Self-adaptive physical sign crowd portrait processing method and device, equipment and storage medium

Abstract

The invention discloses a self-adaptive physical sign crowd portrait processing method, device and equipment and a storage medium, and is applied to a first server, the method comprises the following steps: sending a first calling request to a second server, and receiving a first model sent by the second server in response to the first calling request; generating a second model corresponding to the first terminal based on the first model; a first downloading request sent by a first terminal is received, and the second model is sent to the first terminal in response to the first downloading request; wherein the first model comprises a plurality of first static characteristic parameters of healthy people and a plurality of corresponding first dynamic characteristic parameters, and the plurality of first dynamic characteristic parameters comprise a plurality of first gait characteristic parameters; the second model comprises a plurality of second static feature parameters and a plurality of corresponding second dynamic feature parameters of the user corresponding to the first terminal, and the plurality of second dynamic feature parameters comprise a plurality of second gait feature parameters.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data services, and particularly relates to an adaptive sign population portrait processing method and device, equipment and a storage medium. BACKGROUND

[0002] With the development of body area networks, the sensors carried by the human body area bring more real-time and comprehensive data for the digital twin of a person. In addition to being able to analyze the traditional static portrait of a population, the real-time dynamic portrait of a person can also be more accurately analyzed. However, the data collected by the body area network in real time has the characteristics of large data volume and dynamics, and the population user portrait in the related art basically faces traditional static signs and has a certain update cycle, and cannot accurately determine the dynamic sign health range under different motion states. SUMMARY

[0003] To solve the above technical problems, the present application provides an adaptive sign population portrait processing method, device, equipment and storage medium.

[0004] The present application provides an adaptive sign population portrait processing method, applied to a first server, comprising:

[0005] sending a first call request to a second server and receiving a first model sent by the second server in response to the first call request; wherein the first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of corresponding first dynamic characteristic parameters, and the plurality of first dynamic characteristic parameters comprises a plurality of first gait characteristic parameters;

[0006] generating a second model corresponding to a first terminal based on the first model; wherein the second model comprises a plurality of second static characteristic parameters of a user corresponding to the first terminal and a plurality of corresponding second dynamic characteristic parameters, and the plurality of second dynamic characteristic parameters comprises a plurality of second gait characteristic parameters;

[0007] receiving a first download request sent by a first terminal and sending the second model to the first terminal in response to the first download request.

[0008] The present application provides an adaptive sign population portrait processing method, applied to a second server, comprising:

[0009] Determine a first model based on static feature data of a healthy population and corresponding dynamic feature data, wherein the static feature data includes a plurality of initial static feature parameters, the dynamic feature data includes a plurality of initial dynamic feature parameters, and the plurality of initial dynamic feature parameters includes a plurality of initial gait feature parameters; the first model includes a plurality of first static feature parameters of the healthy population and corresponding a plurality of first dynamic feature parameters, and the plurality of first dynamic feature parameters includes a plurality of first gait feature parameters;

[0010] Receive a first calling request sent by a first server, and send the first model to the first server in response to the first calling request.

[0011] Embodiments of the present application provide a self-adaptive population portrait processing method, applied to a first terminal, including:

[0012] Send a first download request to a first server, and receive a second model sent by the first server in response to the first download request; wherein the second model is a model corresponding to the first terminal generated by the first server according to the first model; the first model is obtained by the first server requesting calling from a second server; the first model includes a plurality of first static feature parameters of a healthy population and corresponding a plurality of first dynamic feature parameters, and the plurality of first dynamic feature parameters includes a plurality of first gait feature parameters; the second model includes a plurality of second static feature parameters of a user corresponding to the first terminal and corresponding a plurality of second dynamic feature parameters, and the plurality of second dynamic feature parameters includes a plurality of second gait feature parameters;

[0013] Generate a third model corresponding to the first terminal based on the second model; wherein the third model includes a plurality of third static feature parameters of the user corresponding to the first terminal after correction and corresponding a plurality of third dynamic feature parameters, and the plurality of third dynamic feature parameters includes a plurality of third gait feature parameters;

[0014] Receive a second download request sent by a second terminal, and send the third model to the second terminal in response to the second download request.

[0015] Embodiments of the present application provide a self-adaptive population portrait processing method, applied to a second terminal, including:

[0016] sending a second download request to the first terminal, and receiving a third model sent by the first terminal in response to the second download request; wherein the third model is a model corresponding to the second terminal and generated by the first terminal based on a second model; the second model is a model corresponding to the first terminal and generated by a first server; the first model is obtained by the first server calling a second server; the third model includes a plurality of third static characteristic parameters and a plurality of third dynamic characteristic parameters after being corrected by a user corresponding to the first terminal, and the plurality of third dynamic characteristic parameters include a plurality of third gait characteristic parameters;

[0017] generating a fourth model corresponding to the second terminal based on the third model; wherein the fourth model includes a plurality of fourth static characteristic parameters and a plurality of fourth dynamic characteristic parameters after being corrected by a user corresponding to the second terminal, and the plurality of fourth dynamic characteristic parameters include a plurality of fourth gait characteristic parameters.

[0018] Embodiments of the present application provide a self-adaptive physical population portrait processing device, applied to a first server, comprising:

[0019] a first sending unit configured to send a first calling request to a second server, and receive a first model sent by the second server in response to the first calling request; wherein the first model includes a plurality of first static characteristic parameters and a plurality of first dynamic characteristic parameters of a healthy population, and the plurality of first dynamic characteristic parameters include a plurality of first gait characteristic parameters;

[0020] a first generating unit configured to generate a second model corresponding to a first terminal based on the first model; wherein the second model includes a plurality of second static characteristic parameters and a plurality of second dynamic characteristic parameters of a user corresponding to the first terminal, and the plurality of second dynamic characteristic parameters include a plurality of second gait characteristic parameters;

[0021] a first receiving unit configured to receive a first download request sent by the first terminal, and send the second model to the first terminal in response to the first download request.

[0022] Embodiments of the present application provide a self-adaptive physical population portrait processing device, applied to a second server, comprising:

[0023] The second determining unit is configured to determine a first model based on static feature data of a healthy population and corresponding dynamic feature data, wherein the static feature data comprises a plurality of initial static feature parameters, the dynamic feature data comprises a plurality of initial dynamic feature parameters, the plurality of initial dynamic feature parameters comprises a plurality of initial gait feature parameters; the first model comprises a plurality of first static feature parameters and corresponding a plurality of first dynamic feature parameters of the healthy population, and the plurality of first dynamic feature parameters comprises a plurality of first gait feature parameters.

[0024] The second receiving unit is configured to receive a first calling request sent by the first server, and send the first model to the first server in response to the first calling request.

[0025] The embodiment of the present application provides a kind of adaptive sign population portrait processing device, applied to first terminal, comprising:

[0026] The third sending unit is configured to send a first download request to the first server, and receive a second model sent by the first server in response to the first download request;Wherein the second model is the first model generated by the first server corresponding to the first terminal model;The first model is obtained by the first server requesting calling to the second server;The first model comprises a plurality of first static feature parameters and corresponding a plurality of first dynamic feature parameters of a healthy population, and the plurality of first dynamic feature parameters comprises a plurality of first gait feature parameters;The second model comprises a plurality of second static feature parameters and corresponding a plurality of second dynamic feature parameters of the user corresponding to the first terminal, and the plurality of second dynamic feature parameters comprises a plurality of second gait feature parameters;

[0027] The third generating unit is configured to generate a third model corresponding to the first terminal based on the second model;Wherein the third model comprises a plurality of third static feature parameters and corresponding a plurality of third dynamic feature parameters of the user corresponding to the first terminal after correction, and the plurality of third dynamic feature parameters comprises a plurality of third gait feature parameters;

[0028] The third receiving unit is configured to receive a second download request sent by the second terminal, and send the third model to the second terminal in response to the second download request.

[0029] The embodiment of the present application provides a kind of adaptive sign population portrait processing device, applied to second terminal, comprising:

[0030] a fourth sending unit, configured to send a second download request to the first terminal, and receive a third model sent by the first terminal in response to the second download request; the third model is a model corresponding to the second terminal and generated by the first terminal based on a second model; the second model is a model corresponding to the first terminal and generated by a first server; the first model is obtained by the first server requesting and calling a second server; the third model comprises a plurality of third static characteristic parameters and a plurality of third dynamic characteristic parameters after being corrected by a user corresponding to the first terminal, and the plurality of third dynamic characteristic parameters comprise a plurality of third gait characteristic parameters;

[0031] a fourth generating unit, configured to generate a fourth model corresponding to the second terminal based on the third model; the fourth model comprises a plurality of fourth static characteristic parameters and a plurality of fourth dynamic characteristic parameters after being corrected by a user corresponding to the second terminal, and the plurality of fourth dynamic characteristic parameters comprise a plurality of fourth gait characteristic parameters.

[0032] The processing device provided by the embodiments of the present application comprises a processor and a memory, the memory is configured to store a computer program, and the processor is configured to invoke and run the computer program stored in the memory to execute any one of the adaptive body sign population portrait processing methods.

[0033] The computer readable storage medium provided by the embodiments of the present application is configured to store a computer program, and the computer program enables a computer to execute any one of the adaptive body sign population portrait processing methods.

[0034] The computer program product provided by the embodiments of the present application comprises computer program instructions, and the computer program instructions enable a computer to execute any one of the adaptive body sign population portrait processing methods.

[0035] In the technical solution of the embodiment of the application, the first server sends a first calling request to the second server, receives a first model sent by the second server in response to the first calling request, then the first server generates a second model corresponding to the first terminal based on the first model, receives a first downloading request sent by the first terminal, and sends the second model to the first terminal in response to the first downloading request, wherein the first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of first dynamic characteristic parameters corresponding thereto, the plurality of first dynamic characteristic parameters comprise a plurality of first gait characteristic parameters, the second model comprises a plurality of second static characteristic parameters of a user corresponding to the first terminal and a plurality of second dynamic characteristic parameters corresponding thereto, and the plurality of second dynamic characteristic parameters comprise a plurality of second gait characteristic parameters. In this way, the cloud server can adapt the body area network convergence center through the clustering model of the cloud clustering middleware, send the adapted clustering model to the body area network convergence center, and enable the body area network convergence center to correct errors based on the dynamic and static parameters corresponding thereto through the adapted clustering model, so as to form a self-adaptive individual twin image model, thereby enabling the self-adaptive individual twin image model to accurately determine the dynamic sign health range of the user in different motion states. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is an architecture schematic diagram of a body area network provided by the embodiment of the application;

[0037] Figure 2 is a system link schematic diagram of a clustering capability facing a hospital provided by the embodiment of the application;

[0038] Figure 3 is a process schematic diagram of calculating gait parameters through a foot micro long standby sensor provided by the embodiment of the application;

[0039] Figure 4 is a distribution schematic diagram of gait parameters calculated through a foot micro long standby sensor provided by the embodiment of the application;

[0040] Figure 5 is a display schematic diagram of a certain type of static user image of a sign provided by the embodiment of the application;

[0041] Figure 6 is a flow schematic diagram of an adaptive sign population image processing method applied to a first server provided by the embodiment of the application;

[0042] Figure 7 is a flow schematic diagram of an adaptive sign population image processing method applied to a second server provided by the embodiment of the application;

[0043] Figure 8is a flowchart of an adaptive sign population portrait processing method applied to a first terminal provided by an embodiment of the present application;

[0044] Figure 9 is a flowchart of an adaptive sign population portrait processing method applied to a second terminal provided by an embodiment of the present application;

[0045] Figure 10 is a schematic diagram of a user performing a motor function assessment provided by an embodiment of the present application;

[0046] Figure 11 is a principle schematic diagram of a spatial posture solving algorithm provided by an embodiment of the present application;

[0047] Figure 12 is a schematic diagram of gait cycle division and zero speed interval updating provided by an embodiment of the present application;

[0048] Figure 13 is a schematic diagram of a user straight walking trajectory analysis provided by an embodiment of the present application;

[0049] Figure 14 is a schematic diagram of geometric modeling of a foot in a stage one push-off phase provided by an embodiment of the present application;

[0050] Figure 15 is a schematic diagram of geometric modeling of a foot in a stage three landing phase provided by an embodiment of the present application;

[0051] Figure 16 is a detailed flowchart of an adaptive sign population portrait processing method provided by an embodiment of the present application;

[0052] Figure 17 is a structural schematic diagram of an adaptive sign population portrait processing device applied to a first server provided by an embodiment of the present application;

[0053] Figure 18 is a structural schematic diagram of an adaptive sign population portrait processing device applied to a second server provided by an embodiment of the present application;

[0054] Figure 19 is a structural schematic diagram of an adaptive sign population portrait processing device applied to a first terminal provided by an embodiment of the present application;

[0055] Figure 20 is a structural schematic diagram of an adaptive sign population portrait processing device applied to a second terminal provided by an embodiment of the present application;

[0056] Figure 21 is a structural schematic diagram of a processing device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0057] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described hereinafter. Obviously, the described embodiments are only some of the embodiments of the present application, but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0058] In the description of the embodiments of the present application, the term "corresponding" can represent a direct or indirect corresponding relationship between the two, can also represent an associated relationship between the two, or can be indicative of the relationship between the indicated and the configured.

[0059] In order to facilitate the understanding of the technical solutions of the embodiments of the present application, the related technologies of the embodiments of the present application are described as follows. The following related technologies can be combined with the technical solutions of the embodiments of the present application in any way, and all of them belong to the protection scope of the embodiments of the present application.

[0060] A body area network is a short-range, low-power, high-rate wireless communication network, mainly used to connect sensors and medical devices carried on the body. With the development of body area networks, the sensors carried on the human body provide more real-time and comprehensive data for digital twins. In addition to analyzing the traditional static portrait of the crowd (relatively stable age, height and other behavior habits within a period of time), it can also more accurately analyze the real-time dynamic portrait of the person (fine signs under different motion states). Figure 1 As shown in the figure, the body area network is an application network composed of a body area soft gateway, a terminal connected to the body area soft gateway, and a core node. The body area soft gateway is an important part of the body area network, mainly responsible for processing and forwarding data from connected terminal devices. The terminal connected to the body area soft gateway is usually various sensors and medical devices carried on the body, such as heart rate detectors, blood pressure meters, and blood glucose meters. These terminal devices are connected to the soft gateway through wireless communication to realize data transmission and exchange. The core node refers to a node with core functions in the body area network, usually including the body area soft gateway and some other key devices. The core node is responsible for processing data streams in the data, realizing data routing, protocol conversion, data encryption, etc., to ensure the normal operation of the body area network. In some application scenarios, the core node can also be interconnected with external networks (such as hospital information systems, home networks, etc.) to realize data exchange and information sharing.

[0061] However, with the generation of massive data, while bringing the accuracy of human data twins, it also brings new load and challenges to the hospital system relying on the body area network, such as Figure 2The system link diagram shown is the hospital-oriented clustering capability, and the adaptive open clustering middleware based on the daily home sign population portrait can realize the health detection function of the user at home. The doctor can remotely analyze the daily home sign data of the user to quickly understand the health status of the user and provide personalized medical advice for the user.

[0062] Among them, the massive data collected by the body area network in real time has the characteristics of large quantity and dynamics. Taking gait data as an example, through calibration, preprocessing, feature extraction, coordinate system conversion, and attitude solution from the sensor, 164 gait parameters need to be calculated, such as Figure 3 The process diagram shown is the process of calculating gait parameters by the foot micro long standby sensor, such as Figure 4 The distribution diagram shown is the gait parameters calculated by the foot micro long standby sensor, which is divided into four types of gait parameters in time domain, space domain, energy domain, and frequency domain. Among them, each large type of gait parameter is further divided into several small types. The dynamic indicators based on wearable systems are closely related to the scene, for example, the step length indicator is different in different motion scenes such as running, walking, and going up and down stairs. In the hospital, the doctor must measure the dynamic indicators according to different scenes, and re-measure the dynamic indicators that do not meet the requirements of the motion scene. The existing home information system lacks a method mechanism to determine the health range of dynamic parameters according to the scene. If the motion scene is not combined, it may bring dirty data, thereby affecting the calculation accuracy of the dynamic parameters.

[0063] As shown in Figure 5 The page diagram shown is a static user portrait in a hospital pilot. The existing hospital information system basically faces traditional static signs, including blood pressure, blood sugar, and CT physical and biochemical indicators, and has a certain update cycle. The user portrait middleware in a certain motion state may not be applicable to the next motion state. If it is directly applied, it will definitely affect the data accuracy, and at the same time, it will also cause redundant data to waste system resources.

[0064] To solve the above technical problems, the present application proposes an open integration method for adaptive sign population portrait based on a wearable system, which combines the data capabilities of future mobile operator networks to assist medical researchers in determining the health boundaries of dynamic signs, thereby taking into account the characteristics of large quantity and dynamics of dynamic parameters.

[0065] In order to facilitate the understanding of the technical solutions of the embodiments of the present application, the technical solutions of the present application are described in detail below through specific embodiments. The above related technologies can be combined with the technical solutions of the embodiments of the present application as optional schemes, which all belong to the protection scope of the embodiments of the present application. The embodiments of the present application include at least part of the following contents.

[0066] Embodiments of the present application provide a self-adaptive physical population portrait processing method, Figure 6 is a flowchart of a self-adaptive physical population portrait processing method applied to a first server provided by the embodiments of the present application, as Figure 6 shown, the method comprises the following steps:

[0067] Step 601: sending a first calling request to a second server, and receiving a first model sent by the second server in response to the first calling request.

[0068] The first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of first dynamic characteristic parameters corresponding thereto, and the plurality of first dynamic characteristic parameters comprises a plurality of first gait characteristic parameters.

[0069] In the embodiments of the present application, a user logs in the first server, inputs corresponding population static characteristics and dynamic characteristics (motion state) according to corresponding health demand information, and then sends a first calling request of a corresponding portrait to the second server through the first server, and receives a first model sent by the second server in response to the first calling request. The first model is generated by the second server based on the static portrait characteristics and the corresponding dynamic portrait characteristics of the healthy population, and the healthy population can include people who are healthy in static physical dimensions such as different genders, different ages, and different heights, and the embodiments of the present application do not make any limitation.

[0070] Here, the static characteristics, i.e. the stability characteristics, refer to certain characteristics in a person's appearance and physical condition that are not easy to change, such as age, gender, height, BMI, etc., the dynamic characteristics refer to the distinctive characteristics exhibited by a person in a dynamic situation, such as walking, running, eating, and speaking, and the gait characteristics refer to the characteristics of the size, direction, and action point of the force reflected in a person's footprint when walking, which is a reflection of the walking habits of the person in the stages of foot landing, foot lifting, and support swing.

[0071] Here, the first server is specifically a cloud application server, and the second server is specifically a cloud clustering middleware server.

[0072] Step 602: generating a second model corresponding to a first terminal based on the first model.

[0073] The second model comprises a plurality of second static characteristic parameters of a user corresponding to the first terminal and a plurality of second dynamic characteristic parameters corresponding thereto, and the plurality of second dynamic characteristic parameters comprises a plurality of second gait characteristic parameters.

[0074] In the embodiments of the present application, after the first server receives the first model sent by the second server in response to the first calling request, the first server can configure the first model according to the user features (dynamic and static features) of the first terminal (body area network convergence center) to generate a second model. The second model includes the static portrait features and the corresponding dynamic portrait features of the user corresponding to the first terminal.

[0075] Here, the process of configuring the first model according to the dynamic and static features of the user of the first terminal can specifically include: first collecting the static features and the dynamic features of the user of the first terminal, and adaptively processing and converting the collected user features, and then configuring the first model according to the processed user features, which includes but is not limited to adjusting the parameters of the first model, using feature enhancement technology, and optimizing the first model using the user features, thereby generating a second model.

[0076] Step 603: receiving the first download request sent by the first terminal, and sending the second model to the first terminal in response to the first download request.

[0077] In the embodiments of the present application, after the first server generates the second model corresponding to the first terminal based on the first model, the first terminal sends a first download request to the first server, the first server receives the first download request sent by the first terminal, and in response to the first download request, the first server sends the second model to the first terminal.

[0078] Based on this, in some embodiments, after the first terminal receives the second model sent by the first server in response to the first download request, the first terminal can generate a third model corresponding to the first terminal (human body area terminal) based on the second model. Specifically, the first terminal corrects the dynamic and static features in the second model based on the second model to obtain a third model. The third model includes the corrected static portrait features and the corresponding dynamic portrait features of the user corresponding to the first terminal.

[0079] Based on this, in some embodiments, after the first terminal corrects the dynamic and static features in the second model based on the second model to obtain a third model, the second terminal sends a second download request to the first terminal, the first terminal receives the second download request sent by the second terminal, and in response to the second download request, the first terminal sends the third model to the second terminal.

[0080] Based on this, in some embodiments, after the second terminal receives the third model sent by the first terminal in response to the second download request, the second terminal generates a fourth model corresponding to the second terminal based on the third model. The fourth model includes the corrected static portrait features and the corresponding dynamic portrait features of the user corresponding to the second terminal.

[0081] Based on this, in some embodiments, after the second terminal generates the fourth model corresponding to the second terminal based on the third model, the second terminal also needs to determine the healthy gait feature center value and the healthy gait feature boundary value, and send the healthy gait feature center value and the healthy gait feature boundary value to the first terminal.

[0082] Based on this, after the first terminal receives the healthy gait feature center value and the healthy gait feature boundary value sent by the second terminal, the first terminal also needs to determine the healthy gait feature value range of the user corresponding to the first terminal, and send the healthy gait feature value range to the first server.

[0083] Based on this, in some embodiments, after the first server receives the healthy gait feature value range sent by the first terminal, the first server further includes the following steps:

[0084] Step S1: Load non-healthy population data, and determine the non-healthy gait feature value range based on the non-healthy population data.

[0085] Step S2: Determine the distribution of the gait feature data of the user corresponding to the first terminal based on the healthy gait feature value range and the non-healthy gait feature value range.

[0086] Step S3: Determine the gait health score of the user corresponding to the first terminal based on the distribution.

[0087] Here, first, load the non-healthy population data, since the non-healthy population data includes the static feature parameters of the non-healthy population, therefore, based on the age parameter included in the static feature parameters of the non-healthy population data, the non-healthy gait feature value range for the user corresponding to the first terminal can be determined, and based on the healthy gait feature value range and the non-healthy gait feature value range, the distribution of the gait feature data of the user corresponding to the first terminal can be determined, so that based on the distribution, the gait health score of the user corresponding to the first terminal can be determined.

[0088] Based on this, in some embodiments, for determining the non-healthy gait feature value range based on the non-healthy population data, it can specifically include:

[0089] Based on the age parameter of the non-healthy population, extract the non-healthy population data with the same age as the user corresponding to the first terminal from the non-healthy population data;

[0090] Determine the non-healthy gait feature value range based on the non-healthy population data with the same age as the user corresponding to the first terminal.

[0091] Here, first, according to the age data in the non-healthy population data, the non-healthy population data corresponding to the same user age as the first terminal is screened out from the non-healthy population data, and then, based on the non-healthy population data corresponding to the same user age as the first terminal, and referring to the calculation formula of the gait feature mean value, standard error and 95% confidence interval in the process of determining the healthy gait feature central value and the healthy gait feature boundary value of the second terminal, the non-healthy gait feature numerical range of the user of the first terminal is determined.

[0092] Based on this, in some embodiments, for determining the distribution of the gait feature data of the user corresponding to the first terminal based on the healthy gait feature numerical range and the non-healthy gait feature numerical range, can specifically include:

[0093] If the gait feature data of the user corresponding to the first terminal is located in the healthy gait feature numerical range, it is determined that the user corresponding to the first terminal is in a healthy gait state; or,

[0094] If the gait feature data of the user corresponding to the first terminal is located in the non-healthy gait feature numerical range, it is determined that the user corresponding to the first terminal is in a non-healthy gait state; or,

[0095] If the gait feature data of the user corresponding to the first terminal is located in the intersection of the healthy gait feature numerical range and the non-healthy gait feature numerical range, it is determined that the user corresponding to the first terminal is in a healthy gait state and a non-healthy gait state at the same time.

[0096] Based on this, in some embodiments, for determining the gait health score of the user corresponding to the first terminal based on the distribution, can specifically include:

[0097] If the user corresponding to the first terminal is in a healthy gait state, the risk degree score of the user corresponding to the first terminal in the healthy gait is determined; or,

[0098] If the user corresponding to the first terminal is in a non-healthy gait state, the severity score of the user corresponding to the first terminal in the non-healthy gait is determined; or,

[0099] If the user corresponding to the first terminal is in a healthy gait state and a non-healthy gait state at the same time, the risk degree score of the user corresponding to the first terminal in the healthy gait and the severity score of the user corresponding to the first terminal in the non-healthy gait are determined.

[0100] Here, the distribution of the gait feature data of the user corresponding to the first terminal can be determined based on the non-healthy gait feature numerical range and the healthy gait feature numerical range of the user of the first terminal, and for each case, the gait health score of the user corresponding to the first terminal is determined, and for the distribution, there are three cases:

[0101]

[0101] (1) If the gait feature data of the user corresponding to the first terminal is located in the healthy gait feature value range, it is determined that the user corresponding to the first terminal is in a healthy gait state, and the gait health score (risk degree score) of the user in the healthy gait is calculated. The calculation formula of the risk degree score is:

[0102]

[0103] wherein Y is the gait feature data of the user corresponding to the first terminal, is the maximum value of the healthy gait feature value range, μ H is the center value of the healthy gait feature value range.

[0104] (2) If the gait feature data of the user corresponding to the first terminal is located in the non-healthy gait feature value range, it is determined that the user corresponding to the first terminal is in a non-healthy gait state, and the gait health score (severity score) of the user in the non-healthy gait is calculated. The calculation formula of the severity score is:

[0105]

[0106] wherein Y is the gait feature data of the user corresponding to the first terminal, is the maximum value of the non-healthy gait feature value range, μ D is the center value of the non-healthy gait feature value range.

[0107] (3) If the gait feature data of the user corresponding to the first terminal is located in the intersection of the healthy gait feature value range and the non-healthy gait feature value range, it is determined that the user corresponding to the first terminal is in a healthy gait state and a non-healthy gait state at the same time. At this time, the gait health score (risk degree score) of the user in the healthy gait and the gait health score (severity score) of the user in the non-healthy gait are calculated at the same time, so as to realize more comprehensive evaluation of the gait health condition of the user.

[0108] In the technical scheme provided by the embodiments of the present application, the first server sends a first calling request to the second server, receives a first model sent by the second server in response to the first calling request, then the first server generates a second model corresponding to a first terminal based on the first model, receives a first downloading request sent by the first terminal, and sends the second model to the first terminal in response to the first downloading request, wherein the first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of first dynamic characteristic parameters corresponding thereto, the plurality of first dynamic characteristic parameters comprise a plurality of first gait characteristic parameters, the second model comprises a plurality of second static characteristic parameters of a user corresponding to the first terminal and a plurality of second dynamic characteristic parameters corresponding thereto, and the plurality of second dynamic characteristic parameters comprise a plurality of second gait characteristic parameters. In this way, the cloud server can call a clustering model of a cloud clustering middleware, adapt the body area network convergence center through the clustering model, and send the adapted clustering model to the body area network convergence center, so that the body area network convergence center can correct errors based on the dynamic and static parameters corresponding thereto through the adapted clustering model to form a self-adaptive individual twin image model, thereby accurately determining the dynamic sign health range of the user in different motion states through the self-adaptive individual twin image model.

[0109] The embodiments of the present application also provide a self-adaptive sign population image processing method, Figure 7 is a flowchart of a self-adaptive sign population image processing method applied to a second server provided by the embodiments of the present application, as Figure 7 shown, the method comprises the following steps:

[0110] Step 701: determining a first model based on static characteristic data of a healthy population and corresponding dynamic characteristic data.

[0111] The static characteristic data comprises a plurality of initial static characteristic parameters, the dynamic characteristic data comprises a plurality of initial dynamic characteristic parameters, and the plurality of initial dynamic characteristic parameters comprise a plurality of initial gait characteristic parameters. The first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of first dynamic characteristic parameters corresponding thereto, and the plurality of first dynamic characteristic parameters comprise a plurality of first gait characteristic parameters.

[0112] In the embodiments of the present application, the second server extracts static feature data and corresponding dynamic feature data of the healthy population from the database, and pre-processes the extracted related data (including but not limited to data cleaning, removing outliers, etc.), then selects a first most relevant parameter and a first key gait feature parameter with a first correlation coefficient greater than a first threshold value with the first most relevant parameter from the processed related data, and selects a suitable machine learning algorithm to train and evaluate the model based on the first most relevant parameter and the first key gait feature parameter. If the effect evaluation coefficient of the model is greater than a preset evaluation threshold, the model is the first model. The second server is specifically a cloud clustering middleware server.

[0113] Based on this, in some embodiments, for the static feature data and corresponding dynamic feature data of the healthy population, determining the first model can specifically include:

[0114] Determining the first most relevant parameter based on the correlation degree between each initial static feature parameter and the plurality of initial gait feature parameters;

[0115] Determining the first key gait feature parameter based on the first most relevant parameter;

[0116] Determining the first model based on the first most relevant parameter and the first key gait feature parameter.

[0117] Here, first, the initial static feature parameters of the healthy population are traversed respectively, the first correlation coefficient between each initial static feature parameter and each initial gait feature parameter is calculated, then for each initial static feature parameter, the first correlation coefficients between each initial static feature parameter and all initial gait feature parameters are summarized, and the first correlation coefficient mean of each initial static feature parameter is calculated, so as to select the initial static feature parameter corresponding to the maximum first correlation coefficient mean as the first most relevant parameter. Under the first most relevant parameter, the initial gait feature parameter with a first correlation coefficient greater than a first threshold (for example, 0.3) with the first most relevant parameter is selected as the first key gait feature parameter. Then, based on the first most relevant parameter and the first key gait feature parameter, the clustering class number and the maximum iteration number corresponding to the data of the first most relevant parameter are set, a suitable machine learning algorithm is selected to train the model, and the healthy population with similar static features and similar gait features is clustered, thereby generating the first model.

[0118] Based on this, in some embodiments, for determining the first most relevant parameter based on the correlation degree between each initial static feature parameter and the plurality of initial gait feature parameters, can specifically include:

[0119] calculate a first correlation coefficient between each initial static feature parameter and the plurality of initial gait feature parameters respectively;

[0120] Based on the determined plurality of first correlation coefficients, determine a first correlation coefficient mean of each initial static feature parameter, and select the initial static feature parameter corresponding to the maximum first correlation coefficient mean as the first most relevant parameter.

[0121] Here, the formula for calculating the first correlation coefficient between the initial static feature parameter and the initial gait feature parameter is:

[0122]

[0123] wherein X i,k is the kth data of the ith initial static feature parameter, Y j,k is the kth data of the jth initial gait feature parameter, is the average value of the ith initial static feature parameter data, is the average value of the jth initial gait feature parameter data, and N is the number of data in the database, r i,j is the first correlation coefficient between the ith initial static feature parameter and the jth initial gait feature parameter.

[0124] Here, after all the first correlation coefficients are calculated, for each initial static feature parameter, the formula for calculating the first correlation coefficient mean of each initial static feature parameter is:

[0125]

[0126] wherein L is the number of initial gait feature parameters, is the first correlation coefficient mean of the ith initial static feature parameter.

[0127] Step 702: receiving the first call request sent by the first server, and sending the first model to the first server in response to the first call request.

[0128] In the embodiment of the present application, after the second server determines the first model based on the static feature data and the corresponding dynamic feature data of the healthy population, the first server sends a first call request to the second server, the second server receives the first call request sent by the first server, and sends the first model to the first server in response to the first call request, so that the second server generates the second model corresponding to the first terminal based on the first model.

[0129] The embodiment of the present application also provides a self-adaptive physical population portrait processing method, Figure 8 is a flowchart of the self-adaptive physical population portrait processing method applied to the first terminal provided by the embodiment of the present application, as shown inFigure 8 As shown, the method comprises the following steps:

[0130] Step 801: sending a first download request to a first server, and receiving a second model sent by the first server in response to the first download request.

[0131] The second model is a model corresponding to the first terminal generated by the first server based on a first model; the first model is obtained by the first server requesting and calling a second server; the first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of first dynamic characteristic parameters corresponding thereto, and the plurality of first dynamic characteristic parameters comprise a plurality of first gait characteristic parameters; the second model comprises a plurality of second static characteristic parameters of a user corresponding to the first terminal and a plurality of second dynamic characteristic parameters corresponding thereto, and the plurality of second dynamic characteristic parameters comprise a plurality of second gait characteristic parameters.

[0132] In the embodiments of the present application, after the first server generates the second model corresponding to the first terminal based on the first model, the first terminal sends a first download request to the first server, and the first server receives and responds to the first download request sent by the first terminal, so that the first terminal receives the second model sent by the first server in response to the first download request. The first server is specifically a cloud application server, the second server is specifically a cloud clustering middleware server, and the first terminal is specifically a human domain convergence center (such as a mobile terminal such as a mobile phone).

[0133] Step 802: generating a third model corresponding to the first terminal based on the second model.

[0134] The third model comprises a plurality of third static characteristic parameters of the user corresponding to the first terminal after correction and a plurality of third dynamic characteristic parameters corresponding thereto, and the plurality of third dynamic characteristic parameters comprise a plurality of third gait characteristic parameters.

[0135] In the embodiments of the present application, after the first terminal receives the second model sent by the first server in response to the first download request, the first terminal can generate an updated third model corresponding to the first terminal based on the second model by correcting the parameters in the second model.

[0136] Based on this, in some embodiments, for generating the third model corresponding to the first terminal based on the second model, it can specifically comprise:

[0137] determining a correction error based on the plurality of first gait characteristic parameters and the plurality of second gait characteristic parameters;

[0138] determining a correction error prediction value based on the correction error and a correction error model;

[0139] determining a loss value based on the correction error and the correction error prediction value;

[0140] updating the correction error model based on the loss value;

[0141] correcting the second model based on the updated correction error model to obtain a third model.

[0142] Here, the first terminal inputs corresponding demographic information static features (such as age, height, gender, BMI), downloads the corresponding second model of the first server, and takes the first model as an individual twin image template, calculates a first gait feature comprehensive parameter according to all gait feature data (first gait feature parameters) in the individual twin image template, then calculates a correction error according to the second gait feature parameters and the first gait feature comprehensive parameter, inputs the correction error into the deep learning error correction model to obtain a correction error prediction value, calculates a loss value according to the correction error and the correction error prediction value, then updates the correction error model according to the loss value, and corrects the second model according to the updated correction error model to obtain a third model.

[0143] Based on this, in some embodiments, for determining the correction error based on the plurality of first gait feature parameters and the plurality of second gait feature parameters, can specifically include:

[0144] determining a first gait feature comprehensive parameter based on the plurality of first gait feature parameters;

[0145] determining the correction error based on the first gait feature comprehensive parameter and the plurality of second gait feature parameters.

[0146] Here, the formula for calculating the first gait feature comprehensive parameter is:

[0147]

[0148] wherein c j,i is the i-th data of the j-th first gait feature parameter, and N is the number of first gait feature parameter data.

[0149] Here, the formula for calculating the correction error between the first gait feature comprehensive parameter and the plurality of second gait feature parameters is:

[0150]

[0151] wherein Δ j is the correction error between the j-th second gait feature parameter and the first gait feature comprehensive parameter.

[0152] Here, the formula for calculating the correction error prediction value is:

[0153]

[0154] wherein xj is the jth second static feature parameter, F(·) is a deep learning error correction algorithm, is the correction error prediction value corresponding to the jth second gait feature parameter.

[0155] Here, the formula for calculating the loss value according to the correction error and the correction error prediction value is:

[0156]

[0157] wherein, represents the jth second gait feature parameter in the correction error prediction value, Δ j represents the real correction error corresponding thereto, K is the number of clustering clusters, N is the number of second gait feature parameters, and L is the final loss value.

[0158] Here, in the process of updating the correction error model using the loss value, if the loss value decreases, it indicates that the model is still converging, and the parameters of the correction error model are updated by the gradient descent method. In order to avoid too long model training time, the maximum training round is set to 1000, that is, the training is stopped after the maximum training round is reached.

[0159] If the loss value no longer decreases in 10 calculation processes, it indicates that the model can no longer continue to converge, and the model update is stopped:

[0160]

[0161] wherein, L t is the loss value of the current training calculation, is the average of the previous 10 loss values, and count is the continuous count of the loss value in the training process that no longer decreases, or the stop condition is triggered after 10.

[0162] In some embodiments, in the updating process of the correction error model and in the process of correcting the second model using the updated correction error model, the amount of data of the newly added healthy population is recorded and saved to the database. When the amount of newly added data reaches 200, the model updating mechanism is started, the correction error model and the second model are retrained using the current entire data set to realize the optimization of the algorithm. With the increase of data amount, the population gait portrait is made more refined through iteration, and the prediction error of gait features is continuously reduced.

[0163] In some embodiments, after the first terminal generates the third model corresponding to the first terminal based on the second model, the health gait feature central value and the health gait feature boundary value sent by the second terminal can also be received, and the health gait feature value range of the user corresponding to the first terminal is determined based on the third model and the health gait feature boundary value, and the health gait feature value range is sent to the first server.

[0164] Step 803: receiving the second download request sent by the second terminal, and sending the third model to the second terminal in response to the second download request.

[0165] In the embodiments of the present application, after the first terminal generates the third model corresponding to the first terminal based on the second model, the second terminal sends a second download request for the third model to the first terminal, and the first terminal receives the second download request sent by the second terminal and sends the third model to the second terminal in response to the second download request.

[0166] In some embodiments, after the first terminal receives the second model sent by the first server in response to the first download request, the following steps can also be included:

[0167] Step Z1: determining a plurality of first weight coefficients based on a plurality of first static feature parameters and a plurality of first gait feature parameters.

[0168] Step Z2: determining a plurality of second weight coefficients based on the plurality of first gait feature parameters.

[0169] Step Z3: determining a plurality of fusion weight coefficients based on the plurality of first weight coefficients and the plurality of second weight coefficients.

[0170] Step Z4: determining a health risk comprehensive score based on the plurality of fusion weight coefficients.

[0171] Here, the first static feature parameter includes the age parameter of the healthy population, the second correlation coefficient between the age parameter of the healthy population and each first gait feature parameter is calculated respectively, and then the absolute value of each calculated second correlation coefficient is obtained to obtain a plurality of first weight coefficients; at the same time, based on the plurality of first gait feature parameters, a plurality of clustering cluster data under each first gait feature parameter is determined, and based on the similarity coefficient within each clustering cluster data in the plurality of clustering cluster data and the dissimilarity coefficient between each clustering cluster data, a plurality of second weight coefficients are determined; then based on the plurality of first weight coefficients and the plurality of second weight coefficients, a plurality of fusion weight coefficients are determined, so as to determine a health risk comprehensive score based on the plurality of fusion weight coefficients.

[0172] Based on this, in some embodiments, for determining a plurality of first weight coefficients based on a plurality of first static feature parameters and a plurality of first gait feature parameters, it can specifically include:

[0173] calculate a second correlation coefficient between the age parameter of the healthy population and each of the first gait characteristic parameters;

[0174] determine a plurality of first weight coefficients based on the determined plurality of second correlation coefficients.

[0175] Here, the formula for calculating the second correlation coefficient between the age parameter of the healthy population and the first gait characteristic parameters is:

[0176]

[0177] wherein A k is the kth data of the age parameter of the healthy population, B j,k is the kth data of the jth first gait characteristic parameter, is the average value of the age parameter data of the healthy population, is the average value of the jth first gait characteristic parameter, N is the number of data in the database, D j is the second correlation coefficient between the age parameter of the healthy population and the jth first gait characteristic parameter.

[0178] After calculating the second correlation coefficient between the age parameter of the healthy population and each of the first gait characteristic parameters, the absolute values of each of the calculated second correlation coefficients are sorted from small to large, and a plurality of first weight coefficients corresponding to the order are obtained, and the calculation formula is:

[0179] [D (1) ,D (2) ,…,D (M) ]=Sort([|D1|,|D2|,…,|D M |]) (11)

[0180]

[0181] wherein |D j | is the absolute value of the second correlation coefficient, because the value range of the correlation coefficient is [-1, 1], and the correlation coefficient less than zero indicates that the two data are negatively correlated, and the sign of the correlation is not considered here, only the degree of correlation is considered, therefore the absolute value of the correlation coefficient is used; Sort(·) represents sorting the input numerical sequence from small to large, and the result of the sorting is [D (1) ,D (2) ,…,D (M) ], wherein D (1) is the minimum value in the original sequence, D (M) is the maximum value, i.e. D (1) ≤D (2) ≤...≤D (M)M is the number of first gait feature parameters; is the first gait feature parameter corresponding to the first gait feature parameter at the (i)th position in the sorted correlation sequence, and the value is

[0182] Based on this, in some embodiments, based on the plurality of gait feature parameters, the plurality of secondary weight coefficients is determined, which can specifically include:

[0183] Determine a plurality of cluster data corresponding to each first gait feature parameter;

[0184] Based on the plurality of cluster data, determine the similarity coefficient within each cluster data and the dissimilarity coefficient between each cluster data;

[0185] Based on the plurality of similarity coefficients and the plurality of dissimilarity coefficients determined, determine a plurality of contour coefficients;

[0186] Based on the plurality of contour coefficients, determine a plurality of secondary weight coefficients.

[0187] Here, under the mth first gait feature parameter, P m clusters are obtained according to the clustering model training, and the cluster data corresponding to each cluster is obtained, and P m cluster centers and cluster data are traversed, the similarity coefficient in each cluster is calculated and averaged, and the calculation formula is:

[0188]

[0189] Where C m,i is the number of data in the ith cluster under the mth first gait feature parameter, B m,i,j is the jth data in the ith cluster under the mth first gait feature parameter, and a m is the overall average value of the cluster similarity coefficient of the mth first gait feature parameter calculated.

[0190] After calculating the average value of the cluster similarity coefficient of each first gait feature parameter, all first gait feature parameters are traversed, and the dissimilarity coefficient between the clusters in each first gait feature parameter is calculated, and the calculation formula is:

[0191]

[0192] Where C nearest represents the number of cluster data closest to the cluster to which the ith first gait feature parameter belongs, B m,i represents the ith data under the mth first gait feature parameter, B m,j represents the jth data in its adjacent cluster under the mth first gait feature parameter, and bm a calculated dissimilarity coefficient between clusters of the mth first gait feature parameter.

[0193] According to the dissimilarity coefficient between clustering clusters in each first gait feature parameter and the average of the similarity coefficient within the clustering cluster in each first gait feature parameter, a clustering silhouette coefficient is calculated, and the calculation formula is:

[0194]

[0195] wherein max(·) is the maximum value of the input data, s m is the silhouette coefficient under the mth first gait feature parameter.

[0196] After calculating the silhouette coefficient under each first gait feature parameter, the calculated silhouette coefficients are sorted from small to large, and a plurality of secondary weight coefficients corresponding to the order are obtained, and the calculation formula is:

[0197] [s (1) ,s (2) ,...,s (M) ]=Sort([r1,r2,…,r M ]) (16)

[0198]

[0199] wherein [s (1) ,s (2) ,...,s (M) ] is the sequence after sorting the silhouette coefficients of each first gait feature parameter from small to large, wherein s (1) is the minimum value in the original sequence, s (M) is the maximum value, that is, s (1) ≤s (2) ≤...≤s (M) ; is the secondary weight coefficient corresponding to the first gait feature parameter at the (i)th position in the sorted silhouette coefficient sequence, and the value is

[0200] Based on this, in some embodiments, for determining a health risk comprehensive score based on a plurality of fusion weight coefficients, can specifically include:

[0201] Converting and normalizing each fusion weight coefficient to obtain a processed each fusion weight coefficient;

[0202] Determining a health risk comprehensive score based on the processed each fusion weight coefficient.

[0203] Here, after the plurality of primary weight coefficients and the plurality of secondary weight coefficients are calculated, the primary weight coefficients and the secondary weight coefficients under each first gait feature parameter are fused to obtain a plurality of fused weight coefficients, and the calculation formula is:

[0204]

[0205] wherein, and respectively represent the primary weight coefficient and the secondary weight coefficient under the mth first gait feature parameter, is the fused weight coefficient calculated, and the value range is

[0206] Here, each fused weight coefficient calculated is in a direct proportional relationship with the degree of influence of age on the first gait feature parameter, in order to obtain the first gait feature parameter with low correlation with age, the direct proportional relationship needs to be converted into an inverse proportional relationship, and the fused weight coefficients are sorted, according to the value range of the fused weight coefficient, the fused weight coefficient is converted in a negative logarithm manner, and the calculation formula is:

[0207]

[0208] wherein, is the fused weight coefficient under the mth first gait feature parameter after logarithmic conversion, and the value range is [0, 2logM], at this time, the smaller the value of the fused weight coefficient, the smaller the influence of the age factor, and the larger the value, the greater the influence.

[0209] After the fused weight coefficient under each first gait feature parameter is logarithmically converted, the converted fused weight coefficient is sorted in descending order, and the converted fused weight coefficient is normalized to regulate its value range to the interval [0, 1], and the calculation formula is:

[0210]

[0211] wherein, 2logM is the maximum value of , and is the normalized fused weight coefficient under the mth first gait feature parameter.

[0212] After the converted fused weight coefficient is normalized, the fused weight coefficients (normalized fused weight coefficients) corresponding to all first gait feature parameters are added to obtain a health risk comprehensive score, and the calculation formula is:

[0213]

[0214] wherein, is the normalized fusion weight coefficient of the mth first gait feature parameter, Score disease is the calculated health risk comprehensive score.

[0215] The embodiment of the present application also provides a self-adaptive sign population portrait processing method, Figure 9 is a flowchart of the self-adaptive sign population portrait processing method applied to the second terminal provided by the embodiment of the present application, as Figure 9 indicated, the method comprises the following steps:

[0216] Step 901: sending a second download request to the first terminal, and receiving a third model sent by the first terminal in response to the second download request.

[0217] The third model is a model corresponding to the second terminal and generated by the first terminal based on a second model; the second model is a model corresponding to the first terminal and generated by the first server; the first model is obtained by requesting and calling the first server by the second server; the third model comprises a plurality of third static feature parameters and a plurality of third dynamic feature parameters corresponding to a user of the first terminal, and the plurality of third dynamic feature parameters comprise a plurality of third gait feature parameters.

[0218] In the embodiment of the present application, after the first terminal generates the third model corresponding to the first terminal based on the second model, the second terminal sends a second download request to the first terminal, the first terminal receives and responds to the second download request sent by the second terminal, so that the second terminal receives the third model sent by the first terminal in response to the second download request. Wherein, the first terminal is specifically a human domain convergence center (such as a mobile terminal such as a mobile phone), and the second terminal is specifically a human domain terminal collection terminal (such as a wearable device).

[0219] Step 902: generating a fourth model corresponding to the second terminal based on the third model.

[0220] The fourth model comprises a plurality of fourth static feature parameters and a plurality of fourth dynamic feature parameters corresponding to a user of the second terminal, and the plurality of fourth dynamic feature parameters comprise a plurality of fourth gait feature parameters.

[0221] In the embodiment of the present application, after the second terminal receives the third model sent by the first terminal in response to the second download request, the second terminal can generate the fourth model corresponding to the second terminal by modifying the third model.

[0222] Based on this, in some embodiments, for generating the fourth model corresponding to the second terminal based on the third model, it can specifically comprise:

[0223] determine a third most relevant parameter based on a correlation degree between each third dynamic characteristic parameter and the plurality of third gait characteristic parameters;

[0224] determine a third key gait characteristic parameter based on the third most relevant parameter;

[0225] correct the third model based on the third most relevant parameter and the third key gait characteristic parameter to obtain a fourth model.

[0226] Here, first, the third static characteristic parameters are traversed respectively, the third correlation coefficients between the third dynamic characteristic parameters corresponding to each third static characteristic parameter and each third gait characteristic parameter are calculated, then for each third dynamic characteristic parameter, the third correlation coefficients between each third dynamic characteristic parameter and all third gait characteristic parameters are summarized, and the third correlation coefficient mean of each third dynamic characteristic parameter is calculated to select the third dynamic characteristic parameter corresponding to the maximum third correlation coefficient mean as the third most relevant parameter. Under the third most relevant parameter, the third gait characteristic parameter with the third correlation coefficient greater than a third threshold (for example, 0.5) between the third most relevant parameter is selected as the third key gait characteristic parameter. Then, based on the third most relevant parameter and the third key gait characteristic parameter, the parameters in the third model are corrected and updated to obtain the fourth model.

[0227] In some embodiments, for determining the third most relevant parameter based on the correlation degree between each third dynamic characteristic parameter and the plurality of third gait characteristic parameters, can specifically include:

[0228] respectively calculate the third correlation coefficients between each third dynamic characteristic parameter and the plurality of third gait characteristic parameters;

[0229] determine the third correlation coefficient mean of each third dynamic characteristic parameter based on the determined plurality of third correlation coefficients, and select the third dynamic characteristic parameter corresponding to the maximum third correlation coefficient mean as the third most relevant parameter.

[0230] Here, the formula for calculating the third correlation coefficient between the third dynamic characteristic parameter and the third gait characteristic parameter is:

[0231]

[0232] wherein Z i,k is the kth data of the ith third dynamic characteristic parameter, M j,k is the kth data of the jth third gait characteristic parameter, is the average value of the ith third dynamic characteristic parameter data, is the average value of the jth third gait characteristic parameter data, N is the number of data in the database, R i,jThe third correlation coefficient between the i-th third dynamic feature parameter and the j-th third gait feature parameter.

[0233] Here, after calculating all the third correlation coefficients, for each third dynamic feature parameter, the formula for calculating the third correlation coefficient mean of each third dynamic feature parameter is:

[0234]

[0235] wherein L is the number of third gait feature parameters, The first correlation coefficient mean of the i-th initial static feature parameter.

[0236] In some embodiments, after the second terminal receives the third model sent by the first terminal in response to the second download request, the gait feature type of the user corresponding to the first terminal is determined based on the age parameter of the user corresponding to the first terminal; the first gait feature parameter of the same type as the gait feature type is selected from the plurality of first gait feature parameters of the healthy population, and the value range of the first gait feature parameter of the same type as the gait feature type is set in the first interval; the healthy gait feature center value and the healthy gait feature boundary value are determined based on the parameter value in the first interval; and the healthy gait feature center value and the healthy gait feature boundary value are sent to the first terminal.

[0237] Here, first, the gait feature parameter corresponding to the age parameter of the user corresponding to the first terminal is extracted according to the age parameter of the user corresponding to the first terminal, and the user corresponding to the first terminal is classified into different age groups (such as teenagers, adults, and the elderly), wherein the gait feature parameters of the users corresponding to each age group have similarity, and then the K-Means clustering model is trained based on the different age groups and the gait feature parameters corresponding to each age group, so that the trained model can be used to determine the gait feature type of the user corresponding to the first terminal according to the age parameter of the user corresponding to the first terminal.

[0238] Here, after determining the gait feature type of the user corresponding to the first terminal, the first gait feature parameter of the same type as the gait feature type of the user corresponding to the first terminal is selected from the plurality of first gait feature parameters of the healthy population, and the value range of the first gait feature parameter of the same type as the gait feature type of the user corresponding to the first terminal is set in the first interval (such as saving the first gait feature parameter within the 95% confidence interval), wherein the center value of the parameter in the first interval is the healthy gait feature center value, and the upper and lower limits of the first interval are the healthy gait feature boundary value.

[0239] Here, the formula for calculating the mean of all first gait feature parameters of the same type as the gait feature type of the user corresponding to the first terminal is:

[0240]

[0241] wherein Y i is the i-th gait feature parameter of the same gait feature type of the user corresponding to the first terminal, N is the number of the gait feature parameters of the same gait feature type of the user corresponding to the first terminal, and μ is the mean of all the gait feature parameters of the same gait feature type of the user corresponding to the first terminal.

[0242] Here, the formula for calculating the variance of the gait feature parameters of the same gait feature type of the user corresponding to the first terminal is:

[0243]

[0244] wherein σ 2 is the variance of all the gait feature parameters of the same gait feature type of the user corresponding to the first terminal, and σ is the standard error.

[0245] Here, the formula for calculating the gait feature parameters within the 95% confidence interval is:

[0246]

[0247] wherein 1.96 is the coefficient of the standard error corresponding to the 95% confidence interval, and Confidence_Interval 0.95 represents the 95% confidence interval of all the gait feature parameters of the same gait feature type of the user corresponding to the first terminal.

[0248] Based on the above method, in some embodiments, after the second terminal receives the third model sent by the first terminal in response to the second download request, the following steps can be further included:

[0249] Step Y1: determining whether the acceleration is abnormal, and if so, correcting the acceleration to obtain a corrected acceleration.

[0250] Step Y2: determining a plurality of corrected third dynamic feature parameters based on the corrected acceleration.

[0251] wherein the plurality of third dynamic feature parameters include the acceleration.

[0252] Here, the user corresponding to the second terminal walks in a free state (e.g., first walks straight and then turns), and the gait acceleration and other data of the user corresponding to the second terminal are synchronously collected. If it is determined that the acceleration data of the user is abnormal, the acceleration data needs to be corrected to a normal value, so as to determine a plurality of corrected fourth gait feature parameters based on the corrected acceleration data.

[0253] Here, the third dynamic characteristic parameters also include angular velocity, attitude angle and Euler angle, and the third gait characteristic parameters include support phase and swing phase.

[0254] Based on this, in some implementations, correcting the acceleration to obtain the corrected acceleration may specifically include:

[0255] Based on multiple third dynamic characteristic parameters, the gait cycle is divided into the support phase and the swing phase; among them, the support phase includes the single support phase and the double support phase; the single support phase includes the push-off phase from heel lift-off to ipsilateral toe lift-off, and the landing phase from toe-off to ipsilateral heel-off.

[0256] Calculate the horizontal distance between the two feet corresponding to a single support;

[0257] If there are data points that meet the abnormal conditions corresponding to the distance between the two feet on the horizontal plane, correct the acceleration corresponding to the data points.

[0258] The first error correction parameter corresponding to the double support, the second error correction parameter corresponding to the extension period, the third error correction parameter corresponding to the landing period, and the fourth error correction parameter corresponding to the swing are obtained respectively.

[0259] Based on the first error correction parameter, the second error correction parameter, the third error correction parameter, and the fourth error correction parameter, the acceleration is corrected to obtain the corrected acceleration.

[0260] Based on this, in some implementations, the division of the support phase and swing phase of the gait cycle based on multiple third dynamic characteristic parameters may specifically include:

[0261] Based on Euler angles, identify toe-off gait events and heel-strike gait events;

[0262] Based on posture angles, identify heel-off gait events and flatfoot gait events;

[0263] Based on toe-off gait events, heel-off gait events, heel-off gait events, and flat-foot gait events, the gait cycle is divided into the support phase and the swing phase.

[0264] like Figure 10 The diagram illustrates a user's motor function assessment. The user first performs a static test, standing still with feet 15 centimeters apart. During the walking test, the user performs the following... Figure 4The user is asked to perform the task of "straight-walk-turn- straight-walk-turn-straight-walk" and repeat several laps (for example, 20 laps) at a self-comfortable speed, with a path length of 7 meters, wherein the user needs to make a 180-degree turn during walking, and the acceleration, angular velocity, and attitude angle data are synchronously collected by an inertial measurement unit (IMU) during the user's walking.

[0265] According to the data collected by the IMU, the quaternions of the spatial attitude changes of the IMU coordinates and the earth coordinate system are calculated, the Euler angles and the global acceleration of the non-gravitational earth coordinate system are calculated, wherein the acceleration and angular velocity data are fused into the quaternion mode, the real-time changes of speed and direction are selected, and the coordinate system is unified to the earth coordinate system, which can reduce the time overhead and conversion error of multiple coordinate system conversion, such as Figure 11 The principle diagram of the specific spatial attitude calculation algorithm is shown.

[0266] Here, the quaternion The direction of the IMU relative to the earth coordinate system can be obtained by integrating the quaternion change rate, and the calculation formula is:

[0267]

[0268] Where the quaternion change rate The normalized quaternion The gyroscope measurement value and the gained acceleration error term are calculated.

[0269] The initial is a fixed value, which is not accurate at the initial calculation. By dynamically converging the gain coefficient K from a larger initial value to a regular value at the initialization time t init , the slowly converges to the accurate value. The calculation formula of the gain coefficient K is:

[0270]

[0271] Where t init is the initialization time in seconds, K init is the initial value of the gain coefficient, K normal is the normal gain coefficient of the gain coefficient when calculating the quaternion normally, and K init >K normal , t is the actual sampling time point.

[0272] The acceleration error term is determined by the acceleration measurement value a, and the calculation formula is as follows:

[0273]

[0274] where q x is the normalized quaternion y . z ω is the element of the normalized quaternion .

[0275] The gyroscope will produce linear error during the measurement process, which can be subtracted from the estimated error value ω bias before the complementary filter estimation calculation, thereby correcting the gyroscope error. The ω in the quaternion rate of change calculation uses the corrected gyroscope reading ω′, and its calculation formula is:

[0276] ω′=ω-ω bias (30)

[0277] where the estimated error value ω bias is calculated using the low-pass filter result, and its calculation formula is:

[0278] ω bias =2πf c ∫pω·dt (31)

[0279]

[0280] Zero gravity acceleration a zero is the acceleration measurement value a after removing gravity, which is obtained by subtracting the component of gravity in the IMU coordinate system from the acceleration measurement value a, and its calculation formula is:

[0281]

[0282] By normalizing the change of the quaternion, the global acceleration in the earth coordinate system can be calculated, and its calculation formula is:

[0283]

[0284] Linear acceleration suppression reduces errors caused by linear and rotational motion acceleration. The working principle of acceleration suppression is to compare the instantaneous inclination measurement value provided by the accelerometer with the current inclination measurement value output by the algorithm, and if the angle difference between the two inclinations is greater than the threshold t a , then the accelerometer measurement value will be ignored in this algorithm update, and its calculation formula is:

[0285]

[0286] Converting a quaternion to Euler angles is a common attitude representation conversion, which is used to convert the attitude information of a quaternion to a more understandable Euler angle representation. The normalized quaternion is converted to Euler angles, and its calculation formula is:

[0287] ​

[0288] where q w, q x, q y q z are elements of the standard quaternion, and the results of arctan and arcsin are This does not cover all orientations (for θ pitch angle has already been satisfied, so atan2 is used instead of arctan.

[0289] Here, according to the posture angle data collected by the IMU, peak detection is used to identify the toe-off (TO) and heel-strike (HS) gait events, and according to the posture angle data collected by the IMU, first-order differentiation is performed to obtain the pitch angle velocity; a third-order zero-lag high-pass filter is used to filter the pitch angle velocity to obtain the filtered pitch angle velocity; threshold detection and peak detection are used on the filtered pitch angle velocity to identify the heel-off (HO) and foot-flat (FF) gait events; according to the HO, TO, HS, and FF gait events, output gait time parameters are calculated, and then the gait cycle is divided, and the support phase time and swing phase time of the gait cycle are calculated.

[0290] According to the divided gait cycle, a zero position fusion filtering scheme is adopted, and the small change in the horizontal plane distance between the two feet is used as the basis for detecting abnormal values in the acceleration data.

[0291] As shown in Figure 12 is a gait cycle division and zero velocity interval update schematic diagram, according to the gait events, the foot stationary period is defined as: toe strike (TS)→HO, and the foot movement period is defined as: HO→TS, which can be refined into three stages: stage one HO→TO, stage two TO→HS, and stage three HS→TS.

[0292] As shown in Figure 13 is a user straight line walking trajectory analysis schematic diagram, when the user is in a walking state, one side foot is in stage one HO→TO of the movement period, and the other side foot is in stage three HS→TS of the movement period, so the horizontal plane distance amplitude change of the two feet in the two-foot support phase is small.

[0293] Here, first, the two-foot position in the earth coordinate system is calculated, and the calculation formula is:

[0294]

[0295] where tstart is the start time of walking, t end is the end time of walking, p0 represents the initial position when walking.

[0296] Then, the horizontal distance of the biped within the biped support phase is calculated, and the calculation formula is:

[0297]

[0298] Where, T DLS is the duration of the biped support phase, P represents the data of the position, R / L respectively represent the right foot and the left foot, D t represents the horizontal distance between the biped.

[0299] Here, for the calculated horizontal distance of the biped within the biped support phase, the absolute median deviation method can be used to detect the outliers, first, the median of the horizontal distance data of the biped within the biped support phase is calculated, and the calculation formula is:

[0300]

[0301] If the number n of all distance data points is odd, the median is the middle number, if n is even, the median is the average of the middle two numbers, where X represents all sorted distance data points, median() represents the method of solving the median.

[0302] Secondly, the absolute deviation of each distance data point from the median is calculated, and the calculation formula is:

[0303] Deviation i = |X i -median(X) | (40)

[0304] Where, X i represents the i-th distance data point.

[0305] Then, the absolute median deviation MAD of all the above deviations is calculated, and the calculation formula is:

[0306] MAD = median(Deviation1, Deviation2, …, Deviation n ) (41)

[0307] Finally, the corrected Z-Score corresponding to each distance data point is calculated, and the calculation formula is:

[0308]

[0309] where the above formula converts MAD to a scale similar to standard deviation with a constant factor of 0.6745.

[0310] If the modified Z-Score calculated from the ith distance data exceeds theta DLS = 2.5, the distance data point X (i) is considered as an outlier, then the acceleration data of the biped corresponding to the data point is modified to its corresponding last normal value.

[0311] After determining the outlier data points and modifying them, the foot movement trajectories in the two time phases before and after the flat foot phase are collectively modeled based on the zero velocity joint zero position scheme, and the corresponding cumulative errors are calculated respectively, and the support phase cumulative errors before and after the flat foot phase are solved.

[0312] First, the user's foot is stationary during the stationary period, and the linear velocity of the foot is zero. The velocity during the motion period is calculated by integrating the acceleration. Thus, the velocity of the biped during the motion period in the earth coordinate system is calculated by integrating the acceleration in the zero gravity earth coordinate system, and the calculation formula is:

[0313]

[0314] where a global (t) is the global zero gravity acceleration in the earth coordinate system, t HO is the time point when the heel leaves the ground, and t TS is the time point when the toe lands.

[0315] Second, since there is an error in the acceleration measurement, the velocity estimated by the above formula may not be zero when the foot is in the stationary period. Therefore, the difference between the actual velocity (which is known to be zero) and the velocity obtained by integration can be used to correct the acceleration error. By calculating the cumulative error during the motion period (HO→TS), the error caused by the integration of acceleration is further eliminated.

[0316] The acceleration measurement value during the motion period is:

[0317]

[0318] where, is the actual acceleration value measured by the IMU, e represents the earth coordinate system, is the true acceleration value generated by motion, and ε is the bias error. T is the duration of the motion period, t HO is the time point when the heel leaves the ground, and t TS is the time point when the toe lands.

[0319] The bias error is a variable ε(t) that fluctuates slightly over time in a short time period of the motion phase. And at the beginning of the motion phase, the foot velocity is zero, based on which the above formula is rewritten as:

[0320]

[0321] wherein, The velocity calculated by integration is composed of the actual velocity and the cumulative error function e(t), t∈[0, T].

[0322] At the end of the motion phase, t=T, when the foot sole fully contacts the ground, the actual velocity is zero. Therefore, the bias error in the acceleration measurement can be calculated by the following formula:

[0323]

[0324] That is, the value of is a cumulative error value for the acceleration integration in the foot motion phase. Based on this, the cumulative error of the IMU accelerometer between two full-foot landing phases is The calculation formula is:

[0325]

[0326] wherein, the cumulative error is composed of three stages: stage one, the extension period HO→TO, stage two, the swing period TO→HS, and stage three, the landing period HS→TS.

[0327] Since the bias error is a variable ε(t) that fluctuates slightly over time, the accumulated error values in the three stages of the motion phase are different, and the bias error mean value needs to be solved for each stage respectively, and then the error compensation is performed, which can more accurately correct the bias error of the IMU measurement unit.

[0328] In the process of calculating the bias error of each stage, first, the user gender and height data are obtained, and the foot length is calculated in proportion according to Table 1, and the average value is calculated for the user data that cannot be obtained, such as the L tiptoe and L heel two parameters in the following stage one and stage three.

[0329] Table 1

[0330] ​ Gender Height Foot length Male 170 24.5 Female 158 22.5

[0331] For the bias error of the calculation phase one pedal extension period HO→TO, as Figure 14 The geometric modeling diagram of the foot in the phase one pedal extension period is shown, the changing trajectory of the foot is compared with a circle, and the distance changes of the x, y, and z axes in the phase one are calculated through geometric analysis. According to Figure 14 It can be known that the distance calculation formula of the x, y, and z axes in the phase one pedal extension period is:

[0332]

[0333]

[0334]

[0335] Wherein, Δz1, Δy1, and Δx1 respectively represent the distances of the z, y, and x axes in the phase one, L tiptoe is the distance between the IMU placement position and the insole tip, represents the pitch angle corresponding to the TO gait event, represents the yaw angle corresponding to the TO gait event.

[0336] It is known that the motion equation calculation formula according to the distance is:

[0337]

[0338] Given Δz1, Δy1, and Δx1, the real average acceleration value in the phase one can be solved, that is,

[0339] According to the above formula, the average value of the IMU measurement value in the phase one can be calculated as:

[0340]

[0341] Wherein, is the average value of the IMU measurement value in the phase one, is the average value of the real acceleration in the phase one, is the average error value in the phase one.

[0342] Based on this, the cumulative error calculation formula of the phase one is:

[0343]

[0344] Wherein e(T1) represents the cumulative error in the phase one, and T1 is the duration of the phase one.

[0345] For the bias error of the calculation phase three landing period HS→TS, asFigure 15 The geometric modeling diagram of the foot in the third phase landing period is shown, and the changing trajectory of the foot is compared with a circle. The distance changes of the x, y, and z axes in the third phase are calculated through geometric analysis. According to the geometric modeling diagram of the foot in the third phase landing period, the distance changes of the x, y, and z axes in the third phase are calculated. Figure 15 It can be seen that the distance calculation formula of the x, y, and z axes in the third phase landing period is:

[0346]

[0347]

[0348]

[0349] wherein Δz3, Δy3, and Δx3 represent the distances of the z, y, and x axes in the third phase, respectively, L heel represents the distance between the IMU placement position and the heel of the insole, represents the pitch angle corresponding to the HS gait event, represents the yaw angle corresponding to the HS gait event.

[0350] The third phase movement process is the reverse process of the first phase, and the speed gradually increases from zero in the first phase, while the speed gradually decreases to zero in the third phase. Therefore, the distance motion equation of the third phase is also the same as that of the first phase as follows:

[0351]

[0352] Given Δz3, Δy3, and Δx3, the real average acceleration value in the third phase can be solved, that is,

[0353] According to the above formula, the average value of the IMU measurement value in the third phase can be calculated as:

[0354]

[0355] wherein is the average value of the IMU measurement value in the third phase, is the average value of the real acceleration in the third phase, is the average error value in the third phase.

[0356] Based on this, the cumulative error calculation formula of the third phase is:

[0357]

[0358] wherein e(T3) represents the cumulative error in the third phase, and T3 is the duration of the third phase.

[0359] After the cumulative errors of the two phases before and after the flat-foot period are obtained, the cumulative error in the swing phase can be obtained according to the obtained cumulative errors of the two phases before and after the flat-foot period, and the acceleration data in the three stages is corrected by using the average error values of the three stages in turn, so as to more accurately correct the double-foot bottom IMU measurement bias error.

[0360] According to the total cumulative error e(T) of the motion period, the cumulative error e(T1) of the first stage of the motion period and the cumulative error e(T3) of the third stage of the motion period, the cumulative error of the second stage of the motion period can be calculated as follows:

[0361]

[0362] Wherein, e(T2) represents the cumulative error in the second stage, and T2 is the duration of the second stage.

[0363] The bias average error ε1 in the first stage obtained by solving the above formula, the bias average error ε2 in the second stage obtained by solving the above formula, and the bias average error ε3 in the third stage obtained by solving the above formula can be used to compensate the cumulative error caused by acceleration integration to solve the speed in the three stages of the motion period.

[0364] According to the compensated speed of the foot motion period geometric modeling analysis, the compensated speed is more close to the linear speed generated by the foot during walking, so the step length and position data calculated by the integral method are more accurate.

[0365] Here, the step length is calculated by integrating the speed in the second stage of the foot motion period, and the calculation formula is:

[0366]

[0367] Wherein, The speed after error correction is represented by t HS is the time point of the HS gait event, t TO is the time point of the TO gait event.

[0368] The traditional zero speed update uses the average error of the entire motion period to compensate the cumulative error of the speed in the second stage, which is not accurate. According to the geometric modeling analysis of the foot motion period, the average error of the second stage of the motion period is used to compensate the cumulative error of the speed in the second stage, so that the compensated speed of the second stage is more consistent with the true value, and the calculated step length is more accurate.

[0369] Here, the position at each time is calculated by integrating the speed in the walking time, and the calculation formula is:

[0370]

[0371] wherein t start is the start time of walking, t end is the end time of walking, and p0 represents the initial position at the time of walking.

[0372] According to the geometric modeling analysis of the foot movement period, the average error of each stage in the movement period is used to compensate for the cumulative error of the speed of each stage, and the compensated speed is closer to the true speed value of each stage in the movement period. The accuracy of the position data calculated at each time is higher. The error compensation is performed for each step, and the speed compensation accuracy of each step is improved compared with the traditional zero speed update. The final position data is the integral of the speed of the entire walking process, and therefore the accuracy of the finally calculated position data is higher than that of the traditional zero speed update.

[0373] As Figure 16 shown is a detailed flowchart of an adaptive demographic portrait processing method provided by an embodiment of the application, which is divided into nine levels, and specifically includes:

[0374] The first level (cloud clustering middleware server)

[0375] Step 1601: Extract initial static feature data and corresponding initial dynamic feature data of a healthy population in a database, and the initial dynamic feature data includes initial gait feature data.

[0376] Step 1602: Perform feature selection based on the initial static feature data and the initial gait feature data, select a first most relevant parameter and a first key gait feature parameter with a first correlation coefficient between the first most relevant parameter greater than a first threshold value, and set the clustering class number and the maximum iteration number with the data corresponding to the first most relevant parameter.

[0377] Step 1603: Select a suitable machine learning algorithm to perform model training, and cluster the healthy populations with similar static features and similar gait features.

[0378] Step 1604: Determine whether the model can stop training, if yes, proceed to step 1605; otherwise, return to step 1603.

[0379] Step 1605: Calculate the effect evaluation coefficient of the model to evaluate the performance of the model.

[0380] Step 1606: If the effect evaluation coefficient of the model is greater than 0.3, output the model as a first model; otherwise, retrain the model.

[0381] The second level (cloud application server)

[0382] Step 1607: The doctor logs in the application server.

[0383] Step 1608: According to the department and the corresponding disease demand information, the static characteristics and dynamic characteristics (motion state) of the patient population are input.

[0384] Step 1609: A first call request of a first model is sent to the middleware server.

[0385] Step 1610: The first model sent by the middleware server in response to the first call request is received, and the first model is downloaded.

[0386] Step 1611: The first model is configured to the mobile terminal (human body domain mobile convergence center) of the user based on the first model, and a second model corresponding to the mobile terminal is generated.

[0387] Third level (human body domain mobile convergence center)

[0388] Step 1612: The user fills in the demographic information static characteristic data.

[0389] Step 1613: A first download request is sent to the cloud application server, a second model sent by the cloud application server in response to the first download request is received, and the second model is downloaded.

[0390] Step 1614: A second static characteristic parameter corresponding to a second gait characteristic parameter is selected based on the second model.

[0391] Fourth level (human body domain mobile convergence center)

[0392] Step 1615: A correction error is determined based on the second gait characteristic parameter and the first gait characteristic parameter in the first model.

[0393] Step 1616: The correction error is input into a deep learning error correction model to determine a correction error prediction value.

[0394] Step 1617: A loss value is determined based on the correction error and the correction error prediction value.

[0395] Step 1618: Deep learning error correction is performed based on the loss value, and the correction error model is updated.

[0396] Step 1619: It is judged whether the loss value in the correction error model updating process is no longer reduced, if not, it indicates that the model is still converging, and step 1620 is entered; otherwise, step 1621 is entered.

[0397] Step 1620: It is judged whether the number of model update training times is less than 1000, if yes, the model parameters are continuously updated through the gradient descent algorithm; otherwise, step 1621 is entered.

[0398] Step 1621: correcting the second model based on the updated correction error model to generate a third model.

[0399] Step 1622: receiving the health gait feature center value and the health gait feature boundary value sent by the human domain terminal, and dividing the user into a dynamic clustering category (health gait feature value range) based on the health gait feature boundary value according to the dynamic characteristics of the user.

[0400] The fifth level (human domain terminal)

[0401] Step 1623: sending a second download request to the human domain mobile aggregation center, receiving the third model sent by the human domain mobile aggregation center, and downloading the third model.

[0402] Step 1624: performing feature selection based on the third dynamic characteristic parameter and the third gait characteristic parameter, selecting the third most relevant parameter and the third key gait characteristic parameter with a third correlation coefficient greater than a third threshold value, and setting the clustering class number and the maximum iteration number corresponding to the third most relevant parameter.

[0403] Step 1625: selecting a suitable machine learning algorithm to train the model, and clustering users with similar dynamic characteristics and similar gait characteristics.

[0404] Step 1626: determining whether the model can stop training, if yes, entering step 1606; otherwise, returning to step 1625.

[0405] Step 1627: calculating the effect evaluation coefficient of the model to evaluate the performance of the model.

[0406] Step 1628: if the effect evaluation coefficient of the model is greater than 0.5, outputting the model as a fourth model; otherwise, retraining the model.

[0407] Step 1629: calculating the health gait feature center value and the health gait feature boundary value, and sending the health gait feature center value and the health gait feature boundary value to the human domain mobile aggregation center.

[0408] The sixth level (human domain terminal)

[0409] Step 1630: the user walks in a free state, and synchronously collects acceleration, angular velocity and other data of the user.

[0410] Step 1631: converting the acceleration and angular velocity data from the IMU coordinate system to the earth coordinate system, and calculating the quaternion.

[0411] Step 1632: converting the quaternion to Euler angle.

[0412] Step 1633: Based on the Euler angles, divide the gait cycle of the user.

[0413] Step 1634: Based on the acceleration data, preliminarily calculate the position information of the user.

[0414] Step 1635: Calculate the horizontal plane distance of the user's two feet in the double support phase.

[0415] Step 1636: If the data points of the horizontal plane distance of the two feet in the double support phase are abnormal, correct the acceleration data of the two feet corresponding to the data points to obtain the pre-processed acceleration.

[0416] Step 1637: Based on the zero velocity update of the flat foot phase, update the total cumulative error of the foot movement period.

[0417] Step 1638: Geometrically analyze the foot movement trajectory of the stage one push-off phase and the stage three landing phase to determine the cumulative error of the stage one push-off phase and the cumulative error of the stage three landing phase.

[0418] Step 1639: Based on the total cumulative error of the foot movement period, the cumulative error of the stage one push-off phase, and the cumulative error of the stage three landing phase, calculate the cumulative error of the stage two swing phase.

[0419] Step 1640: Based on the cumulative error of the stage one push-off phase, the cumulative error of the stage two swing phase, and the cumulative error of the stage three landing phase, correct the acceleration data to obtain the corrected acceleration data.

[0420] Step 1641: Based on the corrected acceleration data, calculate the stride and position of the user.

[0421] Step 1642: Based on the stride and position of the user, determine the corrected third dynamic feature parameter.

[0422] Seventh level (cloud application server)

[0423] Step 1643: Determine in which gait range the user is located.

[0424] Step 1644: If the user is located in the healthy gait feature value range, generate a reverse linear score function according to the healthy gait feature center value and the boundary value.

[0425] Step 1645: Put the gait parameters of the user into the score function to obtain the risk degree score of the user.

[0426] Step 1646: If the user is located in the overlapping region of the healthy gait feature value range and the non-healthy gait feature value range, a reverse linear score function is generated according to the center value and the boundary value of the healthy and non-healthy gait features.

[0427] Step 1647: The gait parameters of the user are brought into the score function to obtain the risk degree score of the user.

[0428] Step 1648: If the user is located in the non-healthy gait feature value range, a positive linear score function is generated according to the center value and the boundary value of the non-healthy gait feature.

[0429] Step 1649: The gait parameters of the user are brought into the score function to obtain the severity score of the user.

[0430] Step 1650: The score of the user is converted according to the healthy cluster center, the healthy and non-healthy boundary, and the non-healthy cluster center.

[0431] Eighth level (human domain mobile convergence center)

[0432] Step 1651: Load the healthy population data, which includes the first static feature parameters (including the age parameter) and the first dynamic feature parameters (including the first gait feature parameters) of the healthy population.

[0433] Step 1652: Traverse the first static feature parameter data and the first gait feature parameter data to calculate the second correlation between the age parameter in the first static feature parameter and the first gait feature parameter.

[0434] Step 1653: Traverse the first gait feature parameter data to calculate the profile coefficient of the first gait feature parameter.

[0435] Step 1654: Determine the primary weight coefficient based on the second correlation between the age parameter and the first gait feature parameter, and determine the secondary weight coefficient based on the profile coefficient of the first gait feature parameter.

[0436] Step 1655: Fuse the primary weight coefficient and the secondary weight coefficient to obtain a fused weight coefficient, and eliminate the first gait feature parameters that are more affected by age in the fused weight coefficient.

[0437] Step 1656: Sort the eliminated fused weight coefficients in order and perform normalization processing.

[0438] Step 1657: Add the converted fused weight coefficients to obtain a healthy risk comprehensive score.

[0439] Ninth level (cloud domain collaboration)

[0440] ​​​​​​​​​​​​​​Step 1658: In the updating process of the correction error model, if the loss value reaches convergence, if the user data is added to the human body domain mobile convergence center database.

[0441] Step 1659: Determine whether the amount of new user data reaches 200, if yes, go to step 1660; otherwise, end the updating process.

[0442] Step 1660: Start the model updating mechanism, and retrain the correction error model and the third model.

[0443] Embodiments of the present application also provide a self-adaptive physical population portrait processing device, Figure 18 is a structural schematic diagram of a self-adaptive physical population portrait processing device applied to a first server provided by the embodiments of the present application, as Figure 17 shown, the device comprises:

[0444] The first sending unit 1701 is configured to send a first calling request to a second server, and receive a first model sent by the second server in response to the first calling request; wherein the first model comprises a plurality of first static characteristic parameters of a healthy population and a plurality of first dynamic characteristic parameters corresponding thereto, and the plurality of first dynamic characteristic parameters comprises a plurality of first gait characteristic parameters.

[0445] The first generating unit 1702 is configured to generate a second model corresponding to a first terminal based on the first model; wherein the second model comprises a plurality of second static characteristic parameters of a user corresponding to the first terminal and a plurality of second dynamic characteristic parameters corresponding thereto, and the plurality of second dynamic characteristic parameters comprises a plurality of second gait characteristic parameters.

[0446] The first receiving unit 1703 is configured to receive a first download request sent by the first terminal, and send the second model to the first terminal in response to the first download request.

[0447] In some embodiments, the first receiving unit 1703 is further configured to receive a healthy gait characteristic numerical range sent by the first terminal.

[0448] In some embodiments, the device further comprises:

[0449] The first determining unit is configured to load non-healthy population data, and determine a non-healthy gait characteristic numerical range based on the non-healthy population data; determine a distribution of gait characteristic data of the user corresponding to the first terminal based on the healthy gait characteristic numerical range and the non-healthy gait characteristic numerical range; and determine a gait health score of the user corresponding to the first terminal based on the distribution.

[0450] In some embodiments, the first determining unit is specifically configured to extract, from the non-healthy population data, non-healthy population data of the same age as the user corresponding to the first terminal based on an age parameter of the non-healthy population; and determine the non-healthy gait feature value range based on the non-healthy population data of the same age as the user corresponding to the first terminal. The non-healthy population data includes an age parameter of the non-healthy population.

[0451] In some embodiments, the first determining unit is further specifically configured to determine that the user corresponding to the first terminal is in a healthy gait state if the gait feature data of the user corresponding to the first terminal is located in the healthy gait feature value range; or determine that the user corresponding to the first terminal is in a non-healthy gait state if the gait feature data of the user corresponding to the first terminal is located in the non-healthy gait feature value range; or determine that the user corresponding to the first terminal is in both the healthy gait state and the non-healthy gait state if the gait feature data of the user corresponding to the first terminal is located in the intersection of the healthy gait feature value range and the non-healthy gait feature value range.

[0452] In some embodiments, the first determining unit is further specifically configured to determine a risk degree score of the user corresponding to the first terminal in the healthy gait state if the user corresponding to the first terminal is in the healthy gait state; or determine a severity score of the user corresponding to the first terminal in the non-healthy gait state if the user corresponding to the first terminal is in the non-healthy gait state; or determine the risk degree score of the user corresponding to the first terminal in the healthy gait state and the severity score of the user corresponding to the first terminal in the non-healthy gait state if the user corresponding to the first terminal is in both the healthy gait state and the non-healthy gait state.

[0453] The application further provides a self-adaptive population portrait processing device, Figure 18 is a structural schematic diagram of a self-adaptive population portrait processing device applied to a second server provided by the application, as Figure 18 shown, the device comprises:

[0454] The second determining unit 1801 is configured to determine the first model based on the static feature data and the corresponding dynamic feature data of the healthy population. The static feature data includes a plurality of initial static feature parameters, and the dynamic feature data includes a plurality of initial dynamic feature parameters. The plurality of initial dynamic feature parameters includes a plurality of initial gait feature parameters. The first model includes a plurality of first static feature parameters and a plurality of first dynamic feature parameters of the healthy population. The plurality of first dynamic feature parameters includes a plurality of first gait feature parameters.

[0455] The second receiving unit 1802 is configured to receive the first calling request sent by the first server, and send the first model to the first server in response to the first calling request.

[0456] In some embodiments, the second determining unit 1801 is specifically configured to determine a first most relevant parameter based on a degree of correlation between each initial static feature parameter and the plurality of initial gait feature parameters; determine a first key gait feature parameter based on the first most relevant parameter; and determine the first model based on the first most relevant parameter and the first key gait feature parameter.

[0457] In some embodiments, the second determining unit 1801 is further specifically configured to calculate a first correlation coefficient between each initial static feature parameter and the plurality of initial gait feature parameters respectively; determine a first correlation coefficient mean of each initial static feature parameter based on the plurality of determined first correlation coefficients, and select an initial static feature parameter corresponding to a maximum first correlation coefficient mean as the first most relevant parameter.

[0458] In some embodiments, the second determining unit 1801 is further specifically configured to select an initial gait feature parameter with a first correlation coefficient greater than a first threshold value with the first most relevant parameter as the first key gait feature parameter.

[0459] Embodiments of the present application also provide an adaptive physical population portrait processing device, Figure 19 is a structural schematic diagram of an adaptive physical population portrait processing device applied to a first terminal provided by the embodiments of the present application, as Figure 19 shown, the device comprises:

[0460] The third sending unit 1901 is configured to send a first download request to a first server, and receive a second model sent by the first server in response to the first download request; wherein the second model is a model corresponding to the first terminal generated by the first server according to a first model; the first model is obtained by the first server requesting and calling a second server; the first model comprises a plurality of first static feature parameters of a healthy population and a plurality of first dynamic feature parameters corresponding thereto, and the plurality of first dynamic feature parameters comprise a plurality of first gait feature parameters; the second model comprises a plurality of second static feature parameters of a user corresponding to the first terminal and a plurality of second dynamic feature parameters corresponding thereto, and the plurality of second dynamic feature parameters comprise a plurality of second gait feature parameters.

[0461] The third generating unit 1902 is configured to generate a third model corresponding to the first terminal based on the second model; wherein the third model comprises a plurality of third static feature parameters of the user corresponding to the first terminal after correction and a plurality of third dynamic feature parameters corresponding thereto, and the plurality of third dynamic feature parameters comprise a plurality of third gait feature parameters.

[0462] The third receiving unit 1903 is configured to receive a second download request sent by a second terminal, and send the third model to the second terminal in response to the second download request.

[0463] In some embodiments, the third generation unit 1902 is specifically configured to determine a correction error based on the plurality of first gait feature parameters and the plurality of second gait feature parameters; determine a correction error prediction value based on the correction error and a correction error model; determine a loss value based on the correction error and the correction error prediction value; update the correction error model based on the loss value; and correct the second model based on the updated correction error model to obtain a third model.

[0464] In some embodiments, the third generation unit 1902 is further specifically configured to determine a first gait feature comprehensive parameter based on the plurality of first gait feature parameters; and determine the correction error based on the first gait feature comprehensive parameter and the plurality of second gait feature parameters.

[0465] In some embodiments, the third receiving unit 1903 is further configured to receive the health gait feature center value and the health gait feature boundary value sent by the second terminal.

[0466] In some embodiments, the apparatus further includes:

[0467] The third determination unit is configured to determine a health gait feature numerical value range of the user corresponding to the first terminal based on the third model and the health gait feature boundary value.

[0468] In some embodiments, the third sending unit 1901 is further configured to send the health gait feature numerical value range to the first server.

[0469] In some embodiments, the third determination unit is further configured to determine a plurality of first-level weight coefficients based on the plurality of first static feature parameters and the plurality of first gait feature parameters; determine a plurality of second-level weight coefficients based on the plurality of first gait feature parameters; determine a plurality of fusion weight coefficients based on the plurality of first-level weight coefficients and the plurality of second-level weight coefficients; and determine the health risk comprehensive score based on the plurality of fusion weight coefficients.

[0470] In some embodiments, the third determination unit is further specifically configured to calculate a plurality of second correlation coefficients between an age parameter of a healthy population and the plurality of first gait feature parameters respectively; and determine the plurality of first-level weight coefficients based on the plurality of determined second correlation coefficients; wherein the plurality of first static feature parameters include the age parameter of the healthy population.

[0471] In some embodiments, the third determination unit is further specifically configured to determine a plurality of clustering cluster data corresponding to each first gait feature parameter; determine a similarity coefficient within each clustering cluster data and a dissimilarity coefficient between each clustering cluster data based on the plurality of clustering cluster data; determine a plurality of contour coefficients based on the plurality of determined similarity coefficients and the plurality of dissimilarity coefficients; and determine the plurality of second-level weight coefficients based on the plurality of contour coefficients.

[0472] In some embodiments, the third determining unit is further configured to convert and normalize each fusion weight coefficient to obtain a processed each fusion weight coefficient; and determine the health risk comprehensive score based on the processed each fusion weight coefficient.

[0473] The embodiment of the present application further provides a self-adaptive physical population portrait processing device, Figure 20 is a structural schematic diagram of a self-adaptive physical population portrait processing device applied to a second terminal provided by the embodiment of the present application, as Figure 20 shown, the device comprises:

[0474] The fourth sending unit 2001 is configured to send a second download request to the first terminal, and receive a third model sent by the first terminal in response to the second download request; wherein the third model is a model corresponding to the second terminal and generated by the first terminal based on a second model; the second model is a model corresponding to the first terminal and generated by the first server; the first model is obtained by the first server requesting and calling the second server; the third model comprises a plurality of third static feature parameters and a plurality of third dynamic feature parameters corresponding to a user of the first terminal after correction, and the plurality of third dynamic feature parameters comprise a plurality of third gait feature parameters.

[0475] The fourth generating unit 2002 is configured to generate a fourth model corresponding to the second terminal based on the third model; wherein the fourth model comprises a plurality of fourth static feature parameters and a plurality of fourth dynamic feature parameters corresponding to a user of the second terminal after correction, and the plurality of fourth dynamic feature parameters comprise a plurality of fourth gait feature parameters.

[0476] In some embodiments, the fourth generating unit 2002 is specifically configured to determine a third most relevant parameter based on a correlation degree between each third dynamic feature parameter and the plurality of third gait feature parameters; determine a third key gait feature parameter based on the third most relevant parameter; and correct the third model based on the third most relevant parameter and the third key gait feature parameter to obtain the fourth model.

[0477] In some embodiments, the fourth generating unit 2002 is further configured to calculate a third correlation coefficient between each third dynamic feature parameter and the plurality of third gait feature parameters respectively; determine a third correlation coefficient mean of each third dynamic feature parameter based on the determined plurality of third correlation coefficients, and select a third dynamic feature parameter corresponding to a maximum third correlation coefficient mean as the third most relevant parameter.

[0478] In some embodiments, the fourth generating unit 2002 is further configured to select a third gait feature parameter with a third correlation coefficient greater than a third threshold value between the third most relevant parameter as the third key gait feature parameter.

[0479] In some embodiments, the apparatus further comprises:

[0480] The fourth determining unit is configured to determine a gait feature type of the user corresponding to the first terminal based on an age parameter of the user corresponding to the first terminal; select a first gait feature parameter of the same type as the gait feature type from a plurality of first gait feature parameters of the healthy population, and set a value range of the first gait feature parameter of the same type as the gait feature type in a first interval; determine a healthy gait feature center value and a healthy gait feature boundary value based on the parameter value in the first interval; and the plurality of third static feature parameters include the age parameter of the user corresponding to the first terminal.

[0481] In some embodiments, the fourth sending unit 2001 is further configured to send the healthy gait feature center value and the healthy gait feature boundary value to the first terminal.

[0482] In some embodiments, the apparatus further comprises:

[0483] The fourth correcting unit is configured to determine whether the acceleration is abnormal, and if so, correct the acceleration to obtain a corrected acceleration; and the plurality of third dynamic feature parameters include the acceleration.

[0484] In some embodiments, the fourth determining unit is further configured to determine the plurality of third dynamic feature parameters after correction based on the corrected acceleration.

[0485] In some embodiments, the fourth correcting unit is specifically configured to divide a support phase and a swing phase of a gait cycle based on the plurality of third dynamic feature parameters; the support phase includes a single support phase and a double support phase; the single support phase includes a push-off period from a heel off to a same-side toe off, and a landing period from a toe on to a same-side heel on; calculate a double-foot horizontal plane distance corresponding to the single support phase; if a data point corresponding to the double-foot horizontal plane distance satisfies an abnormal condition, correct the acceleration corresponding to the data point; obtain a first error correction parameter corresponding to the double support phase, a second error correction parameter corresponding to the push-off period, a third error correction parameter corresponding to the landing period, and a fourth error correction parameter corresponding to the swing phase, respectively; and correct the acceleration based on the first error correction parameter, the second error correction parameter, the third error correction parameter, and the fourth error correction parameter to obtain the corrected acceleration.

[0486] In some embodiments, the fourth correction unit is further configured to identify a toe-off gait event and a heel-strike gait event based on the Euler angles; identify a heel-off gait event and a flat-foot gait event based on the attitude angles; and divide a support phase and a swing phase of a gait cycle based on the toe-off gait event, the heel-strike gait event, the heel-off gait event, and the flat-foot gait event. The third dynamic feature parameters include angular velocities, the attitude angles, and the Euler angles, and the third gait feature parameters include the support phase and the swing phase.

[0487] Those skilled in the art should understand that, Figure 17 、 Figure 18 、 Figure 19 、 Figure 20 The implementation functions of each unit in the adaptive sign population portrait processing apparatus shown in the foregoing method can be understood with reference to the related descriptions of the foregoing method. Figure 17 、 Figure 18 、 Figure 19 、 Figure 20 The functions of each unit in the adaptive sign population portrait processing apparatus shown in the foregoing method can be implemented by a program running on a processor, or by a specific logic circuit.

[0488] Figure 21 is a structural schematic diagram of a processing device provided by an embodiment of the present application. The processing device can be a terminal device or a network device, Figure 21 The processing device shown in the foregoing method can include a processor 2101. The processor 2101 can invoke and run a computer program from a memory to implement the method in the embodiments of the present application.

[0489] Optionally, as shown in the foregoing method, the processing device can further include a memory 2102. The processor 2101 can invoke and run a computer program from the memory 2102 to implement the method in the embodiments of the present application. Figure 21 The memory 2102 can be a separate device independent of the processor 2101, or can be integrated in the processor 2101.

[0490] Optionally, as shown in the foregoing method, the processing device can further include a transceiver 2103. The processor 2101 can control the transceiver 2103 to communicate with other devices, specifically, can send information or data to other devices, or receive information or data sent by other devices.

[0491] Figure 21 The transceiver 2103 can include a transmitter and a receiver. The transceiver 2103 can further include an antenna, and the number of antennas can be one or more.

[0492]

[0493] ​​The processing device can be specifically an adaptive sign population portrait processing device of the embodiments of the present application, and the processing device can implement corresponding processes of various method implementations of the embodiments of the present application. For brevity, details are not repeated here.

[0494] It should be understood that the processor of the embodiments of the present application can be an integrated circuit chip with a signal processing capability. In the implementation process, each step of the above method embodiments can be completed by integrated logic circuits of hardware in the processor or instructions in the form of software. The processor mentioned above can be a general processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. Each method, step and logic block diagram disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as a hardware code processor for execution, or a combination of hardware and software modules in the code processor for execution. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register or other mature storage medium in the art. The storage medium is located in the memory, and the processor reads the information in the memory, and combines the hardware to complete the steps of the above method.

[0495] It is to be understood that the memory in the embodiments of the present application can be a volatile memory or a nonvolatile memory, or can include both volatile and nonvolatile memory. Among them, the nonvolatile memory can be a read-only memory (Read-Only Memory, ROM), a programmable read-only memory (Programmable ROM, PROM), an erasable programmable read-only memory (Erasable PROM, EPROM), an electrically erasable programmable read-only memory (Electrically EPROM, EEPROM) or a flash memory. The volatile memory can be a random access memory (Random Access Memory, RAM) used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (Static RAM, SRAM), dynamic random access memory (Dynamic RAM, DRAM), synchronous dynamic random access memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (Synchlink DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM). It should be noted that the memory of the system and method described herein is intended to include, but not limited to, these and any other suitable types of memory.

[0496] It should be understood that the above-mentioned memory is exemplary but not limiting, for example, the memory in the embodiments of the present application can also be static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM) and the like. That is, the memory in the embodiments of the present application is intended to include, but not limited to, these and any other suitable types of memory.

[0497] The embodiment of the present application further provides a computer readable storage medium for storing a computer program. The computer readable storage medium can be applied to the processing device in the embodiment of the present application, and the computer program causes the computer to execute the corresponding process realized by the adaptive sign population portrait processing device in each method of the embodiment of the present application. For brevity, details are not repeated here.

[0498] The embodiment of the present application further provides a computer program product comprising computer program instructions. The computer program product can be applied to the processing device in the embodiment of the present application, and the computer program instructions cause the computer to execute the corresponding process realized by the adaptive sign population portrait processing device in each method of the embodiment of the present application. For brevity, details are not repeated here.

[0499] Those skilled in the art can understand that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0500] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, and details are not repeated here.

[0501] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0502] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0503] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit.

[0504] The functions, if implemented in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in part, or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and various other media that can store program codes.

[0505] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which shall be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An adaptive vital sign-based population profiling method, characterized in that, Applied to a first server, the method includes: Send a first call request to a second server and receive a first model sent by the second server in response to the first call request; wherein, the first model includes multiple first static feature parameters of healthy people and multiple corresponding first dynamic feature parameters, the multiple first dynamic feature parameters including multiple first step feature parameters; A second model corresponding to the first terminal is generated based on the first model; wherein, the second model includes multiple second static feature parameters of the user corresponding to the first terminal and multiple corresponding second dynamic feature parameters, the multiple second dynamic feature parameters including multiple second gait feature parameters; The system receives a first download request from a first terminal and, in response to the first download request, sends the second model to the first terminal.

2. The method according to claim 1, characterized in that, The method further includes: Receive the range of health gait feature values ​​sent by the first terminal; Load data on unhealthy individuals and, based on this data, determine the numerical range of unhealthy gait characteristics. Based on the range of healthy gait features and the range of unhealthy gait features, the distribution of gait feature data of the user corresponding to the first terminal is determined; Based on the distribution, the gait health score of the user corresponding to the first terminal is determined.

3. The method according to claim 2, characterized in that, The data on unhealthy individuals includes the age parameters of those who are unhealthy. The determination of the numerical range of unhealthy gait characteristics based on the unhealthy population data includes: Based on the age parameters of the unhealthy population, extract unhealthy population data that are the same age as the user corresponding to the first terminal from the unhealthy population data; Based on data from unhealthy individuals of the same age as the user corresponding to the first terminal, the numerical range of the unhealthy gait characteristics is determined.

4. The method according to claim 2, characterized in that, The step of determining the distribution of gait feature data of the user corresponding to the first terminal based on the range of healthy gait feature values ​​and the range of unhealthy gait feature values ​​includes: If the gait feature data of the user corresponding to the first terminal is within the range of the healthy gait feature values, then it is determined that the user corresponding to the first terminal is in a healthy gait state; or, If the gait feature data of the user corresponding to the first terminal is within the range of the unhealthy gait feature values, then it is determined that the user corresponding to the first terminal is in an unhealthy gait state; or, If the gait feature data of the user corresponding to the first terminal is located within the intersection of the healthy gait feature value range and the unhealthy gait feature value range, then it is determined that the user corresponding to the first terminal is simultaneously in the healthy gait state and the unhealthy gait state.

5. The method according to claim 4, characterized in that, The step of determining the gait health score of the user corresponding to the first terminal based on the distribution includes: If the user corresponding to the first terminal is in the healthy gait state, then determine the risk level score of the user corresponding to the first terminal in the healthy gait state; or, If the user corresponding to the first terminal is in the unhealthy gait state, then determine the severity score of the unhealthy gait state for the user corresponding to the first terminal; or, If the user corresponding to the first terminal is simultaneously in the healthy gait state and the unhealthy gait state, then the risk level score of the user corresponding to the first terminal in the healthy gait state and the severity score in the unhealthy gait state are determined.

6. An adaptive vital sign-based population profiling method, characterized in that, Applied to a second server, the method includes: A first model is determined based on static feature data and corresponding dynamic feature data of healthy individuals; wherein, the static feature data includes multiple initial static feature parameters, the dynamic feature data includes multiple initial dynamic feature parameters, and the multiple initial dynamic feature parameters include multiple initial gait feature parameters; the first model includes multiple first static feature parameters and corresponding multiple first dynamic feature parameters of healthy individuals, and the multiple first dynamic feature parameters include multiple first gait feature parameters. Receive a first call request sent by a first server, and in response to the first call request send the first model to the first server.

7. The method according to claim 6, characterized in that, The first model is determined based on static feature data and corresponding dynamic feature data of healthy individuals, including: Based on the correlation between each initial static feature parameter and the plurality of initial gait feature parameters, the first most relevant parameter is determined; Based on the first most relevant parameter, the first key gait feature parameter is determined; The first model is determined based on the first most relevant parameter and the first key gait feature parameter.

8. The method according to claim 7, characterized in that, The determination of the first most relevant parameter based on the correlation between each initial static feature parameter and the plurality of initial gait feature parameters includes: Calculate the first correlation coefficient between each initial static feature parameter and the plurality of initial gait feature parameters respectively; Based on the determined multiple first correlation coefficients, the mean of the first correlation coefficients for each initial static feature parameter is determined, and the initial static feature parameter corresponding to the largest mean of the first correlation coefficients is selected as the first most relevant parameter.

9. The method according to claim 8, characterized in that, The determination of the first key gait feature parameters based on the first most relevant parameter includes: An initial gait feature parameter whose first correlation coefficient with the first most relevant parameter is greater than a first threshold is selected as the first key gait feature parameter.

10. An adaptive method for processing population profiles based on vital signs, characterized in that, Applied to a first terminal, the method includes: A first download request is sent to a first server, and a second model is received from the first server in response to the first download request; wherein, the second model is a model corresponding to the first terminal generated by the first server based on the first model; the first model is obtained by the first server requesting and calling the second server; the first model includes multiple first static feature parameters and corresponding multiple first dynamic feature parameters of healthy individuals, the multiple first dynamic feature parameters including multiple first gait feature parameters; the second model includes multiple second static feature parameters and corresponding multiple second dynamic feature parameters of the user corresponding to the first terminal, the multiple second dynamic feature parameters including multiple second gait feature parameters. A third model corresponding to the first terminal is generated based on the second model; wherein, the third model includes multiple third static feature parameters and multiple third dynamic feature parameters corresponding to the first terminal after user correction, and the multiple third dynamic feature parameters include multiple third gait feature parameters. The system receives a second download request from the second terminal and, in response to the second download request, sends the third model to the second terminal.

11. The method according to claim 10, characterized in that, Based on the second model, a third model corresponding to the first terminal is generated, including: Based on the plurality of first step characteristic parameters and the plurality of second step characteristic parameters, the correction error is determined; Based on the aforementioned correction error and correction error model, the predicted value of the correction error is determined; The loss value is determined based on the correction error and the predicted value of the correction error; The correction error model is updated based on the loss value; The second model is corrected based on the updated correction error model to obtain the third model.

12. The method according to claim 11, characterized in that, The step of determining the correction error based on the plurality of first step characteristic parameters and the plurality of second step characteristic parameters includes: Based on the multiple first-step feature parameters, determine the first-step feature comprehensive parameters; The correction error is determined based on the first step feature synthesis parameters and the plurality of second step feature parameters.

13. The method according to claim 10, characterized in that, The method further includes: Receive the center value and boundary value of the healthy gait feature sent by the second terminal; Based on the third model and the healthy gait feature boundary values, the range of healthy gait feature values ​​for the user corresponding to the first terminal is determined. Send the range of healthy gait feature values ​​to the first server.

14. The method according to claim 10, characterized in that, The method further includes: Based on the plurality of first static feature parameters and the plurality of first step state feature parameters, a plurality of first-level weight coefficients are determined; Based on the aforementioned multiple first-step feature parameters, multiple second-level weight coefficients are determined; Based on the multiple primary weight coefficients and the multiple secondary weight coefficients, multiple fusion weight coefficients are determined; Based on the aforementioned multiple fusion weight coefficients, a comprehensive health risk score is determined.

15. The method according to claim 14, characterized in that, The plurality of first static feature parameters include the age parameter of healthy individuals; The determination of multiple first-level weight coefficients based on the multiple first static feature parameters and the multiple first-step feature parameters includes: Calculate the second correlation coefficient between the age parameter of healthy individuals and the plurality of first-step characteristic parameters; Based on the determined multiple second correlation coefficients, the multiple first-level weight coefficients are determined.

16. The method according to claim 14, characterized in that, The determination of multiple secondary weight coefficients based on the multiple gait feature parameters includes: Determine the multiple cluster data corresponding to each step state feature parameter; Based on the multiple cluster data, determine the similarity coefficient within each cluster data and the dissimilarity coefficient between each cluster data; Based on the determined multiple similarity coefficients and multiple dissimilarity coefficients, multiple contour coefficients are determined; Based on the multiple contour coefficients, the multiple secondary weight coefficients are determined.

17. The method according to claim 14, characterized in that, The determination of the comprehensive health risk score based on the multiple fusion weight coefficients includes: Each fusion weight coefficient is transformed and normalized to obtain the processed fusion weight coefficient. The comprehensive health risk score is determined based on each fusion weight coefficient after the processing.

18. An adaptive vital sign-based population profiling method, characterized in that, Applied to a second terminal, the method includes: A second download request is sent to a first terminal, and a third model is received from the first terminal in response to the second download request; wherein, the third model is a model corresponding to the second terminal generated by the first terminal based on the second model; the second model is a model corresponding to the first terminal generated by the first server; the first model is obtained by the first server requesting and calling the second server; the third model includes multiple third static feature parameters and multiple corresponding third dynamic feature parameters corresponding to the first terminal after user correction, and the multiple third dynamic feature parameters include multiple third gait feature parameters; A fourth model corresponding to the second terminal is generated based on the third model; wherein, the fourth model includes multiple fourth static feature parameters and multiple fourth dynamic feature parameters corresponding to the second terminal after user correction, and the multiple fourth dynamic feature parameters include multiple fourth gait feature parameters.

19. The method according to claim 18, characterized in that, The step of generating the fourth model corresponding to the second terminal based on the third model includes: Based on the degree of correlation between each third dynamic feature parameter and the plurality of third gait feature parameters, the third most relevant parameter is determined; Based on the third most relevant parameter, the third key gait feature parameter is determined; Based on the third most relevant parameter and the third key gait feature parameter, the third model is modified to obtain the fourth model.

20. The method according to claim 19, characterized in that, The determination of the third most relevant parameter based on the correlation between each third dynamic feature parameter and the plurality of third gait feature parameters includes: Calculate the third correlation coefficient between each of the third dynamic feature parameters and the plurality of third gait feature parameters; Based on the determined multiple third correlation coefficients, the mean value of the third correlation coefficient for each third dynamic feature parameter is determined, and the third dynamic feature parameter corresponding to the largest mean value of the third correlation coefficient is selected as the third most relevant parameter.

21. The method according to claim 20, characterized in that, The determination of the third key gait feature parameter based on the third most relevant parameter includes: The third gait feature parameter whose third correlation coefficient with the third most relevant parameter is greater than the third threshold is selected as the third key gait feature parameter.

22. The method according to any one of claims 18 to 21, characterized in that, The plurality of third static feature parameters include the age parameter of the user corresponding to the first terminal; the method further includes: Based on the age parameter of the user corresponding to the first terminal, determine the gait feature type of the user corresponding to the first terminal. Select the first step gait feature parameter that is the same as the gait feature type from multiple first step gait feature parameters of healthy people, and set the value range of the first step gait feature parameter that is the same as the gait feature type to the first interval; Based on the parameter values ​​within the first interval, determine the center value and boundary value of the healthy gait feature; The healthy gait feature center value and the healthy gait feature boundary value are sent to the first terminal.

23. The method according to any one of claims 18 to 21, characterized in that, The plurality of third dynamic characteristic parameters include acceleration; the method further includes: Determine if the acceleration is abnormal; if it is abnormal, correct the acceleration to obtain the corrected acceleration. Based on the corrected acceleration, a number of corrected third dynamic characteristic parameters are determined.

24. The method according to claim 23, characterized in that, The step of correcting the acceleration to obtain the corrected acceleration includes: Based on the aforementioned multiple third dynamic characteristic parameters, the gait cycle is divided into a support phase and a swing phase; wherein, the support phase includes a single support phase and a double support phase; the single support phase includes the push-off phase from heel lift-off to toe lift-off on the same side, and the landing phase from toe-off to heel-off on the same side. Calculate the distance between the two feet on the horizontal plane corresponding to the single support; If there are data points corresponding to the distance between the two feet on the horizontal plane that meet the abnormal conditions, the acceleration corresponding to the data points is corrected. The first error correction parameter corresponding to the dual support, the second error correction parameter corresponding to the extension period, the third error correction parameter corresponding to the landing period, and the fourth error correction parameter corresponding to the swing are obtained respectively. Based on the first error correction parameter, the second error correction parameter, the third error correction parameter, and the fourth error correction parameter, the acceleration is corrected to obtain the corrected acceleration.

25. The method according to claim 24, characterized in that, The plurality of third dynamic characteristic parameters also include angular velocity, attitude angle and Euler angle, and the plurality of third gait characteristic parameters include support phase and swing phase; The process of dividing the gait cycle into support and swing phases based on the multiple third dynamic characteristic parameters includes: Based on the Euler angles, toe-off gait events and heel-off gait events are identified; Based on the posture angle, heel-off gait events and flatfoot gait events are identified; Based on toe-off gait events, heel-off gait events, heel-off gait events, and flat-foot gait events, the gait cycle is divided into the support phase and the swing phase.

26. An adaptive vital sign-based crowd profiling device, characterized in that, Applied to a first server, the device includes: The first sending unit is configured to send a first call request to the second server and receive a first model sent by the second server in response to the first call request; wherein, the first model includes multiple first static feature parameters of healthy people and multiple corresponding first dynamic feature parameters, the multiple first dynamic feature parameters including multiple first step feature parameters; The first generation unit is used to generate a second model corresponding to the first terminal based on the first model; wherein the second model includes multiple second static feature parameters of the user corresponding to the first terminal and multiple corresponding second dynamic feature parameters, the multiple second dynamic feature parameters including multiple second gait feature parameters; The first receiving unit is configured to receive a first download request sent by the first terminal, and in response to the first download request, send the second model to the first terminal.

27. An adaptive vital sign-based crowd profiling device, characterized in that, Applied to a second server, the device includes: The second determining unit is used to determine a first model based on the static feature data and corresponding dynamic feature data of a healthy population; wherein the static feature data includes multiple initial static feature parameters, the dynamic feature data includes multiple initial dynamic feature parameters, and the multiple initial dynamic feature parameters include multiple initial gait feature parameters; the first model includes multiple first static feature parameters and corresponding multiple first dynamic feature parameters of a healthy population, and the multiple first dynamic feature parameters include multiple first gait feature parameters. The second receiving unit is configured to receive a first call request sent by the first server, and in response to the first call request, send the first model to the first server.

28. An adaptive vital sign-based crowd profiling device, characterized in that, Applied to a first terminal, the device includes: The third sending unit is configured to send a first download request to the first server and receive a second model sent by the first server in response to the first download request; wherein the second model is a model corresponding to the first terminal generated by the first server based on the first model; the first model is obtained by the first server requesting and calling the second server; the first model includes multiple first static feature parameters and corresponding multiple first dynamic feature parameters of healthy individuals, the multiple first dynamic feature parameters including multiple first gait feature parameters; the second model includes multiple second static feature parameters and corresponding multiple second dynamic feature parameters of the user corresponding to the first terminal, the multiple second dynamic feature parameters including multiple second gait feature parameters. The third generation unit is used to generate a third model corresponding to the first terminal based on the second model; wherein, the third model includes multiple third static feature parameters and multiple third dynamic feature parameters corresponding to the first terminal after user correction, and the multiple third dynamic feature parameters include multiple third gait feature parameters. The third receiving unit is used to receive the second download request sent by the second terminal, and in response to the second download request, send the third model to the second terminal.

29. An adaptive vital sign crowd profiling processing device, characterized in that, Applied to a second terminal, the device includes: The fourth sending unit is configured to send a second download request to the first terminal and receive a third model sent by the first terminal in response to the second download request; wherein, the third model is a model corresponding to the second terminal generated by the first terminal based on the second model; the second model is a model corresponding to the first terminal generated by the first server; the first model is obtained by the first server requesting and calling the second server; the third model includes multiple third static feature parameters and multiple corresponding third dynamic feature parameters corresponding to the first terminal after user correction, and the multiple third dynamic feature parameters include multiple third gait feature parameters; The fourth generation unit is used to generate a fourth model corresponding to the second terminal based on the third model; wherein the fourth model includes multiple fourth static feature parameters and multiple fourth dynamic feature parameters corresponding to the second terminal after user correction, and the multiple fourth dynamic feature parameters include multiple fourth gait feature parameters.

30. A processing apparatus, characterized in that, include: A processor and a memory for storing a computer program, the processor for calling and running the computer program stored in the memory to perform the method as described in any one of claims 1 to 25.

31. A computer-readable storage medium, characterized in that, Used to store a computer program that causes a computer to perform the method as described in any one of claims 1 to 25.

32. A computer program product, characterized in that, It includes computer program instructions that cause a computer to perform the method as described in any one of claims 1 to 25.

Images