module, which is used to mark the current oral development period of the child through information identification and behavior analysis, and preliminarily evaluate the cleaning behavior score of the child according to the behavior habits in the oral development period.

[0051] A health assessment module, which is used to generate oral assessment indicators through data processing and feature extraction of oral examination data, and input them into the constructed multi-instance scoring model to evaluate the oral health score.

[0052] A scoring display module is used to integrate the cleaning behavior score and the oral health score, comprehensively output the oral cleaning score, and generate a scoring evaluation report according to the scoring result. The oral cleaning score of children is displayed through a data visualization tool, and personalized cleaning improvement suggestions are provided.

[0053] Among them, the data acquisition module, the recognition and analysis module, the health assessment module, and the scoring display module are connected in sequence.

[0054] The beneficial effects of the present invention are as follows:

[0055] 1. Through targeted information recognition and behavior analysis, the cleaning behavior can be personalized evaluated according to the oral development stage of children, avoiding the insufficient applicability of traditional scoring methods to children in different development stages; by processing oral examination data through a multi-instance scoring model, ensuring that the health status of each tooth is carefully evaluated, and then integrating the cleaning behavior score and the oral health score, the overall oral cleaning level of children can be more comprehensively reflected; in addition, a dual scoring mechanism combining behavior data and health data is adopted, and personalized cleaning improvement suggestions are provided by using a data visualization tool, greatly improving the practicality and pertinence of the evaluation report, being more suitable for the special oral development characteristics and behavior habits of children, and being able to provide more accurate and operable suggestions for parents and doctors.

[0056] 2. Oral evaluation indicators are generated through data processing and feature extraction of oral examination data, and oral health scores are combined with a multi-instance scoring model, which can perform targeted processing on different types of oral data, ensuring the flexibility and accuracy of the data processing process; secondly, through effective feature extraction, the comprehensiveness and accuracy of oral evaluation indicators are guaranteed; by using a multi-instance scoring model, independent scoring can be performed on each tooth, and then the overall oral health score is comprehensively calculated, avoiding the problem of ignoring individual differences in traditional evaluation methods. By introducing a multi-instance model with data type division and binary input, higher scoring accuracy and individual difference analysis are provided; at the same time, by integrating multiple scores, a more comprehensive oral health assessment result is provided for doctors and patients, thus forming a more accurate diagnosis and personalized nursing suggestions for oral problems, with stronger practicality and guidance. Description of the Drawings

[0057] The drawings described herein are used to provide a further understanding of the present invention, form a part of the present invention, and the schematic embodiments of the present invention and their descriptions are used to explain the present invention, and do not constitute an improper limitation to the present invention. In the drawings:

[0058] Figure 1 is a flowchart of a method for scoring children's oral cleaning according to an embodiment of the present invention;

[0059] Figure 2 It is a schematic block diagram of a children's oral cleaning scoring system according to an embodiment of the present invention.

[0060] Reference numerals in the attached drawings: 1. Data acquisition module; 2. Identification and analysis module; 3. Health assessment module; 4. Scoring display module. Specific embodiments

[0061] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the attached drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0062] Please refer to Figure 1 , and a children's oral cleaning scoring method is provided. The method includes the following steps:

[0063] S1. Obtain the basic identity information and cleaning behavior information of children participating in oral cleaning scoring, and detect the oral examination data of the children through an oral health detection device.

[0064] Specifically, an electronic health record system can be used to automatically collect children's basic identity information, such as name, age, gender, oral health history, etc. Standardized questionnaires or behavior record forms filled out by parents are used to collect cleaning behavior information. For example, the brushing frequency (times per day), brushing duration (minutes per time), types of toothpaste and toothbrush used, and eating habits (frequency of sweet food intake, etc.).

[0065] Through an oral health detection device, obtain the oral examination data of the children, including image or numerical data such as the health status, cleanliness, caries or dental calculus conditions of the teeth. Numerical and image-based oral data are obtained through conventional oral examination equipment in oral hospitals.

[0066] Among them, the oral health detection device includes related devices such as an oral scanner, dental X-ray machine, oral endoscope, periodontal probe, laser caries detector, oral CT, and digital dental camera, which are used to obtain images or other detection parameters inside the oral cavity, and understand information such as the internal structure of the oral cavity, tooth position, tooth morphology, tooth surface information, caries, periodontitis, and gingival status.

[0067] S2. Through information identification and behavior analysis, mark the current oral development stage of the child, and preliminarily evaluate the cleaning behavior score of the child according to the behavior habits in the oral development stage.

[0068] In the description of the present invention, through information identification and behavior analysis, marking the current oral development stage of the child, and preliminarily evaluating the cleaning behavior score of the child according to the behavior habits in the oral development stage includes the following steps:

[0069] S21. Mark the current oral development stage of the child by extracting and identifying the age information, gender, and tooth growth records in the basic identity information. The oral development stage includes the primary tooth eruption stage, the completion stage of primary tooth occlusion, the mixed dentition stage, and the permanent dentition stage.

[0070] Specifically, based on the collected identity and behavior information, determine the current oral development stage of the child:

[0071] 1. Primary tooth eruption stage (0 - 3 years old): Mainly for the eruption of primary teeth, pay attention to the number and health status of primary teeth.

[0072] 2. Completion stage of primary tooth occlusion (3 - 6 years old): The occlusion of primary teeth is gradually completed. At this stage, pay attention to whether the child's oral hygiene habits are guided by parents.

[0073] 3. Mixed dentition stage (6 - 12 years old): Primary teeth are gradually replaced, and permanent teeth begin to erupt. At this stage, pay attention to the child's self - cleaning ability and the process of tooth replacement.

[0074] 4. Permanent dentition stage (12 years old and above): Mainly permanent teeth, pay attention to the child's independent cleaning ability and the cultivation of good oral hygiene habits.

[0075] S22. By identifying the cleaning behavior information, extract the cleaning frequency, cleaning duration, cleaning tools, and eating habits of the child, and match the corresponding cleaning scoring criteria according to the current oral development stage.

[0076] Specifically, according to the characteristics of different oral development stages, set the corresponding scoring criteria. The set scoring items and score values are shown in the following examples.

[0077] 1. Primary tooth eruption stage: Mainly evaluate the participation of parents, whether they have the habit of helping the child brush teeth, and give a score. The scoring criteria can be set as follows:

[0078] 1.1. Brushing frequency: Once a day (2 points), twice a day (4 points);

[0079] 1.2. Parent participation: Yes (3 points), No (0 points).

[0080] 2. Completion stage of primary tooth occlusion: Pay attention to whether the child can brush teeth independently under the guidance of parents, and evaluate the child's understanding of cleaning behavior. The scoring criteria can be set as follows:

[0081] 2.1. Brushing frequency: Once a day (2 points), twice a day (4 points);

[0082] 2.2. Brushing time: < 2 minutes (1 point), 2 - 3 minutes (3 points), > 3 minutes (5 points);

[0083] 2.3. Parent guidance: Yes (3 points), No (0 points);

[0084] 2.4. Toothbrushing tools: using a dedicated toothbrush or toothpaste (3 points), not using a dedicated toothbrush or toothpaste (0 points).

[0085] 3. Replacement dentition period: assess the child's independent brushing ability and brushing skills, such as whether the child can use the toothbrush correctly and whether the brushing is comprehensive. The scoring criteria can be set as follows:

[0086] 3.1. Toothbrushing frequency: once a day (2 points), twice a day (4 points);

[0087] 3.2. Brushing time: <2 minutes (1 point), 2-3 minutes (3 points), >3 minutes (5 points);

[0088] 3.3. Toothbrushing skills (e.g., whether all teeth can be cleaned): Yes (5 points), No (0 points).

[0089] 4. Permanent dentition stage: Emphasize the self-management ability of children, and pay attention to whether they can brush their teeth and use dental floss every day. The scoring criteria can be set as follows:

[0090] 4.1. Toothbrushing frequency: once a day (2 points), twice a day (4 points);

[0091] 4.2. Brushing time: <2 minutes (1 point), 2-3 minutes (3 points), >3 minutes (5 points);

[0092] 4.3. Use of dental floss: Yes (5 points), No (0 points).

[0093] S23. According to the cleaning scoring standard, the actual scores of each data in the cleaning behavior information are evaluated, and the calculation results are used as the cleaning behavior scores of the preliminary evaluation by cumulative summation.

[0094] Specifically, the questionnaire results can be quantified to generate a cleaning behavior score. Specific scoring criteria need to be set for each question, and finally the scores are added up to get a comprehensive score. The behavior score is then weighted and adjusted according to the child's developmental stage to ensure the fairness and effectiveness of the score.

[0095] For example, suppose a child in the replacement dentition period brushes twice a day (score 4 points), brushes for 2-3 minutes (score 3 points), and can clean all teeth (score 5 points). And the scoring standard can be set according to the highest score in each oral development stage: for example, for the replacement dentition period, the full score is 20 points, and 12 points can be considered "good", but still needs improvement.

[0096] Ultimately, the specific profiles and scores for each developmental period can be combined to provide a more in-depth oral health analysis and identify areas for improvement in a child's cleaning habits.

[0097] S3. Generate oral evaluation indicators through data processing and feature extraction of oral examination data, and input them into the constructed multi-instance scoring model to evaluate the oral health score.

[0098] Specifically, the present invention uses a multi-instance scoring model built with a multi-instance learning algorithm to evaluate the oral health score of children. Multi-instance Learning (MIL) is a special machine learning framework suitable for dealing with data with incomplete or unclear labels. In traditional supervised learning, each training sample has a clear label, while in multi-instance learning, the data is organized in the form of bags, each bag contains multiple instances, and each bag has only one label, rather than the label of a single instance.

[0099] In other words, a bag is labeled as positive or negative, but it is not clear whether each instance in the bag is positive or negative. Multi-instance learning is a powerful tool that can effectively handle complex and unclear-labeled data. By constructing a multi-instance scoring model, the information of multiple instances can be integrated to output meaningful scoring results. Therefore, it also plays a key role in the oral cleaning scoring scenario of the present invention.

[0100] In the description of the present invention, generating oral evaluation indicators through data processing and feature extraction of oral examination data, and inputting them into the constructed multi-instance scoring model to evaluate the oral health score includes the following steps:

[0101] S31. Divide the oral examination data into image data and numerical data according to the detection method, and adopt corresponding data processing methods according to the data type to perform data preprocessing.

[0102] Specifically, the image data includes images of the internal oral structure, such as dental X-rays, intraoral camera images, tooth surface images, etc. These data are mainly used for visual recognition and image analysis.

[0103] The preprocessing of the image data includes: 1. Denoising processing, applying filters (such as Gaussian filter, mean filter, etc.) to remove the noise in the image to improve the image quality. 2. Image enhancement: Using methods such as contrast enhancement and histogram equalization to improve the visibility of the image, making the teeth and other structures more obvious. 3. Standardization processing: Unifying the image size, for example, adjusting all images to the same resolution (such as 256×256 pixels) for subsequent analysis. 4. Image segmentation: Using deep learning models (such as U-Net, FCN, etc.) for image segmentation to accurately identify and extract teeth and other related structures.

[0104] Numerical data includes quantitative data such as the depth of teeth, cleaning frequency, dental health score, age and gender of children, etc. These data are used for statistical analysis and feature extraction.

[0105] The preprocessing of numerical data includes: 1. Data cleaning: Check for missing values, outliers, and duplicate data in the dataset and take corresponding measures (such as filling missing values, deleting outliers, etc.) to ensure the integrity and accuracy of the data. 2. Standardization and normalization: Perform standardization (such as Z-score standardization) or normalization (scale the data to the [0, 1] interval) on numerical data to eliminate the impact of different dimensions on the analysis results. 3. Feature selection: Select features related to oral health for further analysis based on correlation analysis and statistical tests (such as analysis of variance, correlation coefficient, etc.) to reduce redundant data.

[0106] S32. Extract effective features of various types of oral examination data to generate oral evaluation indicators.

[0107] In the description of the present invention, extracting effective features of various types of oral examination data to generate oral evaluation indicators includes the following steps:

[0108] S321. Extract the image data after data preprocessing, use a deep learning model to automatically segment teeth, identify the position and shape of each tooth, and use a classification model for identification according to the current oral development stage to label the type and number of each tooth.

[0109] S322. Detect the health status of each tooth, extract cleaning effect features by analyzing the tooth surface image, and use image analysis methods to quantify the cleaning effect to form feature indicators.

[0110] Specifically, detect health problems such as tooth defects, dental caries, and discoloration through image analysis, and a classifier (such as a support vector machine or a deep learning model) can be trained to judge the health status of teeth, such as normal, mild dental caries, moderate dental caries, etc.

[0111] In the process of quantifying the cleaning effect, image processing techniques (such as edge detection, texture analysis) need to be used to extract the cleaning features on the tooth surface. The following features can be considered:

[0112] 1. Cleanliness index: Analyze the difference between the cleaned image and the image before cleaning to quantify the effect of removing dirt or plaque.

[0113] 2. Surface smoothness: Evaluate the texture features of the tooth surface and calculate the smoothness index using texture analysis algorithms (such as LBP, Gabor filter).

[0114] 3. Color difference analysis: Calculate the color change on the tooth surface before and after cleaning using a color space (such as HSV or Lab) to reflect the cleaning effect.

[0115] Quantify the extracted cleaning effect features into numerical indicators (such as cleanliness score, smoothness score, color difference score, etc.), and integrate them according to the set standards to generate a comprehensive cleaning effect feature index.

[0116] S323. Extract the numerical data after preprocessing the data, match each type of numerical data with the tooth positions in the image data, and map them to the corresponding tooth numbers.

[0117] S324. Integrate the feature indicators extracted from the image data and numerical data according to the tooth position distribution, and combine with the tooth numbers to generate oral evaluation indicators.

[0118] Specifically, the oral evaluation indicators may include: 1. Tooth position and shape: Identify the exact position, shape, and type of each tooth, including whether the tooth arrangement is neat, whether there are missing teeth or malformations, etc. 2. Health status score: Evaluate the health status of each tooth, usually by observing the appearance, color, wear degree of the tooth, and whether there are diseases such as dental caries and periodontal disease for scoring. 3. Cleaning effect features: Quantify the oral cleaning effect of children, which may include the dirt degree on the tooth surface, the thickness of dental plaque, etc., which can be achieved through image analysis methods. 4. Oral hygiene behavior score: Evaluate the oral hygiene behavior of children according to their brushing habits, use of dental floss, eating habits, etc.

[0119] S33. Set the value range of the indicators and convert the oral evaluation indicators into multiple groups of binary inputs.

[0120] In the description of the present invention, setting the value range of the indicators and converting the oral evaluation indicators into multiple groups of binary inputs includes the following steps:

[0121] S331. Obtain the quantified oral evaluation indicators, divide the value range of each oral evaluation indicator into (1, C n -1) according to the nine-point system, and set the classification threshold. For each classification threshold, generate a binary input, and C n represents the total score, and the classification threshold ∈ (1, C n -1).

[0122] Specifically, in the previous steps, the various oral evaluation indicators have been quantified to obtain the corresponding scoring data. For example, the evaluation indicators may include the cleanliness, health status, caries risk of the teeth, etc. According to the set nine-point scoring system, divide the value range of each evaluation indicator into (1, C n -1), where C nRepresents the total score of all metrics (e.g., 9), and the value range is from 1 to 8. For each evaluation metric, set the classification threshold according to the actual situation. For example, the threshold can be set to 3, 5, 7, etc., and the specific choice of the threshold should be determined based on clinical experience, literature research, or expert opinions.

[0123] For each classification threshold, generate a binary input. That is, when the oral evaluation metric is less than or equal to the threshold, the binary result is 0 (indicating not meeting the standard). When the oral evaluation metric is greater than the threshold, the binary result is 1 (indicating meeting the standard). For example, for a metric with a score of 6, if the set threshold is 5, the binary result is 1.

[0124] S332. For the divided classification thresholds, construct binary classifiers. Each type of oral evaluation metric corresponds to a binary classifier. Use the corresponding oral evaluation metric of each tooth as the input of the binary classifier and output the corresponding binary result.

[0125] Specifically, the present invention divides the multi - value classification problem into multiple binary classifiers. Specifically, for each example (such as each tooth), divide its multi - value score of the health status (such as a 9 - point scale score) into several binary classifiers. Each classifier corresponds to a binary decision problem, that is, to judge whether the score of the example is above or below a certain threshold.

[0126] For each type of oral evaluation metric (such as cleanliness, health status, etc.), construct a corresponding binary classifier respectively. The task of each classifier is to judge the state of the tooth based on the input oral evaluation metric. Use the corresponding oral evaluation metric of each tooth as the input data and provide it to the corresponding binary classifier. For example, if the cleanliness score of a certain tooth is 7, then use it as the input and send it to the cleanliness classifier for judgment.

[0127] S333. Integrate the binary results corresponding to each tooth to form multiple groups of binary inputs.

[0128] Specifically, each classifier generates a binary output according to the input score, respectively representing the state of the tooth on a specific oral evaluation metric. Finally, a set of binary results will be obtained. For example:

[0129] Cleanliness classifier output: 1 (meeting the standard);

[0130] Health status classifier output: 0 (not meeting the standard);

[0131] Dental caries risk classifier output: 1 (meeting the standard).

[0132] S34. By defining packages and examples, establish a multi - example scoring model, calculate the individual scores of each example respectively, and then integrate and calculate the overall score of the package as the oral health score.

[0133] In the description of the present invention, by defining packages and examples, a multi-example scoring model is established. The individual scores of each example are calculated respectively, and then the overall score of the package is calculated through integration, which is used as the oral health score, including the following steps:

[0134] S341. Define the entire oral cavity of a child as a package, which includes all oral evaluation indicators, and then define each tooth as an example, and each example includes the oral evaluation indicators under the corresponding number.

[0135] S342. Integrate the binary classifier and the Gaussian model, adjust the weight values of different binary classifiers to form a combined classifier, and calculate the individual scores of each example.

[0136] Specifically, in order to have different weight allocations for different classifiers, the Gaussian model is used to assign weights to each classifier according to the position of the classifier (i.e., the value of the classification threshold) and the deviation degree from the middle score. The core idea of the Gaussian model is to give higher weights to the classifiers close to the middle value, and lower weights to the extreme classifiers at both ends.

[0137] In the description of the present invention, integrating the binary classifier and the Gaussian model, adjusting the weight values of different binary classifiers to form a combined classifier, and calculating the individual scores of each example includes the following steps:

[0138] S3421. For each example, use the oral evaluation indicators of each tooth as input features.

[0139] S3422. According to the position of the binary classifier and the deviation degree from the middle score, use the Gaussian model to assign weight values to each binary classifier, adjust the contribution degree of each binary classifier to the individual score, and form a combined classifier by integrating all binary classifiers.

[0140] S3423. Input the input features into the combined classifier and calculate the individual scores of each example.

[0141] In the description of the present invention, the calculation formula for the individual score of each example is:

[0142] ;

[0143] In the formula, S i represents the individual score of the i th example; j represents the serial number of the binary classifier; M represents the number of binary classifiers; g (·) represents the weight function; C i represents thei The historical average score value of one example; C n Indicates the total score value; f ij Indicates the i th j score value of the

[0144] S343. Integrate the individual scores of all examples, combine the distribution of pathological individuals, construct a multi-instance scoring model, and through model calculation, output the oral health score of the entire oral cavity.

[0145] In the description of the present invention, integrating the individual scores of all examples, combining the distribution of pathological individuals, constructing a multi-instance scoring model, and through model calculation, outputting the oral health score of the entire oral cavity includes the following steps:

[0146] S3431. Obtain the individual scores of all examples in the oral cavity. At the same time, sort out the types of teeth corresponding to each example and assign individual weights to the examples.

[0147] S3432. According to the examination results of the oral cavity and dental health status, count the types and quantities of pathological teeth in the oral cavity, and set a penalty factor to represent the non-linear negative impact.

[0148] S3433. Integrate the scores of all examples and the penalties for pathological individuals, construct a multi-instance scoring model, and through model calculation, obtain the oral health score of the child's oral cavity.

[0149] In the description of the present invention, the expression of the multi-instance scoring model is:

[0150] ;

[0151] In the formula, Q Indicates the oral health score; S i Indicates the i th α i Indicates the i th β Indicates the penalty factor; N Indicates the number of examples; U Indicates the number of pathological teeth in the oral cavity.

[0152] Suppose there is a child's oral health assessment data, involving three teeth (A, B, C), and each tooth has different oral assessment indicators and corresponding cleaning behavior scores.

[0153] I. The oral examination data can be shown as:

[0154] Tooth A: Health status score (8 / 10), cleaning effect score (7 / 10);

[0155] Tooth B: Health status score (5 / 10), cleaning effect score (6 / 10);

[0156] Tooth C: Health status score (9 / 10), cleaning effect score (8 / 10);

[0157] II. For the processing of image data, automatic segmentation of teeth is performed through a deep learning model, and the numbers and positions of Teeth A, B, and C are marked. Assuming that the image analysis results show that the surfaces of Teeth A and C are relatively smooth, while the surface of Tooth B has slight dirt. The extracted features include health status and cleaning effect scores, which are mapped to the corresponding tooth numbers.

[0158] III. Generating oral evaluation indicators from the processed data:

[0159] Tooth A: Health status indicator = 8, cleaning effect indicator = 7;

[0160] Tooth B: Health status indicator = 5, cleaning effect indicator = 6;

[0161] Tooth C: Health status indicator = 9, cleaning effect indicator = 8;

[0162] IV. Converting the indicators into multiple groups of binary inputs, setting the value range (1 - 10) for each oral evaluation indicator and converting it into binary inputs. For example:

[0163] Health status threshold: Tooth A: 8 (high), Tooth B: 5 (medium), Tooth C: 9 (high);

[0164] Cleaning effect threshold: Tooth A: 7 (medium), Tooth B: 6 (medium), Tooth C: 8 (high);

[0165] Converted to:

[0166] Tooth A: (high, medium), Tooth B: (medium, medium), Tooth C: (high, high)

[0167] V. Calculating the scores using a combined classifier

[0168] Input feature preparation:

[0169] The feature input of Tooth A is (8, 7), Tooth B is (5, 6), and Tooth C is (9, 8).

[0170] Weight assignment and construction of the combined classifier. Assuming that after calculation through the Gaussian model, the weights corresponding to each tooth are obtained: Weight of Tooth A: 0.4, Weight of Tooth B: 0.3, Weight of Tooth C: 0.3. Calculation of individual scores:

[0171] Individual score of Tooth A = 0.4×8 + 0.4×7 = 3.2 + 2.8 = 6;

[0172] Individual score of Tooth B = 0.3×5 + 0.3×6 = 1.5 + 1.8 = 3.3;

[0173] Individual score of Tooth C = 0.3×9 + 0.3×8 = 2.7 + 2.4 = 5.1.

[0174] VI. Integrate all individual scores to obtain the overall oral health score.

[0175] S4. Integrate the cleaning behavior score and the oral health score, comprehensively output the oral cleaning score, and generate a score evaluation report according to the score result. Display the oral cleaning score of children through a data visualization tool and provide personalized cleaning improvement suggestions.

[0176] Specifically, the comprehensive score can be calculated by the weighted average method. For example, set the same weight (50%) for the cleaning behavior score and the oral health score. Generate a detailed score evaluation report according to the calculation result, and then use a data visualization tool to present the score result in the form of a chart.

[0177] Please refer to Figure 2 , and also provide a children's oral cleaning score system, which includes:

[0178] Data acquisition module 1, used to obtain the basic identity information and cleaning behavior information of children participating in the oral cleaning score, and detect the oral examination data of the child through an oral health detection device.

[0179] Identification and analysis module 2, used to mark the current oral development period of the child through information identification and behavior analysis, and preliminarily evaluate the cleaning behavior score of the child according to the behavior habits in the oral development period.

[0180] Health assessment module 3, used to generate oral assessment indicators through data processing and feature extraction of oral examination data, input them into the constructed multi-instance scoring model, and evaluate the oral health score.

[0181] Score display module 4, used to integrate the cleaning behavior score and the oral health score, comprehensively output the oral cleaning score, and generate a score evaluation report according to the score result. Display the oral cleaning score of children through a data visualization tool and provide personalized cleaning improvement suggestions.

[0182] Among them, data acquisition module 1, identification and analysis module 2, health assessment module 3 and score display module 4 are connected in sequence.

[0183] In summary, by means of the above technical solutions of the present invention, through targeted information recognition and behavior analysis, it is possible to individually evaluate the cleaning behavior according to the oral development stage of children, avoiding the insufficient applicability of traditional scoring methods to children in different development stages; by processing oral examination data through a multi-instance scoring model, ensuring that the health status of each tooth is carefully evaluated, and then integrating the cleaning behavior score and the oral health score, it is possible to more comprehensively reflect the overall oral cleaning level of children; in addition, a dual scoring mechanism combining behavior data and health data is adopted, and personalized cleaning improvement suggestions are provided using data visualization tools, greatly enhancing the practicality and pertinence of the evaluation report, being more suitable for the special oral development characteristics and behavior habits of children, and being able to provide more accurate and operable suggestions for parents and doctors.

[0184] By processing and feature extraction of oral examination data to generate oral evaluation indicators, and combining with a multi-instance scoring model for oral health scoring, it is possible to perform targeted processing on different types of oral data, ensuring the flexibility and accuracy of the data processing process; secondly, through effective feature extraction, the comprehensiveness and accuracy of oral evaluation indicators are guaranteed; using the multi-instance scoring model, it is possible to independently score each tooth, and then comprehensively calculate the overall oral health score, avoiding the problem of ignoring individual differences in traditional evaluation methods. By introducing a multi-instance model with data type division and binary input, higher scoring accuracy and individual difference analysis are provided; at the same time, by integrating multiple scores, a more comprehensive oral health evaluation result is provided for doctors and patients, thus forming a more accurate diagnosis of oral problems and personalized nursing suggestions, with stronger practicality and guidance.

[0185] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the indication of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same moment, but can be executed at different moments, and their execution order is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

Claims

1. A method for scoring children's oral cavity cleaning, characterized in that, Including: S1. Obtain the basic identity information and cleaning behavior information of children participating in oral cleaning score, and detect the oral examination data of the children through an oral health detection device; S2. Mark the current oral development stage of the children through information recognition and behavior analysis, and preliminarily evaluate the cleaning behavior score of the children according to the behavior habits in the current oral development stage; S3. Generate oral evaluation indicators through data processing and feature extraction of oral examination data, input them into the constructed multi-instance scoring model, and evaluate the oral health score; S4. Integrate the cleaning behavior score and the oral health score, comprehensively output the oral cleaning score, generate a score evaluation report according to the score result, display the oral cleaning score of the children through a data visualization tool, and provide personalized cleaning improvement suggestions; Among them, the step of marking the current oral development stage of the children through information recognition and behavior analysis, and preliminarily evaluating the cleaning behavior score of the children according to the behavior habits in the current oral development stage includes: S21. Mark the current oral development stage of the children by extracting and recognizing the age information, gender and tooth growth record in the basic identity information, and the oral development stage includes the deciduous tooth eruption stage, the deciduous tooth occlusion completion stage, the mixed dentition stage and the permanent dentition stage; S22. Identify the cleaning frequency, cleaning duration, cleaning tools and eating habits of the children by recognizing the cleaning behavior information, and match the corresponding cleaning score standards according to the current oral development stage; S23. Evaluate the actual scores of each item of data in the cleaning behavior information according to the cleaning score standards, and use the cumulative summation method to take the calculation result as the preliminarily evaluated cleaning behavior score; The step of generating oral evaluation indicators through data processing and feature extraction of oral examination data, inputting them into the constructed multi-instance scoring model, and evaluating the oral health score includes: S31. Divide the oral examination data into image data and numerical data according to the detection method of the oral examination data, and perform data preprocessing according to the data type using the corresponding data processing method; S32. Extract the effective features of each type of oral examination data to generate oral evaluation indicators; S33. Set the index value range, and convert the oral evaluation indicators into multiple groups of binary inputs; S34. Establish a multi-instance scoring model by defining packages and examples, calculate the individual scores of each example respectively, and then integrate and calculate the overall score of the package as the oral health score; The step of extracting the effective features of each type of oral examination data to generate oral evaluation indicators includes: S321. Extract the image data after data preprocessing, automatically segment the teeth using a deep learning model, identify the position and shape of each tooth, and use a classification model for recognition according to the current oral development stage to mark the type and number of each tooth; S322. Detect the health status of each tooth, extract the cleaning effect features by analyzing the tooth surface image, and use an image analysis method to quantify the cleaning effect to form feature indicators; S323. Extract the numerical data after preprocessing the data, match each type of the numerical data with the tooth positions in the image data, and map them to the corresponding tooth numbers; S324. Integrate the feature indicators extracted from the image data and the numerical data according to the tooth position distribution, and combine with the tooth numbers to generate oral evaluation indicators.

2. The method for scoring children's oral cavity cleaning according to claim 1, wherein The steps for converting the oral evaluation indicators into multiple groups of binary inputs by setting the value range of the indicators are as follows: S331. Obtain the quantified oral assessment indicators. According to the nine-point system, divide the value range of each of the oral assessment indicators into (1, C n -1), and set a classification threshold. For each of the classification thresholds, generate a binary input, and C n represents the total score, and the classification threshold ∈ (1, C n -1); S332. For the divided classification thresholds, construct binary classifiers. Each type of oral evaluation indicator corresponds to a binary classifier. Use the oral evaluation indicators corresponding to each tooth as the input of the binary classifier, and output the corresponding binary results; S333. Integrate the binary results corresponding to each tooth to form multiple groups of binary inputs.

3. The method for scoring children's oral cavity cleaning according to claim 2, characterized in that, The steps for establishing a multi-instance scoring model by defining bags and examples, calculating the individual scores of each example respectively, and then integrating and calculating the overall score of the bag as the oral health score are as follows: S341. Define the entire oral cavity of the child as a bag, which contains all oral evaluation indicators. Then define each tooth as an example, and each example contains the oral evaluation indicators under the corresponding number; S342. Integrate the binary classifier and the Gaussian model, adjust the weight values of different binary classifiers to form a combined classifier, and calculate the individual scores of each example; S343. Integrate the individual scores of all examples, combine with the distribution of pathological individuals, construct a multi-instance scoring model, and through model calculation, output the oral health score of the entire oral cavity.

4. The method for scoring children's oral cavity cleaning according to claim 3, characterized in that, The steps for integrating the binary classifier and the Gaussian model, adjusting the weight values of different binary classifiers to form a combined classifier, and calculating the individual scores of each example are as follows: S3421. For each example, use the oral evaluation indicators of each tooth as input features; S3422. According to the deviation degree between the position of the binary classifier and the intermediate score, use the Gaussian model to assign weight values to each binary classifier, adjust the contribution degree of each binary classifier to the individual score, and form a combined classifier by integrating all binary classifiers; S3423. Input the input features into the combined classifier to calculate the individual scores of each example.

5. The method for scoring children's oral cleaning according to claim 4, wherein The calculation formula for calculating the individual scores of each example is: ； In the formula, S i represents the i individual score of the i -th example; j Indicates the binary classifier serial number; M Indicates the number of binary classifiers; g (·) represents a weight function; C i Indicates the i historical average rating value of the C n Indicates the total score value; f ij Indicates the i score of the j th binary classifier in the th example.

6. The method for scoring children's oral cavity cleaning according to claim 4, characterized in that The steps for integrating the individual scores of all examples, combining with the distribution of pathological individuals, constructing a multi-instance scoring model, and through model calculation, outputting the oral health score of the entire oral cavity are as follows: S3431. Obtain the individual scores of all examples in the oral cavity. At the same time, sort out the types of teeth corresponding to each example and assign individual weights to the examples; S3432. According to the inspection results of the oral cavity and tooth health status, count the types and quantities of pathological teeth in the oral cavity, and set a penalty factor to represent the non-linear negative impact; S3433. Combine the scores of all examples and the penalties of pathological individuals to construct a multi-instance scoring model, and through model calculation, obtain the oral health score of the child's oral cavity.

7. A method for scoring children's oral hygiene, according to claim 6, characterized in that The expression of the multi-instance scoring model is: ； In the formula, Q represents the oral health score; S i Indicates the individual score of the i th example; α i Indicates the individual weight of the i-th example; β denotes a penalty factor; N Indicates the number of examples; U Indicates the number of diseased teeth in the oral cavity.

8. A children's oral cavity cleaning scoring system for implementing the children's oral cavity cleaning scoring method described in any one of claims 1-7, characterized in that, The system includes: A data acquisition module, which is used to acquire the basic identity information and cleaning behavior information of children participating in oral cleaning scoring, and through an oral health detection device, detect the oral examination data of the child; An identification and analysis module, which is used to mark the current oral development stage of the child through information identification and behavior analysis, and preliminarily evaluate the cleaning behavior score of the child according to the behavior habits in the oral development stage; A health assessment module, which is used to generate oral assessment indicators through data processing and feature extraction of oral examination data, input them into a constructed multi-instance scoring model, and evaluate the oral health score; A scoring display module, which is used to integrate the cleaning behavior score and the oral health score, comprehensively output the oral cleaning score, and generate a scoring evaluation report according to the scoring result, display the oral cleaning score of the child through a data visualization tool, and provide personalized cleaning improvement suggestions; Among them, the data acquisition module, the identification and analysis module, the health assessment module, and the scoring display module are connected in sequence.

Images