Intelligent oral health assessment and personalized nursing system and method thereof

Abstract

The invention relates to the field of oral care, in particular to an intelligent oral health assessment and personalized care system and method. The oral image data of the user is acquired and preprocessed through the data acquisition and processing module, and the oral disease recognition module is used for analyzing the data based on a deep learning algorithm to recognize diseases. The system further combines a personalized nursing recommendation module and a personalized product recommendation module, and generates and recommends personalized nursing schemes and products according to the disease recognition result and the user information. And the health monitoring and feedback module monitors oral health of the user in real time and dynamically adjusts a nursing scheme and product recommendation. According to the system, the intelligent, personalized and dynamic oral health management level is effectively improved through technical innovation of an innovative framework, an advanced algorithm, multi-modal data fusion, a dynamic adjustment mechanism and the like, the problems of an existing system are solved, and comprehensive and efficient oral health nursing is achieved.

Description

Technical Field

[0001] The present invention relates to the field of oral care, and in particular to an intelligent oral health assessment and personalized care system and method thereof. Background Art

[0002] As people's attention to oral health continues to grow, the demand for intelligent, personalized oral health management is becoming increasingly prominent. Traditional oral health management methods rely primarily on regular checkups and standardized care recommendations, which are difficult to meet the unique needs of different individuals. In recent years, with the rapid development of artificial intelligence technology, intelligent oral health assessment and personalized care systems have gradually become a research hotspot.

[0003] Existing intelligent oral health management systems typically include functional modules such as oral image acquisition, disease identification, and care recommendation generation. For oral image acquisition, commonly used technologies include intraoral cameras, X-rays, and CT scans. These technologies can acquire two-dimensional or three-dimensional oral imaging data, providing a foundation for subsequent disease identification. For disease identification, traditional methods rely primarily on manually defined rules and feature extraction. Recent research has begun to explore the application of deep learning algorithms, such as convolutional neural networks (CNNs), to automatically identify common oral diseases such as caries and periodontal disease. For care recommendation generation, most systems employ a rule-based approach, providing appropriate care recommendations based on preset conditions.

[0004] However, existing intelligent oral health management systems still have some significant problems and limitations:

[0005] 1. Insufficient data integration: Most systems rely on a single data source (such as oral images), ignoring multi-dimensional data such as users' personal information and living habits, resulting in one-sided and lack of comprehensiveness in the evaluation results.

[0006] 2. Limited accuracy in disease identification: Although some systems have begun to apply deep learning algorithms, the model designs are often simple and fail to fully consider the particularities of oral imaging, resulting in low recognition accuracy, especially for early or mild oral problems.

[0007] 3. Insufficient personalization: The care recommendations of existing systems are usually generated based on simple rules, which are difficult to adapt to the individual differences of different users and cannot provide truly personalized care plans.

[0008] 4. Lack of dynamic adjustment mechanism: Most systems are static and cannot be dynamically adjusted according to the user's real-time health status and feedback, making it difficult to adapt to changes in the user's oral health status.

[0009] 5. Low user engagement: Existing systems often lack effective user feedback mechanisms and easy-to-use interactive interfaces, resulting in low user engagement and difficulty in continuously and effectively managing oral health.

[0010] 6. Insufficient security and privacy protection: Considering the sensitivity of oral health data, the existing systems’ measures for data security and user privacy protection are often not comprehensive and strict enough. Summary of the Invention

[0011] To address these issues, the present invention proposes an intelligent oral health assessment and personalized care system and method. This system, through the collaborative operation of multiple modules, enables intelligent management of the entire process, from data collection and disease identification to personalized care recommendations. This system not only accurately identifies oral diseases but also provides dynamically adjusted care plans and product recommendations based on the user's individual characteristics and real-time health status, while also prioritizing data security and user privacy.

[0012] The present invention proposes an intelligent oral health assessment and personalized care system and method thereof, including:

[0013] Data acquisition and processing module, used for:

[0014] Obtain the user's oral imaging data;

[0015] Preprocessing the oral image data, including noise reduction and image enhancement;

[0016] The oral disease identification module is in communication with the data acquisition and processing module and is used to:

[0017] Receiving the pre-processed oral image data sent by the data acquisition and processing module;

[0018] Based on the preprocessed oral image data, identifying oral diseases through a deep learning algorithm;

[0019] The personalized care recommendation module is in communication with the oral disease identification module and is used to:

[0020] Receiving the disease identification result sent by the oral disease identification module;

[0021] Generate a personalized care plan based on the disease identification results, user personal information and lifestyle data;

[0022] A personalized product recommendation module, in communication with the personalized care recommendation module, is used to:

[0023] Receiving the personalized care plan sent by the personalized care recommendation module;

[0024] Recommending oral care products suitable for the user based on the personalized care plan;

[0025] The health monitoring and feedback module is in communication with the personalized care recommendation module and the personalized product recommendation module and is used to:

[0026] Real-time monitoring of users’ oral health status;

[0027] Dynamically adjust personalized care plans and product recommendations based on monitoring results.

[0028] Preferably, the data acquisition and processing module includes:

[0029] An image acquisition unit, used to obtain the user's oral imaging data through X-ray or oral CT equipment;

[0030] a data preprocessing unit, communicating with the image acquisition unit and configured to perform noise reduction, image enhancement, and standardization processing on the oral image data;

[0031] An image segmentation unit is communicatively connected to the data preprocessing unit and is used to:

[0032] Perform binarization on the pre-processed oral image data;

[0033] Based on edge detection and connected region analysis methods, oral images are segmented into different anatomical regions.

[0034] Preferably, the oral disease identification module includes:

[0035] Convolutional neural network units, used to:

[0036] Build a deep learning model with multi-layer convolution and pooling structures;

[0037] Perform feature extraction and disease identification on oral imaging data through the deep learning model;

[0038] An attention mechanism unit, in communication with the convolutional neural network unit, for:

[0039] Introducing an attention mechanism into the deep learning model;

[0040] Enhance the model's ability to identify key lesion areas;

[0041] A model optimization unit, in communication with the convolutional neural network unit and the attention mechanism unit, is configured to:

[0042] Use residual optimization technology to improve model training efficiency and performance;

[0043] Dynamically adjust model parameters based on validation set data.

[0044] Preferably, the personalized care recommendation module includes:

[0045] Data fusion unit, used to integrate oral disease identification results, user personal information and lifestyle data;

[0046] A random forest algorithm unit is communicatively connected to the data fusion unit and is used to:

[0047] constructing a random forest model based on the integrated data;

[0048] generating a personalized care plan through the random forest model;

[0049] A solution optimization unit, in communication with the random forest algorithm unit, is configured to:

[0050] Evaluate the generated nursing plan according to the predefined nursing effect evaluation indicators;

[0051] Optimize and adjust the nursing plan based on the evaluation results.

[0052] Preferably, the personalized product recommendation module includes:

[0053] Deep Learning Recommendation Unit, used for:

[0054] Build a product recommendation model based on deep neural networks;

[0055] Using user historical data and product feature information to train the recommendation model;

[0056] A personalized matching unit, in communication with the deep learning recommendation unit, is configured to:

[0057] Based on the user's personalized care plan and oral health status, select the most suitable product from the recommendation model output;

[0058] The recommendation interpretation unit is in communication with the personalized matching unit and is used to:

[0059] Generate explanatory information for product recommendations, including reasons for the recommendation and expected results.

[0060] Preferably, the health monitoring and feedback module includes:

[0061] Real-time data acquisition unit, used to periodically obtain the user's oral health data;

[0062] A health status assessment unit is communicatively connected to the real-time data acquisition unit and is used to:

[0063] Evaluate the user's oral health status based on real-time collected data;

[0064] Detect abnormal changes in oral health status;

[0065] A dynamic adjustment unit, in communication with the health status assessment unit, is configured to:

[0066] Dynamically adjust personalized care plans based on health status assessment results;

[0067] Update personalized product recommendation list.

[0068] Preferably, a user feedback module is also included, for:

[0069] Collect users' evaluations of the care plans and products recommended by the system;

[0070] Based on user feedback, optimize the system's recommendation algorithm and product library.

[0071] Preferably, a data security and privacy protection module is also included, which is used to:

[0072] Encrypted storage of users' oral health data and personal information;

[0073] Implement identity authentication and permission control for data access;

[0074] Use security protocols to protect data during transmission.

[0075] Preferably, a mobile application module is also included for:

[0076] Provides a mobile user interface that allows users to access system functions through smartphones or tablet devices; enables visual display and interactive query of oral health data;

[0077] Send oral care reminders and health advice notifications.

[0078] Intelligent oral health assessment and personalized care method, including the following steps:

[0079] S1, collect user's oral image data:

[0080] Obtain the user's oral image through X-ray or oral CT equipment;

[0081] Preprocess the acquired image data, including noise reduction, enhancement and standardization;

[0082] S2, Oral Disease Identification:

[0083] Using pre-processed image data, oral diseases are identified through deep learning models;

[0084] The deep learning model includes a multi-layer convolutional neural network, an attention mechanism, and a residual optimization structure;

[0085] S3, generate personalized care plan:

[0086] Integrate oral disease identification results, user personal information and lifestyle data;

[0087] Generate personalized care plans using the random forest algorithm;

[0088] Optimize care plans based on predefined evaluation indicators;

[0089] S4, recommend personalized oral care products:

[0090] Build a deep learning product recommendation model;

[0091] Filter the most suitable products from the recommendation model output based on personalized care plans and user’s oral health status;

[0092] Generate explanatory information for product recommendations;

[0093] S5, health monitoring and dynamic adjustment:

[0094] Periodically collect users' oral health data;

[0095] Assess the user's oral health status and detect abnormal changes;

[0096] Dynamically adjust care plans and product recommendations based on assessment results;

[0097] S6, collect user feedback and optimize the system:

[0098] Obtain user reviews of care plans and recommended products;

[0099] Optimize the system's algorithms and product library based on user feedback.

[0100] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0101] First, from a macro perspective, this invention builds a closed-loop intelligent oral health management ecosystem. Through the organic integration of a data acquisition and processing module, an oral disease identification module, a personalized care recommendation module, a personalized product recommendation module, and a health monitoring and feedback module, it achieves intelligent management of the entire process, from data input to personalized output. This system architecture not only improves the efficiency and accuracy of oral health management, but also enables continuous optimization and adjustment to adapt to the dynamic changes in the user's oral health status.

[0102] Secondly, in terms of data integration, this invention breaks through the limitations of a single data source and integrates multiple dimensions of data, including oral imaging data, user personal information, and lifestyle habits. This fusion analysis of multimodal data enables the system to more comprehensively assess the user's oral health status, laying a solid foundation for subsequent personalized recommendations.

[0103] For disease recognition, this invention utilizes advanced deep learning algorithms, specifically a convolutional neural network with an attention mechanism and residual optimization architecture. This innovative model design significantly improves the accuracy of oral disease recognition, particularly for early-stage or mild oral problems.

[0104] In terms of personalized recommendations, this invention combines a random forest algorithm with a deep learning recommendation model to achieve highly personalized care plans and product recommendations. The system not only considers the user's oral health status, but also incorporates factors such as personal characteristics and lifestyle habits, thereby providing more accurate and effective care suggestions.

[0105] Another key feature of this invention is its dynamic adjustment mechanism. Through its health monitoring and feedback module, the system monitors the user's oral health status in real time and dynamically adjusts care plans and product recommendations based on the results. This closed-loop design ensures continuous system optimization, consistently providing the most appropriate health management solution for the user.

[0106] In terms of user experience, the present invention provides an intuitive and easy-to-use interactive interface through the mobile application module, greatly improving user engagement. In particular, the application of augmented reality (AR) technology provides users with an intuitive visual experience of oral health, further enhancing their self-management capabilities.

[0107] Finally, in terms of data security and privacy protection, this invention adopts a multi-layered security strategy, including data encryption, identity authentication, permission control, and secure transmission protocols. These measures comprehensively protect users' sensitive data, effectively enhancing the credibility of the system and user confidence.

[0108] In summary, the intelligent oral health assessment and personalized care system and method of the present invention, through a series of technological innovations such as innovative system architecture design, advanced algorithm application, multimodal data fusion, and dynamic adjustment mechanisms, effectively solves the problems existing in existing systems and realizes truly intelligent, personalized, and dynamic oral health management. This not only greatly improves the efficiency and accuracy of oral health management, but also significantly enhances the user experience. It is expected to play a significant role in improving the public's oral health level and pave a new path for the intelligent development of oral healthcare. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] Figure 1 It is the overall architecture diagram of the system of the present invention;

[0110] Figure 2 This is a diagram of the data acquisition and processing module of the present invention;

[0111] Figure 3 This is a diagram of the oral disease identification module of the present invention;

[0112] Figure 4 This is a diagram of a personalized care recommendation module of the present invention;

[0113] Figure 5 This is a diagram of the personalized product recommendation module of the present invention.

[0114] Figure 6 This is the health monitoring and feedback module diagram of the present invention DETAILED DESCRIPTION

[0115] Please refer to the attached Figure 1-6 The present invention provides an intelligent oral health assessment and personalized care system and method. The system includes multiple functional modules and can achieve intelligent management of the entire process from oral data collection to personalized care plan recommendation.

[0116] Specifically, the intelligent oral health assessment and personalized care system of the present invention includes a data acquisition and processing module 1, an oral disease identification module 2, a personalized care recommendation module 3, a personalized product recommendation module 4, and a health monitoring and feedback module 5. These modules are connected through communication to form a complete system architecture that can work together to achieve intelligent oral health management.

[0117] Data acquisition and processing module 1 is used to acquire and preprocess the user's oral image data. In a preferred embodiment of the present invention, this module can acquire high-quality oral images using equipment such as X-rays or oral CT scans. The acquired raw image data typically contains some noise and interference, necessitating preprocessing. This preprocessing process includes noise reduction and image enhancement, which helps improve the accuracy of subsequent disease identification. For example, median filtering can be used for noise reduction, and histogram equalization can be used for image enhancement to make oral structures more clearly visible.

[0118] Oral disease identification module 2 is communicatively connected to data acquisition and processing module 1, and is configured to receive preprocessed oral image data and identify oral diseases based on this data. The present invention utilizes deep learning algorithms to achieve high-precision disease identification. Preferably, a convolutional neural network (CNN) can be used as the underlying model, as it performs well in image recognition tasks. For example, a classic CNN architecture such as ResNet-50 or DenseNet can be employed, appropriately adjusted and optimized based on the characteristics of oral images.

[0119] The personalized care recommendation module 3 is in communication with the oral disease identification module 2, receiving disease identification results and, based on the user's personal information and lifestyle data, generating a personalized care plan. In one embodiment of the present invention, this module can use a random forest algorithm to generate care plans. The random forest algorithm has excellent generalization and noise immunity, making it suitable for processing multi-source heterogeneous data. For example, the disease identification results, user age, dietary habits, brushing frequency, and other features can be used as inputs to predict the most suitable care plan through the random forest model.

[0120] The personalized product recommendation module 4 is in communication with the personalized care recommendation module 3 and recommends oral care products suitable for the user based on the generated personalized care plan. The present invention preferably employs a recommendation algorithm based on deep learning, such as a hybrid recommendation model combining collaborative filtering and deep neural networks. This approach fully leverages historical user behavior data and product feature information to provide more accurate product recommendations.

[0121] The health monitoring and feedback module 5 communicates with the personalized care recommendation module 3 and the personalized product recommendation module 4 to monitor the user's oral health status in real time and dynamically adjust care plans and product recommendations based on the monitoring results. This closed-loop design ensures that the system can continuously optimize and adapt to changes in the user's oral health. For example, a smart toothbrush or mobile app can collect a user's daily oral care data and, combined with regular oral examination results, evaluate the effectiveness of the care plan and make necessary adjustments.

[0122] The data acquisition and processing module 1 further includes an image acquisition unit 11, a data preprocessing unit 12, and an image segmentation unit 13. Image acquisition unit 11 is responsible for acquiring oral imaging data from the user using X-rays or oral CT equipment. In practical applications, appropriate imaging equipment can be selected based on different diagnostic needs. For example, for early diagnosis of dental caries, a digital X-ray imaging system can be used, as it has high spatial resolution and can capture subtle lesions.

[0123] The data preprocessing unit 12 is in communication with the image acquisition unit 11 and performs noise reduction, image enhancement, and normalization on the acquired oral image data. Noise reduction can utilize methods such as Gaussian filtering or wavelet transform to remove random noise from the image. Image enhancement can utilize the contrast-limited adaptive histogram equalization (CLAHE) algorithm to improve image contrast and detail visibility. Normalization includes image resizing and pixel value normalization to ensure consistency in subsequent processing.

[0124] The image segmentation unit 13 is in communication with the data preprocessing unit 12 and is responsible for binarizing the preprocessed oral image data and segmenting the oral image into different anatomical regions based on edge detection and connected region analysis. Binarization can employ the Otsu algorithm to adaptively select a threshold. Edge detection can employ the Canny algorithm, which effectively detects the boundaries of oral structures. Connected region analysis can employ region growing or watershed algorithms to segment different anatomical structures (such as teeth, gums, and oral soft tissue).

[0125] Oral disease recognition module 2 includes a convolutional neural network unit 21, an attention mechanism unit 22, and a model optimization unit 23. Convolutional neural network unit 21 is used to construct a deep learning model with a multi-layer convolution and pooling structure, which is used to extract features and identify diseases from oral image data. In a preferred embodiment of the present invention, the ResNet-101 architecture can be used as the base network. This network performs well on the ImageNet dataset and has strong feature extraction capabilities.

[0126] Specifically, the ResNet-101 network contains 101 convolutional layers, which are organized in the form of residual blocks. Each residual block can be expressed as:

[0127] F(x)=H(x)-x,

[0128] Where F(x) represents the residual function, H(x) represents the desired underlying mapping, and x represents the input. This structure can effectively alleviate the vanishing gradient problem of deep networks, allowing the model to learn deeper features.

[0129] The attention mechanism unit 22 is connected to the convolutional neural network unit 21 to introduce the attention mechanism into the deep learning model to enhance the model's ability to identify key lesion areas. The present invention adopts a spatial attention mechanism, which can be expressed as:

[0130]

[0131] Among them, M represents the attention weight matrix, σ represents the sigmoid activation function, represents a 7×7 convolution operation, AvgPool and MaxPool represent average pooling and maximum pooling operations, respectively, and F represents the input feature map. This attention mechanism helps the model focus on abnormal areas in oral images and improves the accuracy of disease recognition.

[0132] The model optimization unit 23 is connected to the convolutional neural network unit 21 and the attention mechanism unit 22, and is used to improve the training efficiency and performance of the model using residual optimization technology, and dynamically adjust the model parameters based on the validation set data. Residual optimization technology can alleviate the degradation problem of deep networks, making it easier for the model to learn the identity mapping. During the training process, the Adam optimizer can be used, with the initial learning rate set to 0.001 and a learning rate decay strategy employed. For example, the learning rate can be reduced to 0.1 times the original value every 50 epochs. In addition, an early stopping strategy can be used to prevent overfitting. Specifically, the training process is terminated when the performance on the validation set does not improve for five consecutive epochs.

[0133] Through the collaborative work of these modules, the intelligent oral health assessment and personalized care system of the present invention can achieve highly accurate oral disease identification and personalized care plan recommendations. This system has the following advantages: First, the integration of multimodal data improves the accuracy and comprehensiveness of disease identification; second, the application of deep learning algorithms enables the system to automatically learn complex features, reducing the workload of manual feature engineering; third, the personalized recommendation algorithm considers individual user differences, providing more accurate and effective care recommendations; finally, real-time monitoring and feedback mechanisms ensure that the system can continuously optimize and adapt to the dynamic changes in the user's oral health status.

[0134] The personalized care recommendation module 3 includes a data fusion unit 31, a random forest algorithm unit 32 and a solution optimization unit 33. The data fusion unit 31 is used to integrate oral disease identification results, user personal information and lifestyle data. In a preferred embodiment of the present invention, the data fusion unit 31 adopts a feature-level fusion method to convert data from different sources into a unified feature vector. For example, the oral disease identification result can be represented as a multidimensional vector, where each dimension corresponds to a possible oral disease and its severity; user personal information may include age, gender, past medical history, etc.; lifestyle data may include brushing frequency, eating habits, smoking status, etc. These heterogeneous data are properly encoded and normalized to form a comprehensive feature vector.

[0135] The random forest algorithm unit 32 is in communication with the data fusion unit 31 and is used to construct a random forest model based on the integrated data and generate a personalized care plan through the model. The random forest algorithm was selected because it has excellent generalization ability and noise resistance, and is particularly suitable for processing high-dimensional feature spaces and complex nonlinear relationships. In the implementation of the present invention, the random forest model can include 100 decision trees, and the maximum depth of each tree is set to 10 to balance the complexity and generalization ability of the model. Feature selection uses the Gini index as the partitioning criterion, which helps to select the most discriminative features for decision making.

[0136] The mathematical expression of the random forest model can be summarized as follows:

[0137]

[0138] Where H(x) represents the output of the random forest model, K is the number of decision trees, and h k (x) represents the prediction result of the kth decision tree. Each decision tree is trained on a bootstrap sample of the original training set. This method is called bagging and can effectively reduce the risk of overfitting.

[0139] The plan optimization unit 33 is in communication with the random forest algorithm unit 32 and is used to evaluate the generated care plan based on predefined care effect evaluation indicators and optimize and adjust the care plan based on the evaluation results. The present invention adopts a multi-objective optimization method, simultaneously considering the effectiveness, feasibility, and user acceptance of the care plan. Evaluation indicators may include the expected degree of improvement in oral health, the difficulty of implementing the plan, the time required, and the cost. The optimization process can use a genetic algorithm to find the optimal care plan combination through iterative evolution.

[0140] The personalized product recommendation module 4 includes a deep learning recommendation unit 41, a personalized matching unit 42, and a recommendation interpretation unit 43. The deep learning recommendation unit 41 is used to build a product recommendation model based on a deep neural network and train the model using user historical data and product feature information. The present invention utilizes a neural collaborative filtering model based on an attention mechanism, which effectively captures the complex interactions between user preferences and product features.

[0141] The core part of the model can be expressed as:

[0142]

[0143] in, represents the predicted preference score of user u for product i, u u and v i are the embedding vectors of users and products respectively, e uis the user’s additional feature vector, Attention(·) is the attention mechanism function, and W1 and b1 are learnable parameters.

[0144] The personalized matching unit 42 is in communication with the deep learning recommendation unit 41 and is used to select the most suitable products from the recommendation model output based on the user's personalized care plan and oral health status. This process not only takes into account the user's preferences predicted by the model but also incorporates the user's current oral health needs. For example, for users with periodontitis, the system will prioritize toothpaste and mouthwash containing anti-inflammatory ingredients.

[0145] The recommendation explanation unit 43 is in communication with the personalized matching unit 42 and is responsible for generating explanations for product recommendations, including the rationale and expected results. This function enhances the system's interpretability and user trust. The explanations are generated using a template- and rule-based approach that integrates product features, user needs, and model predictions.

[0146] The health monitoring and feedback module 5 includes a real-time data acquisition unit 51, a health status assessment unit 52, and a dynamic adjustment unit 53. The real-time data acquisition unit 51 is used to periodically obtain the user's oral health data. In one embodiment of the present invention, this can be achieved using a smart toothbrush, an oral scanner, or a mobile application. For example, a smart toothbrush can record brushing time, force, and area covered each time it is used; an oral scanner can perform a 3D scan of the oral cavity once a month; and a mobile application can allow users to regularly record their subjective feelings about their oral health.

[0147] The health status assessment unit 52 is in communication with the real-time data acquisition unit 51 and is used to assess the user's oral health status based on the real-time data collected and detect abnormal changes in oral health status. The assessment process uses a combination of a rule-based expert system and a machine learning model. For example, some basic rules can be set, such as brushing for less than one minute for three consecutive days will trigger a warning; at the same time, a support vector machine (SVM) model is used to detect abnormal changes in oral health indicators.

[0148] The dynamic adjustment unit 53 is in communication with the health status assessment unit 52 and is used to dynamically adjust the personalized care plan and update the personalized product recommendation list based on the health status assessment results. This process adopts the concept of reinforcement learning, using the user's oral health improvement as a reward signal to continuously optimize the care plan and product recommendation strategy. Specifically, the Q-learning algorithm can be used, and its core update formula is:

[0149]

[0150] Among them, Q(s t ,at ) means in state s t Take action t The value of , α is the learning rate, r t is the immediate reward, and γ is the discount factor.

[0151] The system of the present invention also includes a user feedback module 6 for collecting user evaluations of the system's recommended care plans and products and optimizing the system's recommendation algorithm and product library based on user feedback. The introduction of the user feedback module 6 forms a complete closed loop for the system, enabling continuous self-optimization and adaptation to changing user needs.

[0152] In a preferred embodiment of the present invention, user feedback utilizes a multi-dimensional rating system, encompassing satisfaction, usability, and effectiveness. This feedback data is used as additional features in the recommendation algorithm and incorporated into the deep learning recommendation model through transfer learning. The system also employs collaborative filtering technology, leveraging feedback from similar users to improve personalized recommendations.

[0153] Through the collaborative work of the above modules, the intelligent oral health assessment and personalized care system of the present invention can provide comprehensive, accurate, and dynamic oral health management services. The system can not only accurately identify oral diseases but also provide personalized care advice and product recommendations based on the user's individual characteristics and real-time health status. Furthermore, through continuous health monitoring and user feedback, the system can continuously optimize its recommendation strategies, providing users with services that are increasingly tailored to their individual needs. This intelligent, personalized, and dynamic oral health management approach is expected to significantly improve users' oral health and quality of life.

[0154] The system of the present invention also includes a data security and privacy protection module 7, which is used to encrypt and store users' oral health data and personal information, implement identity authentication and permission control for data access, and use security protocols to protect data during transmission. Given the increasing emphasis on data security and user privacy, the introduction of this module is crucial to system compliance and user trust.

[0155] In a preferred embodiment of the present invention, the data security and privacy protection module 7 adopts a multi-layered security strategy. First, for data storage, the system uses the AES-256 encryption algorithm to encrypt sensitive data. AES-256 is a widely recognized high-strength encryption standard that can effectively prevent unauthorized access. The encryption process can be expressed as:

[0156] C=E k (P),

[0157] Among them, C is the ciphertext, P is the plaintext, and E kis an encryption function using the key k.

[0158] Secondly, for identity authentication and permission control, the system uses a role-based access control (RBAC) model combined with a multi-factor authentication mechanism. The RBAC model allows for flexible definition of access rights for different user roles, while multi-factor authentication (such as passwords, SMS verification codes, and biometrics) further enhances system security.

[0159] The system uses TLS 1.3 for data transmission encryption. Compared to previous protocols, TLS 1.3 offers faster handshake speeds and stronger security. Furthermore, the system implements web security best practices, such as protection against SQL injection and cross-site scripting (XSS), to comprehensively protect user data.

[0160] The system of the present invention also includes a mobile application module 8, which provides a mobile user interface. This allows users to access system functions via smartphones or tablet devices, enabling visual display and interactive query of oral health data, as well as the delivery of oral care reminders and health advice notifications. The introduction of the mobile application module 8 significantly improves the system's accessibility and user experience.

[0161] In one embodiment of the present invention, the mobile application module 8 adopts a responsive design that can adapt to mobile device screens of different sizes. The interface design follows the principles of Material Design to provide an intuitive and consistent user experience. The core functions of the application include:

[0162] 1. Health data display: Use interactive charts (such as line charts and radar charts) to display users' oral health trends.

[0163] 2. Personalized care plan: displays care tasks in calendar format and supports setting reminders.

[0164] 3. Product recommendations: Display personalized oral care products and provide detailed descriptions.

[0165] 4. Real-time consultation: Integrate intelligent chatbots to answer users’ oral health questions.

[0166] 5. Health education: Provide oral health knowledge base and customized health advice.

[0167] The mobile app also integrates augmented reality (AR) technology, allowing users to view their oral health in real time through the camera and receive overlaid diagnostic information and care recommendations. This innovative feature greatly enhances user engagement and self-management capabilities.

[0168] The present invention also provides an intelligent oral health assessment and personalized care method, comprising the following steps:

[0169] S1. Acquire the user's oral image data: Obtain the user's oral images using X-ray or oral CT equipment, and preprocess the acquired image data, including noise reduction, enhancement, and normalization. In a preferred embodiment of the present invention, cone-beam CT (CBCT) technology is used to acquire three-dimensional oral images. This technology has lower radiation dose and higher spatial resolution than traditional CT. The preprocessing stage uses wavelet transform for noise reduction and adaptive histogram equalization for enhancement. Finally, the image size is uniformly adjusted to 512x512 pixels, and the pixel values ​​are normalized to the range of 0-1.

[0170] S2, Oral Disease Identification: Oral diseases are identified using a deep learning model using preprocessed image data. The deep learning model employed in this invention includes a multi-layer convolutional neural network, an attention mechanism, and a residual optimization structure. Specifically, the model is based on the ResNet-101 architecture, but incorporates a spatial attention mechanism within each residual block. The model's loss function uses weighted cross entropy to address class imbalance:

[0171]

[0172] Where C is the number of disease categories, w i is the weight of the i-th class, y i is the true label, is the predicted probability.

[0173] S3, generate personalized care plans: integrate oral disease identification results, user personal information and lifestyle data, use the random forest algorithm to generate personalized care plans, and optimize the care plans based on predefined evaluation indicators. In this step, the random forest model contains 200 decision trees, and the maximum depth of each tree is 15. The feature importance analysis uses the mean impurity reduction (Mean Decrease Impurity) method to identify the most critical influencing factors. The optimization process uses a genetic algorithm with a population size of 100, an evolutionary generation of 50, a crossover probability of 0.8, and a mutation probability of 0.1.

[0174] S4, recommending personalized oral care products: Build a deep learning product recommendation model. Based on the personalized care plan and the user's oral health status, filter the most suitable products from the recommendation model output and generate explanatory information for the product recommendations. The recommendation model uses the Neural Collaborative Filtering (NCF) framework, combining General Matrix Factorization (GMF) and Multilayer Perceptron (MLP):

[0175]

[0176] Among them, σ is the sigmoid activation function, represents vector concatenation, GMF(u,i) and MLP(u,i) are the outputs of GMF and MLP respectively.

[0177] S5, Health Monitoring and Dynamic Adjustment: Periodically collects users' oral health data, assesses their oral health status, detects abnormal changes, and dynamically adjusts care plans and product recommendations based on the assessment results. This step uses a time series anomaly detection algorithm based on a long short-term memory (LSTM) network, which can capture long-term trends and short-term fluctuations in users' oral health status. The core structure of the LSTM model can be expressed as:

[0178] f t =σ(W f ·[h t-1 ,x t ]+b f ),

[0179] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0180] o t =σ(W o ·[h t-1 ,x t ]+b o ),

[0181]

[0182] Among them, f t 、i t 、o t They are forget gate, input gate and output gate, c t is the cell state, h t It is a hidden state.

[0183] S6, Collect User Feedback and Optimize the System: Obtain user feedback on care plans and recommended products, and optimize the system's algorithms and product library based on user feedback. This step uses an online learning algorithm that can update model parameters in real time. Specifically, the stochastic gradient descent (SGD) method is used to update the weights of the deep learning model:

[0184]

[0185] Among them, w t is the current weight, η is the learning rate, is the gradient of the loss function.

[0186] Through the above steps, the intelligent oral health assessment and personalized care method of the present invention achieves intelligent management of the entire process, from data collection to personalized services. This method not only accurately identifies oral diseases but also provides dynamically adjusted care plans and product recommendations based on the user's individual characteristics and real-time health status. This intelligent, personalized, and dynamic oral health management approach is expected to significantly improve users' oral health and quality of life, providing a new paradigm for the intelligent development of oral healthcare.

[0187] It should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. Intelligent oral health assessment and personalized care system, characterized by ,include: Data acquisition and processing module, used for: Obtain the user's oral imaging data; Preprocessing the oral image data, including noise reduction and image enhancement; The oral disease identification module is in communication with the data acquisition and processing module and is used to: Receiving the pre-processed oral image data sent by the data acquisition and processing module; Based on the preprocessed oral image data, identifying oral diseases through a deep learning algorithm; The personalized care recommendation module is in communication with the oral disease identification module and is used to: Receiving the disease identification result sent by the oral disease identification module; Generate a personalized care plan based on the disease identification results, user personal information and lifestyle data; A personalized product recommendation module, in communication with the personalized care recommendation module, is used to: Receiving the personalized care plan sent by the personalized care recommendation module; Recommending oral care products suitable for the user based on the personalized care plan; The health monitoring and feedback module is in communication with the personalized care recommendation module and the personalized product recommendation module and is used to: Real-time monitoring of users’ oral health status; Dynamically adjust personalized care plans and product recommendations based on monitoring results.

2. The system according to claim 1, characterized in that , the data acquisition and processing module includes: An image acquisition unit, used to obtain the user's oral imaging data through X-ray or oral CT equipment; a data preprocessing unit, communicating with the image acquisition unit and configured to perform noise reduction, image enhancement, and standardization processing on the oral image data; An image segmentation unit is communicatively connected to the data preprocessing unit and is used to: Perform binarization on the pre-processed oral image data; Based on edge detection and connected region analysis methods, oral images are segmented into different anatomical regions.

3. The system according to claim 1, characterized in that , the oral disease recognition module includes: Convolutional neural network units, used to: Build a deep learning model with multi-layer convolution and pooling structures; Perform feature extraction and disease identification on oral imaging data through the deep learning model; An attention mechanism unit, in communication with the convolutional neural network unit, for: Introducing an attention mechanism into the deep learning model; Enhance the model's ability to identify key lesion areas; A model optimization unit, in communication with the convolutional neural network unit and the attention mechanism unit, is configured to: Use residual optimization technology to improve model training efficiency and performance; Dynamically adjust model parameters based on validation set data.

4. The system according to claim 1, characterized in that , the personalized care recommendation module includes: Data fusion unit, used to integrate oral disease identification results, user personal information and lifestyle data; A random forest algorithm unit is communicatively connected to the data fusion unit and is used to: constructing a random forest model based on the integrated data; generating a personalized care plan through the random forest model; A solution optimization unit, in communication with the random forest algorithm unit, is configured to: Evaluate the generated nursing plan according to the predefined nursing effect evaluation indicators; Optimize and adjust the nursing plan based on the evaluation results.

5. The system according to claim 1, characterized in that , the personalized product recommendation module includes: Deep Learning Recommendation Unit, used for: Build a product recommendation model based on deep neural networks; Using user historical data and product feature information to train the recommendation model; A personalized matching unit, in communication with the deep learning recommendation unit, is configured to: Based on the user's personalized care plan and oral health status, select the most suitable product from the recommendation model output; The recommendation interpretation unit is in communication with the personalized matching unit and is used to: Generate explanatory information for product recommendations, including reasons for the recommendation and expected results.

6. The system according to claim 1, characterized in that , the health monitoring and feedback module includes: Real-time data acquisition unit, used to periodically obtain the user's oral health data; A health status assessment unit is communicatively connected to the real-time data acquisition unit and is used to: Evaluate the user's oral health status based on real-time collected data; Detect abnormal changes in oral health status; A dynamic adjustment unit, in communication with the health status assessment unit, is configured to: Dynamically adjust personalized care plans based on health status assessment results; Update personalized product recommendation list.

7. The system according to claim 1, characterized in that , also includes a user feedback module for: Collect users' evaluations of the care plans and products recommended by the system; Based on user feedback, optimize the system's recommendation algorithm and product library.

8. The system according to claim 1, characterized in that , also includes data security and privacy protection modules for: Encrypted storage of users' oral health data and personal information; Implement identity authentication and permission control for data access; Use security protocols to protect data during transmission.

9. The system according to claim 1, characterized in that , also includes a mobile application module for: Provide a mobile user interface that allows users to access system functions through smartphones or tablet devices; Realize visual display and interactive query of oral health data; Send oral care reminders and health advice notifications.

10. An intelligent oral health assessment and personalized care method, using the system according to any one of claims 1 to 9, characterized in that , including the following steps: S1, collect user's oral image data: Obtain the user's oral image through X-ray or oral CT equipment; Preprocess the acquired image data, including noise reduction, enhancement and standardization; S2, Oral Disease Identification: Using pre-processed image data, oral diseases are identified through deep learning models; The deep learning model includes a multi-layer convolutional neural network, an attention mechanism, and a residual optimization structure; S3, generate personalized care plan: Integrate oral disease identification results, user personal information and lifestyle data; Generate personalized care plans using the random forest algorithm; Optimize care plans based on predefined evaluation indicators; S4, recommend personalized oral care products: Build a deep learning product recommendation model; Filter the most suitable products from the recommendation model output based on personalized care plans and user’s oral health status; Generate explanatory information for product recommendations; S5, health monitoring and dynamic adjustment: Periodically collect users' oral health data; Assess the user's oral health status and detect abnormal changes; Dynamically adjust care plans and product recommendations based on assessment results; S6, collect user feedback and optimize the system: Obtain user reviews of care plans and recommended products; Optimize the system's algorithms and product library based on user feedback.

Images