Mobile health supervision system based on traditional Chinese medicine multi-diagnosis combined ginseng

Abstract

The invention provides a mobile health supervision system based on traditional Chinese medicine multi-diagnosis combined parameters, and the system comprises the steps: collecting the multi-modal health data of a patient, and recognizing a plurality of single-modal health features of the multi-modal health data according to the traditional Chinese medicine knowledge, the single-mode health features are subjected to weighted fusion to obtain a plurality of compact fusion features, the inquiry contribution degrees of the compact fusion features are calculated, the multi-mode health data are enhanced according to each inquiry contribution degree, and a health information value corresponding to each single-mode health feature is obtained. According to the compact fusion features and the health information values, suspected diseases of the patients and disease type grades corresponding to the suspected diseases are analyzed, supervision diagnosis reports and rehabilitation suggestions are generated and displayed, and the problems that traditional Chinese medicine diagnosis is high in subjectivity, insufficient in standardization and difficult to popularize on a large scale are solved through technical innovation; various scenes such as family health maintenance, community medical service and auxiliary diagnosis of traditional Chinese medicine diagnosis and treatment institutions are realized, and health management requirements of different crowds are met.

Description

Technical Field

[0001] This invention relates to the intersection of traditional Chinese medicine diagnosis and mobile health technology, and in particular to a mobile health monitoring system based on the integration of multiple diagnostic methods in traditional Chinese medicine. Background Technology

[0002] Traditional Chinese medicine (TCM) diagnosis centers on the four diagnostic methods of inspection, auscultation and olfaction, inquiry, and palpation, emphasizing the comprehensive assessment of a person's health status through multi-dimensional physical signs. However, traditional TCM diagnosis heavily relies on the individual experience of physicians, resulting in inconsistent diagnostic standards, strong subjectivity, and difficulties in large-scale implementation. With the development of mobile internet and artificial intelligence technologies, mobile health monitoring systems are gradually becoming an important vehicle for health management; however, existing systems still have many shortcomings. First, the collection of multimodal health data lacks the guidance of traditional Chinese medicine theory, the data types are singular or have low correlation with traditional Chinese medicine syndrome differentiation, making it difficult to support the needs of multi-diagnosis and comprehensive reference. Second, the data processing did not fully integrate traditional Chinese medicine knowledge, and the feature extraction and fusion lacked specificity, resulting in insufficient diagnostic accuracy. Third, the lack of effective data augmentation and optimization mechanisms and the weak ability to process low-quality data affect the reliability of subsequent diagnostic results. Fourth, disease diagnosis and symptom classification are disconnected and lack a synergistic relationship; moreover, intervention recommendations are mostly general templates, lacking personalization and adaptability to traditional Chinese medicine. Fifth, the model lacks flexibility in deployment and upgrades, making it difficult to adapt to the health monitoring needs of different scenarios.

[0003] Therefore, there is an urgent need for a standardized, intelligent, and mobile TCM diagnosis and treatment system. This invention provides a mobile health monitoring system based on multi-diagnosis integration in TCM. Summary of the Invention

[0004] This invention provides a mobile health monitoring system based on multi-diagnosis and integrated treatment in traditional Chinese medicine. It enables comprehensive collection and precise processing of multimodal health data, improves the accuracy of disease diagnosis and symptom classification through collaborative algorithms, generates rehabilitation suggestions tailored to individual patients, and promotes the modernization and popularization of traditional Chinese medicine health services.

[0005] This invention provides a mobile health monitoring system based on multi-diagnosis and comprehensive consultation in Traditional Chinese Medicine, comprising: The consultation preparation module is used to collect patients' multimodal health data and identify several single-modal health features of the multimodal health data based on traditional Chinese medicine knowledge. The fusion analysis module is used to weight and fuse different numbers of the single-modal health features to obtain several compact fusion features, and to calculate the consultation contribution of each compact fusion feature. The enhanced analysis module is used to enhance the multimodal health data based on the contribution of each consultation, and obtain the health information value corresponding to each single-modal health feature; The supervised analysis module is used to analyze the patient's suspected diseases and the symptom type classification of each suspected disease based on the compact fusion features and health information values, and generate a supervised diagnosis report. The intervention suggestion module is used to perform rehabilitation simulation training on the supervised diagnosis report using genetic principles, obtain rehabilitation suggestions for the patient, and display them.

[0006] In one implementable manner, the consultation preparation module includes: The multi-mode acquisition unit is used to acquire visual image information and infrared thermal imaging information of the patient's first designated organ, acquire pulse signal of the patient's second designated organ, as well as the patient's response text to the specified questions and the patient's self-described symptoms in voice. The information fusion unit is used to construct an equivalent patient model of the patient based on the visual image information, the infrared thermal imaging information, the pulse signal, the response text, and the patient's self-reported symptoms, and to introduce several model life backgrounds into the equivalent patient model to obtain the patient's scene health characteristics in different scenarios. The knowledge decomposition unit is used to find several pieces of traditional Chinese medicine knowledge corresponding to each health feature of the scenario, construct the patient's health knowledge graph, and deduce the health level corresponding to each health feature of the scenario in the health knowledge graph. The single-modal analysis unit is used to perform single-modal feature learning on each of the health features of the scenario to obtain the single-modal health information of the patient in each single-modal dimension, and to construct the single-modal health feature of the patient in each single-modal dimension by combining the corresponding health level.

[0007] In one implementable manner, the fusion analysis module includes: A standard processing unit is used to simultaneously map each of the single-modal health features into the same dimensional space to obtain the feature correlation properties between different single-modal health features, and to generate several standardized health feature vectors with the same dimension by combining the mapping results corresponding to each single-modal health feature. The network construction unit is used to organize the TCM knowledge using a multi-layer perception mechanism, construct an attention network by combining prior knowledge in the field of TCM diagnosis with clinical data patterns, and use the attention network to analyze the attention score corresponding to each of the standardized health feature vectors. The weighted execution unit is used to treat each attention score as a weighting coefficient, perform weighted summation on the standardized health feature vector to obtain an initial fusion feature vector, filter the effective information contained in the initial fusion feature vector, and perform dimensionality reduction on the initial fusion feature vector to obtain several compact fusion features. The contribution analysis unit is used to identify the proportion of effective information contained in each compact fusion feature, and to determine the consultation contribution of each compact fusion feature in descending order of the proportion of effective information.

[0008] In one implementable manner, the enhancement analysis module includes: The spatial construction unit is used to perform quality evaluation and integrity evaluation on each of the compact fusion features respectively, determine the matching parameters between each of the compact fusion features and traditional Chinese medicine theory based on the evaluation results, construct the state space in combination with the single-modal health features, and construct the action space according to the binary selection rule. The data augmentation unit is used to input the multimodal health data into the state space and the action space respectively for state augmentation and action augmentation, and to record several data augmentation nodes of the multimodal health data, as well as the augmentation information corresponding to each data augmentation node; The feature comparison unit is used to identify the health feature data corresponding to each single-modal health feature in the multimodal health data, and to determine several data augmentation nodes and augmentation information contained in each single-modal health feature data, and to construct single-modal augmentation information corresponding to each single-modal health feature. The numerical determination unit is used to acquire the initial information corresponding to each of the single-modal health features, and to determine the health information value corresponding to each of the single-modal health features by combining the corresponding single-modal enhancement information.

[0009] One feasible approach also includes: The high-risk early warning module is used to determine whether the patient has a high-risk disease based on the health information value; If so, identify the symptoms corresponding to the high-risk disease and issue an emergency warning.

[0010] In one implementable manner, the supervised analysis module includes: The collaborative processing unit is used to acquire feature difference information and feature-related information between each of the compact fusion features and each of the single-modal health features, construct an information value association matrix based on the feature-related information, and perform collaborative processing on the health information values ​​corresponding to the compact fusion features and each of the single-modal health features based on the feature difference information until the difference metric value corresponding to the feature difference information is less than a preset threshold. A multi-level processing unit is used to identify several missing matrix elements contained in the information value association matrix based on the health information value corresponding to each single-modal health feature, and to use the knowledge of traditional Chinese medicine to perform element compensation on the missing matrix elements to generate several matrix feature vectors. The disease assessment unit is used to derive diseases from the feature vectors of each matrix using a preset disease inference model, to obtain the patient's suspected disease and the symptom type classification corresponding to each suspected disease, and at the same time, to search for disease evidence corresponding to each suspected disease in the multimodal health data and generate a supervised diagnosis report.

[0011] One feasible approach also includes: The model update unit is used to construct a corresponding model inference layer based on the single-mode dimension and single-mode function corresponding to each single-mode health feature; Based on the knowledge of traditional Chinese medicine, the model inference layers are combined to obtain the model framework, and several disease inference bases are constructed based on the knowledge of traditional Chinese medicine. Each of the disease derivation criteria is mapped into the model framework to generate a preset disease inference model.

[0012] In one implementable manner, the intervention recommendation module includes: The report training unit is used to construct the core logic of each suspected disease based on the supervised diagnosis report, with the training objective of eliminating the suspected disease and the training basis of the core logic of each disease to carry out rehabilitation simulation training. The training supervision unit is used to record the training process, determine several effective and ineffective training contents corresponding to each suspected disease, construct rehabilitation suggestions based on the effective training contents corresponding to each suspected disease, and display them.

[0013] One feasible approach also includes: It is suggested that the module be adjusted to obtain health advice uploaded by doctors and adjust the rehabilitation advice accordingly.

[0014] One feasible approach also includes: A pre-training unit is used to acquire the patient's health knowledge graph and use the health knowledge graph to update the TCM knowledge base of the pre-trained model. The updated pre-trained model is used to extract the TCM-related features corresponding to each of the aforementioned single-modal health features; Screen out single-model health features to be adjusted where the ratio of TCM-related features to non-TCM-related features is lower than a specified threshold. Based on the updated TCM knowledge base, identify TCM-related information contained in the corresponding non-TCM-related features, and use the TCM-related information to adjust the corresponding single-mode health features to be adjusted.

[0015] The beneficial effects achievable by this invention are as follows: To enable health monitoring and diagnostic assessment anytime and anywhere, significantly lower the access threshold for TCM health services, and promote the modernization and popularization of TCM health management, the invention first comprehensively collects multimodal health data corresponding to the four diagnostic methods of TCM (inspection, auscultation, inquiry, and palpation). Simultaneously, relying on TCM knowledge, it accurately identifies single-modal health characteristics, ensuring both the comprehensiveness and relevance of data collection and ensuring that the extracted features deeply align with TCM diagnostic logic, effectively avoiding interference from irrelevant data. This provides high-quality, highly relevant foundational data support for subsequent fusion analysis and diagnostic assessment. Then, by weighted fusion of different numbers of single-modal health features, multi-dimensional health information can be flexibly integrated to generate compact fusion features that are both complete and targeted. Simultaneously, the contribution of each fusion feature to the consultation is accurately calculated, clearly defining the weight of each single-modal feature in the diagnosis. Furthermore, using the consultation contribution as the core guide, the multimodal health data is targeted for enhancement, prioritizing the focus on... High-value data is optimized to effectively improve the usability of low-quality data and reduce noise interference in subsequent analysis. Simultaneously, quantified health information values ​​are generated, transforming abstract TCM characteristics into intuitive and comparable numerical indicators. Further combining globally correlated information with compactly integrated features and quantified health information values ​​enables two-dimensional collaborative analysis of patients' suspected diseases and corresponding syndrome classifications. This significantly improves the accuracy of disease identification and the precision of syndrome classification. Finally, by utilizing genetic principles to conduct multiple rounds of rehabilitation simulation training on the supervised diagnostic report, highly personalized rehabilitation suggestions are generated based on the patient's specific disease type, syndrome classification, and abnormal health information values. This completely eliminates the limitations of generic templates. Through technological innovation, this addresses the problems of strong subjectivity, insufficient standardization, and difficulty in large-scale promotion in traditional TCM diagnosis. It enables various scenarios such as family health maintenance, community medical services, and auxiliary diagnosis in TCM clinics, meeting the health management needs of different population groups.

[0016] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram illustrating the composition of a mobile health monitoring system based on multiple diagnostic methods in Traditional Chinese Medicine, as described in an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the composition of the enhanced analysis module in a mobile health monitoring system based on multiple diagnostic methods in traditional Chinese medicine, as described in an embodiment of the present invention. Detailed Implementation

[0019] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0020] Example 1: This example provides a mobile health monitoring system based on multi-diagnosis and comprehensive consultation in Traditional Chinese Medicine, such as... Figure 1 As shown, it includes: The consultation preparation module is used to collect patients' multimodal health data and identify several single-modal health features of the multimodal health data based on traditional Chinese medicine knowledge. The fusion analysis module is used to weight and fuse different numbers of the single-modal health features to obtain several compact fusion features, and to calculate the consultation contribution of each compact fusion feature. The enhanced analysis module is used to enhance the multimodal health data based on the contribution of each consultation, and obtain the health information value corresponding to each single-modal health feature; The supervised analysis module is used to analyze the patient's suspected diseases and the symptom type classification of each suspected disease based on the compact fusion features and health information values, and generate a supervised diagnosis report. The intervention suggestion module is used to perform rehabilitation simulation training on the supervised diagnosis report using genetic principles, obtain rehabilitation suggestions for the patient, and display them.

[0021] In this example, the multimodal health data represents the data obtained after examining the patient through observation, auscultation, inquiry, and palpation. In this example, unimodal health features represent the characteristics of the data presented by a patient under a certain consultation method, such as the characteristics of the patient's pulse data presented during palpation; In this example, compact fusion features represent the result of fusing two or more single-modal health features; In this example, the consultation contribution rate represents the proportion of effective information provided by a compact fusion feature in this consultation. In this example, the health information value represents a numerical representation of the patient's health status derived from a single-modal health feature; In this example, the fusion analysis module not only makes feature fusion more logical and accurate, reducing the computational burden caused by redundant information, but also makes the diagnostic process traceable and interpretable, making it easier for users and medical personnel to understand the impact of each diagnostic dimension on the diagnostic results and increasing their trust in the diagnostic conclusions. The enhanced analysis module not only achieves a reasonable allocation of data resources, but also provides accurate quantitative evidence for subsequent disease diagnosis and syndrome differentiation, making the diagnostic results more scientific and convincing. In this example, the symptom type classification represents the result of classifying the severity level of suspected diseases. The symptom type classification result can clearly present the severity of the condition, helping users and medical personnel to quickly grasp the level of health risk. The generated monitoring and diagnosis report standardizes and integrates key information such as diagnostic basis, results and risk warnings, providing a comprehensive and systematic reference for subsequent rehabilitation intervention, and also providing standardized diagnosis and treatment data for medical docking. In this example, rehabilitation simulation training represents the process of analyzing how a patient can recover their health in a virtual space; In this example, the rehabilitation recommendations cover multiple dimensions such as traditional Chinese medicine conditioning, acupuncture treatment, diet and exercise, and are displayed in an intuitive and easy-to-understand way, making it convenient for users to implement them directly. At the same time, it supports dynamic adjustment based on subsequent health monitoring data, forming a closed-loop health management system to help users efficiently improve their health status and control the progression of their condition.

[0022] The working principle and beneficial effects of the above technical solution are as follows: To enable health monitoring and diagnostic assessment anytime and anywhere, significantly lower the access threshold for TCM health services, and promote the modernization and popularization of TCM health management, the solution first comprehensively collects multimodal health data corresponding to the four diagnostic methods of TCM (inspection, auscultation, inquiry, and palpation). Simultaneously, relying on TCM knowledge, it accurately identifies single-modal health characteristics, ensuring both the comprehensiveness and relevance of data collection and ensuring that the extracted features deeply align with TCM diagnostic logic. This effectively avoids interference from irrelevant data, providing high-quality, highly relevant foundational data support for subsequent fusion analysis and diagnostic assessment. Then, by weighted fusion of different numbers of single-modal health features, multi-dimensional health information can be flexibly integrated to generate compact fusion features that are both complete and targeted. Simultaneously, the consultation contribution of each fusion feature is accurately calculated, clearly defining the weight of each single-modal feature in diagnosis. Furthermore, the multimodal health data is targeted and enhanced based on the consultation contribution, prioritizing... By focusing on optimizing high-value data, the usability of low-quality data is effectively improved, noise interference in subsequent analysis is reduced, and quantified health information values ​​are generated. Abstract TCM characteristics are transformed into intuitive and comparable numerical indicators. Furthermore, by combining globally correlated information of compactly integrated features with quantified health information values, a two-dimensional collaborative analysis of patients' suspected diseases and corresponding syndrome classifications is conducted, significantly improving the accuracy of disease identification and the precision of syndrome classification. Finally, by utilizing genetic laws to conduct multiple rounds of rehabilitation simulation training on the supervised diagnosis report, highly personalized rehabilitation suggestions can be generated by fully combining the patient's specific disease type, syndrome classification, and abnormal health information values, completely breaking away from the limitations of generalized templates. In this way, technological innovation solves the problems of strong subjectivity, insufficient standardization, and difficulty in large-scale promotion of traditional TCM diagnosis, realizing multiple scenarios such as family health maintenance, community medical services, and auxiliary diagnosis in TCM clinics, meeting the health management needs of different groups.

[0023] Example 2: Based on Example 1, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine includes a consultation preparation module comprising: The multi-mode acquisition unit is used to acquire visual image information and infrared thermal imaging information of the patient's first designated organ, acquire pulse signal of the patient's second designated organ, as well as the patient's response text to the specified questions and the patient's self-described symptoms in voice. The information fusion unit is used to construct an equivalent patient model of the patient based on the visual image information, the infrared thermal imaging information, the pulse signal, the response text, and the patient's self-reported symptoms, and to introduce several model life backgrounds into the equivalent patient model to obtain the patient's scene health characteristics in different scenarios. The knowledge decomposition unit is used to find several pieces of traditional Chinese medicine knowledge corresponding to each health feature of the scenario, construct the patient's health knowledge graph, and deduce the health level corresponding to each health feature of the scenario in the health knowledge graph. The single-modal analysis unit is used to perform single-modal feature learning on each of the health features of the scenario to obtain the single-modal health information of the patient in each single-modal dimension, and to construct the single-modal health feature of the patient in each single-modal dimension by combining the corresponding health level.

[0024] In this example, the first designated organ is the face, and the second designated organ is the wrist; In this example, the specified question refers to the question the doctor asks the patient; In this example, the patient's self-reported symptoms refer to the voice of the patient describing their own symptoms. In this example, unimodal feature learning represents the process of enhancing the presentation features of scene health features in a unimodal dimension; In this example, scenario-based health features represent the physical health status of patients when living in different scenarios; In this example, the model's living context represents a virtual space used to simulate the patient's different actions in different scenarios, such as sleeping scenarios and exercise scenarios. In this example, the health knowledge graph represents the result of integrating the patient's relevant TCM knowledge. By deriving the health level through the knowledge graph, the abstract TCM characteristics are transformed into clear and quantifiable level indicators, which solves the problem of the lack of standardized assessment standards for traditional TCM characteristics. This makes the determination of health status more scientific and consistent, and also clarifies the core direction for the subsequent construction of single-modal health characteristics.

[0025] The working principle and beneficial effects of the above technical solution are as follows: First, it accurately collects visual image information, infrared thermal imaging information, pulse signals, response text, and self-reported symptoms, covering key data dimensions corresponding to the four diagnostic methods of traditional Chinese medicine (TCM): observation, auscultation, inquiry, and palpation. The collection targets are clearly defined, the data types are comprehensive and highly targeted, effectively avoiding redundant irrelevant information and ensuring that the acquired raw data can directly meet the needs of TCM diagnosis. Then, it integrates the collected data from multiple types to construct an equivalent patient model, fully replicating the patient's physiological characteristics and health status. Simultaneously, it introduces various model life backgrounds to simulate and cover different life scenarios of the patient. Finally, based on the TCM knowledge system, it matches corresponding health characteristics for each scenario. Traditional Chinese medicine (TCM) knowledge is used to construct a systematic health knowledge graph, deeply binding scenario-based health characteristics with TCM theory. This ensures that feature analysis has a solid theoretical basis in TCM. Finally, specialized monomodal feature learning is conducted for each scenario's health characteristics to deeply explore the core health information under each monomodal dimension, ensuring the depth and accuracy of monomodal health information. At the same time, monomodal health information is organically combined with corresponding health levels. The constructed monomodal health features not only contain specific monomodal data details but also have a clear health status orientation. This provides high-quality and highly relevant core inputs for subsequent system fusion analysis and disease diagnosis, significantly improving the efficiency and accuracy of subsequent diagnosis and treatment.

[0026] Example 3: Based on Example 1, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine, wherein the fusion analysis module includes: A standard processing unit is used to simultaneously map each of the single-modal health features into the same dimensional space to obtain the feature correlation properties between different single-modal health features, and to generate several standardized health feature vectors with the same dimension by combining the mapping results corresponding to each single-modal health feature. The network construction unit is used to organize the TCM knowledge using a multi-layer perception mechanism, construct an attention network by combining prior knowledge in the field of TCM diagnosis with clinical data patterns, and use the attention network to analyze the attention score corresponding to each of the standardized health feature vectors. The weighted execution unit is used to treat each attention score as a weighting coefficient, perform weighted summation on the standardized health feature vector to obtain an initial fusion feature vector, filter the effective information contained in the initial fusion feature vector, and perform dimensionality reduction on the initial fusion feature vector to obtain several compact fusion features. The contribution analysis unit is used to identify the proportion of effective information contained in each compact fusion feature, and to determine the consultation contribution of each compact fusion feature in descending order of the proportion of effective information.

[0027] In this example, the feature correlation property represents the influence relationship between different unimodal health features; In this example, the standardized health feature vector represents a vector of a single modality of health under the same dimensional standard. The standardized health feature vector ensures the consistency of all single modality features in terms of data format, allowing the features corresponding to the four diagnostic methods of traditional Chinese medicine (inspection, auscultation, inquiry, and palpation) to be compared and integrated within the same framework. This avoids the loss of effective information due to differences in data format and provides standardized and unified input data for subsequent weighted fusion and contribution analysis, greatly improving the smoothness and accuracy of the entire fusion process. In this example, the attention network can accurately match the logic of TCM diagnosis. By analyzing the attention scores corresponding to the standardized health feature vectors, it can automatically identify the importance of each single modality feature in TCM diagnosis. This ensures the professionalism of the score calculation and avoids the drawback of traditional fusion algorithms lacking domain specificity. It provides a scientific basis for subsequent weighted fusion that meets the needs of TCM diagnosis and treatment, making the fusion process more in line with the core idea of ​​TCM multi-diagnosis and comprehensive reference. In this example, the compact fusion feature balances the integrity and condensation of information with efficient computational adaptability, enabling it to quickly meet the diagnostic needs of subsequent supervised analysis modules.

[0028] The working principle and beneficial effects of the above technical solution are as follows: First, different single-modal health features are uniformly mapped to the same dimensional space, effectively solving the problem of inconsistent dimensions of single-modal features and the difficulty in direct collaborative analysis. Simultaneously, by mining the correlation properties between different single-modal features, the potential associations between modalities are accurately captured, providing a logical foundation for subsequent fusion analysis. Then, a multi-layer perception mechanism is used to systematically organize traditional Chinese medicine knowledge, transforming scattered TCM theories into structured knowledge that can support algorithm operation. Furthermore, prior knowledge and clinical data patterns in the field of TCM diagnosis are deeply integrated, giving the constructed attention network a distinct TCM orientation, rather than simply relying on general algorithm modalities. Furthermore, attention scores are used as weighting coefficients to give higher weight to highly important single-modal features during the fusion process. This ensures that the fused features can focus on the key information of TCM diagnosis, improving the relevance and effectiveness of the fusion results. Finally, by identifying the proportion of effective information in compact fused features, the actual value of each compact fused feature is accurately quantified. The contribution of the consultation is determined by ranking the proportion of effective information, clearly showing the weight of different fused features in the diagnostic process. This allows for targeted enhancement of multimodal data corresponding to high-contribution fused features, and also makes the entire diagnostic process more interpretable, helping the system to continuously improve diagnostic accuracy.

[0029] Example 4: Based on Example 1, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine, such as... Figure 2 As shown, the enhancement analysis module includes: The spatial construction unit is used to perform quality evaluation and integrity evaluation on each of the compact fusion features respectively, determine the matching parameters between each of the compact fusion features and traditional Chinese medicine theory based on the evaluation results, construct the state space in combination with the single-modal health features, and construct the action space according to the binary selection rule. The data augmentation unit is used to input the multimodal health data into the state space and the action space respectively for state augmentation and action augmentation, and to record several data augmentation nodes of the multimodal health data, as well as the augmentation information corresponding to each data augmentation node; The feature comparison unit is used to identify the health feature data corresponding to each single-modal health feature in the multimodal health data, and to determine several data augmentation nodes and augmentation information contained in each single-modal health feature data, and to construct single-modal augmentation information corresponding to each single-modal health feature. The numerical determination unit is used to acquire the initial information corresponding to each of the single-modal health features, and to determine the health information value corresponding to each of the single-modal health features by combining the corresponding single-modal enhancement information.

[0030] In this example, the binary selection rule represents the rule for retaining or discarding a data segment of a certain modality; In this example, the state space represents the various forms in which compact fusion features can be presented, and the action space represents the space in which compact fusion features can be retained or selected. The constructed state space integrates fusion features, matching parameters, and single-modal health features, while the action space clarifies the direction of data processing based on binary selection rules. These two spaces provide a structured and standardized operating framework for subsequent data augmentation, making the data augmentation process systematic and targeted. This solves the problem of traditional data augmentation lacking theoretical constraints and a unified framework, laying a solid foundation for improving data quality and the accuracy of subsequent analysis.

[0031] The working principle and beneficial effects of the above technical solution are as follows: First, a dual evaluation of the quality and integrity of the compact fusion features is conducted to accurately select high-quality and highly complete core fusion features, effectively eliminating low-quality and redundant data. Simultaneously, the matching parameters between the fusion features and traditional Chinese medicine (TCM) theory are quantified to ensure that all analyzed data closely adheres to the TCM diagnostic logic and avoids deviating from the core orientation of multi-diagnosis and comprehensive reference. Then, multimodal health data is imported into the state space and action space, and state enhancement and action enhancement are performed respectively to achieve targeted optimization of the data. This not only strengthens the effective information related to TCM diagnosis in the data but also compensates for any possible deficiencies or omissions in the original data. Furthermore, detailed records of data enhancement nodes and corresponding enhancement information are kept, making the entire data enhancement process traceable and verifiable. This facilitates subsequent tracking of data optimization trajectories, verification of the rationality of enhancement effects, and further accurate identification of single modalities. The health feature data corresponding to the health features are targeted and associated with the enhancement nodes and enhancement information in the data enhancement process to construct exclusive single-modal enhancement information. This avoids confusion of enhancement information for different modal features and ensures that the enhancement of each single-modal feature has clear basis and complete record. Finally, based on the initial information of the single-modal health features, the targeted enhancement information is fully integrated. This retains the core attributes of the original features and absorbs the effective content supplemented during the enhancement process. The final generated health information value has both the authenticity of the original data and the richness of the enhanced data. This solves the problem of the difficulty in accurately quantifying traditional Chinese medicine features, makes the health status corresponding to each single-modal feature clearly identifiable, greatly improves the scientificity and accuracy of the diagnostic results, and meets the core needs of mobile health monitoring systems for rapid and accurate data analysis.

[0032] Example 5: Based on Example 1, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine further includes: The high-risk early warning module is used to determine whether the patient has a high-risk disease based on the health information value; If so, identify the symptoms corresponding to the high-risk disease and issue an emergency warning.

[0033] The working principle and beneficial effects of the above technical solution are as follows: it provides emergency warnings for emergencies, further protecting the personal safety of patients.

[0034] Example 6: Based on Example 1, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine, wherein the monitoring and analysis module includes: The collaborative processing unit is used to acquire feature difference information and feature-related information between each of the compact fusion features and each of the single-modal health features, construct an information value association matrix based on the feature-related information, and perform collaborative processing on the health information values ​​corresponding to the compact fusion features and each of the single-modal health features based on the feature difference information until the difference metric value corresponding to the feature difference information is less than a preset threshold. A multi-level processing unit is used to identify several missing matrix elements contained in the information value association matrix based on the health information value corresponding to each single-modal health feature, and to use the knowledge of traditional Chinese medicine to perform element compensation on the missing matrix elements to generate several matrix feature vectors. The disease assessment unit is used to derive diseases from the feature vectors of each matrix using a preset disease inference model, to obtain the patient's suspected disease and the symptom type classification corresponding to each suspected disease, and at the same time, to search for disease evidence corresponding to each suspected disease in the multimodal health data and generate a supervised diagnosis report.

[0035] In this example, the method of eliminating feature difference information not only retains the core effective information of various features, but also achieves collaborative data adaptation, providing a high-quality data foundation with strong consistency and high correlation for subsequent matrix construction and disease inference, and greatly reducing the risk of diagnostic bias caused by data differences; In this example, the information value association matrix represents the matrix obtained by decomposing feature-related information into several independent sub-information and then inputting them into a matrix of specified size; In this example, the matrix eigenvector represents the vector generated from the matrix rows and columns of the correlation matrix based on the compensated information values. The matrix eigenvector integrates complete correlation information and TCM-guided compensation data, achieving comprehensive information coverage and making the feature presentation more regular and logical. It can directly meet the input requirements of the disease inference model, providing structurally standardized, content-complete, and TCM-inspired feature support for subsequent disease assessment, thus improving the smoothness and accuracy of model derivation.

[0036] The working principle and beneficial effects of the above technical solution are as follows: First, it captures the difference and related information between compact fusion features and single-modal health features. By constructing an information value association matrix, the system integrates the association logic of the two types of features, allowing the scattered feature data to form an organic whole. At the same time, it carries out targeted collaborative processing for feature difference information, continuously optimizing compact fusion features and health information values ​​until the differences are eliminated. This effectively solves the inconsistency problem of feature data from different sources, ensuring that the data input into the diagnostic process is highly consistent in logic and format. Then, based on the single-modal health information values, it accurately identifies missing matrix elements in the information value association matrix, avoiding information gaps caused by missing data, and ensuring the integrity and continuity of the matrix. Simultaneously, relying on the knowledge system of Traditional Chinese Medicine (TCM) to provide targeted compensation for missing elements, rather than using a general data completion algorithm, ensures that the compensation results strictly adhere to the TCM diagnostic logic, guaranteeing the professionalism and rationality of the matrix data. This solves the problems of traditional data completion lacking domain specificity and easily deviating from diagnostic and treatment needs. Finally, a pre-set disease inference model is used to professionally derive diseases from the matrix feature vectors. The model is based on the TCM multi-diagnosis and multi-parameter theory and trained and optimized with a large amount of clinical data. It can accurately mine the correspondence between feature vectors and diseases and syndrome types, efficiently identify patients' suspected diseases, and scientifically classify the syndrome types corresponding to each disease, clearly presenting the severity of the condition. This solves the problems of traditional TCM... This addresses the issues of subjective medical diagnosis and inconsistent criteria for syndrome differentiation. It also traces disease evidence corresponding to suspected diseases from multimodal health data, making diagnostic results traceable and verifiable, significantly improving the credibility and persuasiveness of diagnostic conclusions. Through professional deduction and evidence tracing, it enhances the scientific rigor and reliability of diagnoses. The generated supervisory diagnostic reports not only meet the needs of ordinary users for a clear understanding of their health status but also provide standardized data support for professional medical scenarios. It effectively connects family health management with clinical diagnosis and treatment services, promoting the precision and standardization of TCM mobile health supervision and further expanding the system's applicable scenarios and practical value.

[0037] Example 7: Based on Example 6, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine further includes: The model update unit is used to construct a corresponding model inference layer based on the single-mode dimension and single-mode function corresponding to each single-mode health feature; Based on the knowledge of traditional Chinese medicine, the model inference layers are combined to obtain the model framework, and several disease inference bases are constructed based on the knowledge of traditional Chinese medicine. Each of the disease derivation criteria is mapped into the model framework to generate a preset disease inference model.

[0038] The working principle and beneficial effects of the above technical solution are as follows: a pre-built disease inference model is constructed to assist the supervision process, improve the efficiency and quality of supervision, and ensure the safety of patients' lives.

[0039] Example 8: Based on Example 1, the intervention suggestion module of the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine includes: The report training unit is used to construct the core logic of each suspected disease based on the supervised diagnosis report, with the training objective of eliminating the suspected disease and the training basis of the core logic of each disease to carry out rehabilitation simulation training. The training supervision unit is used to record the training process, determine several effective and ineffective training contents corresponding to each suspected disease, construct rehabilitation suggestions based on the effective training contents corresponding to each suspected disease, and display them.

[0040] The working principle and beneficial effects of the above technical solution are as follows: First, based on the supervised diagnosis report, the core logic of each suspected disease is constructed, so that rehabilitation simulation training has a clear goal orientation and scientific basis, avoiding the blindness of rehabilitation training. Then, the training process is recorded to accurately distinguish between effective and ineffective training content, realizing the empirical screening of rehabilitation training effects. This avoids the waste of resources and inefficient intervention caused by including ineffective training content in rehabilitation recommendations. In this way, the accurate screening of effective training content can further ensure the practicality of rehabilitation recommendations, help users improve their health status through targeted training, and strengthen the intervention value of health supervision.

[0041] Example 9: Based on Example 1, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine further includes: It is suggested that the module be adjusted to obtain health advice uploaded by doctors and adjust the rehabilitation advice accordingly.

[0042] The working principle and beneficial effects of the above scheme are as follows: After the doctor uploads suggestions, the rehabilitation suggestions are modified based on the doctor's suggestions to provide patients with effective and actionable rehabilitation suggestions.

[0043] Example 10: Based on Example 6, the mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine further includes: A pre-training unit is used to acquire the patient's health knowledge graph and use the health knowledge graph to update the TCM knowledge base of the pre-trained model. The updated pre-trained model is used to extract the TCM-related features corresponding to each of the aforementioned single-modal health features; Screen out single-model health features to be adjusted where the ratio of TCM-related features to non-TCM-related features is lower than a specified threshold. Based on the updated TCM knowledge base, identify TCM-related information contained in the corresponding non-TCM-related features, and use the TCM-related information to adjust the corresponding single-mode health features to be adjusted.

[0044] In this example, the pre-trained model represents a model used to identify TCM-related theories contained in health characteristics; In this example, after adjusting the health features of the single modality to be adjusted, the original features are replaced with the adjusted features.

[0045] The working principle and beneficial effects of the above technical solution are as follows: First, the TCM knowledge base of the pre-trained model is updated to enhance the TCM adaptability of feature extraction. At the same time, single-modality health features to be adjusted are selectively screened if the ratio of TCM-related features to non-TCM-related features is lower than a specified threshold. Adjustments are made using TCM-related information, thereby improving the quality and TCM relevance of single-modality health features from the data source. This provides higher-quality feature data support for the collaborative processing, missing element compensation, and disease inference of the subsequent supervised analysis module. This not only further improves the accuracy of disease assessment but also strengthens the TCM multi-diagnosis and intelligent data processing capabilities of the entire mobile health supervision system.

[0046] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in Traditional Chinese Medicine, characterized in that, include: The consultation preparation module is used to collect patients' multimodal health data and identify several single-modal health features of the multimodal health data based on traditional Chinese medicine knowledge. The fusion analysis module is used to weight and fuse different numbers of the single-modal health features to obtain several compact fusion features, and to calculate the consultation contribution of each compact fusion feature. The enhanced analysis module is used to enhance the multimodal health data based on the contribution of each consultation, and obtain the health information value corresponding to each single-modal health feature; The supervised analysis module is used to analyze the patient's suspected diseases and the symptom type classification of each suspected disease based on the compact fusion features and health information values, and generate a supervised diagnosis report. The intervention suggestion module is used to perform rehabilitation simulation training on the supervised diagnosis report using genetic principles, obtain rehabilitation suggestions for the patient, and display them.

2. The mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, The consultation preparation module includes: The multi-mode acquisition unit is used to acquire visual image information and infrared thermal imaging information of the patient's first designated organ, acquire pulse signal of the patient's second designated organ, as well as the patient's response text to the specified questions and the patient's self-described symptoms in voice. The information fusion unit is used to construct an equivalent patient model of the patient based on the visual image information, the infrared thermal imaging information, the pulse signal, the response text, and the patient's self-reported symptoms, and to introduce several model life backgrounds into the equivalent patient model to obtain the patient's scene health characteristics in different scenarios. The knowledge decomposition unit is used to find several pieces of traditional Chinese medicine knowledge corresponding to each health feature of the scenario, construct the patient's health knowledge graph, and deduce the health level corresponding to each health feature of the scenario in the health knowledge graph. The single-modal analysis unit is used to perform single-modal feature learning on each of the health features of the scenario to obtain the single-modal health information of the patient in each single-modal dimension, and to construct the single-modal health feature of the patient in each single-modal dimension by combining the corresponding health level.

3. The mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, The fusion analysis module includes: A standard processing unit is used to simultaneously map each of the single-modal health features into the same dimensional space to obtain the feature correlation properties between different single-modal health features, and to generate several standardized health feature vectors with the same dimension by combining the mapping results corresponding to each single-modal health feature. The network construction unit is used to organize the TCM knowledge using a multi-layer perception mechanism, construct an attention network by combining prior knowledge in the field of TCM diagnosis with clinical data patterns, and use the attention network to analyze the attention score corresponding to each of the standardized health feature vectors. The weighted execution unit is used to treat each attention score as a weighting coefficient, perform weighted summation on the standardized health feature vector to obtain an initial fusion feature vector, filter the effective information contained in the initial fusion feature vector, and perform dimensionality reduction on the initial fusion feature vector to obtain several compact fusion features. The contribution analysis unit is used to identify the proportion of effective information contained in each compact fusion feature, and to determine the consultation contribution of each compact fusion feature in descending order of the proportion of effective information.

4. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, The enhanced analysis module includes: The spatial construction unit is used to perform quality evaluation and integrity evaluation on each of the compact fusion features respectively, determine the matching parameters between each of the compact fusion features and traditional Chinese medicine theory based on the evaluation results, construct the state space in combination with the single-modal health features, and construct the action space according to the binary selection rule. The data augmentation unit is used to input the multimodal health data into the state space and the action space respectively for state augmentation and action augmentation, and to record several data augmentation nodes of the multimodal health data, as well as the augmentation information corresponding to each data augmentation node; The feature comparison unit is used to identify the health feature data corresponding to each single-modal health feature in the multimodal health data, and to determine several data augmentation nodes and augmentation information contained in each single-modal health feature data, and to construct single-modal augmentation information corresponding to each single-modal health feature. The numerical determination unit is used to acquire the initial information corresponding to each of the single-modal health features, and to determine the health information value corresponding to each of the single-modal health features by combining the corresponding single-modal enhancement information.

5. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, Also includes: The high-risk early warning module is used to determine whether the patient has a high-risk disease based on the health information value; If so, identify the symptoms corresponding to the high-risk disease and issue an emergency warning.

6. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, The supervisory analysis module includes: The collaborative processing unit is used to acquire feature difference information and feature-related information between each of the compact fusion features and each of the single-modal health features, construct an information value association matrix based on the feature-related information, and perform collaborative processing on the health information values ​​corresponding to the compact fusion features and each of the single-modal health features based on the feature difference information until the difference metric value corresponding to the feature difference information is less than a preset threshold. A multi-level processing unit is used to identify several missing matrix elements contained in the information value association matrix based on the health information value corresponding to each single-modal health feature, and to use the knowledge of traditional Chinese medicine to perform element compensation on the missing matrix elements to generate several matrix feature vectors. The disease assessment unit is used to derive diseases from the feature vectors of each matrix using a preset disease inference model, to obtain the patient's suspected disease and the symptom type classification corresponding to each suspected disease, and at the same time, to search for disease evidence corresponding to each suspected disease in the multimodal health data and generate a supervised diagnosis report.

7. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in Traditional Chinese Medicine as described in claim 6, characterized in that, Also includes: The model update unit is used to construct a corresponding model inference layer based on the single-mode dimension and single-mode function corresponding to each single-mode health feature; Based on the knowledge of traditional Chinese medicine, the model inference layers are combined to obtain the model framework, and several disease inference bases are constructed based on the knowledge of traditional Chinese medicine. Each of the disease derivation criteria is mapped into the model framework to generate a preset disease inference model.

8. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, The intervention suggestion module includes: The report training unit is used to construct the core logic of each suspected disease based on the supervised diagnosis report, with the training objective of eliminating the suspected disease and the training basis of the core logic of each disease to carry out rehabilitation simulation training. The training supervision unit is used to record the training process, determine several effective and ineffective training contents corresponding to each suspected disease, construct rehabilitation suggestions based on the effective training contents corresponding to each suspected disease, and display them.

9. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 1, characterized in that, Also includes: It is suggested that the module be adjusted to obtain health advice uploaded by doctors and adjust the rehabilitation advice accordingly.

10. A mobile health monitoring system based on multi-diagnosis and comprehensive consultation in traditional Chinese medicine as described in claim 6, characterized in that, Also includes: A pre-training unit is used to acquire the patient's health knowledge graph and use the health knowledge graph to update the TCM knowledge base of the pre-trained model. The updated pre-trained model is used to extract the TCM-related features corresponding to each of the aforementioned single-modal health features; Screen out single-model health features to be adjusted where the ratio of TCM-related features to non-TCM-related features is lower than a specified threshold. Based on the updated TCM knowledge base, identify TCM-related information contained in the corresponding non-TCM-related features, and use the TCM-related information to adjust the corresponding single-mode health features to be adjusted.

Images