Method and system for coating fluorine on tooth surface
Through the combination of detection module and control module, electrochemical sensors and machine learning models, precise spraying of fluoride on the surface of tooth is achieved, solving the problems of uneven fluorine coating and unstable effect in the prior art, ensuring that fluorine is applied every time Achieve the best protection effect.
Patent Information
- Application Number
- CN202411041804.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-11-15
AI Technical Summary
It is difficult for existing dental care products to accurately control the fluoride content during the fluoride application process, resulting in uneven fluoride application and unstable effect, making it difficult for users to evaluate the actual fluoride content, which may lead to insufficient or excessive fluoride use.
The detection module is used to detect the temperature, humidity and fluorine content of the tooth surface, and the data is processed through the control module, and the spraying instructions are generated. The fluorine coating module is used for precise spraying, including electrochemical sensor adjustment, filtering algorithm denoising, feature extraction and machine learning model. Predictions to ensure optimal results for each fluorine application.
It realizes precise control and automatic adjustment of fluoride on the surface of the tooth, ensuring that each fluorine application achieves the best protective effect, real-time detection of fluorine content, and improving the prevention of tooth corrosion and enhancing the fluorine application effect.
Smart Images

Figure CN118948479A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a tooth fluoride coating technology, in particular to a method and system for coating fluoride on the surface of teeth, belonging to the technical field of tooth care. Background Art
[0002] Fluoride application is an effective method commonly used in dentistry to prevent dental caries. Human teeth are divided into crowns, necks and roots. The crowns are located in the oral cavity and can be seen with the naked eye. The outermost part of the crown is enamel, followed by dentin, and the innermost part is the pulp cavity. There are some inorganic salt crystals in the enamel, which are easily corroded by the acid produced by the decomposition of food by bacteria in the mouth, causing calcium ions to escape, collapse and form dental cavities. After fluoride application, fluoride ions will form a more stable structure with inorganic salt crystals, which is not easily corroded by acidic substances and can prevent dental caries. Therefore, fluoride application is an effective method for preventing dental caries.
[0003] Currently, there are various dental care products on the market, such as toothpaste and oral spray. Although these products can apply fluoride manually, users often cannot accurately control the required fluoride content, resulting in uneven tooth fluoridation and unstable effects. In addition, it is difficult for users to evaluate the actual fluoride content after fluoridation, which may lead to insufficient or excessive use of fluoride. Summary of the invention
[0004] The technical problem to be solved by the present invention is to provide a method and system for applying fluoride on the tooth surface, which solves the problems of excessive or insufficient fluoride application in the prior art.
[0005] The technical problem to be solved by the present invention is achieved by adopting the following technical solutions:
[0006] The present invention provides a method for applying fluoride to a tooth surface, comprising the following steps:
[0007] Step S1, the detection module performs detection to obtain detection data;
[0008] Step S2, the control module processes the detection data obtained in step S1 to obtain tooth fluoride supplementary information data;
[0009] Step S3, the control module controls the fluorine coating module to perform fluorine coating treatment.
[0010] As a preferred technical solution of the present invention, in step S1, the detection data obtained include the temperature, humidity and fluoride content of the oral cavity;
[0011] The step S2 includes the following sub-steps:
[0012] Step S21, receiving data and performing denoising;
[0013] Step S22, calculating the fluoride deficiency information of each part of the tooth according to the algorithm model;
[0014] Step S23, generating tooth fluoride supplementary information data;
[0015] The step S3 includes the following sub-steps:
[0016] Step S31, the control module generates a spraying instruction according to the tooth fluoride supplementary information data;
[0017] Step S32, the nozzle of the fluorine coating module completes the fluorine coating process according to the instruction.
[0018] As a preferred technical solution of the present invention, in step S1, the electrochemical sensor of the monitoring module is used to detect the fluorine content, which is characterized in that:
[0019] Before testing the fluoride content, the control module adjusts and calibrates the electrochemical sensor according to the temperature and humidity in the oral cavity.
[0020] As a preferred technical solution of the present invention, in the step S2, in its sub-step S21, denoising includes using a practical filtering algorithm to remove random noise from the sensor and / or the environment;
[0021] The sub-step S21 also includes outlier detection.
[0022] As a preferred technical solution of the present invention, step S22 includes the following sub-steps:
[0023] Step S221, feature extraction, first feature recognition is performed based on the information detected in step S1, including local fluoride concentration mean, fluoride concentration standard deviation, fluoride concentration maximum and minimum, fluoride concentration gradient, and spatial pattern recognition;
[0024] Step S222, performing feature conversion on the feature recognition information in step S221 to generate a feature vector;
[0025] Step S223, inputting the feature vector in step S222 into the machine learning model to generate tooth fluoride supplementary information data.
[0026] Preferably, in step S223, the training of the machine learning model includes the following steps:
[0027] Step J1, select one of random forest, support vector machine or neural network for training;
[0028] Step J2, training the model using the historical data set and adjusting the model parameters to optimize the prediction performance;
[0029] Step J3, receiving feedback data, importing it into step J2, and updating the model.
[0030] As a preferred technical solution of the present invention, in the step S3, in its sub-step S32, the adjustment of the fluorine coating module includes adjusting the nozzle position and angle, controlling the spraying amount, and controlling the spraying time and intensity.
[0031] The present invention also provides a system for applying fluoride to tooth surface, which is used to implement the above-mentioned method for applying fluoride to tooth surface, including a control module, a fluoride application module, a detection module and an interaction module, wherein:
[0032] The control module controls the fluorine coating module and the interaction module according to the data of the detection module.
[0033] The beneficial effects of the present invention are as follows: the present invention can accurately control and automatically adjust the amount of fluoride applied to the tooth surface, and simultaneously detect and display the fluorine content on the tooth surface, ensuring that each fluoride application can achieve the best protective effect; the present invention can automatically adjust the amount of fluoride applied according to the detected fluorine content on the tooth surface, which is achieved through advanced sensors and control algorithms, ensuring that each fluoride application can achieve the optimal fluoride concentration required by the individual; real-time detection of the fluorine content on the tooth surface can instantly evaluate the fluorine coverage on the tooth surface, which is the key to preventing tooth corrosion and enhancing the effect of fluoride application. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 This is a schematic diagram of the structure of a fluoride coating system for tooth surface according to the present invention;
[0035] In the figure: 1. Control module; 2. Fluorine coating module; 3. Detection module; 4. Interaction module. DETAILED DESCRIPTION
[0036] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.
[0037] Example 1
[0038] This embodiment provides a method for applying fluoride to the surface of teeth, comprising the following steps:
[0039] Step S1, the detection module 3 performs detection, obtains detection data, and transmits it to the control module 1;
[0040] Step S2, the control module 1 processes the detection data obtained in step S1 to obtain tooth fluoride supplementary information data;
[0041] Step S3, the control module 1 controls the fluorine coating module 2 to perform fluorine coating treatment.
[0042] Specifically, in step S1, the detection data obtained include the temperature, humidity, and fluoride content of the oral cavity. Preferably, the temperature, humidity, and fluoride content of the oral cavity are detected by a sensor built into an electrochemical sensor. Preferably, the electrochemical sensor model used is an Ion Sensitive Field Effect Transistor (ISFET), which is specially designed for detecting fluoride ions.
[0043] Step S2 includes the following sub-steps:
[0044] Step S21, receiving data and performing denoising;
[0045] Step S22, calculating the fluoride deficiency information of each part of the tooth according to the algorithm model;
[0046] Step S23, generating tooth fluoride supplementary information data;
[0047] Step S3 includes the following sub-steps:
[0048] Step S31, the control module 1 generates a spraying instruction according to the tooth fluoride supplementary information data;
[0049] Step S32, the nozzle of the fluorine coating module 2 completes the fluorine coating process according to the instruction.
[0050] Preferably, in step S1, the electrochemical sensor of the monitoring module 3 for detecting the fluoride content has an ion-selective electrode specially designed for detecting fluoride ions, and works by the selective response of its electrode to fluoride ions. The potential change caused by fluoride ions on the electrode surface is proportional to its concentration. Before performing the fluoride content detection, the control module 1 adjusts and calibrates the electrochemical sensor according to the temperature and humidity in the oral cavity.
[0051] The preferred implementation process of adjusting and calibrating the electrochemical sensor includes:
[0052] Regarding adjustment, when the device is started, it first measures the current oral environment through the built-in temperature and humidity sensor, and automatically adjusts the sensor's operating parameters, such as the electrode potential and the chemical composition of the sensitive layer, based on the measurement results. This allows the sensor to be adjusted to a state suitable for the current oral environment, optimizing its performance and measurement accuracy.
[0053] Regarding calibration, in order to ensure that the sensor provides accurate fluoride ion concentration readings in actual use. The calibration performed includes calibration solution preparation, preparation of fluoride ion standard solution with known concentration; immersion calibration: immersing the sensor electrode in the standard solution, adjusting and recording the calibration curve according to the electrode reaction; and parameter setting: setting the analytical equation or conversion parameter of the sensor output based on the calibration curve. The accurate calibration of the sensor can provide reliable data in the subsequent practical application.
[0054] Preferably, the electrode potential is also adjusted to adapt to the pH and temperature changes of the oral environment to ensure that the electrode responds best to fluoride ions, and the operation is as follows:
[0055] 1. Measurement: Use built-in pH and temperature sensors to obtain current oral environment data.
[0056] 2. Calculation: Calculate the theoretical ideal potential according to the Nernst equation:
[0057]
[0058] Where E 0 is the standard electrode potential, R is the gas constant, T is the absolute temperature (in Kelvin), n is the number of electrons transferred by the fluoride ion in the reaction (1 for F^-), F is the Faraday constant, [F - ] is the activity of the fluoride ion, and [ref] is the activity of the reference electrode.
[0059] 3. Adjustment: Adjust the electrode potential according to the calculation results to ensure that the electrode has the best response sensitivity and specificity in the specific oral environment.
[0060] Results: After adjusting the electrode potential, the sensor can more accurately detect the fluoride ion concentration in the oral environment.
[0061] 3. Adjustment: Adjust the electrode potential according to the calculation results to ensure that the electrode has the best response sensitivity and specificity in the specific oral environment. After the electrode potential is adjusted, the sensor can more accurately detect the fluoride ion concentration in the oral environment.
[0062] Preferably, the chemical composition of the sensitive layer is adjusted to improve its selectivity and stability to fluoride ions.
[0063] Preferably, in other preferred embodiments of the present invention, the optimization process of the electrochemical sensor includes:
[0064] 1. Adjust the working parameters of the sensor
[0065] Measure environmental data: Use the built-in temperature and humidity sensor to measure the temperature and humidity of the oral environment in real time to obtain the current environmental conditions.
[0066] Automatic parameter adjustment: Based on the measurement results, the sensor's operating parameters, such as the electrode potential and the chemical composition of the sensitive layer, are automatically adjusted. These adjustments ensure that the sensor can still provide accurate readings under different environmental conditions. For example, the electrode's operating potential may need to be adjusted under low temperature conditions to ensure that its sensitivity to fluoride ions is not affected.
[0067] 2. Calibration of electrochemical sensors
[0068] Preparation of calibration solution: Use a fluoride ion standard solution of known concentration for calibration. The preparation of the standard solution needs to strictly follow laboratory standards to ensure the accuracy of its concentration.
[0069] Immersion calibration: Immerse the sensor electrode in the standard solution and record the potential change of the electrode. Adjust and record the calibration curve according to the potential change to establish a relationship model between fluoride ion concentration and potential.
[0070] Parameter setting: According to the calibration curve, set the output conversion parameters of the sensor. This step ensures that the sensor can accurately reflect the concentration of fluoride ions in actual detection.
[0071] 3. Adjust electrode potential
[0072] Measure oral environment data: Obtain the current pH value and temperature data in the oral cavity through the built-in pH and temperature sensors.
[0073] Calculate theoretical potential: Use the Nernst equation to calculate the ideal electrode potential under current environmental conditions. The Nernst equation is used to describe the response characteristics of electrochemical sensors.
[0074] Adjusting electrode potential: Based on the calculation results, the control module adjusts the electrode potential to keep it in the best response state at the current pH value and temperature. This adjustment ensures the sensitivity and selectivity of the electrode to fluoride ions.
[0075] 4. Optimize the chemical composition of the sensitive layer
[0076] Analyze the sensitive layer materials: Evaluate the performance of currently used sensitive layer materials (such as LaF3 crystals), especially their selectivity and anti-interference ability for fluoride ions under different oral environments.
[0077] Optimize the sensitive layer: Add or adjust the ingredients in the sensitive layer, such as adding silver, copper and other plasma deactivators. These additives can improve the selectivity of the sensitive layer to fluoride ions and reduce the interference of other ions.
[0078] Adjust the preparation process: Based on the test results, optimize the preparation process of the sensitive layer, such as changing the doping ratio, re-coating or improving the curing conditions. These adjustments ensure that the sensitive layer can provide stable and accurate test results in various oral environments.
[0079] 5. Denoising and Outlier Detection
[0080] Denoising: Use Kalman filtering or moving average filtering to remove random noise from sensor data. These filtering algorithms can smooth the data and reduce errors caused by fluctuations in the sensor itself or the external environment.
[0081] Kalman filter: Recursively estimates the system state by combining a dynamic model of the system and a model of measurement noise.
[0082] Outlier detection: Apply statistical tests (Z-score method) to identify and remove outliers. The specific steps of outlier detection include:
[0083] Z-score method: The Z-score is calculated for each data point, and points with a Z-score greater than 3 or less than 3 are considered outliers.
[0084] The operation is as follows:
[0085] 1. Analysis: Evaluate the performance of the current sensitive layer material (such as LaF3 crystal) and analyze it based on the interfering ions (such as OH^-) that may exist in the oral cavity.
[0086] 2. Optimization: Add or adjust plasma deactivators such as silver and copper. These additives can improve the selectivity and anti-interference ability of the crystal.
[0087] 3. Implementation: Adjust the preparation process of the sensitive layer according to the test results, such as changing the doping ratio, re-coating or improving the curing conditions.
[0088] The result is a tuned sensitive layer with higher selectivity and lower interference response to fluoride ions, improving the overall measurement accuracy and reliability.
[0089] In step S2, in its sub-step S21, denoising includes using a practical filtering algorithm to remove random noise of the sensor and / or the environment;
[0090] Sub-step S21 also includes outlier detection, such as applying statistical tests (such as box plot analysis or Z-score method) to identify and remove outliers. The processed data is cleaner, reduces errors, and improves the accuracy of subsequent analysis.
[0091] The above denoising and outlier detection are data cleaning, which specifically ensures the quality and availability of the raw data obtained from the sensor. In this process, we need to process the raw data, identify and remove noise and outliers.
[0092] More specifically, raw data definition: Raw data refers to the fluoride concentration data on the tooth surface directly measured and recorded by the fluoride ion selective electrode sensor. These data usually appear in the form of voltage or current readings, where the change in the reading is proportional to the concentration of fluoride ions. Raw data includes the potential values of each measurement point, which are collected continuously or intermittently at different locations on the tooth surface. The data set may contain data from hundreds to thousands of measurement points, depending on the coverage and fineness of the scan.
[0093] Noise data definition: Noise data refers to non-representative or misleading data caused by various interference factors during the measurement process. Noise may come from the internal circuits of electronic devices, transient interference on the surface of electrodes, small fluctuations in the measurement environment, or errors in operation.
[0094] Noise data type:
[0095] 1. Electronic noise: Random electronic fluctuations from the device’s own circuits.
[0096] 2. Environmental noise: sensor response affected by changes in temperature and humidity.
[0097] 3. Operational noise: data fluctuations caused by improper user operation (such as uneven movement speed, unstable contact between sensor and teeth).
[0098] Outlier definition: Outliers are data points that deviate significantly from other data points and do not conform to the expected measurement pattern. These values may be caused by equipment failure, sensor damage, or extreme external conditions.
[0099] Outlier identification method:
[0100] 1. Statistical methods: Use statistical tests (such as Z-score, IQR method) to identify and remove outliers. For example, data points with Z-scores exceeding ±3 are usually considered outliers.
[0101] 2. Graphical method: Visually identify outliers by drawing a box plot of the data distribution. Points outside the box plot are usually considered outliers.
[0102] Operation process
[0103] Denoising:
[0104] 1. Apply filtering techniques: Use Kalman filtering or moving average filtering methods to remove random noise from the data. These filters can effectively smooth the data series and reduce fluctuations.
[0105] 2. Adjust parameters: Adjust the parameters of the filter based on experimental data to ensure that no important signals are lost.
[0106] Outlier handling:
[0107] 1. Statistical analysis: Calculate statistical indicators of data points, such as mean, standard deviation, and use Z-score method to identify data points that deviate from the mean by multiple standard deviations.
[0108] 2. Eliminate or correct: Data points identified as abnormal are eliminated or appropriately corrected based on the context.
[0109] result:
[0110] 1. Improved data quality: By removing noise and outliers, the overall quality and reliability of the dataset is improved.
[0111] 2. Analysis preparation: The cleaned data provides a solid foundation for subsequent feature extraction and machine learning analysis.
[0112] This detailed data cleaning process is an important step to ensure analysis accuracy and final application effect, and is particularly critical for smart devices based on sensor data.
[0113] Step S22 includes the following sub-steps:
[0114] Step S221, feature extraction, first feature recognition is performed based on the information detected in step S1, including local fluoride concentration mean, fluoride concentration standard deviation, fluoride concentration maximum and minimum, fluoride concentration gradient, and spatial pattern recognition;
[0115] Step S222, performing feature conversion on the feature recognition information in step S221 to generate a feature vector;
[0116] Step S223, inputting the feature vector in step S222 into the machine learning model to generate tooth fluoride supplementary information data.
[0117] Specifically, regarding feature extraction and its application in machine learning, when processing data from fluorine content detection, feature extraction is a key step in converting raw data into a form that facilitates analysis and prediction. This step involves identifying and constructing features that can represent the essential attributes of the data so that the machine learning model can learn and predict more effectively.
[0118] Feature definition: Features are data attributes extracted or derived from raw data and used as input to machine learning models. In fluorine content detection, the following are several key features:
[0119] 1. Local fluoride concentration mean: the average value of fluoride concentration in each test area.
[0120] 2. Standard deviation of fluoride concentration: measures the degree of dispersion of fluoride concentration distribution in each area.
[0121] 3. Maximum and minimum fluoride concentration: reflects the extreme values of fluoride concentration in each detection area.
[0122] 4. Fluoride concentration gradient: The rate of change of fluoride concentration between adjacent areas, used to evaluate the uniformity of fluoride distribution.
[0123] 5. Spatial pattern recognition: Identify the distribution pattern of fluoride on the tooth surface through image processing technology.
[0124] Feature conversion process, goal: convert raw data into a form that helps the model recognize and predict.
[0125] operate:
[0126] 1. Data aggregation: Aggregate the fluoride concentrations of the original measurement points and calculate the mean and standard deviation of each area.
[0127] 2. Gradient calculation: Calculate the difference in fluoride concentration between adjacent measurement points to estimate the concentration gradient.
[0128] 3. Pattern analysis: Apply image processing or spatial analysis techniques, such as edge detection or cluster analysis, to identify patterns in concentration distribution.
[0129] 4. Data normalization: Normalize feature data, such as minimum-maximum normalization or Z-score normalization, to eliminate dimensionality effects and improve the stability and effectiveness of model training.
[0130] Feature Dimensionality Reduction:
[0131] 1. Principal Component Analysis (PCA): PCA technique is used to reduce the dimensionality of the data and extract the principal components that best represent the variability of the data.
[0132] 2. Feature selection: Determine the most informative features for prediction through model selection algorithms (such as recursive feature elimination).
[0133] result:
[0134] 1. Feature vector: A set of feature vectors is formed to help the machine learning model recognize and predict, and each feature reflects an important dimension of the data.
[0135] 2. Data preparation is complete: The completed feature set provides preparation for the next step of machine learning model training and prediction.
[0136] In machine learning applications, feature vectors are a crucial concept, especially when dealing with complex data such as fluoride content testing. A feature vector is a combination of multiple eigenvalues that together describe a sample in the dataset. In this case, each feature vector represents an area or point on the tooth surface, and these vectors are used to train the model to predict fluoride gaps.
[0137] Description of the feature vector in this embodiment:
[0138] 1. Local fluoride concentration average:
[0139] Description: This feature represents the average fluoride concentration measured in a specific area of the tooth.
[0140] Importance: Helping the model understand the overall fluoride levels in the area is critical to determining whether fluoride supplementation is needed.
[0141] 2. Fluoride concentration standard deviation:
[0142] Description: A measure of the variability or dispersion of fluoride concentrations within an area.
[0143] Importance: A high standard deviation indicates that fluoride is unevenly distributed in this area and may require special attention.
[0144] 3. Maximum and minimum fluoride concentrations:
[0145] Description: These two features capture the extreme high and low values of fluoride concentration in each area, respectively.
[0146] Importance: Extreme values provide additional information beyond the average and help the model identify areas that require more focused fluoride application.
[0147] 4. Fluoride concentration gradient:
[0148] Description: Calculates the fluoride concentration difference between adjacent detection points to evaluate the spatial rate of change of concentration.
[0149] Importance: Areas with larger gradients indicate that the spatial distribution of fluoride is very heterogeneous and may indicate that some areas are under-fluoridated.
[0150] 5. Spatial pattern recognition:
[0151] Description: Analyze the spatial pattern of fluoride distribution by image processing techniques such as edge detection or clustering.
[0152] Importance: Pattern recognition can help the model understand the overall layout of fluoride on the tooth surface and identify typical distribution patterns or abnormal patterns.
[0153] About feature conversion and feature vector utilization:
[0154] Data transformation: After extracting features, some transformation processing is usually required, such as normalization or standardization, to make the feature value range consistent and enhance the learning efficiency and prediction performance of the model. For example, you can apply minimum-maximum normalization to scale all features to the range of 0 to 1, or use Z-score normalization to ensure that the features have zero mean and unit variance.
[0155] Application of feature vectors: The feature vectors formed will be used as input to the machine learning model. These vectors are used during the model training process to learn how to predict fluoride gaps in various areas of the teeth based on the given features. In actual application, when new measurement data is input into the system, the corresponding feature vectors will be generated and processed through the trained model, and the model will output the corresponding fluoride supplementation recommendation.
[0156] By carefully designing and selecting feature vectors, the prediction accuracy and efficiency of the model can be significantly improved, making fluoride application more precise and effective. This approach not only improves the quality of oral health care, but also reduces resource waste by precisely controlling fluoride use.
[0157] In step S223, the training of the machine learning model includes the following steps:
[0158] Step J1, select one of random forest, support vector machine or neural network for training;
[0159] Step J2, training the model using the historical data set and adjusting the model parameters to optimize the prediction performance;
[0160] Step J3, receiving feedback data, importing it into step J2, and updating the model.
[0161] Specifically, it includes the following processes or steps:
[0162] The application of machine learning models is to analyze fluoride concentration data and predict the fluoride needs of various areas of the teeth. This step includes selecting a suitable model, training the model, and optimizing the model parameters to achieve the best prediction performance.
[0163] In this application scenario, we can consider using the following two types of models:
[0164] 1. Random Forest:
[0165] Advantages: Able to handle nonlinear relationships, insensitive to outliers, and not prone to overfitting.
[0166] Uses: Suitable for handling complex regression problems with multiple input features, such as predicting the fluoride gap.
[0167] 2. Neural Networks:
[0168] Advantages: Extremely strong nonlinear modeling capabilities, and can flexibly handle various data relationships by adjusting the network structure.
[0169] Uses: Good for discovering and modeling complex patterns and relationships in data, especially for large datasets.
[0170] What is included in the preset historical data set?
[0171] 1. Feature data: This includes the feature vectors mentioned earlier, such as the mean, standard deviation, maximum, minimum, gradient, and spatial pattern of fluoride concentration.
[0172] 2. Target variable: The actual fluoride gap corresponding to each eigenvector, that is, the actual amount of fluoride that needs to be supplemented.
[0173] 3. Additional information: May also include data such as tooth area identification, timestamp (test time), and the patient's oral health status.
[0174] Regarding adjusting model parameters to optimize prediction performance, the parameter adjustment process for random forest and neural network is as follows:
[0175] Random Forest
[0176] -Parameter adjustment:
[0177] 1. Number of trees (n_estimators): Increasing the number of trees can improve the stability and accuracy of the model, but it will also increase the computational load.
[0178] 2. Maximum depth of the tree (max_depth): Control overfitting and prevent the model from over-learning on the training data.
[0179] 3. Minimum number of split samples (min_samples_split): Adjust this parameter to avoid the model being overly sensitive to small fluctuations in the data.
[0180] -Optimization strategy:
[0181] 1. Cross-validation: Use K-fold cross-validation to evaluate the performance of different parameter configurations.
[0182] 2. Grid Search: Systematically test various parameter combinations to find the best model configuration.
[0183] For Neural Networks:
[0184] -Parameter adjustment:
[0185] 1. Number of layers and neurons: Increasing the number of layers and neurons can improve the learning ability of the model, but it may also lead to overfitting.
[0186] 2. Activation function: Select an appropriate activation function (such as ReLU, sigmoid, tanh) to optimize nonlinear learning effects.
[0187] 3. Learning rate: Adjust the learning rate of the optimization algorithm to ensure a stable and effective learning process.
[0188] -Optimization strategy:
[0189] 1. Early Stopping: Monitor the performance of the validation set and stop training when the performance stops improving to avoid overfitting.
[0190] 2. Regularization techniques (such as L1, L2 regularization):
[0191] Reduce model complexity and improve generalization ability.
[0192] Algorithm processing and optimization process
[0193] -Processing: The data is fed into the model, which makes predictions based on the rules learned during training and outputs the predicted fluoride gap value for each tooth area.
[0194] -Optimization process: Ensure that the model performs best on new data through continuous monitoring of model performance and adjustment of parameters. Regularly retrain the model with newly collected data to ensure that the model updates are consistent with the actual situation.
[0195] These detailed steps ensure that the machine learning model can accurately predict fluoride needs and provide more precise personalized oral care plans.
[0196] In step S3 and its sub-step S32, the adjustment of the fluorine coating module 2 includes adjusting the nozzle position and angle, controlling the spraying amount, and controlling the spraying time and intensity.
[0197] Specifically, it is used to achieve precise positioning of the nozzle so that it can cover all designated tooth areas.
[0198] Here are the steps:
[0199] 1. Receive coordinate data: Receive the precise coordinates of each area based on the tooth area requirements output by the prediction model.
[0200] 2. Calculate the path: Use a path planning algorithm (such as Dijkstra) to calculate the optimal path for the nozzle to move.
[0201] 3. Adjust the nozzle: Based on the calculated path, the position and angle of the nozzle are adjusted by the stepper motor.
[0202] This allows the nozzle to be precisely positioned to each tooth area that requires fluoride application.
[0203] Regarding controlling the spraying amount, that is, adjusting the amount of fluoride sprayed according to the specific needs of the tooth area, the steps are as follows:
[0204] 1. Read demand data: Get the amount of fluoride required for each area from the forecast model.
[0205] 2. Calculate spraying parameters: Calculate the duration and intensity of the spraying based on the required amount of fluoride and the performance characteristics of the nozzle.
[0206] 3. Perform spraying: Accurately control the amount of fluoride sprayed by controlling the speed of the pump and the opening and closing of the nozzle.
[0207] Ensures that each area receives the precise amount of fluoride required.
[0208] Regarding calculating the duration and intensity of the spray, in the automatic fluoride application equipment, calculating the duration and intensity of the spray based on the required amount of fluoride and the performance characteristics of the nozzle is a key step in achieving accurate fluoride application. This process ensures that each tooth area receives the right amount of fluoride to achieve the best care effect.
[0209] Parameter Definition
[0210] Amount of fluoride required: The amount of fluoride (usually in milligrams) predicted by the machine learning model based on the specific needs of the tooth area.
[0211] Nozzle performance characteristics: including nozzle flow rate (usually measured in ml / min), injection pressure, injection accuracy and other parameters.
[0212] Calculation method
[0213] Step 1: Determine the nozzle flow rate
[0214] Objective: Determine the flow rate of a nozzle at a given pressure.
[0215] operate:
[0216] 1. Flow Test: Test the actual flow rate of the nozzle at different pressure settings in a controlled environment.
[0217] 2. Flow record: record flow data and establish a relationship model between pressure and flow.
[0218] The result: an accurate flow model for calculating flow at a specific pressure.
[0219] Step 2: Calculate the volume of fluoride solution required
[0220] Objective: Convert the amount of fluoride required into the corresponding volume of liquid.
[0221] operate:
[0222] 1. Concentration conversion: Use the concentration of the fluoride solution (mg / ml) to convert the mass of fluoride into liquid volume.
[0223] 2. Volume calculation:
[0224] 2. Volume calculation: Result: The actual volume of liquid that needs to be sprayed is obtained.
[0225] Step 3: Calculate spraying duration
[0226] Objective: Calculate the spraying time required to achieve a desired volume of liquid.
[0227] operate:
[0228] Duration formula:
[0229] 1. Duration formula:
[0230] 2. Time calculation: Calculate the spray duration using the flow rate determined in step 1 and the volume calculated in step 2.
[0231] Result: The required spraying duration for each tooth area is obtained.
[0232] Step 4: Set the spray intensity
[0233] Goal: Adjust the spray pressure to control spray coverage and depth based on the needs of the tooth area.
[0234] operate:
[0235] 1. Pressure selection: Select the appropriate spray pressure based on the size and shape of the spray area.
[0236] 2. Adjust the nozzle: Adjust the nozzle setting according to the selected pressure to ensure that the fluorine coating is even and comprehensive.
[0237] Result: Every tooth area is fluoridated with the right amount of pressure and intensity.
[0238] In this embodiment, feedback control and optimization are also implemented:
[0239] Goal: Optimize the spraying process through feedback mechanisms.
[0240] operate:
[0241] 1. Real-time monitoring: Use sensors to monitor the flow rate and coating effect during the spraying process.
[0242] 2. Adjustment and optimization: adjust the spraying according to the monitoring results
[0243] Parameters such as duration and pressure to optimize spraying results.
[0244] The result: a more precise and efficient spraying process and more economical and effective use of fluoride.
[0245] Through these detailed calculation and adjustment steps, the automatic fluoride application equipment can accurately control the fluoride spraying of each tooth area, ensuring that each treatment achieves the expected health care effect while reducing waste and improving treatment efficiency.
[0246] Goal: Optimize spraying accuracy and efficiency through feedback mechanisms.
[0247] operate:
[0248] 1. Collect feedback: Use sensors to detect fluoride concentration after spraying and compare it with predicted demand.
[0249] 2. Data analysis: Analyze the difference between the spraying effect and the expected effect, and identify any systematic deviations or errors.
[0250] 3. Parameter adjustment: Adjust the parameters of the nozzle control algorithm, such as flow rate, spray pattern and path planning, based on the analysis results.
[0251] The result: Through continuous feedback and optimization, the accuracy and efficiency of the spraying system have been significantly improved.
[0252] Precise nozzle control is achieved through integrated control software, which includes path planning, spray volume calculation and feedback control modules.
[0253] Optimization process:
[0254] 1. Continuous learning: The system uses new spraying data to continuously learn and optimize, and adjust path planning and spraying parameters.
[0255] 2. Algorithm iteration: Regularly update control algorithms and introduce more advanced optimization technologies, such as machine learning and artificial intelligence, to further enhance the level of intelligent operation.
[0256] Through these detailed steps, the fluoride spray adjustment system can operate very precisely according to the actual dental care needs, ensuring that each tooth area can be properly treated, while continuously improving its performance and efficiency through intelligent feedback and optimization mechanisms. This highly accurate and responsive system design is an important advancement in modern oral care technology.
[0257] The use of the present invention can significantly improve the convenience and effect of tooth care, ensure that the drug dosage can be accurately controlled each time fluoride is applied, effectively prevent tooth corrosion and caries, and avoid excessive use of fluoride through the automatic detection function to ensure oral health;
[0258] The following is a typical implementation and calculation example:
[0259] 1. Entrance detection
[0260] The entrance detection module consists of an electrochemical sensor and a temperature and humidity sensor, which detects the temperature, humidity and fluoride ion concentration of the oral environment in real time.
[0261] 2. Data Preprocessing
[0262] The test data, including temperature, humidity and fluoride ion concentration, are preprocessed through the following steps:
[0263] Denoising: Use a Kalman filter to remove random noise from the data.
[0264] Outlier Detection: Identify and remove outliers using the Z-score method.
[0265] 3. Feature extraction
[0266] Extract features from the preprocessed data, including:
[0267] Mean local fluoride concentration
[0268] Fluoride concentration standard deviation
[0269] Maximum and minimum fluoride concentrations
[0270] Fluoride concentration gradient
[0271] 4. Deep learning model training
[0272] Select the neural network model for training. The model structure is as follows:
[0273] Input layer: accepts the extracted feature vector;
[0274] Hidden layer: two layers, each layer contains 64 neurons, and the activation function is ReLU;
[0275] Output layer: outputs the amount of fluoride required for each tooth area, and the activation function is a linear function;
[0276] Training steps:
[0277] 1. Dataset preparation: Use the historical dataset for training, which contains feature vectors and target variables (the amount of fluoride required for each tooth area).
[0278] 2. Model training: Use mean square error (MSE) as the loss function and Adam optimizer for training.
[0279]
[0280] Among them, y i is the actual value, is the predicted value.
[0281] 5. Spray fluoride
[0282] Use the trained deep learning model to predict the fluoride requirement for each tooth area and control the spraying module to perform the spraying operation.
[0283] Spraying parameter calculation:
[0284] Spray volume: Calculate the required spray liquid volume based on the predicted fluoride demand.
[0285]
[0286] Where V is the liquid volume, M is the mass of fluoride, and C is the solution concentration.
[0287] Spraying time: Calculate the spraying time based on the nozzle flow rate.
[0288]
[0289] Where t is the spraying time, V is the liquid volume, and Q is the nozzle flow rate.
[0290] 6. Feedback optimization
[0291] Use feedback mechanisms to optimize spraying results:
[0292] 1. Real-time monitoring: Use sensors to detect the fluoride concentration after spraying.
[0293] 2. Feedback data processing: compare with the predicted value and calculate the error.
[0294]
[0295] Among them, e is the error, y is the actual value, is the predicted value.
[0296] 3. Model update: Use feedback data to update the deep learning model for incremental learning.
[0297]
[0298] Among them, θ is the model parameter, η is the learning rate, and L(θ) is the loss function.
[0299] Through the above steps, a complete process from entrance detection, data preprocessing, feature extraction, deep learning model training to fluoride spraying and feedback optimization is realized. This process ensures the accurate application of fluoride on the tooth surface and improves the effect of oral care.
[0300] About specific example applications:
[0301] Suppose we have an entry detection dataset as follows:
[0302] Temperature: 37℃
[0303] Humidity: 80%
[0304] Fluoride ion concentration: 0.8ppm
[0305] Data preprocessing:
[0306] Denoising: The above data were processed using a Kalman filter to obtain a smooth fluoride ion concentration curve.
[0307] Outlier Detection: Use box plot analysis to identify and remove extreme values.
[0308] Feature extraction:
[0309] Average local fluoride concentration: 0.8ppm
[0310] Fluoride concentration standard deviation: 0.05ppm
[0311] Maximum and minimum fluoride concentration: 0.85ppm and 0.75ppm
[0312] Fluoride concentration gradient: 0.02ppm / cm
[0313] Deep learning model training:
[0314] The neural network model is trained using the above feature vectors.
[0315] Fluoride spray:
[0316] According to the model prediction, 0.05 ml of fluoride solution needs to be sprayed on the tooth area.
[0317] According to the nozzle flow rate (0.01ml / s), the spraying time is calculated to be 5 seconds.
[0318] Feedback optimization:
[0319] The fluoride ion concentration after spraying was monitored in real time and the actual concentration was found to be 0.82 ppm.
[0320] The calculated error is 0.02ppm.
[0321] Update model parameters to optimize subsequent spraying process.
[0322] Example 2
[0323] like Figure 1 As shown, this embodiment provides a system for applying fluoride to a tooth surface, which is used to implement a method for applying fluoride to a tooth surface in embodiment 1, and includes a control module 1, a fluoride applying module 2, a detection module 3 and an interaction module 4, wherein the control module 1 controls the fluoride applying module 2 and the interaction module 4 according to data from the detection module 3;
[0324] Specifically, the fluoride coating module 2 includes a liquid storage container for storing fluoride solution, and a nozzle, and the solution spraying amount is controlled by a precision pump; the detection module 3 uses a sensor to detect the fluoride content on the tooth surface and feeds the data back to the control system; the control module 1 automatically adjusts the amount of fluoride sprayed by the nozzle based on the detection data to ensure that each fluoride coating is uniform and sufficient; the interactive module 4 allows the user to monitor the current fluoride content of the teeth and adjust the fluoride coating preference settings.
[0325] Preferably, the device is designed to be portable, and the user only needs to put the device in the mouth, and the device will automatically scan the tooth surface and apply fluoride. The control system inside the device automatically adjusts the amount of fluoride released by the nozzle according to the detected fluoride content. The user can view the current fluoride content and adjust the frequency and dosage of fluoride application through the mobile application connected to the device.
[0326] The present invention can accurately control and automatically adjust the amount of fluoride applied to the tooth surface, and simultaneously detect and display the fluorine content on the tooth surface, ensuring that each fluoride application can achieve the best protective effect; it can automatically adjust the amount of fluoride applied according to the detected fluorine content on the tooth surface, which is achieved through advanced sensors and control algorithms, ensuring that each fluoride application can achieve the optimal fluoride concentration required by the individual; it can detect the fluorine content on the tooth surface in real time, and can instantly evaluate the fluorine coverage on the tooth surface, which is the key to preventing tooth corrosion and enhancing the effect of fluoride application.
[0327] The basic principles, main features and advantages of the present invention are shown and described above. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of the present invention is defined by the attached claims and their equivalents.
Claims
1. A method for applying fluoride to a tooth surface, It is characterized in that The steps include: Step S1, the detection module (3) performs detection to obtain detection data; Step S2, the control module (1) processes the detection data obtained in step S1 to obtain tooth fluoride supplementary information data; Step S3, the control module (1) controls the fluorine coating module (2) to perform fluorine coating treatment.
2. A method for applying fluoride to a tooth surface according to claim 1, It is characterized in that In step S1, the detection data obtained include the temperature, humidity and fluoride content of the oral cavity; The step S2 includes the following sub-steps: Step S21, receiving data and performing denoising; Step S22, calculating the fluoride deficiency information of each part of the tooth according to the algorithm model; Step S23, generating tooth fluoride supplementary information data; The step S3 includes the following sub-steps: Step S31, the control module (1) generates a spraying instruction according to the tooth fluoride supplementary information data; Step S32, the nozzle of the fluorine coating module (2) completes the fluorine coating process according to the instruction.
3. A method for applying fluoride to tooth surface according to claim 1, It is characterized in that In the step S1, the electrochemical sensor of the monitoring module (3) is used to detect the fluorine content, which is characterized in that: Before the fluoride content is detected, the control module (1) adjusts and calibrates the electrochemical sensor according to the temperature and humidity in the oral cavity.
4. A method for applying fluoride to tooth surface according to claim 1, It is characterized in that In the step S2, in its sub-step S21, denoising includes using a practical filtering algorithm to remove random noise from the sensor and / or the environment; The sub-step S21 also includes outlier detection.
5. A method for applying fluoride to tooth surface according to claim 1, It is characterized in that The step S22 includes the following sub-steps: Step S221, feature extraction, firstly feature recognition is performed based on the information detected in step S1, including local fluoride concentration mean, fluoride concentration standard deviation, maximum and minimum fluoride concentration, fluoride concentration gradient, and spatial pattern recognition; Step S222, performing feature conversion on the feature recognition information in step S221 to generate a feature vector; Step S223, inputting the feature vector in step S222 into the machine learning model to generate tooth fluoride supplementary information data.
6. A fluoride coating system for tooth surface according to claim 5, It is characterized in that In step S223, the training of the machine learning model includes the following steps: Step J1, select one of random forest, support vector machine or neural network for training; Step J2, training the model using the historical data set and adjusting the model parameters to optimize the prediction performance; Step J3, receiving feedback data, importing it into step J2, and updating the model.
7. A fluoride coating system for tooth surface according to claim 5, It is characterized in that In the step S3, in its sub-step S32, the adjustment of the fluorine coating module (2) includes adjusting the nozzle position and angle, controlling the spraying amount, and controlling the spraying time and intensity.
8. A system for applying fluoride to a tooth surface, used to implement a method for applying fluoride to a tooth surface as claimed in any one of claims 1 to 7, It is characterized in that It comprises a control module (1), a fluorine coating module (2), a detection module (3) and an interaction module (4), wherein: The control module (1) controls the fluorine coating module (2) and the interaction module (4) according to the data of the detection module (3).
