Federal learning-based medical health analysis bracelet control method and medical health analysis bracelet

Abstract

The invention provides a federated learning-based medical health analysis bracelet control method and a medical health analysis bracelet. The federated learning-based medical health analysis bracelet control method comprises the steps of collecting physiological data and activity information of a wearer to form first data information; performing local preprocessing on the first data information to form second data information; performing local training on the second data information by using a federal learning model and converting the second data information into third data information; the third data information is transmitted to a cloud server, and the cloud server uses a federal learning algorithm to carry out aggregation updating processing on the third data information to form a global model; and the cloud server analyzes the third data information by using the global model and feeds back a data analysis result. After the collected information is preprocessed, the accuracy and rationality of the data are ensured, the global model of the cloud analyzes and processes the data, and then the analysis result is transmitted to the bracelet, so that the accuracy of data analysis is ensured, and the detection effect of the bracelet on each index of the human body is improved.

Description

Technical Field

[0001] The present invention relates to the field of medical health monitoring, and in particular to a control method for a medical health analysis bracelet based on federated learning and a medical health analysis bracelet. Background Art

[0002] With the improvement of health awareness and the growth of demand for personal health management, wearable devices such as smart bracelets and watches have become important tools for monitoring personal health. Medical bracelets are wearable portable medical or health electronic devices that can sense, record, analyze, regulate, intervene and even treat diseases or maintain health status with software support. These devices are usually divided into two types: commercial consumer grade and professional medical grade. Commercial consumer grade devices may include daily-wear bracelets, watches, etc., while professional medical grade wearable medical devices can monitor human physiological characteristics such as blood pressure, blood sugar, heart rate, etc. in real time, and can be used for disease treatment.

[0003] Although existing medical-grade wristband technology can provide basic physiological data monitoring, such as heart rate and activity, its accuracy and data in-depth analysis capabilities are limited. These wristbands usually directly use sensors to collect data and perform preliminary analysis through integrated software. However, this method is often limited by the accuracy of sensors and the complexity of data processing algorithms, resulting in insufficient accuracy and reliability of monitoring data, making it difficult to meet the needs of precision medicine. Summary of the invention

[0004] In view of this, the present invention discloses a control method of a medical health analysis bracelet based on federated learning and a medical health analysis bracelet, providing an innovative solution to improve the accuracy of data processing and the depth of analysis while protecting user privacy, thereby providing users with more accurate health monitoring and analysis.

[0005] The present application discloses a control method for a medical health analysis bracelet based on federated learning, including: Used to collect physiological data and activity information of the wearer to form first data information; locally preprocessing the first data information to form second data information; Using a federated learning model to perform local training on the second data information and convert it into third data information; The third data information is transmitted to a cloud server, and the cloud server aggregates and updates the third data information using a federated learning algorithm to form a global model; The cloud server uses the global model to analyze the third data information and feeds back the data analysis results to the bracelet terminal.

[0006] Furthermore, The wristband is integrated with a variety of sensors; The sensors collect heart rate, blood oxygen saturation, steps, and sleep information to form physiological data; Collect information such as accelerometers through sensors to form activity information; The physiological data and the activity information are integrated to form the first data information.

[0007] Furthermore, the local preprocessing includes: The first data information is cleaned, denoised and standardized to filter out waste information, and converted into second data information that matches the federated learning model.

[0008] Furthermore, Setting a standard value range, integrating the first data information, and filtering out the first data information within the value range; The first filtered data information is standardized by filtering, data conversion, etc. to form second data information. The federated learning model performs local training on the second data in the bracelet and converts it into third data information.

[0009] Furthermore, Using the federated learning model, the trained model is shared and the original data is kept confidential; The third data information trained by the federated learning model is transmitted to the cloud server in a secure manner, and the cloud server is used to process the third data information.

[0010] Furthermore, The cloud server uses a federated learning algorithm, in conjunction with data security technologies such as homomorphic encryption and secure multi-party computing, to update and fuse the third data information to form a global model; Perform data analysis on the global model and output analysis results.

[0011] Furthermore, Using the global model trained by federated learning, comparison data is set and compared with the third-party data information received from the cloud server to predict potential health risks.

[0012] Furthermore, Feedback is provided to users on the analysis results of the federated learning model, and the federated learning model can be adjusted based on the feedback results.

[0013] The present application also discloses a medical health analysis bracelet, including the above-mentioned control method.

[0014] Compared with the prior art, the technical solution disclosed in this application has the following beneficial effects: By preprocessing the collected information, the accuracy and rationality of the data are guaranteed. The data is then analyzed and processed by the global model in the cloud, and the analysis results are then transmitted to the bracelet, ensuring the accuracy of the data analysis and improving the bracelet's detection effect on various human body indicators. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Schematic diagram of the control method. DETAILED DESCRIPTION

[0016] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. It should be noted that when a component is considered to be "connected" to another component, it can be directly connected to another component, or there may be a central component at the same time. Unless otherwise defined, all technical and scientific terms used in this article have the same meaning as those generally understood by technicians in the technical field of the present invention. It should also be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection, or a connection between the two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0017] It should also be noted that in the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0018] The present application discloses a control method for a medical health analysis bracelet based on federated learning, which can more accurately monitor and analyze various indicators of the human body.

[0019] First, the user wears the electronic bracelet on the body, preferably, on the wrist, and the bracelet is used to collect various physiological data and activity information of the human body.

[0020] The method includes: collecting physiological data and activity information of the wearer to form first data information; locally preprocessing the first data information to form second data information; Using a federated learning model to perform local training on the second data information and convert it into third data information; The third data information is transmitted to a cloud server, and the cloud server aggregates and updates the third data information using a federated learning algorithm to form a global model; The cloud server uses the global model to analyze the third-party data information and provides feedback on the data analysis results.

[0021] Specifically, after the bracelet is worn on a person's wrist, the bracelet can monitor various physiological data and activity information of the human body. The collected physiological data and activity information are temporarily defined as the first data information; then the bracelet processes the first data information according to its own processing module to obtain more accurate information, which is defined as the second data information. The bracelet is integrated with a federated learning model, which performs local training on the second data information and converts it into third data information that can be processed by a cloud server. After the cloud server collects the third data information, it integrates and fuses the information, uses a federated learning algorithm to analyze and process the data, and transmits the processing results to the user.

[0022] Through the above method, the acquired information can be accurately selected, and the data can be processed by cloud-based algorithms to improve the accuracy of the data, thereby enabling more accurate knowledge of various data indicators of the human body.

[0023] Furthermore, the wristband is integrated with a variety of sensors; Sensors are used to collect information such as heart rate, blood oxygen saturation, steps, and sleep to form physiological data; Collect information such as accelerometers through sensors to form activity information; The physiological data and the activity information are integrated to form the first data information.

[0024] Specifically, a variety of sensors can be set on the wristband to collect different data indicators of the human body. The type and quantity of the sensors are set according to the data information to be obtained. If any type of data is needed, the corresponding sensor is integrated on the wristband. This application does not limit it in detail.

[0025] Furthermore, the local preprocessing includes: The first data information is cleaned, denoised and standardized to filter out waste information, and converted into second data information that matches the federated learning model.

[0026] Specifically, after obtaining the first data information, it is necessary to perform precision operations on the first data information according to the data situation. Since the bracelet worn by the human body may be affected by the environment, the way the human body wears it, and some information with large errors may be obtained. The bracelet can intercept data that conforms to the actual situation based on the data information in the medical field. The subsequent steps only process and analyze the data within this range, and abandon other data content that does not conform to scientific basis, so as to ensure the accuracy of subsequent operations and thus improve the accuracy of analysis.

[0027] A standard value range is set, the first data information is integrated, and the first data information within the value range is screened out; the first data information after screening is standardized by filtering, data conversion, etc. to form second data information, and the federated learning model performs local training on the second data in the bracelet and converts it into third data information. The second data information is used for local training of the federated learning model to generate model parameters, and these parameters are securely transmitted to the cloud server as third data information.

[0028] Furthermore, the federated learning model is used to share the trained model and keep the original data confidential; the third data information trained by the federated learning model is transmitted to the cloud server in a secure manner, and the cloud server is used to process the third data information. After receiving the third data information from multiple users, the cloud server uses the federated learning algorithm to aggregate and update it to form a global model.

[0029] In this application, since the detected human body-related vital signs data needs to be uploaded to the cloud server, in order to ensure personal privacy, only the processed third data information is transmitted in this application, and the original data is kept confidential and stored by the storage module of the bracelet.

[0030] The cloud server uses a federated learning algorithm, in conjunction with data security technologies such as homomorphic encryption and secure multi-party computing, to update and fuse the third data information to form a global model; performs data analysis on the global model and outputs the analysis results.

[0031] During the application, Kaiyuan's data model was used and the data was encrypted to meet relevant requirements.

[0032] Using the global model trained by federated learning, comparison data is set and compared with the third-party data information received from the cloud server to predict potential health risks.

[0033] Specifically, the federated average (FedAvg) or weighted average algorithm.

[0034] The federated average algorithm mentioned in this application first adopts an adaptive privacy budget allocation based on data characteristics, dynamically adjusts the privacy budget according to the characteristics and distribution of the data, so as to maximize the availability of data while protecting privacy. Secondly, a privacy budget allocation method based on model correlation is adopted, which uses the model correlation difference to adaptively allocate privacy budgets to users and dynamically adjust the privacy budget to reduce the error of the noise mechanism and improve the model accuracy and convergence speed.

[0035] The federated learning model mentioned in this application is also known as federated learning, joint learning, and federated learning. Federated machine learning is a machine learning framework that can effectively help multiple institutions use data and conduct machine learning modeling while meeting the requirements of user privacy protection, data security, and government regulations.

[0036] After the information is analyzed and the results are obtained through the cloud model, the analysis results of the federated learning model are fed back to the user, and the federated learning model can be adjusted according to the feedback results to make it more in line with personal characteristics and improve the monitoring effect. In summary, in specific use, when the user wears the bracelet on the wrist, the bracelet begins to collect the user's physiological data and activity information through its integrated sensors. These data include but are not limited to heart rate, blood oxygen saturation, number of steps, sleep quality, etc., and these raw data are defined as first data information. The local processing module of the bracelet processes these first data information in real time, including data cleaning, denoising and standardization processing to filter out invalid data and improve the accuracy of the data. The processed data is defined as second data information. These second data information are then used for local training of the federated learning model to generate model parameters, which are securely transmitted to the cloud server as third data information. After the cloud server receives the third data information from multiple users, it uses the federated learning algorithm to aggregate and update to form a global model. The global model can provide more in-depth data analysis, such as health trend prediction, disease risk assessment, etc., and feed back these analysis results to the user's bracelet so that the user can understand their health status in real time and take corresponding health measures.

[0037] The present application also discloses a medical health analysis bracelet, including the above-mentioned control method.

[0038] The present invention can be implemented in various forms and modifications without departing from the broad spirit and scope of the present invention. The above-described embodiments are intended to illustrate the invention, but are not intended to limit the scope of the present invention.

Claims

1. A control method for a medical health analysis bracelet based on federated learning, characterized in that: include: Used to collect physiological data and activity information of the wearer to form first data information; locally preprocessing the first data information to form second data information; Using a federated learning model to perform local training on the second data information and convert it into third data information; The third data information is transmitted to a cloud server, and the cloud server aggregates and updates the third data information using a federated learning algorithm to form a global model; The cloud server uses the global model to analyze the third data information and feeds back the data analysis results to the bracelet terminal.

2. According to the control method of a medical health analysis bracelet based on federated learning according to claim 1, it is characterized in that: The wristband is integrated with a variety of sensors; The sensors collect heart rate, blood oxygen saturation, steps, and sleep information to form physiological data; Collect information such as accelerometers through sensors to form activity information; The physiological data and the activity information are integrated to form the first data information.

3. According to a control method of a medical health analysis bracelet based on federated learning according to claim 1, it is characterized in that: The local preprocessing includes: The first data information is cleaned, denoised and standardized to filter out waste information, and converted into second data information that matches the federated learning model.

4. According to the control method of a medical health analysis bracelet based on federated learning according to claim 3, it is characterized in that: Setting a standard value range, integrating the first data information, and filtering out the first data information within the value range; The first filtered data information is standardized by filtering, data conversion, etc. to form second data information. The federated learning model performs local training on the second data in the bracelet and converts it into third data information.

5. A control method for a medical health analysis bracelet based on federated learning according to claim 4, characterized in that: Using the federated learning model, the trained model is shared and the original data is kept confidential; The third data information trained by the federated learning model is transmitted to the cloud server in a secure manner, and the cloud server is used to process the third data information.

6. A control method for a medical health analysis bracelet based on federated learning according to claim 5, characterized in that: The cloud server uses a federated learning algorithm, in conjunction with data security technologies such as homomorphic encryption and secure multi-party computing, to update and fuse the third data information to form a global model; Perform data analysis on the global model and output analysis results.

7. A control method for a medical health analysis bracelet based on federated learning according to claim 6, characterized in that: Using the global model trained by federated learning, comparison data is set and compared with the third-party data information received from the cloud server to predict potential health risks.

8. The control method of a medical health analysis bracelet based on federated learning according to claim 7 is characterized in that: The federated learning algorithm includes: FedAvg or weighted average algorithm.

9. The control method of a medical health analysis bracelet based on federated learning according to claim 7 is characterized in that: Feedback is provided to users on the analysis results of the federated learning model, and the federated learning model can be adjusted based on the feedback results.

10. A medical health analysis bracelet, characterized in that: The method comprises the control method as claimed in any one of claims 1 to 9.

Images