Remote medical management system and method based on cloud platform

Abstract

The invention relates to the field of remote medical management, and particularly discloses a remote medical management system and method based on a cloud platform. The method comprises the following steps: firstly, acquiring basic information of a diabetic patient acquired by a database, doctor's advice information of the diabetic patient acquired by the database and blood glucose values of the diabetic patient acquired by a portable device at a plurality of preset time points, then performing feature extraction and correlation analysis on the three information by using a deep learning technology, and finally, generating the blood glucose values of the diabetic patient through a generator. Therefore, a doctor is helped to quickly know the current health condition and the treatment effect of the patient, the treatment scheme can be conveniently and timely adjusted, personalized medical treatment is realized, and continuous optimization of long-term health management is promoted.

Description

Technical Field

[0001] The present application relates to the field of telemedicine management, and more specifically, to a cloud platform-based telemedicine management system and method thereof. Background Art

[0002] Diabetes usually requires long-term management and continuous attention. This management involves the implementation of plans in multiple aspects, including medication, lifestyle adjustments (such as diet and exercise), and rehabilitation therapy. In the traditional medical process, diabetic patients need to go to the hospital for follow-up regularly so that doctors can monitor the progress of the disease and adjust the treatment plan. This method is time-consuming, labor-intensive, and inconvenient.

[0003] With the development of telemedicine technology, doctors can track and monitor the health status and key physiological indicators of diabetic patients in real time through the Internet. This technology allows both doctors and diabetic patients to avoid leaving home and complete the necessary follow-up process, thus improving the convenience and efficiency of medical services.

[0004] Therefore, a cloud-based remote medical management system and method thereof are desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a remote medical management system based on a cloud platform and a method thereof, which first obtains the basic information of diabetic patients collected by a database, the medical advice information of diabetic patients collected by a database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by a portable device, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally passes a generator to generate a doctor-side patient health status assessment report, thereby helping doctors quickly understand the patient's current health status and treatment effect, facilitating timely adjustment of treatment plans, realizing personalized medicine, and promoting continuous optimization of long-term health management.

[0006] According to one aspect of the present application, a remote medical management system based on a cloud platform is provided, which includes:

[0007] A diabetic patient medical data acquisition module, used to acquire basic information of diabetic patients collected from a database, medical advice information of diabetic patients collected from a database, and blood sugar values ​​of diabetic patients at multiple predetermined time points collected from a portable device;

[0008] A diabetic patient medical data extraction module, used to extract a diabetic patient information multi-dimensional feature vector and a patient blood glucose feature vector from the diabetic patient basic information collected by the database, the diabetic patient medical advice information collected by the database, and the diabetic patient blood glucose values ​​at multiple predetermined time points collected by the portable device;

[0009] The doctor-side evaluation report generation module is used to generate a doctor-side patient health status evaluation report based on the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector.

[0010] According to another aspect of the present application, a remote medical management method based on a cloud platform is provided, comprising:

[0011] Obtaining basic information of the diabetic patient collected from a database, medical advice information of the diabetic patient collected from a database, and blood sugar values ​​of the diabetic patient at multiple predetermined time points collected from a portable device;

[0012] Extracting a multi-dimensional feature vector of diabetic patient information and a patient blood glucose feature vector from the basic information of diabetic patients collected by the database, the medical advice information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by the portable device;

[0013] Based on the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector, a doctor-side patient health status assessment report is generated.

[0014] Compared with the prior art, the present application provides a cloud-based remote medical management system and method thereof, which first obtains the basic information of diabetic patients collected by the database, the medical order information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by a portable device, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally passes the generator to generate a doctor-side patient health status assessment report, thereby helping doctors to quickly understand the patient's current health status and treatment effect, facilitate timely adjustment of treatment plans, realize personalized medicine, and promote continuous optimization of long-term health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 The block diagram is a cloud platform-based telemedicine management system according to an embodiment of the present application.

[0017] Figure 2 The present invention is a block diagram of a medical data extraction module for diabetic patients in a cloud platform-based remote medical management system according to an embodiment of the present application.

[0018] Figure 3 The present invention is a block diagram of a patient basic information feature extraction unit in a cloud platform-based remote medical management system according to an embodiment of the present application.

[0019] Figure 4 This is a block diagram of a doctor-side evaluation report generation module in a cloud platform-based remote medical management system according to an embodiment of the present application.

[0020] Figure 5 The present invention is a flowchart of a remote medical management method based on a cloud platform according to an embodiment of the present application.

[0021] Figure 6 is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0023] Figure 1 FIG. 1 is a block diagram of a cloud-based remote medical management system according to an embodiment of the present application. Figure 1 As shown, according to the cloud platform-based remote medical management system 100 of the embodiment of the present application, it includes: a diabetic patient medical data acquisition module 110, which is used to acquire the basic information of the diabetic patient collected by the database, the medical order information of the diabetic patient collected by the database, and the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device; a diabetic patient medical data extraction module 120, which is used to extract the diabetic patient information multi-dimensional feature vector and the patient blood glucose feature vector from the basic information of the diabetic patient collected by the database, the medical order information of the diabetic patient collected by the database, and the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device; a doctor-side evaluation report generation module 130, which is used to generate a doctor-side patient health status evaluation report based on the diabetic patient information multi-dimensional feature vector and the patient blood glucose feature vector.

[0024] In the above-mentioned cloud-based telemedicine management system 100, the diabetic patient medical data acquisition module 110 is used to obtain the basic information of diabetic patients collected by the database, the medical advice information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by the portable device. It should be understood that diabetes is a chronic disease that requires continuous attention and long-term management, involving complex treatment plans, including drug therapy, lifestyle adjustment, rehabilitation therapy and other aspects of the plan execution. In the traditional medical model, diabetic patients must go to the hospital for follow-up regularly, which is not only time-consuming and labor-intensive, but also inconvenient. With the advancement of telemedicine technology, doctors can now track and monitor the health status and key physiological indicators of diabetic patients in real time through the Internet. This technology enables both doctors and patients to avoid frequent outings while still completing the necessary follow-up process. Through telemedicine, doctors can adjust treatment plans in a timely manner, monitor the progression of the disease, and provide personalized health management suggestions. This progress not only improves the convenience and efficiency of medical services, but also improves patients' treatment experience and quality of life. Patients can more easily participate in their own treatment process, reducing the time and cost of going to and from the hospital. At the same time, telemedicine technology provides doctors with more comprehensive patient data and more timely feedback, thereby helping them make more accurate diagnosis and treatment decisions. Therefore, in the technical solution of the present application, by obtaining the basic information of diabetic patients collected by the database, the medical advice information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by portable devices, and combining deep learning technology, a doctor-side patient health status assessment report is generated to help doctors quickly grasp the patient's current health status and treatment response. This report can provide doctors with detailed data analysis, support them to adjust treatment plans in a timely manner, and realize personalized medical services. In this way, the continuous optimization of long-term health management strategies is promoted to ensure that patients receive continuous high-quality medical care.

[0025] Specifically, the basic information of diabetic patients includes key data such as age, gender, height, weight, and medical history. This information is essential for assessing the overall health status of patients and formulating personalized treatment plans. The acquisition of basic information is achieved through the database, which ensures the accuracy and reliability of the information and provides the medical team with a comprehensive understanding of the patient's background. Doctor's advice information refers to the treatment recommendations and behavioral guidance formulated by doctors for patients' diabetes management. This information covers aspects such as drug use, diet control, and exercise plans. Through the collection and management of the database, the persistence and real-time nature of the doctor's advice information are ensured, helping doctors to effectively track and adjust treatment plans in a remote environment. Blood glucose data is an extremely critical monitoring indicator in diabetes management. The blood glucose values ​​at multiple time points collected regularly by portable devices reflect the blood glucose change trend of patients in different time periods, and are an important basis for evaluating treatment effects and adjusting treatment plans. These data are transmitted and stored in real time through the cloud platform, which not only provides doctors with a detailed understanding of patients' blood glucose control, but also supports remote monitoring and immediate intervention, thereby effectively improving the accuracy and effectiveness of treatment. Obtaining the basic information, medical order information and blood sugar data of diabetic patients not only provides data support and technical guarantee for the operation of the remote medical management system, but more importantly, provides the medical team with the ability to fully grasp the patient's health status and formulate personalized treatment plans, thereby achieving more accurate and efficient diabetes management and monitoring.

[0026] In the above-mentioned cloud-based telemedicine management system 100, the diabetic patient medical data extraction module 120 is used to extract the diabetic patient information multi-dimensional feature vector and the patient blood glucose feature vector from the diabetic patient basic information collected by the database, the diabetic patient medical order information collected by the database, and the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device. It should be understood that the extraction of the diabetic patient information multi-dimensional feature vector and the patient blood glucose feature vector is to deeply understand the patient's health status, personalized needs and treatment response. These vectors combine the patient's basic characteristics, medical guidance and biological indicator data, and provide accurate data support and intelligent health management functions for the telemedicine management system through advanced data analysis and pattern recognition technology, thereby improving the quality of life and treatment effect of diabetic patients.

[0027] Figure 2 FIG. 1 is a block diagram of a diabetic patient medical data extraction module in a cloud platform-based remote medical management system according to an embodiment of the present application. Figure 2As shown, in a specific embodiment of the present application, the diabetic patient medical data extraction module 120 includes: a patient basic information feature extraction unit 121, which is used to perform feature extraction on the diabetic patient basic information collected by the database to obtain a diabetic patient basic information semantic understanding feature vector; a patient medical order information feature extraction unit 122, which is used to perform feature extraction on the diabetic patient medical order information collected by the database to obtain a diabetic patient medical order information semantic understanding feature vector; a patient information multi-dimensional feature association unit 123, which is used to associate the diabetic patient basic information semantic understanding feature vector with the diabetic patient medical order information semantic understanding feature vector to obtain the diabetic patient information multi-dimensional feature vector; a patient blood glucose value feature extraction unit 124, which is used to perform feature extraction on the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device to obtain the patient blood glucose feature vector.

[0028] It should be understood that the basic information of diabetic patients covers multiple key areas, such as personal physical characteristics, living habits, medical history, and possible genetic factors. Through feature extraction, useful features can be mined from this information, such as age group, gender, BMI index, etc. These features are not only the basis for medical decision-making, but also help the system understand the patient's overall health status and potential health risk factors. Feature extraction helps to construct a semantic understanding feature vector of the basic information of diabetic patients, so that the system can better understand and process information. This vectorized expression not only simplifies the storage and processing of data, but also reveals the patterns and laws hidden behind the data through data mining and machine learning techniques. For example, by analyzing the feature vector, the patient's possible lifestyle, such as dietary preferences, exercise habits, etc., can be inferred, thereby providing doctors with more accurate personalized advice and treatment plans.

[0029] Furthermore, feature extraction of the medical advice information of diabetic patients in the database is to provide more effective personalized services and decision support in the process of medical management and treatment. Diabetes is a chronic disease, and the patient's treatment plan usually involves a variety of medical advice and guidance, which needs to be systematically organized and utilized. Through feature extraction, such as extracting key information such as the frequency of drug use, diet control type, exercise advice, etc., the system can analyze and evaluate the patient's treatment compliance, lifestyle preferences, and disease control. These feature vectors not only help doctors identify patients' personalized treatment needs, but also predict future treatment trends based on historical data, thereby optimizing treatment plans and management strategies.

[0030] Furthermore, the semantic understanding feature vector of basic information of diabetic patients and the semantic understanding feature vector of medical order information are associated in order to achieve more comprehensive and accurate personalized services in medical management. The generation of this multi-dimensional feature vector combines the patient's personal characteristics, health status and treatment needs, providing the medical team with deeper insights and effective decision-making basis. The semantic understanding feature vector of basic information of diabetic patients includes key information such as the patient's age, gender, body mass index (BMI), family medical history, etc. These feature vectors help the system understand the basic health status and potential risk factors of each patient, and provide a basis for the formulation of personalized treatment plans. For example, by analyzing the basic information feature vector, the system can identify the different needs that young female patients and middle-aged male patients may have in diabetes management, so as to provide targeted advice and services. The semantic understanding feature vector of medical order information covers specific treatment guidance such as the patient's medication use, diet control, exercise advice, etc. These feature vectors reflect the personalized needs and behavior patterns of each patient during the treatment process. By analyzing the feature vector of medical order information, the system can evaluate the patient's treatment compliance, behavioral preferences and treatment effects, and provide the medical team with real-time feedback and a basis for adjusting treatment strategies. By associating these two types of feature vectors to form a multi-dimensional feature vector, the system can more comprehensively understand and grasp the overall situation and treatment needs of each diabetic patient. For example, based on a comprehensive analysis of basic information and doctor's advice information, the system can infer how a patient's lifestyle preferences affect their disease management, or predict the efficacy of a specific patient group under a specific treatment strategy. The application of this multi-dimensional feature vector not only improves the comprehensiveness and operability of medical data, but also helps achieve the goal of precision medicine, which is to tailor the best health management plan for each patient.

[0031] In particular, feature extraction of blood glucose values ​​of diabetic patients at multiple predetermined time points collected by portable devices is to comprehensively evaluate the blood glucose control of patients and provide a scientific basis for personalized treatment and management. Among them, blood glucose value, as one of the core indicators of diabetes management, reflects the dynamic changes of patients' blood glucose levels. Portable devices can monitor blood glucose values ​​in real time in patients' daily lives and obtain data at different time points. Through feature extraction, such as extracting key features such as average blood glucose value, blood glucose fluctuation range, and blood glucose control index, the system can comprehensively analyze the patient's blood glucose fluctuation pattern and trend changes, and evaluate their disease control level and treatment effect. Through time series analysis and pattern recognition, the system can discover the tendency of hypoglycemia or hyperglycemia events that patients may have, and promptly remind doctors and patients to take necessary intervention measures. For example, the feature vector can show the patient's performance in postprandial blood glucose control, or the law of blood glucose fluctuations in a specific time period, and provide real-time suggestions for adjusting diet and drug treatment plans.

[0032] Figure 3 FIG. 1 is a block diagram of a patient basic information feature extraction unit in a cloud platform-based telemedicine management system according to an embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, the patient basic information feature extraction unit 121 includes: a patient basic information semantic encoding subunit 1211, which is used to obtain a plurality of diabetic patient basic information feature vectors by passing the diabetic patient basic information collected from the database through a diabetic patient basic information semantic encoding model; a patient basic information feature encoding subunit 1212, which is used to perform feature encoding on the plurality of diabetic patient basic information feature vectors to obtain the diabetic patient basic information semantic understanding feature vector.

[0033] It should be understood that the basic information of diabetic patients involves multiple dimensions such as age, gender, height and weight, family history, medical history records, etc. This information is crucial to understanding the individual characteristics and health status of patients. Through the semantic coding model of basic information of diabetic patients, these complex basic information can be converted into structured feature vectors, so that the medical system can more quickly and accurately obtain and analyze the key characteristics of patients. Specifically, the basic information of diabetic patients collected from the database is segmented to obtain a patient basic information word sequence; the embedding layer of the semantic coding model of basic information of diabetic patients is used to map each patient basic information word in the patient basic information word sequence to a patient basic information word embedding vector to obtain a sequence of patient basic information word embedding vectors; the converter-based Bert model of the semantic coding model of basic information of diabetic patients is used to perform global context semantic encoding on the sequence of patient basic information word embedding vectors to obtain multiple diabetic patients' basic information feature vectors.

[0034] Furthermore, feature encoding of multiple diabetic patients' basic information feature vectors is to improve the efficiency of medical data expression and analysis, thereby supporting more accurate and personalized medical decision-making and services. The basic information of diabetic patients involves multiple aspects, such as age, gender, BMI, family medical history, medication status, etc. This information is crucial for understanding the patient's overall health status, risk of disease, and the formulation of treatment plans. Through feature encoding, these complex basic information can be converted into structured and standardized feature vectors, allowing the medical system to store, retrieve and analyze this information more effectively. The feature encoding process converts the original multidimensional data into a higher-level semantic understanding feature vector through dimensionality reduction and abstraction. These vectorized features include not only static data of basic information, but also may integrate dynamic factors such as the patient's lifestyle and socioeconomic background, thereby more comprehensively reflecting the patient's individual characteristics and health status. For example, by encoding features such as age and BMI into numerical indicators, the system can more conveniently perform statistical analysis and trend identification, providing doctors with more specific and practical clinical advice.

[0035] In a specific embodiment of the present application, the patient basic information feature encoding subunit 1212 includes: arranging the multiple diabetic patients' basic information feature vectors according to the sample dimension to obtain a two-dimensional feature matrix of diabetic patients' basic information; passing the two-dimensional feature matrix of diabetic patients' basic information through a diabetic patients' basic information convolutional neural network as a feature encoder to obtain a semantic understanding feature vector of the diabetic patients' basic information.

[0036] It should be understood that the basic information of diabetic patients covers a variety of key features, such as age, gender, body mass index (BMI), blood pressure, medication status, etc. After each patient's information is quantized through feature vectors, a vector containing multiple features can be formed. By arranging the feature vectors of multiple patients according to the dimensions of the sample (i.e., the patient), a two-dimensional feature matrix can be constructed, in which each row represents a diabetic patient and each column represents a feature. The construction of a two-dimensional feature matrix of the basic information of diabetic patients makes medical data more operational and capable of visual analysis. In this form of a matrix, medical professionals can view the key features of each patient at a glance, such as age distribution, gender ratio, distribution of different BMI intervals, etc. This intuitive data display method helps to quickly identify common characteristics and potential health risks of patient groups.

[0037] Furthermore, the two-dimensional feature matrix of the basic information of diabetic patients is passed through the convolutional neural network of the basic information of diabetic patients as a feature encoder, aiming to achieve higher-level feature extraction and semantic understanding through deep learning technology, thereby supporting more accurate and intelligent diabetes management and personalized medical decision-making. The two-dimensional feature matrix of the basic information of diabetic patients contains a large amount of structured data, such as multi-dimensional features such as age, gender, BMI, and blood pressure. Although these data have been organized into a matrix form for easy analysis, it is still necessary to further extract higher-level and more abstract features to reveal the deep patterns and associations behind the health status of patients. Convolutional neural network (CNN) plays a key role as a feature encoder. CNN can effectively process two-dimensional data structures and extract the spatial hierarchy and pattern of features through multi-layer convolution and pooling operations. For the basic information matrix of diabetic patients, CNN can identify and strengthen the complex relationships between features while retaining the data structure, such as the interaction between different features and the impact of specific feature combinations on the health status of patients. The semantic understanding feature vector of the basic information of diabetic patients generated by CNN not only contains the high-level semantic information hidden in the original data, but also effectively reduces the data dimension and retains key information. Specifically, each layer of the convolutional neural network of the basic information of diabetic patients used as a feature encoder performs convolution processing, mean pooling processing based on the local feature matrix and nonlinear activation processing on the input data in the forward pass of the layer, so that the last layer of the convolutional neural network of the basic information of diabetic patients used as a feature encoder outputs the semantic understanding feature vector of the basic information of diabetic patients, wherein the input of the convolutional neural network of the basic information of diabetic patients used as a feature encoder is the two-dimensional feature matrix of the basic information of diabetic patients.

[0038] In a specific embodiment of the present application, the patient medical order information feature extraction unit 122 includes: passing the diabetes patient medical order information collected from the database through a diabetes patient medical order information word embedding model to obtain multiple diabetes patient medical order information data items; passing the multiple diabetes patient medical order information data items through a diabetes patient medical order information context semantic encoder to obtain the diabetes patient medical order information semantic understanding feature vector.

[0039] It should be understandable that the medical advice information for diabetic patients covers many aspects such as medication recommendations, dietary guidance, and exercise recommendations. This information is crucial for patient health management, but it usually exists in an unstructured form and is difficult to use directly for data analysis and model training. Through the medical advice information word embedding model, these complex natural language descriptions can be converted into numerical vector representations, allowing the medical system to understand and use this information more efficiently. The medical advice information word embedding model uses deep learning technology to convert medical advice text into feature vectors with semantic understanding capabilities. These vectors can not only capture key words and context in the medical advice content, but also reflect the similarities and differences between different medical advices. For example, for similar treatment recommendations, the model can generate close embedding vectors to help doctors identify effective treatment patterns and best practices. Based on the extraction and embedding representation of medical advice information data items, the structured and intelligent application of medical data is supported.

[0040] Furthermore, multiple data items of medical order information of diabetic patients are converted into feature vectors of semantic understanding of medical order information through the context semantic encoder of medical order information of diabetic patients, aiming to realize high-level semantic understanding of medical data through deep learning technology, so as to support more intelligent and personalized disease management and treatment decision-making. Among them, the data items of medical order information of diabetic patients contain various treatment suggestions, medication plans and lifestyle guidance. This information is crucial for the health management of patients, but due to its unstructured and diverse nature, traditional data processing methods often fail to fully utilize its potential. Through the context semantic encoder of medical order information, these information items can be converted into feature vectors of semantic understanding, so that the medical system can more accurately understand and analyze the treatment needs and health status of patients. The context semantic encoder of medical order information uses deep neural network technology, especially models suitable for processing natural language such as recurrent neural network (RNN) or Transformer model, which can effectively capture the semantic context of medical order information. These encoders can understand the meaning of words, the structure of sentences and the logical relationship between different information items in medical order information, so as to generate semantic understanding feature vectors that can accurately reflect the content of medical order. Specifically, the embedding layer of the context semantic encoder of the medical order information for diabetic patients is used to map each of the multiple medical order information data items for diabetic patients into a medical order information word embedding vector to obtain a sequence of medical order information word embedding vectors; the converter-based Bert model of the context semantic encoder of the medical order information for diabetic patients is used to perform global context semantic encoding on the sequence of medical order information word embedding vectors to obtain multiple medical order information semantic feature vectors; and, the multiple medical order information semantic feature vectors are cascaded to obtain the semantic understanding feature vector of the medical order information for diabetic patients.

[0041] In a specific embodiment of the present application, the patient blood glucose value feature extraction unit 124 includes: arranging the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device according to the time dimension to obtain a patient blood glucose input vector; passing the patient blood glucose input vector through a patient blood glucose time series encoder including a fully connected layer and a one-dimensional encoder to obtain the patient blood glucose feature vector.

[0042] It should be understood that blood glucose value is one of the important indicators for assessing the health status of diabetic patients. Portable devices can monitor the patient's blood glucose level at different time points in real time and generate time series data. By arranging these data according to the time dimension, a blood glucose input vector can be formed, in which each time point corresponds to a data point in the vector, reflecting the dynamic changes of the patient's blood glucose value. The construction of the blood glucose input vector helps the medical team to have a more comprehensive understanding of the patient's blood glucose control. By analyzing these vectors, the patient's blood glucose fluctuation pattern, peak hours, and potential hypoglycemic events can be identified. This information is crucial for adjusting drug treatment plans, recommending diet and exercise plans, and evaluating treatment effectiveness.

[0043] Furthermore, the patient blood glucose input vector is converted into a patient blood glucose feature vector through a patient blood glucose time series encoder including a fully connected layer and a one-dimensional encoder, aiming to further extract and characterize the advanced features of blood glucose data through deep learning technology, thereby supporting more accurate and intelligent diabetes management and treatment decisions. The patient blood glucose input vector is blood glucose measurement data arranged in the time dimension, reflecting the changes in the patient's blood glucose level at different time points. Although these data already have certain time series characteristics, they still need further processing and analysis to reveal deeper blood glucose control patterns and dynamic changes in the patient's health status. The patient blood glucose time series encoder is usually composed of a fully connected layer and a one-dimensional encoder. The fully connected layer is used to perform preliminary feature extraction and conversion on the blood glucose input vector, integrating and processing the blood glucose data at each time point with its context. The one-dimensional encoder uses deep neural network technology, such as long short-term memory network (LSTM) or gated recurrent unit (GRU), to specifically process time series data and capture the temporal dependencies and long-term memory therein. The patient blood glucose feature vector generated by the patient blood glucose time series encoder not only contains the important features and trends in the original blood glucose data, but also reduces the dimension of the data and retains key information. These feature vectors can be used to identify the patient's blood sugar control status, such as whether there are diurnal fluctuations, whether hyperglycemia or hypoglycemia events occur frequently, etc. This refined feature extraction helps doctors to more accurately assess the patient's health status and develop personalized treatment plans. Specifically, the patient blood sugar input vector is fully connected using the fully connected layer of the patient blood sugar time series encoder comprising a fully connected layer and a one-dimensional encoder to extract high-dimensional implicit features of the eigenvalues ​​at each position in the patient blood sugar input vector; and, the patient blood sugar input vector is one-dimensionally encoded using the one-dimensional convolutional layer of the patient blood sugar time series encoder comprising a fully connected layer and a one-dimensional encoder to extract high-dimensional implicit correlation features of the correlation between the eigenvalues ​​at each position in the patient blood sugar input vector.

[0044] In the above-mentioned cloud-based telemedicine management system 100, the doctor-side evaluation report generation module 130 is used to generate a doctor-side patient health status evaluation report based on the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector. It should be understood that the multi-dimensional feature vector of the diabetic patient information covers the patient's basic information, medical history, lifestyle, family history and other data. These feature vectors are presented in a structured manner, so that the medical team can quickly understand the patient's overall background and necessary information for chronic disease management. The patient's blood glucose feature vector reflects the patient's blood glucose fluctuations, stability and possible high and low blood glucose events in different time periods. These vectors are generated by advanced time series analysis technology, which can reveal the pattern and trend of the patient's blood glucose control and provide an important basis for assessing the severity and risk of diabetes. Based on these feature vectors, the doctor's health status evaluation report can be combined with clinical guidelines and the latest research results for in-depth analysis and comprehensive evaluation. The report can not only evaluate the patient's current condition and treatment effect, but also predict potential health risks and the possibility of complications. For example, by analyzing the blood glucose feature vector, the report can provide suggestions on improving the treatment plan or adjusting the medication to ensure that the patient's blood glucose is controlled within a safe range.

[0045] Figure 4 FIG. 1 is a block diagram of a doctor-side evaluation report generation module in a cloud platform-based telemedicine management system according to an embodiment of the present application. Figure 4 As shown, in a specific embodiment of the present application, the doctor-side evaluation report generation module 130 includes: a patient health status feature fusion unit 131, which is used to fuse the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector to obtain a patient health status feature vector; a patient health status feature optimization unit 132, which is used to perform back-propagation correction on the patient health status feature vector based on weight feature information analysis to obtain an optimized patient health status feature vector; a doctor-side patient report generation unit 133, which is used to pass the optimized patient health status feature vector through a generator to generate a doctor-side patient health status evaluation report.

[0046] It should be understood that by fusing the multi-dimensional feature vector of diabetic patient information and the patient's blood glucose feature vector, a more comprehensive and comprehensive feature vector of the patient's health status can be obtained. This fusion not only combines the patient's physiological characteristics and medical history, but also reveals the relationship between the two and potential influencing factors. For example, the patient's blood glucose control may be affected by multiple factors such as age, weight, and medication history. The fused feature vector can understand these complex relationships at a higher level. The generated patient health status feature vector can provide an important basis for medical decision-making. The medical team can perform accurate health risk assessment, disease prediction, and personalized treatment plan design based on these feature vectors. Through intelligent data analysis and report generation, doctors can more effectively optimize patients' treatment strategies, improve treatment effects and patients' quality of life.

[0047] In particular, in the technical solution of the present application, if the collection standards or frequencies of basic information, medical advice information and blood sugar values ​​are inconsistent, data inconsistency may result, and outliers may be introduced. The semantic encoding model of basic information of diabetic patients and the contextual semantic encoder of medical advice information may not fully understand all relevant information, especially for complex or ambiguous medical terms. When arranging multiple feature vectors into a two-dimensional feature matrix, if the arrangement method is inappropriate, information overlap or conflict may occur. The convolutional neural network as a feature encoder may not be able to capture all relevant features, especially when processing unstructured text data. The word embedding model of medical advice information may be limited by its expressive ability and cannot fully capture all key features of medical advice information. When associating the basic information semantic understanding feature vector and the medical advice information semantic understanding feature vector, if the association method is inappropriate, it may result in abnormal distribution of multi-dimensional feature vectors. If there is a local abnormal distribution of the patient's health status feature vector, the generator may not be able to correctly learn the feature distribution of the normal health state during the training process, resulting in a bias towards outliers when generating a health status assessment report, that is, category outlier offset. Category outlier offset will affect the prediction accuracy and generalization ability of the generator, so that the generated assessment report may not accurately reflect the actual health status of the patient. In order to solve this problem, the patient health status feature vector is subjected to back-propagation correction based on weight feature information analysis to obtain an optimized patient health status feature vector.

[0048] Among them, the patient health state feature vector is subjected to back-propagation correction based on weight feature information analysis to obtain an optimized patient health state feature vector, including: determining a decoding weight matrix of a decoder; performing matrix decoupling on the decoding weight matrix in units of row vectors to obtain a set of decoding weight vectors; calculating a patient health state metric learning loss information item between each decoding weight vector in the set of decoding weight vectors and the patient health state feature vector to obtain a patient health state metric learning compensation representation vector; inputting the patient health state metric learning compensation representation vector into a Softmax activation function to obtain a probabilistic patient health state metric learning compensation weight vector; calculating the positional point multiplication of the probabilistic patient health state metric learning compensation weight vector and the patient health state feature vector to obtain a point product feature vector, and calculating the weighted sum of the point product feature vector and the patient health state feature vector to obtain the optimized patient health state feature vector.

[0049] The patient health state metric learning loss information item between each decoding weight vector in the set of decoding weight vectors and the patient health state feature vector is calculated by the following optimization formula to obtain a patient health state metric learning compensation representation vector;

[0050] Wherein, the optimization formula is:

[0051]

[0052] in, represents the eigenvalue of the i-th position of the j-th decoding weight vector in the set of decoding weight vectors, represents the eigenvalue of the ith position of the patient's health status eigenvector, L represents the length of the eigenvector, and D j Represents the eigenvalue of the j-th position of the patient health status metric learning compensation representation vector.

[0053] That is, the patient health state feature vector is subjected to back propagation correction based on weight feature information analysis, which first determines the decoding weight matrix of the decoder to understand the decoder distinguishing different categories of features. Subsequently, the decoding weight matrix is ​​decoupled by row vectors to obtain a set of decoding weight vectors to allow the model to independently analyze the weight of each category and provide more refined control for feature compensation. Next, the patient health state metric learning loss information item is calculated. This process involves quantitative analysis of the interaction between each category weight vector and the patient health state feature vector. Through the calculation of a specific loss function, the inconsistency between the feature vector and the category weight is captured, providing a quantitative indicator for compensation. Next, the patient health state metric learning compensation representation vector is input into the activation function, nonlinearly transformed, and converted into a probabilistic form. The application of the activation function introduces nonlinearity, enhances the model's ability to interpret the compensation weight, and makes the compensation weight more consistent with the actual distribution characteristics. Then, the patient health state metric learning compensation representation vector processed by the activation function is used to calculate the probabilistic patient health state metric learning compensation weight vector. This step provides probabilistic guidance for the adjustment of the feature vector in a probabilistic form, making the optimization process more consistent with statistical characteristics, thereby enhancing the accuracy of the model's feature adjustment. Finally, by calculating the dot product of the probabilistic patient health status metric learning compensation weight vector and the patient health status feature vector, and then calculating the weighted sum with the original patient health status feature vector, the optimized patient health status feature vector is obtained. This step combines the original feature information and compensation information to generate an adjusted feature vector, reducing the impact of abnormal distribution and improving the accuracy of decoding.

[0054] Furthermore, the optimized patient health status feature vector reflects the patient's multi-dimensional physiology, medical history, lifestyle and other key information. The optimized patient health status feature vector is obtained through systematic data collection and analysis, including but not limited to the patient's basic information, physiological indicators, disease development history, medication records and any potential health risk factors. The health status assessment report on the doctor's side is generated by the generator, which effectively integrates and presents the patient's complex data. Based on advanced algorithms and models, the generator can automatically analyze and interpret the data in the feature vector and provide a clear and easy-to-understand report format. This automated report generation method not only saves doctors' tedious manual analysis time, but also greatly improves the efficiency and accuracy of the analysis. The health status assessment report on the doctor's side is not only a summary and display of data, but also a comprehensive analysis based on medical knowledge and clinical experience. The report will include key health indicators, trend analysis, risk assessment and treatment recommendations. With the help of these reports, doctors can quickly identify patients' health problems and treatment responses, adjust treatment plans in a targeted manner, and take necessary medical measures in a timely manner. In this way, the generator not only supports the implementation of personalized medicine, but also promotes the optimization of long-term health management strategies. Doctors can develop customized treatment plans and preventive measures based on each patient's unique health status report, thereby maximizing treatment effectiveness and patient quality of life.

[0055] In summary, the embodiment of the present application first obtains the basic information of diabetic patients collected by the database, the medical order information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by the portable device, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally passes the generator to generate a patient health status assessment report on the doctor side, thereby helping doctors to quickly understand the patient's current health status and treatment effect, facilitate timely adjustment of treatment plans, realize personalized medical care, and promote continuous optimization of long-term health management.

[0056] As described above, the cloud-based telemedicine management system 100 according to the embodiment of the present application can be implemented in various terminal devices. In one example, the cloud-based telemedicine management system 100 can be integrated into the terminal device as a software module and / or a hardware module. For example, the cloud-based telemedicine management system 100 can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the cloud-based telemedicine management system 100 can also be one of the many hardware modules of the terminal device.

[0057] Alternatively, in another example, the cloud-based telemedicine management system 100 and the terminal device may be separate devices, and the cloud-based telemedicine management system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0058] Figure 5 FIG. 1 is a flow chart of a remote medical management method based on a cloud platform according to an embodiment of the present application. Figure 5 As shown, according to the cloud platform-based remote medical management method of the embodiment of the present application, it includes: S110, acquiring basic information of diabetic patients collected by a database, medical advice information of diabetic patients collected by a database, and blood glucose values ​​of diabetic patients at multiple predetermined time points collected by a portable device; S120, extracting a multi-dimensional feature vector of diabetic patient information and a patient blood glucose feature vector from the basic information of diabetic patients collected by the database, the medical advice information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by the portable device; S130, generating a doctor-side patient health status assessment report based on the multi-dimensional feature vector of diabetic patient information and the patient blood glucose feature vector.

[0059] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned cloud platform-based remote medical management method have been referred to above. Figures 1 to 4 The description of the cloud platform-based telemedicine management system has been introduced in detail, and therefore, its repeated description will be omitted.

[0060] Below, reference Figure 6 To describe an electronic device according to an embodiment of the present application.

[0061] Figure 6 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0062] like Figure 6 As shown, the electronic device 10 includes an input device 11, an input interface 12, a central processing unit 13, a memory 14, an output interface 15, an output device 16, and a bus 17. The input interface 12, the central processing unit 13, the memory 14, and the output interface 15 are connected to each other through the bus 17, and the input device 11 and the output device 16 are connected to the bus 17 through the input interface 14 and the output interface 15, respectively, and then connected to other components of the electronic device 10.

[0063] Specifically, the input device 11 receives input information from the outside and transmits the input information to the central processing unit 13 through the input interface 12; the central processing unit 13 processes the input information based on the computer executable instructions stored in the memory 14 to generate output information, temporarily or permanently stores the output information in the memory 14, and then transmits the output information to the output device 16 through the output interface 15; the output device 16 outputs the output information to the outside of the electronic device 10 for user use.

[0064] In one embodiment, Figure 6 The electronic device 10 shown can be implemented as a network device, which may include: a memory configured to store programs; a processor configured to run the programs stored in the memory to execute any one of the cloud platform-based remote medical management methods described in the above embodiments.

[0065] According to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network, and / or installed from a removable storage medium.

[0066] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0067] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present application, but the present application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of the present application, and these modifications and improvements are also considered to be within the scope of protection of the present application.

Claims

1. A remote medical management system based on a cloud platform, characterized in that: include: A diabetic patient medical data acquisition module, used to acquire basic information of diabetic patients collected from a database, medical advice information of diabetic patients collected from a database, and blood sugar values ​​of diabetic patients at multiple predetermined time points collected from a portable device; A diabetic patient medical data extraction module, used to extract a diabetic patient information multi-dimensional feature vector and a patient blood glucose feature vector from the diabetic patient basic information collected by the database, the diabetic patient medical advice information collected by the database, and the diabetic patient blood glucose values ​​at multiple predetermined time points collected by the portable device; The doctor-side evaluation report generation module is used to generate a doctor-side patient health status evaluation report based on the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector.

2. The cloud-based remote medical management system according to claim 1, characterized in that: The diabetic patient medical data extraction module comprises: A patient basic information feature extraction unit, used for extracting features from the diabetic patient basic information collected from the database to obtain a semantic understanding feature vector of the diabetic patient basic information; A patient medical advice information feature extraction unit, used for extracting features from the medical advice information of the diabetic patient collected from the database to obtain a semantic understanding feature vector of the medical advice information of the diabetic patient; A patient information multi-dimensional feature association unit, used for associating the semantic understanding feature vector of the diabetic patient basic information with the semantic understanding feature vector of the diabetic patient medical advice information to obtain the diabetic patient information multi-dimensional feature vector; The patient blood glucose value feature extraction unit is used to extract features from the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device to obtain the patient blood glucose feature vector.

3. The cloud-based remote medical management system according to claim 2, characterized in that: The patient basic information feature extraction unit comprises: A patient basic information semantic coding subunit, used for obtaining a plurality of diabetic patient basic information feature vectors by applying the diabetic patient basic information semantic coding model to the diabetic patient basic information collected from the database; The patient basic information feature encoding subunit is used to perform feature encoding on the multiple diabetic patient basic information feature vectors to obtain the diabetic patient basic information semantic understanding feature vector.

4. The cloud-based remote medical management system according to claim 3, characterized in that: The patient basic information feature encoding subunit includes: Arranging the plurality of diabetic patients' basic information feature vectors according to the sample dimension to obtain a two-dimensional feature matrix of diabetic patients' basic information; The two-dimensional feature matrix of the basic information of diabetic patients is passed through a convolutional neural network of the basic information of diabetic patients as a feature encoder to obtain a semantic understanding feature vector of the basic information of diabetic patients.

5. The cloud-based remote medical management system according to claim 4, characterized in that: The patient medical advice information feature extraction unit comprises: Embedding the diabetes patient medical advice information collected from the database into a diabetes patient medical advice information word model to obtain a plurality of diabetes patient medical advice information data items; The plurality of diabetes patient medical order information data items are passed through a diabetes patient medical order information context semantic encoder to obtain the diabetes patient medical order information semantic understanding feature vector.

6. The cloud-based remote medical management system according to claim 5, characterized in that: The patient blood glucose value feature extraction unit comprises: Arranging the blood glucose values ​​of the diabetic patient at a plurality of predetermined time points collected by the portable device according to the time dimension to obtain a patient blood glucose input vector; The patient blood glucose input vector is passed through a patient blood glucose temporal encoder comprising a fully connected layer and a one-dimensional encoder to obtain the patient blood glucose feature vector.

7. The cloud-based remote medical management system according to claim 6, characterized in that: The doctor-side evaluation report generation module includes: A patient health status feature fusion unit, used for fusing the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector to obtain a patient health status feature vector; A patient health status feature optimization unit, configured to perform back-propagation correction on the patient health status feature vector based on weight feature information analysis to obtain an optimized patient health status feature vector; The doctor-side patient report generating unit is used to pass the optimized patient health status feature vector through a generator to generate a doctor-side patient health status assessment report.

8. The cloud-based remote medical management system according to claim 7, characterized in that: The patient health status characteristic optimization unit comprises: Determine a decoding weight matrix of a decoder; Decoupling the decoding weight matrix in units of row vectors to obtain a set of decoding weight vectors; Calculate the patient health state metric learning loss information item between each decoding weight vector in the set of decoding weight vectors and the patient health state feature vector to obtain the patient health state metric learning compensation representation vector Inputting the patient health status metric learning compensation representation vector into an activation function to obtain a probabilistic patient health status metric learning compensation weight vector; Calculate the positional point multiplication of the probabilistic patient health state metric learning compensation weight vector and the patient health state feature vector to obtain a point product feature vector, and calculate the weighted sum of the point product feature vector and the patient health state feature vector to obtain the optimized patient health state feature vector.

9. A remote medical management method based on a cloud platform, characterized in that: include: Obtaining basic information of the diabetic patient collected from a database, medical advice information of the diabetic patient collected from a database, and blood sugar values ​​of the diabetic patient at multiple predetermined time points collected from a portable device; Extracting a multi-dimensional feature vector of diabetic patient information and a patient blood glucose feature vector from the basic information of diabetic patients collected by the database, the medical advice information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by the portable device; Based on the multi-dimensional feature vector of the diabetic patient information and the patient blood glucose feature vector, a doctor-side patient health status assessment report is generated.

10. The cloud platform-based remote medical management method according to claim 9, characterized in that: Extracting a multi-dimensional feature vector of diabetic patient information and a patient blood glucose feature vector from the basic information of diabetic patients collected by the database, the medical advice information of diabetic patients collected by the database, and the blood glucose values ​​of diabetic patients at multiple predetermined time points collected by the portable device, including: Performing feature extraction on the basic information of diabetic patients collected from the database to obtain a semantic understanding feature vector of the basic information of diabetic patients; Extracting features from the medical advice information of diabetic patients collected from the database to obtain a semantic understanding feature vector of the medical advice information of diabetic patients; Associating the semantic understanding feature vector of the basic information of the diabetic patient with the semantic understanding feature vector of the medical advice information of the diabetic patient to obtain the multi-dimensional feature vector of the diabetic patient information; Feature extraction is performed on the blood glucose values ​​of the diabetic patient at multiple predetermined time points collected by the portable device to obtain the patient's blood glucose feature vector.

Images