Health management method and system based on AI digital human
By building AI digital human and neural network, the problem of difficulty in predicting users' health status in the existing technology when environmental changes is changed is solved, health management and early warning in different environments is realized, and the prevention of users' health status is improved. and preparation ability.
Patent Information
- Application Number
- CN202411269247.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-09-11
AI Technical Summary
The prior art is difficult to predict and manage the user's health status in advance when the environment changes, and it is impossible to predict and prepare for health abnormalities caused by travel or environmental changes in advance.
By building AI digital people, using 3D twin technology and neural networks, users' 3D twins are generated, and users' real-time sign data is moved from the local environment information to the travel location environment information through the digital twin platform, realizing visualization and early warning of sign data .
It realizes early prediction and warning of user health status in different environments, helps users prepare for abnormal health status in advance due to environmental changes, and improves the preventiveness and accuracy of health management.
Smart Images

Figure CN118782273B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent health management technology, and in particular to a health management method and system based on AI digital humans. Background Art
[0002] At present, with the improvement of people's living standards, daily physical health testing has become a common means of health maintenance. In the prior art, some health testing equipment, such as blood pressure meters, blood glucose meters, etc., are often combined with some health monitoring platforms or terminals to achieve daily physical health management and maintenance of users.
[0003] However, this method is limited to the current monitoring and evaluation of the user's health data and whether it meets the health standards. It is difficult to make early predictions and warnings of the user's health status in different environments. It is not conducive to the user to be prepared in advance for abnormal health conditions when the environment changes due to travel or other situations. For example, it is not possible to predict in advance that blood pressure or blood sugar will drop sharply after the environment changes, causing physical discomfort, and prepare in advance. Summary of the invention
[0004] The purpose of the present invention is to provide a health management method and system based on AI digital human to solve the technical problem that it is difficult to predict and manage the user's health status in advance under environmental changes in the prior art.
[0005] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:
[0006] A health management method based on AI digital human, comprising the following steps:
[0007] Based on the user's human shape data, a 3D twin of the user is modeled and generated, and imported into the digital twin platform as the user's AI digital human. The user transmits medical records and real-time vital signs data to the AI digital human.
[0008] The digital twin platform obtains the environmental information of the user's location and the user's travel destination, and uses the medical record data through a neural network to migrate the real-time vital sign data of the AI digital person from the environmental information of the location to the environmental information of the travel destination, thereby obtaining the vital sign data of the AI digital person at the travel destination;
[0009] The digital twin platform visualizes the physical sign data of the travel destination at the corresponding position of the AI digital human, and warns the user of abnormal physical signs during travel.
[0010] As a preferred solution of the present invention, the method for constructing the user's AI digital human includes:
[0011] Step S1, using 3D modeling software to model according to human shape data to create a 3D human shape model;
[0012] Step S2, adding texture and mapping to the 3D humanoid model using image processing software;
[0013] Step S3, using the skeleton plug-in to perform skeleton binding on the 3D human body model;
[0014] Step S4, using animation software to make expressions and movements for the 3D human body model;
[0015] Execute steps S1 to S4 in sequence to obtain the user's 3D twin;
[0016] The rendering engine is used to build a digital twin platform, and the 3D twin is rendered through the digital twin platform to obtain the user's AI digital human.
[0017] As a preferred solution of the present invention, the method for generating the travel location physical sign data of the AI digital person includes:
[0018] The environmental information of the user's location and the location of the user's travel, the environmental information of the digital person's location and the location of the travel;
[0019] The real-time vital sign data and medical record data under the environmental information constraints of the location of the AI digital person are used as input items of the first neural network, and the travel location vital sign data and medical record data under the environmental information constraints of the travel location of the AI digital person are used as output items of the first neural network, so that the first neural network is constructed as a conversion network for converting the AI digital person from the location environment to the travel location environment;
[0020] The vital sign data and medical record data of the travel destination under the environmental information constraints of the AI digital person's travel destination are used as input items of the second neural network, and the real-time vital sign data and medical record data under the environmental information constraints of the AI digital person's location are used as output items of the second neural network, so that the second neural network is constructed as a conversion network for the AI digital person to convert from the travel destination environment to the location environment;
[0021] Constructing a loss function for training the first neural network and the second neural network, and training the first neural network and the second neural network based on the loss function to obtain an environment conversion model of the vital sign data;
[0022] The environment conversion model is:
[0023] [(F+A) out ,(F+B2) out ]=N_net1[(D+A),(D+B1)];
[0024] [(D+A) out ,(D+B1) out ]=N_net2[(F+A),(F+B2)];
[0025] In the formula, (F+A) out The medical record data under the environmental information constraints of the travel destination output by the first neural network, (F+B2) out is the physical sign data of the travel destination under the environmental information constraints of the travel destination output by the first neural network, (D+A) is the medical record data under the environmental information constraints of the location, (D+B1) is the real-time physical sign data under the environmental information constraints of the location, (D+A) out The medical record data under the constraints of the local environment information output by the second neural network, (D+B1) out is the real-time vital sign data under the environmental information constraints of the location output by the second neural network, (F+A) is the medical record data under the environmental information constraints of the travel destination, and (F+B2) is the vital sign data of the travel destination under the environmental information constraints of the travel destination.
[0026] As a preferred solution of the present invention, the loss function for training the first neural network and the second neural network is:
[0027] L = Lx + Ly + Lz;
[0028] Where, Lx=||[(F+A) out -F]-[(D+A) out -D]||1;
[0029] Ly=||N_net2([(F+A) out ,(F+B2) out ])-[(D+A),(D+B1)]||1;
[0030] Lz=||N_net1((D+A) out ,(D+B1) out )-[(F+A),(F+B2)]||1;
[0031] Wherein, L is the loss function, Lx is the function that quantifies the medical record data conversion loss in the environmental conversion process between the first neural network and the second neural network, Ly is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the first neural network, Lz is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the second neural network, F is the environmental information of the travel destination, D is the environmental information of the location, and Lx, Ly and Lz are all L1 norm formulas.
[0032] As a preferred solution of the present invention, the environmental information of the travel destination, the environmental information of the current location, the real-time vital sign data and the vital sign data of the travel destination are normalized.
[0033] As a preferred embodiment of the present invention, by (F+B2) out -F method, separates the travel location vital signs data of AI digital people from the output results of the environment conversion model.
[0034] As a preferred solution of the present invention, the physical sign data of the travel destination is monitored using a preset physical sign warning threshold, and a warning is issued for the physical sign data of the travel destination that exceeds the physical sign warning threshold.
[0035] As a preferred solution of the present invention, the first neural network and the second neural network have the same network structure.
[0036] In a second aspect of the present invention, the present invention provides a health management system based on AI digital human, which is applied to the health management method based on AI digital human, and the system includes:
[0037] A data collection unit is used to collect information based on the user's human shape data, medical record data, real-time vital sign data, environmental information of the user's location, and environmental information of the user's travel location;
[0038] The digital human construction unit is used to model and generate a 3D twin of the user, and import it into the digital twin platform as the user's AI digital human, wherein the user transmits medical record data and real-time vital sign data to the AI digital human;
[0039] The health management unit is used to use the digital twin platform to obtain the environmental information of the user's location and the environmental information of the user's travel destination, and use the medical record data through a neural network to migrate the real-time vital sign data of the AI digital person from the environmental information of the location to the environmental information of the travel destination, so as to obtain the vital sign data of the AI digital person at the travel destination;
[0040] And use the digital twin platform to visualize the physical sign data of the travel location at the corresponding position of the AI digital person, and warn the user of abnormal physical signs during travel.
[0041] As a preferred solution of the present invention, the method for the health management unit to generate the travel location physical sign data of the AI digital person includes:
[0042] The environmental information of the user's location and the location of the user's travel, the environmental information of the digital person's location and the location of the travel;
[0043] The real-time vital sign data and medical record data under the environmental information constraints of the location of the AI digital person are used as input items of the first neural network, and the travel location vital sign data and medical record data under the environmental information constraints of the travel location of the AI digital person are used as output items of the first neural network, so that the first neural network is constructed as a conversion network for converting the AI digital person from the location environment to the travel location environment;
[0044] The vital sign data and medical record data of the travel destination under the environmental information constraints of the AI digital person's travel destination are used as input items of the second neural network, and the real-time vital sign data and medical record data under the environmental information constraints of the AI digital person's location are used as output items of the second neural network, so that the second neural network is constructed as a conversion network for the AI digital person to convert from the travel destination environment to the location environment;
[0045] Constructing a loss function for training the first neural network and the second neural network, and training the first neural network and the second neural network based on the loss function to obtain an environment conversion model of the vital sign data;
[0046] The environment conversion model is:
[0047] [(F+A) out ,(F+B2) out ]=N_net1[(D+A),(D+B1)];
[0048] [(D+A) out ,(D+B1) out ]=N_net2[(F+A),(F+B2)];
[0049] In the formula, (F+A) out The medical record data under the environmental information constraints of the travel destination output by the first neural network, (F+B2) out is the physical sign data of the travel destination under the environmental information constraints of the travel destination output by the first neural network, (D+A) is the medical record data under the environmental information constraints of the location, (D+B1) is the real-time physical sign data under the environmental information constraints of the location, (D+A) out The medical record data under the constraints of the local environment information output by the second neural network, (D+B1) out is the real-time vital sign data under the environmental information constraints of the location output by the second neural network, (F+A) is the medical record data under the environmental information constraints of the travel destination, and (F+B2) is the vital sign data of the travel destination under the environmental information constraints of the travel destination.
[0050] In the third aspect of the present invention, the present invention provides a computer-readable storage medium, in which computer execution instructions are stored. When a processor executes the computer execution instructions, the health management method based on AI digital human is implemented.
[0051] Compared with the prior art, the present invention has the following beneficial effects:
[0052] The present invention uses digital twin technology to build an AI digital human to realize the visualization of health status. It also uses a neural network to migrate vital sign data from the environmental information of the local location to the environmental information of the travel location, thereby realizing early prediction and early warning of the user's health status in different environments. This helps the user to be prepared in advance to deal with abnormal health status when the environment changes due to travel or other situations. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.
[0054] Figure 1 A flow chart of a health management method based on AI digital human provided in an embodiment of the present invention;
[0055] Figure 2 A block diagram of a health management system based on AI digital human provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0057] like Figure 1 As shown, the present invention provides a health management method based on AI digital human, comprising the following steps:
[0058] Based on the user's human shape data, a 3D twin of the user is modeled and generated, and imported into the digital twin platform as the user's AI digital human. The user transmits medical records and real-time vital signs data to the AI digital human.
[0059] The digital twin platform obtains the environmental information of the user's location and the user's travel destination. It uses the neural network and medical record data to transfer the real-time vital signs data of the AI digital person from the environmental information of the location to the environmental information of the travel destination, and obtains the vital signs data of the AI digital person at the travel destination.
[0060] The digital twin platform visualizes the vital signs data of the travel location at the corresponding position of the AI digital person, and warns users of abnormal vital signs during travel.
[0061] The present invention uses digital twin technology to construct a twin digital human with the same human appearance as the user. As a visualization carrier of the user's real vital signs data, the real vital signs data of various parts of the user are displayed on the twin digital human, thereby realizing the visualization of health data and more intuitively observing the user's health status.
[0062] The present invention loads a neural network model on the digital twin platform for converting vital sign data when the environment changes. Before the user travels, the vital sign data in the environmental information of the user's current location is converted to the environmental information of the travel destination, so that the vital sign data when facing the environmental information of the travel destination after arriving at the travel destination can be mastered in advance. This helps users to be prepared in advance to deal with abnormal health conditions when the environment changes due to travel or other situations.
[0063] Therefore, the present invention can realize the conversion of vital sign data under any environment conversion. It is foreseeable that in order to make the vital sign data match the user's personal situation, the present invention takes medical record data into consideration during the conversion, that is, the conversion of vital sign data in different environments not only takes into account the normal physiological functions of the user as a human being. For example, if the travel destination is an altitude sickness area, after the migration conversion, no matter who the user is, it will be predicted in advance that the user will experience physical health discomfort such as altitude sickness at the travel destination, and preparations for altitude sickness prevention must be made in advance. The user's own medical history will also be considered. For example, the travel destination is currently in a stage with more seasonal dust, and the user's medical history shows a history of rhinitis, asthma, etc., then it will be known in advance that vital sign data such as shortness of breath and arrhythmia will occur at the travel destination, and preventive preparations must be made in advance (prepare medicines, masks, etc.). If the vital signs are extremely abnormal, the user will be advised to cancel the trip.
[0064] Therefore, the environment conversion of the present invention can grasp the physical sign data of the travel destination in advance based on the common human characteristics and the user's own personality.
[0065] Vital signs include blood pressure, blood sugar, heart rate, respiratory rate, resting tremor, muscle rigidity,
[0066] Sleep disorders, etc. can be added or deleted during actual use.
[0067] Environmental information can be obtained by obtaining meteorological information from weather forecasts, geological information from geological records, etc., and can be added or deleted during actual use.
[0068] In order to realize the conversion of vital sign data by a neural network in any environment, the present invention constructs an environment conversion model of vital sign data through a neural network. Under the training of two neural networks, it is possible to realize the conversion of vital sign data from the travel environment to the current environment, and from the current environment to the travel environment, and the mutual conversion of the two directions. Therefore, the constructed environment conversion model can realize data conversion in any two environments, thereby improving the robustness of the model in data conversion.
[0069] The present invention uses a function that quantifies the medical record data conversion loss during the environmental conversion process of the first neural network and the second neural network, a function that quantifies the vital sign data conversion loss during the environmental conversion process of the first neural network, and a function that quantifies the vital sign data conversion loss during the environmental conversion process of the second neural network as a total loss function, wherein:
[0070] The user's medical record data is stable data. No matter how the environment changes, it will not be affected. Therefore, in the process of converting from the travel environment to the local environment and from the local environment to the travel environment, it is necessary to keep it stable. This constitutes a function that quantifies the medical record data conversion loss during the environmental conversion of the first neural network and the second neural network, ensuring the stability of the medical record data during the environmental conversion.
[0071] The conversion from the local environment to the travel destination environment is to generate the vital sign data of the travel destination based on the vital sign data of the local location. If the conversion is accurate, the vital sign data of the travel destination are generated based on the vital sign data of the local location, and then converted back (that is, the travel destination environment is converted to the local environment). The local vital sign data generated according to "generating the vital sign data of the travel destination based on the vital sign data of the local location" should be consistent with the original local vital sign data, and the consistency represents the accuracy of the conversion of the local environment to the travel destination environment, and the data error or loss is minimized. Therefore, the present invention quantifies the function of the vital sign data conversion loss during the environment conversion process of the first neural network, thereby ensuring the accuracy of the conversion of the local environment to the travel destination environment.
[0072] Similarly, the conversion from the travel destination environment to the current location environment is to generate the vital signs data of the current location based on the vital signs data of the travel destination. If the conversion is accurate, the vital signs data of the current location are generated based on the vital signs data of the travel destination, and then converted back (i.e., the current location environment is converted to the travel destination environment). The vital signs data of the travel destination generated according to "generating the vital signs data of the current location based on the vital signs data of the travel destination" should be consistent with the original vital signs data of the travel destination, and the consistency represents the accuracy of the conversion of the travel destination environment to the current location environment, and the data error or loss is minimized. Therefore, the present invention quantifies the function of the loss of vital signs data conversion during the environment conversion process of the second neural network, thereby ensuring the accuracy of the conversion of the travel destination environment to the current location environment.
[0073] Therefore, the total loss function can ensure the accuracy of the conversion between any two environments, thereby accurately obtaining the vital sign data after the environment conversion.
[0074] The present invention can not only be used for environment conversion between two locations with different positional relationships (spatial environment conversion), but can also be used for environment conversion at the same location, that is, to predict the vital signs data and health status of the user at the location after a period of time. Specifically, the environmental information of the user's location after a period of time is replaced with the environmental information of the travel destination, thereby realizing the conversion of the environment of the current time period to the environment after a period of time, so as to grasp the health status of the user at the current location in the future time period in advance, and realize the temporal environment conversion and vital signs data grasp.
[0075] The methods for building a user's AI digital human include:
[0076] Step S1, using 3D modeling software (Blender, Maya) to model according to humanoid data to create a 3D humanoid model;
[0077] Step S2, using image processing software (Photoshop) to add texture and mapping to the 3D humanoid model;
[0078] Step S3, using the Unity skeleton plug-in to perform skeleton binding on the 3D human body model;
[0079] Step S4, using animation software (Maya, 3ds max) to make expressions and movements for the 3D human model;
[0080] Execute steps S1 to S4 in sequence to obtain the user's 3D twin;
[0081] Use the rendering engine (Unity, Unreal Engine) to build a digital twin platform, and use the digital twin platform to render the 3D twin to obtain the user's AI digital human.
[0082] The method for generating the travel location physical sign data of AI digital human includes:
[0083] The environmental information of the user's location and the location of the user's travel, the environmental information of the digital person's location and the location of the travel;
[0084] The real-time vital sign data and medical record data under the environmental information constraints of the location of the AI digital person are used as input items of the first neural network, and the travel location vital sign data and medical record data under the environmental information constraints of the travel location of the AI digital person are used as output items of the first neural network, so that the first neural network is constructed as a conversion network for converting the AI digital person from the location environment to the travel location environment;
[0085] The vital sign data and medical record data of the travel destination under the environmental information constraints of the AI digital person's travel destination are used as input items of the second neural network, and the real-time vital sign data and medical record data under the environmental information constraints of the AI digital person's location are used as output items of the second neural network, so that the second neural network is constructed as a conversion network for the AI digital person to convert from the travel destination environment to the location environment;
[0086] Constructing a loss function for training the first neural network and the second neural network, and training the first neural network and the second neural network based on the loss function to obtain an environment conversion model of the vital sign data;
[0087] The environment conversion model is:
[0088] [(F+A) out ,(F+B2) out ]=N_net1[(D+A),(D+B1)];
[0089] [(D+A) out ,(D+B1) out ]=N_net2[(F+A),(F+B2)];
[0090] In the formula, (F+A) out The medical record data under the environmental information constraints of the travel destination output by the first neural network, (F+B2) out is the physical sign data of the travel destination under the environmental information constraints of the travel destination output by the first neural network, (D+A) is the medical record data under the environmental information constraints of the location, (D+B1) is the real-time physical sign data under the environmental information constraints of the location, (D+A) out The medical record data under the constraints of the local environment information output by the second neural network, (D+B1) outis the real-time vital sign data under the environmental information constraints of the location output by the second neural network, (F+A) is the medical record data under the environmental information constraints of the travel destination, and (F+B2) is the vital sign data of the travel destination under the environmental information constraints of the travel destination.
[0091] The loss function for training the first neural network and the second neural network is:
[0092] L = Lx + Ly + Lz;
[0093] Where, Lx=||[(F+A) out -F]-[(D+A) out -D]||1;
[0094] Ly=||N_net2([(F+A) out ,(F+B2) out ])-[(D+A),(D+B1)]||1;
[0095] Lz=||N_net1((D+A) out ,(D+B1) out )-[(F+A),(F+B2)]||1;
[0096] Wherein, L is the loss function, Lx is the function that quantifies the medical record data conversion loss in the environmental conversion process between the first neural network and the second neural network, Ly is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the first neural network, Lz is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the second neural network, F is the environmental information of the travel destination, D is the environmental information of the location, and Lx, Ly and Lz are all L1 norm formulas.
[0097] In order to realize the conversion of vital sign data by a neural network in any environment, the present invention constructs an environment conversion model of vital sign data through a neural network. Under the training of two neural networks, it is possible to realize the conversion of vital sign data from the travel environment to the current environment, and from the current environment to the travel environment, and the mutual conversion of the two directions. Therefore, the constructed environment conversion model can realize data conversion in any two environments, thereby improving the robustness of the model in data conversion.
[0098] The present invention uses a function that quantifies the medical record data conversion loss during the environmental conversion process of the first neural network and the second neural network, a function that quantifies the vital sign data conversion loss during the environmental conversion process of the first neural network, and a function that quantifies the vital sign data conversion loss during the environmental conversion process of the second neural network as a total loss function, wherein:
[0099] The user's medical record data is stable data. No matter how the environment changes, it will not be affected. Therefore, in the process of converting from the travel environment to the local environment and from the local environment to the travel environment, it is necessary to keep it stable. This constitutes a function that quantifies the medical record data conversion loss during the environmental conversion of the first neural network and the second neural network, ensuring the stability of the medical record data during the environmental conversion.
[0100] The conversion from the local environment to the travel destination environment is to generate the vital sign data of the travel destination based on the vital sign data of the local location. If the conversion is accurate, the vital sign data of the travel destination are generated based on the vital sign data of the local location, and then converted back (that is, the travel destination environment is converted to the local environment). The local vital sign data generated according to "generating the vital sign data of the travel destination based on the vital sign data of the local location" should be consistent with the original local vital sign data, and the consistency represents the accuracy of the conversion of the local environment to the travel destination environment, and the data error or loss is minimized. Therefore, the present invention quantifies the function of the vital sign data conversion loss during the environment conversion process of the first neural network, thereby ensuring the accuracy of the conversion of the local environment to the travel destination environment.
[0101] Similarly, the conversion from the travel destination environment to the current location environment is to generate the vital signs data of the current location based on the vital signs data of the travel destination. If the conversion is accurate, the vital signs data of the current location are generated based on the vital signs data of the travel destination, and then converted back (i.e., the current location environment is converted to the travel destination environment). The vital signs data of the travel destination generated according to "generating the vital signs data of the current location based on the vital signs data of the travel destination" should be consistent with the original vital signs data of the travel destination, and the consistency represents the accuracy of the conversion of the travel destination environment to the current location environment, and the data error or loss is minimized. Therefore, the present invention quantifies the function of the loss of vital signs data conversion during the environment conversion process of the second neural network, thereby ensuring the accuracy of the conversion of the travel destination environment to the current location environment.
[0102] Therefore, the total loss function can ensure the accuracy of the conversion between any two environments, thereby accurately obtaining the vital sign data after the environment conversion.
[0103] The environmental information of the travel destination, the environmental information of the current location, the real-time vital sign data and the vital sign data of the travel destination are normalized.
[0104] Pass (F+B2) out -F method, separates the travel location vital signs data of AI digital people from the output results of the environment conversion model.
[0105] The vital sign data of the travel destination are monitored using the preset vital sign warning threshold, and a warning is issued for the vital sign data of the travel destination that exceeds the vital sign warning threshold.
[0106] The first neural network and the second neural network have the same network structure.
[0107] like Figure 2 As shown, in the second aspect of the present invention, the present invention provides a health management system based on AI digital human, which is applied to a health management method based on AI digital human, and the system includes:
[0108] A data collection unit is used to collect information based on the user's human shape data, medical record data, real-time vital sign data, environmental information of the user's location, and environmental information of the user's travel location;
[0109] The digital human construction unit is used to model and generate a 3D twin of the user, and import it into the digital twin platform as the user's AI digital human, wherein the user transmits medical record data and real-time vital sign data to the AI digital human;
[0110] The health management unit is used to use the digital twin platform to obtain the environmental information of the user's location and the user's travel destination. Through the neural network and medical record data, the real-time vital sign data of the AI digital person is transferred from the environmental information of the location to the environmental information of the travel destination, thereby obtaining the vital sign data of the AI digital person at the travel destination;
[0111] The digital twin platform can also be used to visualize the vital signs data of the travel location at the corresponding position of the AI digital human, and warn users of abnormal vital signs during travel.
[0112] The method for the health management unit to generate the travel location physical sign data of the AI digital person includes:
[0113] The environmental information of the user's location and the location of the user's travel, the environmental information of the digital person's location and the location of the travel;
[0114] The real-time vital sign data and medical record data under the environmental information constraints of the location of the AI digital person are used as input items of the first neural network, and the travel location vital sign data and medical record data under the environmental information constraints of the travel location of the AI digital person are used as output items of the first neural network, so that the first neural network is constructed as a conversion network for converting the AI digital person from the location environment to the travel location environment;
[0115] The vital sign data and medical record data of the travel destination under the environmental information constraints of the AI digital person's travel destination are used as input items of the second neural network, and the real-time vital sign data and medical record data under the environmental information constraints of the AI digital person's location are used as output items of the second neural network, so that the second neural network is constructed as a conversion network for the AI digital person to convert from the travel destination environment to the location environment;
[0116] Constructing a loss function for training the first neural network and the second neural network, and training the first neural network and the second neural network based on the loss function to obtain an environment conversion model of the vital sign data;
[0117] The environment conversion model is:
[0118] [(F+A) out ,(F+B2) out ]=N_net1[(D+A),(D+B1)];
[0119] [(D+A) out ,(D+B1) out ]=N_net2[(F+A),(F+B2)];
[0120] In the formula, (F+A) out The medical record data under the environmental information constraints of the travel destination output by the first neural network, (F+B2) out is the physical sign data of the travel destination under the environmental information constraints of the travel destination output by the first neural network, (D+A) is the medical record data under the environmental information constraints of the location, (D+B1) is the real-time physical sign data under the environmental information constraints of the location, (D+A) out The medical record data under the constraints of the local environment information output by the second neural network, (D+B1) out is the real-time vital sign data under the environmental information constraints of the location output by the second neural network, (F+A) is the medical record data under the environmental information constraints of the travel destination, and (F+B2) is the vital sign data of the travel destination under the environmental information constraints of the travel destination.
[0121] The neural network of the present invention can select a GAN network, that is, a generative adversarial network, which is a special neural network structure consisting of two main parts: a generator and a discriminator. The two parts work in an adversarial manner, in which the task of the generator is to create new data (such as images or texts), and the task of the discriminator is to distinguish whether the input data is real or generated by the generator. This adversarial training process enables GAN to learn the complex distribution of real data and generate data that is indistinguishable from the real thing. This structure of GAN enables them to perform well in many applications, so the present invention selects it as the data generation network after the environment conversion, so that the vital sign data is integrated into the travel environment, which is the vital sign condition in the travel environment.
[0122] Generator: Responsible for generating new data from random noise. It improves its generation ability by learning the distribution of real data, with the goal of generating data that can deceive the discriminator.
[0123] Discriminator: Responsible for judging whether the input data is real. It improves its classification ability by learning the characteristics of real data, with the goal of correctly distinguishing real data from fake data generated by the generator.
[0124] In the third aspect of the present invention, the present invention provides a computer-readable storage medium, in which computer execution instructions are stored. When a processor executes the computer execution instructions, a health management method based on AI digital human is implemented.
[0125] The present invention uses digital twin technology to build an AI digital human to realize the visualization of health status. It also uses a neural network to migrate vital sign data from the environmental information of the local location to the environmental information of the travel location, thereby realizing early prediction and early warning of the user's health status in different environments. This helps the user to be prepared in advance to deal with abnormal health status when the environment changes due to travel or other situations.
[0126] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.
Claims
1. A health management method based on AI digital human, characterized in that: The following steps are involved: Based on the user's human shape data, a 3D twin of the user is modeled and generated, and imported into the digital twin platform as the user's AI digital human. The user transmits medical records and real-time vital signs data to the AI digital human. The digital twin platform obtains the environmental information of the user's location and the user's travel destination, and uses the medical record data through a neural network to migrate the real-time vital sign data of the AI digital person from the environmental information of the location to the environmental information of the travel destination, thereby obtaining the vital sign data of the AI digital person at the travel destination; The digital twin platform visualizes the physical sign data of the travel location at the corresponding position of the AI digital human, and warns the user of abnormal physical signs during travel; The methods for building a user's AI digital human include: Step S1, using 3D modeling software to model according to human shape data to create a 3D human shape model; Step S2, adding texture and mapping to the 3D humanoid model using image processing software; Step S3, using the skeleton plug-in to perform skeleton binding on the 3D human body model; Step S4, using animation software to make expressions and movements for the 3D human body model; Execute steps S1 to S4 in sequence to obtain the user's 3D twin; The rendering engine is used to build a digital twin platform, and the 3D twin is rendered through the digital twin platform to obtain the user's AI digital human; The method for generating the travel location physical sign data of the AI digital person includes: The environmental information of the user's location and the location of the user's travel, the environmental information of the digital person's location and the location of the travel; The real-time vital sign data and medical record data under the environmental information constraints of the location of the AI digital person are used as input items of the first neural network, and the travel location vital sign data and medical record data under the environmental information constraints of the travel location of the AI digital person are used as output items of the first neural network, so that the first neural network is constructed as a conversion network for converting the AI digital person from the location environment to the travel location environment; The vital sign data and medical record data of the travel destination under the environmental information constraints of the AI digital person's travel destination are used as input items of the second neural network, and the real-time vital sign data and medical record data under the environmental information constraints of the AI digital person's location are used as output items of the second neural network, so that the second neural network is constructed as a conversion network for the AI digital person to convert from the travel destination environment to the location environment; Constructing a loss function for training the first neural network and the second neural network, and training the first neural network and the second neural network based on the loss function to obtain an environment conversion model of the vital sign data; The environment conversion model is: [(F+A) out ,(F+B2) out ]=N_net1[(D+A),(D+B1)]; [(D+A) out ,(D+B1) out ]=N_net2[(F+A),(F+B2)]; In the formula, (F+A) out The medical record data under the environmental information constraints of the travel destination output by the first neural network, (F+B2) out is the physical sign data of the travel destination under the environmental information constraints of the travel destination output by the first neural network, (D+A) is the medical record data under the environmental information constraints of the location, (D+B1) is the real-time physical sign data under the environmental information constraints of the location, (D+A) out The medical record data under the constraints of the local environment information output by the second neural network, (D+B1) out is the real-time vital sign data under the environmental information constraints of the location output by the second neural network, (F+A) is the medical record data under the environmental information constraints of the travel location, and (F+B2) is the travel location vital sign data under the environmental information constraints of the travel location; The loss function for training the first neural network and the second neural network is: L = Lx + Ly + Lz; Where, Lx=||[(F+A) out -F]-[(D+A) out -D]||1; Ly=||N_net2([(F+A) out ,(F+B2) out ])-[(D+A),(D+B1)]||1; Lz=||N_net1((D+A) out ,(D+B1) out )-[(F+A),(F+B2)]||1; Wherein, L is the loss function, Lx is the function that quantifies the medical record data conversion loss in the environmental conversion process between the first neural network and the second neural network, Ly is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the first neural network, Lz is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the second neural network, F is the environmental information of the travel destination, D is the environmental information of the location, and Lx, Ly and Lz are all L1 norm formulas.
2. A health management method based on AI digital human according to claim 1, characterized in that: The environmental information of the travel destination, the environmental information of the current location, the real-time vital sign data and the vital sign data of the travel destination are normalized.
3. The health management method based on AI digital human according to claim 1 is characterized by: Pass (F+B2) out -F method, separates the travel location vital signs data of AI digital people from the output results of the environment conversion model.
4. The health management method based on AI digital human according to claim 1, characterized in that: The vital sign data of the travel destination are monitored using the preset vital sign warning threshold, and a warning is issued for the vital sign data of the travel destination that exceeds the vital sign warning threshold.
5. The health management method based on AI digital human according to claim 1, characterized in that: The first neural network and the second neural network have the same network structure.
6. A health management system based on AI digital human, characterized in that: A health management method based on AI digital human applied to any one of claims 1 to 5, the system comprising: A data collection unit is used to collect information based on the user's human shape data, medical record data, real-time vital sign data, environmental information of the user's location, and environmental information of the user's travel location; The digital human construction unit is used to model and generate a 3D twin of the user, and import it into the digital twin platform as the user's AI digital human, wherein the user transmits medical record data and real-time vital sign data to the AI digital human; The health management unit is used to use the digital twin platform to obtain the environmental information of the user's location and the environmental information of the user's travel destination, and use the medical record data through a neural network to migrate the real-time vital sign data of the AI digital person from the environmental information of the location to the environmental information of the travel destination, so as to obtain the vital sign data of the AI digital person at the travel destination; and using the digital twin platform to visualize the physical sign data of the travel location at the corresponding position of the AI digital human, and to warn the user of abnormal physical signs during travel; The methods for building a user's AI digital human include: Step S1, using 3D modeling software to model according to human shape data to create a 3D human shape model; Step S2, adding texture and mapping to the 3D humanoid model using image processing software; Step S3, using the skeleton plug-in to perform skeleton binding on the 3D human body model; Step S4, using animation software to make expressions and movements for the 3D human body model; Execute steps S1 to S4 in sequence to obtain the user's 3D twin; The rendering engine is used to build a digital twin platform, and the 3D twin is rendered through the digital twin platform to obtain the user's AI digital human; The method for generating the travel location physical sign data of the AI digital person includes: The environmental information of the user's location and the location of the user's travel, the environmental information of the digital person's location and the location of the travel; The real-time vital sign data and medical record data under the environmental information constraints of the location of the AI digital person are used as input items of the first neural network, and the travel location vital sign data and medical record data under the environmental information constraints of the travel location of the AI digital person are used as output items of the first neural network, so that the first neural network is constructed as a conversion network for converting the AI digital person from the location environment to the travel location environment; The vital sign data and medical record data of the travel destination under the environmental information constraints of the AI digital person's travel destination are used as input items of the second neural network, and the real-time vital sign data and medical record data under the environmental information constraints of the AI digital person's location are used as output items of the second neural network, so that the second neural network is constructed as a conversion network for the AI digital person to convert from the travel destination environment to the location environment; Constructing a loss function for training the first neural network and the second neural network, and training the first neural network and the second neural network based on the loss function to obtain an environment conversion model of the vital sign data; The environment conversion model is: [(F+A) out ,(F+B2) out ]=N_net1[(D+A),(D+B1)]; [(D+A) out ,(D+B1) out ]=N_net2[(F+A),(F+B2)]; In the formula, (F+A) out The medical record data under the environmental information constraints of the travel destination output by the first neural network, (F+B2) out is the physical sign data of the travel destination under the environmental information constraints of the travel destination output by the first neural network, (D+A) is the medical record data under the environmental information constraints of the location, (D+B1) is the real-time physical sign data under the environmental information constraints of the location, (D+A) out The medical record data under the constraints of the local environment information output by the second neural network, (D+B1) out is the real-time vital sign data under the environmental information constraints of the location output by the second neural network, (F+A) is the medical record data under the environmental information constraints of the travel location, and (F+B2) is the travel location vital sign data under the environmental information constraints of the travel location; The loss function for training the first neural network and the second neural network is: L = Lx + Ly + Lz; Where, Lx=||[(F+A) out -F]-[(D+A) out -D]||1; Ly=||N_net2([(F+A) out ,(F+B2) out ])-[(D+A),(D+B1)]||1; Lz=||N_net1((D+A) out ,(D+B1) out )-[(F+A),(F+B2)]||1; Wherein, L is the loss function, Lx is the function that quantifies the medical record data conversion loss in the environmental conversion process between the first neural network and the second neural network, Ly is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the first neural network, Lz is the function that quantifies the vital sign data conversion loss in the environmental conversion process between the second neural network, F is the environmental information of the travel destination, D is the environmental information of the location, and Lx, Ly and Lz are all L1 norm formulas.
