Health state detection method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a health state detection method and device, equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring health data of a user, wherein the health data comprises working data, and sleep data and motion data acquired based on wearable equipment; performing time alignment processing on the working data, the sleep data and the motion data to obtain target health data; performing feature extraction processing on the target health data to obtain a health feature sequence; and performing health state detection on the health feature sequence through a health state detection model based on a Transform architecture to obtain a health state index of the user. According to the embodiment of the invention, the accuracy of health state detection can be improved, so that an accurate and quantifiable comprehensive health state index is provided for a user.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a health status detection method, device, equipment and storage medium. Background Art

[0002] As living standards improve, people pay more and more attention to health. Some health management applications have been launched on the market to help users monitor their health status. However, these health management applications simply count and display users' periodic sleep, exercise and other data, and cannot provide users with accurate and quantifiable comprehensive health status data.

[0003] Therefore, how to improve the accuracy of health status detection to provide users with accurate and quantifiable comprehensive health status data has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose a health status detection method, device, equipment and storage medium, aiming to improve the accuracy of health status detection so as to provide users with an accurate and quantifiable comprehensive health status index.

[0005] To achieve the above object, a first aspect of an embodiment of the present application provides a health status detection method, the method comprising:

[0006] Acquiring health data of the user, wherein the health data includes work data, sleep data and exercise data collected based on a wearable device;

[0007] Perform time alignment processing on work data, sleep data and exercise data to obtain target health data;

[0008] Performing feature extraction processing on the target health data to obtain a health feature sequence;

[0009] The health status detection model based on the Transformer architecture is used to perform health status detection on the health feature sequence to obtain the health status index of the user.

[0010] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a health status detection device, the device comprising:

[0011] A health data acquisition module, used to acquire the user's health data, the health data including work data, sleep data and exercise data collected based on wearable devices;

[0012] A time alignment module, used for performing time alignment processing on the work data, the sleep data and the exercise data to obtain target health data;

[0013] A health feature extraction module, used to perform feature extraction processing on the target health data to obtain a health feature sequence;

[0014] The health status detection module is used to perform health status detection on the health feature sequence through a health status detection model based on the Transformer architecture to obtain the health status index of the user.

[0015] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0016] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0017] The health status detection method, device, equipment and storage medium proposed in the present application, the health status detection method obtains the user's health data, the health data includes work data, sleep data and exercise data collected based on wearable devices; time-aligns the work data, sleep data and exercise data to obtain target health data; performs feature extraction on the target health data to obtain a health feature sequence; and performs health status detection on the health feature sequence through a health status detection model based on the Transformer architecture to obtain the user's health status index. In this way, on the one hand, by time-aligning multimodal work data, sleep data, and exercise data, the consistency of multimodal data in time granularity is ensured, which can provide a more comprehensive and accurate analysis basis for health status detection; on the other hand, through the health status detection model based on the Transformer architecture, health status detection can effectively capture the global long-term dependencies in the health feature sequence from the three dimensions of work, sleep, and exercise at the same time, so as to better understand the contribution of health feature vectors at different time steps in the health feature sequence to the overall health status, and realize more refined and faster automated comprehensive health status detection, thereby improving the accuracy of health status detection, so as to provide users with an accurate and quantifiable comprehensive health status index. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of a health status detection method provided by an embodiment of the present application;

[0019] Figure 2 is another flow chart of the health status detection method provided in an embodiment of the present application;

[0020] Figure 3 is another flow chart of the health status detection method provided in the embodiment of the present application;

[0021] Figure 4 is a schematic block diagram of a health status detection device provided in an embodiment of the present application;

[0022] Figure 5 It is a schematic block diagram of the structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0024] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0026] First, some nouns involved in this application are analyzed:

[0027] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0028] Natural language processing (NLP): NLP uses computers to process, understand and apply human languages ​​(such as Chinese, English, etc.). NLP is a branch of artificial intelligence and an interdisciplinary subject between computer science and linguistics. It is often referred to as computational linguistics. Natural language processing includes grammatical analysis, semantic analysis, and text understanding. Natural language processing is often used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information intent recognition, information extraction and filtering, text classification and clustering, public opinion analysis and opinion mining. It involves data mining related to language processing, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research, and linguistic research related to language computing.

[0029] Based on this, the embodiments of the present application provide a health status detection method, device, equipment and storage medium, which realizes more refined automated comprehensive health status detection through a health status detection model based on the Transformer architecture, aiming to improve the accuracy of health status detection and provide users with an accurate and quantifiable comprehensive health status index.

[0030] The health status detection method, device, equipment and storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the health status detection method in the embodiments of the present application is described.

[0031] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0032] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0033] The health status detection method provided in the embodiment of the present application relates to the field of artificial intelligence technology. The health status detection method provided in the embodiment of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a vehicle-mounted terminal, a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the health status detection method, etc., but is not limited to the above forms.

[0034] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer computer devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0035] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0036] Figure 1 is an optional flow chart of the health status detection method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S104.

[0037] Step S101, obtaining the user's health data, the health data including work data, sleep data and exercise data collected based on wearable devices.

[0038] Health data includes but is not limited to multimodal work data, sleep data and exercise data collected by wearable devices.

[0039] Wearable devices include smart bracelets, smart watches, etc. Wearable devices can be equipped with an electrocardiogram (ECG) sensor for detecting heart rate, a temperature sensor for detecting body temperature, a blood pressure sensor for detecting blood pressure, a photoplethysmography (PPG) sensor for detecting blood oxygen, an accelerometer for detecting motion parameters, a flow sensor for detecting respiratory signals, and / or an electroencephalogram (EEG) sensor for detecting brain wave signals.

[0040] Among them, work data includes but is not limited to terminal usage data of the user during work and work physiological response data collected through wearable devices.

[0041] Terminal usage data includes, but is not limited to, work behavior data such as office software usage monitoring values, mouse movement monitoring values, and / or keyboard input monitoring values.

[0042] Work physiological response data include, but are not limited to, physiological response data such as heart rate, body temperature, blood pressure, blood oxygen and / or exercise steps of the user during work. The user can wear a wearable device during work, so as to collect work physiological response data through the corresponding sensors of the wearable device.

[0043] The sleep data includes, but is not limited to, physiological response data such as the user's heart rate, breathing signal, brain wave signal and / or body temperature during sleep. The user may wear a wearable device during sleep, so that the sleep data is collected through the corresponding sensors of the wearable device.

[0044] The motion data includes, but is not limited to, speed data, acceleration data, heart rate signal data, energy consumption data (such as calorie consumption value), and displacement data (such as running distance) during the user's exercise. The user can wear a wearable device during exercise, so that the motion data can be collected through the corresponding sensors of the wearable device.

[0045] In this way, by obtaining detailed and comprehensive health data, a more comprehensive analysis basis can be provided for subsequent health status testing.

[0046] Step S102, time alignment processing is performed on the work data, sleep data and exercise data to obtain target health data.

[0047] Taking into account that work data, sleep data and motion data have different time scales, for example, work data may be collected once an hour, sleep data may be monitored once a minute, and motion data may be monitored once a second, therefore, time alignment processing is performed on the work data, sleep data and motion data in a unified time format, that is, the work data, sleep data and motion data are converted into a unified time granularity (such as per minute) for alignment.

[0048] Exemplarily, the work data, sleep data and motion data can be resampled first to ensure the consistency of the resampled work data, sleep data and motion data in time granularity, and then the resampled work data, sleep data and motion data can be interpolated to ensure the consistency of the interpolated work data, sleep data and motion data in the same time series.

[0049] The work data, sleep data and exercise data after time alignment constitute the target health data.

[0050] It can be understood that the target health data is presented in the form of a time series. Specifically, for the terminal usage data after time alignment processing, each time step in the time series corresponds to an office software usage monitoring value, a mouse movement monitoring value and / or a keyboard input monitoring value, etc.; for the work physiological response data after time alignment processing, each time step in the time series corresponds to a heart rate, a body temperature, a blood pressure, a blood oxygen and / or a number of movement steps, etc. Similarly, for the sleep data after time alignment processing, each time step in the time series corresponds to a heart rate, a breathing signal, a brain wave signal and / or a body temperature, etc. Similarly, for the motion data after time alignment processing, each time step in the time series corresponds to a motion speed, a motion acceleration, a heart rate signal, an energy consumption value and / or a motion displacement.

[0051] The time series is determined by the total time range of health data collection. For example, if health data is continuously collected from 00:00 to 24:00, then the time series is from 00:00 to 24:00. The time step refers to the resampling time window. For example, a resampling time window of every 5 minutes is a time step.

[0052] Step S103, performing feature extraction processing on the target health data to obtain a health feature sequence.

[0053] The target health data can be processed for feature extraction to obtain a health feature sequence to facilitate the understanding and processing of the health status detection model.

[0054] It can be understood that the health feature sequence includes a time series and a health feature vector, and the work feature vector, sleep feature vector and motion feature vector of each time step in the time series constitute a health feature vector.

[0055] Among them, the work characteristic vector includes but is not limited to the terminal use characteristic vector and the work physiological response characteristic vector. The terminal use characteristic vector includes but is not limited to the office software use characteristic vector (such as the office software use time), the mouse movement characteristic vector (such as the number of mouse clicks) and / or the keyboard input characteristic vector (such as the keyboard input frequency), etc.; the work physiological response characteristic vector includes but is not limited to the heart rate characteristic vector (such as the average heart rate), the body temperature characteristic vector (such as the body temperature fluctuation range), the blood pressure characteristic vector (such as the blood pressure fluctuation range), the blood oxygen characteristic vector (such as the blood oxygen fluctuation range) and / or the exercise step characteristic vector (such as the average exercise step number), etc. during work.

[0056] Among them, the sleep feature vector includes but is not limited to a heart rate feature vector (such as average heart rate) during sleep, a breathing feature vector (such as breathing frequency), an EEG feature vector (such as a brain wave frequency band used to reflect brain activity) and / or a body temperature feature vector, etc.

[0057] Among them, the motion characteristic vector includes but is not limited to a speed characteristic vector (such as average speed), an acceleration characteristic vector (such as average acceleration), a heart rate characteristic vector (such as heart rate standard deviation), an energy consumption characteristic vector (such as total calorie consumption) and / or a displacement characteristic vector (such as average motion displacement), etc.

[0058] Step S104, performing health status detection on the health feature sequence through a health status detection model based on the Transformer architecture to obtain a health status index of the user.

[0059] Among them, the health status detection model based on the Transformer architecture is obtained by pre-training the preset large language model.

[0060] For example, the large language model can be a natural language processing (NLP) model that is good at processing complex time series data, such as a BERT model based on the Transformer architecture or a GPT series model (such as the GPT-2 model). The health status detection model can capture the implicit potential associations between the work feature vector, sleep feature vector, and exercise feature vector at each time step in the health feature sequence, and output a clear health status index.

[0061] Based on this, the health status detection model can be used to perform health status detection on the health feature sequence, so that the health status detection model can effectively understand the correlation between the health feature vector of each time step in the health feature sequence and the health feature vectors of other time steps through the multi-head attention mechanism, thereby capturing the global dependency between the health feature vectors of different time steps, and being able to simultaneously consider the work feature vectors, sleep feature vectors and exercise feature vectors of different time steps to judge their importance to the health status, thereby obtaining an accurate health status index.

[0062] Exemplarily, a health status index such as a health status score (eg, 1-10 points) or a health status grade (eg, low, medium, high) is an indicator used to accurately quantify the health status.

[0063] In this way, through the health status detection model based on the Transformer architecture, high-precision health status detection is achieved, which can improve the efficiency, comprehensiveness and convenience of health status detection.

[0064] The health status detection method provided in the above embodiment obtains the health data of the user, and the health data includes work data, sleep data and motion data collected based on wearable devices; performs time alignment processing on the work data, sleep data and motion data to obtain target health data; performs feature extraction processing on the target health data to obtain a health feature sequence; and performs health status detection on the health feature sequence through a health status detection model based on the Transformer architecture to obtain the health status index of the user. In this way, on the one hand, by performing time alignment processing on multimodal work data, sleep data and motion data, the consistency of multimodal data in time granularity is ensured, which can provide a more comprehensive and accurate analysis basis for health status detection; on the other hand, by performing health status detection on a health status detection model based on the Transformer architecture, the global long-term dependency in the health feature sequence can be effectively captured from the three dimensions of work, sleep and motion at the same time, so as to better understand the contribution of health feature vectors at different time steps in the health feature sequence to the overall health status, and realize more refined and faster automated comprehensive health status detection, thereby improving the accuracy of health status detection, so as to provide users with an accurate and quantifiable comprehensive health status index.

[0065] In step S102 of some embodiments, the target health data can be cleaned to obtain clean health data, which includes clean work data, clean sleep data and clean motion data; the clean work data can be feature extracted to obtain a work feature sequence; the clean sleep data can be feature extracted to obtain a sleep feature sequence; the clean motion data can be feature extracted to obtain a motion feature sequence; the work feature sequence, the sleep feature sequence and the motion feature sequence can be fused to obtain a health feature sequence, which includes a time series and a health feature vector.

[0066] In order to ensure the reliability of the input of the health status detection model, the target health data can be cleaned to obtain clean health data, which includes clean work data, clean sleep data and clean exercise data.

[0067] In order to facilitate the understanding of the health status detection model, the cleaning work data can be subjected to feature extraction processing to obtain a work feature sequence, wherein the work feature sequence includes a terminal use feature sequence and a work physiological response feature sequence, the terminal use feature sequence includes a time series and a terminal use feature vector, and the work physiological response feature sequence includes a time series and a work physiological response feature vector.

[0068] The clean sleep data may also be subjected to feature extraction processing to obtain a sleep feature sequence, which includes a time series and a sleep feature vector.

[0069] The clean motion data may also be subjected to feature extraction processing to obtain a motion feature sequence, which includes a time series and a motion feature vector.

[0070] The work feature sequence, the sleep feature sequence and the motion feature sequence are fused to obtain a health feature sequence, which includes a time series and a health feature vector. For example, by concatenating the terminal usage feature vector, the work physiological response feature vector, the sleep feature vector and the motion feature vector at each time step in the time series, a comprehensive health feature vector for each time step in the time series can be obtained.

[0071] Through the above method, important health feature sequences that are more helpful for health status detection can be accurately extracted, so that the detection efficiency and detection accuracy of the health status detection model are improved.

[0072] In some embodiments, the target health data is cleaned to obtain clean health data, and the target health data may be denoised to obtain denoised target health data; the denoised target health data may be filled with missing values ​​to obtain filled target health data; the filled target health data may be corrected for outliers to obtain corrected target health data; and the corrected target health data may be standardized to obtain clean health data.

[0073] Exemplarily, the cleaning process mainly includes denoising, missing value filling, outlier correction and standardization.

[0074] Considering that the target health data is usually affected by the noise of the user's working environment, the target health data may be denoised, for example, by filtering the target health data to obtain the denoised target health data.

[0075] The denoised target health data may also be processed for missing value filling, for example, by using mean, median filling, or interpolation to fill missing values ​​to obtain filled target health data.

[0076] The filled target health data may also be corrected for outliers, for example, by detecting outliers through box plots, standard deviations, etc., and adjusting the outliers to a normal range, or deleting the outliers to obtain corrected target health data.

[0077] The corrected target health data may also be standardized, for example, the corrected target health data may be normalized to obtain target clean health data.

[0078] In this way, by carefully cleaning the target health data, higher quality clean health data can be provided.

[0079] See also Figure 2 In some embodiments, step S104 may include but is not limited to steps S201 to S204.

[0080] Step S201, inputting the health feature sequence into the health status detection model;

[0081] Step S202, encoding the health feature sequence to obtain a coding vector sequence;

[0082] Step S203, using a multi-head attention mechanism, the encoding vector sequence is processed to capture the global feature dependency relationship to obtain the global long-term dependency feature;

[0083] Step S204, performing health status prediction processing on the global long-term dependency features to obtain a health status index.

[0084] First, the health feature sequence is used as the input of the health status detection model; then, in the health status detection model, the health feature sequence is encoded to obtain a coding vector sequence; then, through the multi-head attention mechanism, the coding vector sequence is processed to capture the global feature dependency relationship, so that the health status detection model can simultaneously capture the potential long-term dependency relationship between the work feature vector, the sleep feature vector and the exercise feature vector from the work, sleep and exercise dimensions, thereby obtaining the global long-term dependency feature; finally, the global long-term dependency feature is processed for health status prediction to obtain a health status index with higher accuracy.

[0085] In step S202 of some embodiments, the health feature sequence may be embedded to obtain an embedded vector sequence; and the embedded vector sequence may be positionally encoded to obtain an encoded vector sequence.

[0086] The health feature sequence is encoded. Specifically, the health feature sequence can be first embedded through the embedding layer of the health status detection model, so as to map the health feature sequence to a high-dimensional space and obtain an embedded vector sequence; then the embedded vector sequence is positionally encoded to obtain an encoded vector sequence, thereby ensuring that the health status detection model can understand the temporal relationship of the health feature vectors in the health feature sequence.

[0087] In step S204 of some embodiments, the global long-term dependency features can be transformed to obtain deep-level global long-term dependency features; the deep-level global long-term dependency features can be pooled to obtain high-level global long-term dependency features; the high-level global long-term dependency features can be fully connected to obtain a health status index.

[0088] The health status prediction processing is performed on the global long-term dependent features. Specifically, the global long-term dependent features can be first processed by nonlinear transformation through the feed-forward neural network (FFN) of the health status detection model to obtain more complex deep-level global long-term dependent features; the deep-level global long-term dependent features are then pooled through the pooling layer of the health status detection model to obtain high-level global long-term dependent features that can represent the overall health status; finally, the high-level global long-term dependent features are fully connected through the fully connected layer of the health status detection model to obtain the health status index for output, thereby ensuring the validity and accuracy of the health status index.

[0089] Before step S104 of some embodiments, the large language model can be trained in advance. Taking BERT as an example, the training process of the large language model includes: obtaining health feature sequence samples of health data samples and actual health status index of health data samples; inputting health feature sequence samples into the BERT model for iterative training, effectively learning the predicted health status index of health data samples; calculating the loss between the actual health status index and the predicted health status index, thereby calculating the loss function of the BERT model according to the loss, and then updating the parameters of the BERT model according to the loss function, until the number of iterations is greater than the preset number threshold, stop training the BERT model, and obtain a trained health status detection model, which has strong robustness, adaptability, generalization ability and stability, and is used for more sophisticated and faster automated health status detection.

[0090] See also Figure 3 In some embodiments, after step S104, it may include but is not limited to steps S301 to S302.

[0091] Step S301: Generate a health report for the user, the health report including health status index, health management strategy information and schedule information;

[0092] Step S302: Visually display the health report.

[0093] After the user's health status index is detected by the health status detection model, a personalized health report can be generated for the user, wherein the health report includes the health status index, and health management strategy information and schedule information generated based on health data for the health status index.

[0094] For example, if the user's health status index is 4 points, the health management strategy information may include "it is recommended to do a certain amount of exercise" or "please pay attention to adjusting your work and rest schedule"; the schedule information may include "please try to turn on the computer to work in the morning, and avoid working overtime after 10 o'clock in the evening to avoid affecting your sleep" or "please adjust your exercise time", so that the schedule information is dynamically linked to the health status.

[0095] The health report can be displayed visually in a striking manner, allowing users to intuitively view their health status index so that users can understand their health status in a timely manner and make necessary adjustments in a timely manner.

[0096] Therefore, the health status detection method provided in the above embodiment realizes intelligent detection of health status in an automated manner, can perform continuous and periodic detection, and can continuously and periodically provide users with adapted personalized health reports, thereby helping users maintain their health more effectively.

[0097] See also Figure 4 , Figure 4 It is a schematic block diagram of a health status detection device provided in an embodiment of the present application. The health status detection device can be configured in a computer device to execute the aforementioned health status detection method.

[0098] like Figure 4 As shown, the health status detection device 400 includes: a health data acquisition module 401, a time alignment module 402, a health feature extraction module 403 and a health status detection module 404.

[0099] The health data acquisition module 401 is used to acquire the user's health data, where the health data includes work data, sleep data and exercise data collected based on the wearable device;

[0100] A time alignment module 402 is used to perform time alignment processing on the work data, the sleep data and the exercise data to obtain target health data;

[0101] The health feature extraction module 403 is used to perform feature extraction processing on the target health data to obtain a health feature sequence;

[0102] The health status detection module 404 is used to perform health status detection on the health feature sequence through a health status detection model based on the Transformer architecture to obtain the health status index of the user.

[0103] In one embodiment, the health status detection module 404 is further used to:

[0104] Inputting the health feature sequence into the health status detection model;

[0105] Encoding the health feature sequence to obtain a coding vector sequence;

[0106] Through a multi-head attention mechanism, the encoding vector sequence is processed by capturing the global feature dependency relationship to obtain a global long-term dependency feature;

[0107] The health status prediction process is performed on the global long-term dependent feature to obtain the health status index.

[0108] In one embodiment, the health status detection module 404 is further used to:

[0109] Embedding the health feature sequence to obtain an embedding vector sequence;

[0110] The embedded vector sequence is positionally encoded to obtain an encoded vector sequence.

[0111] In one embodiment, the health status detection module 404 is further used to:

[0112] Transforming the global long-term dependency feature to obtain a deep-level global long-term dependency feature;

[0113] Pooling the deep-level global long-term dependency features to obtain high-level global long-term dependency features;

[0114] The high-level global long-term dependency features are subjected to full connection operation processing to obtain the health status index.

[0115] In one embodiment, the health feature extraction module 403 is further used to:

[0116] Cleaning the target health data to obtain clean health data, where the clean health data includes clean work data, clean sleep data, and clean exercise data;

[0117] Performing feature extraction processing on the cleaning work data to obtain a work feature sequence;

[0118] Performing feature extraction processing on the cleaning sleep data to obtain a sleep feature sequence;

[0119] Performing feature extraction processing on the cleaning motion data to obtain a motion feature sequence;

[0120] The working feature sequence, the sleeping feature sequence and the motion feature sequence are fused to obtain the health feature sequence, which includes a time series and a health feature vector.

[0121] In one embodiment, the health feature extraction module 403 is further used to:

[0122] Performing denoising processing on the target health data to obtain denoised target health data;

[0123] Performing missing value filling processing on the denoised target health data to obtain filled target health data;

[0124] Performing outlier correction processing on the filled target health data to obtain corrected target health data;

[0125] The corrected target health data is standardized to obtain the clean health data.

[0126] In one embodiment, the health status detection device 400 further includes a health report generating module and a health report displaying module.

[0127] The health report generating module is used to generate a health report of the user, wherein the health report includes the health status index, health management strategy information and schedule information;

[0128] The health report display module is used to visually display the health report.

[0129] Among them, each module in the above-mentioned health status detection device 400 corresponds to each step in the above-mentioned health status detection method embodiment, and its functions and implementation processes are no longer repeated here one by one.

[0130] The health status detection device 400 can execute the health status detection method provided in the embodiment of the present application, and therefore can achieve the beneficial effects that can be achieved by the health status detection method provided in the embodiment of the present application. Please refer to the previous embodiment for details and will not be repeated here.

[0131] The method and apparatus of the present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0132] Exemplarily, the above method and apparatus may be implemented in the form of a computer program. The computer program may be implemented in the form of a computer program. Figure 5 Runs on the computer device shown.

[0133] See also Figure 5 , Figure 5 It is a schematic block diagram of the structure of a computer device provided in an embodiment of the present application.

[0134] See also Figure 5 The computer device includes a processor and a memory connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0135] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0136] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any health status detection method.

[0137] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0138] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0139] Acquiring health data of the user, wherein the health data includes work data, sleep data and exercise data collected based on a wearable device;

[0140] Perform time alignment processing on work data, sleep data and exercise data to obtain target health data;

[0141] Performing feature extraction processing on the target health data to obtain a health feature sequence;

[0142] The health status detection model based on the Transformer architecture is used to perform health status detection on the health feature sequence to obtain the health status index of the user.

[0143] In one embodiment, when the processor implements the health status detection model based on the Transformer architecture to perform health status detection on the health feature sequence to obtain the health status index of the user, it is used to implement:

[0144] Inputting the health feature sequence into the health status detection model;

[0145] Encoding the health feature sequence to obtain a coding vector sequence;

[0146] Through a multi-head attention mechanism, the encoding vector sequence is processed by capturing the global feature dependency relationship to obtain a global long-term dependency feature;

[0147] The health state prediction processing is performed on the global long-term dependent feature to obtain the health state index. In one embodiment, when the processor implements the encoding processing of the health feature sequence to obtain the encoding vector sequence, it is used to implement:

[0148] Embedding the health feature sequence to obtain an embedding vector sequence;

[0149] The embedded vector sequence is positionally encoded to obtain an encoded vector sequence.

[0150] In one embodiment, when the processor performs health status prediction processing on the global long-term dependent feature to obtain the health status index, it is used to implement:

[0151] Transforming the global long-term dependency feature to obtain a deep-level global long-term dependency feature;

[0152] Pooling the deep-level global long-term dependency features to obtain high-level global long-term dependency features;

[0153] The high-level global long-term dependency features are subjected to full connection operation processing to obtain the health status index.

[0154] In one embodiment, when the processor performs feature extraction processing on the target health data to obtain a health feature sequence, it is used to implement:

[0155] Cleaning the target health data to obtain clean health data, where the clean health data includes clean work data, clean sleep data, and clean exercise data;

[0156] Performing feature extraction processing on the cleaning work data to obtain a work feature sequence;

[0157] Performing feature extraction processing on the cleaning sleep data to obtain a sleep feature sequence;

[0158] Performing feature extraction processing on the cleaning motion data to obtain a motion feature sequence;

[0159] The working feature sequence, the sleeping feature sequence and the motion feature sequence are fused to obtain the health feature sequence, which includes a time series and a health feature vector.

[0160] In one embodiment, when the processor implements the cleaning process on the target health data to obtain the clean health data, it is used to implement:

[0161] Performing denoising processing on the target health data to obtain denoised target health data;

[0162] Performing missing value filling processing on the denoised target health data to obtain filled target health data;

[0163] Performing outlier correction processing on the filled target health data to obtain corrected target health data;

[0164] The corrected target health data is standardized to obtain the clean health data.

[0165] In one embodiment, after implementing the health status detection model based on the Transformer architecture to perform health status detection on the sleep feature subsequence to obtain the health status index of the user, the processor is further used to implement:

[0166] Generate a health report for the user, the health report including the health status index, health management strategy information and schedule information;

[0167] The health report is visually displayed.

[0168] The computer device can execute the health status detection method provided in the embodiment of the present application, and therefore can achieve the beneficial effects that can be achieved by the health status detection method provided in the embodiment of the present application. Please refer to the previous embodiments for details and will not be repeated here.

[0169] An embodiment of the present application also provides a computer-readable storage medium.

[0170] The computer-readable storage medium of the present application stores a computer program, and when the computer program is executed by a processor, the steps of the health status detection method as described above are implemented.

[0171] The computer-readable storage medium may be an internal storage unit of the health status detection device or computer device described in the aforementioned embodiment, such as a hard disk or memory of the health status detection device or computer device. The computer-readable storage medium may also be an external storage device of the health status detection device or computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital card (Secure Digital Card, SD Card), a flash card (Flash Card), etc., equipped on the health status detection device or computer device.

[0172] Since the computer program stored in the computer-readable storage medium can execute any health status detection method provided in the embodiments of the present application, the beneficial effects that can be achieved by any health status detection method provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0173] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0174] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0175] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0176] The above description is only a specific implementation mode of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or substitutions within the technical scope disclosed in the present application, and these modifications or substitutions should be included in the protection scope of the present application.

Claims

1. A health status detection method, characterized in that: The method comprises: Acquiring health data of the user, wherein the health data includes work data, sleep data and exercise data collected based on a wearable device; Performing time alignment processing on the work data, the sleep data, and the exercise data to obtain target health data; Performing feature extraction processing on the target health data to obtain a health feature sequence; The health status detection model based on the Transformer architecture is used to perform health status detection on the health feature sequence to obtain the health status index of the user.

2. The method according to claim 1, characterized in that The health status detection model based on the Transformer architecture is used to perform health status detection on the health feature sequence to obtain the health status index of the user, including: Inputting the health feature sequence into the health status detection model; Encoding the health feature sequence to obtain a coding vector sequence; Through a multi-head attention mechanism, the encoding vector sequence is processed by capturing the global feature dependency to obtain a global long-term dependency feature; The health status prediction process is performed on the global long-term dependent feature to obtain the health status index.

3. The method according to claim 2, characterized in that The encoding process of the health feature sequence to obtain the encoding vector sequence includes: Embedding the health feature sequence to obtain an embedding vector sequence; The embedded vector sequence is positionally encoded to obtain an encoded vector sequence.

4. The method according to claim 2, characterized in that: The performing health status prediction processing on the global long-term dependent feature to obtain the health status index includes: Transforming the global long-term dependency feature to obtain a deep-level global long-term dependency feature; Performing pooling processing on the deep-level global long-term dependency features to obtain high-level global long-term dependency features; The high-level global long-term dependency features are subjected to full connection operation processing to obtain the health status index.

5. The method according to claim 1, characterized in that: The step of performing feature extraction processing on the target health data to obtain a health feature sequence includes: Cleaning the target health data to obtain clean health data, where the clean health data includes clean work data, clean sleep data, and clean exercise data; Performing feature extraction processing on the cleaning work data to obtain a work feature sequence; Performing feature extraction processing on the cleaning sleep data to obtain a sleep feature sequence; Performing feature extraction processing on the cleaning motion data to obtain a motion feature sequence; The working feature sequence, the sleeping feature sequence and the motion feature sequence are fused to obtain the health feature sequence, which includes a time series and a health feature vector.

6. The method according to claim 5, characterized in that The cleaning process of the target health data to obtain clean health data includes: Performing denoising processing on the target health data to obtain denoised target health data; Performing missing value filling processing on the denoised target health data to obtain filled target health data; Performing outlier correction processing on the filled target health data to obtain corrected target health data; The corrected target health data is standardized to obtain the clean health data.

7. The method according to any one of claims 1 to 6, characterized in that: After performing health status detection on the sleep feature subsequence by using a health status detection model based on the Transformer architecture to obtain the health status index of the user, the method further includes: Generate a health report for the user, the health report including the health status index, health management strategy information and schedule information; The health report is visually displayed.

8. A health status detection device, characterized in that: The device comprises: A health data acquisition module, used to acquire the user's health data, the health data including work data, sleep data and exercise data collected based on wearable devices; A time alignment module, used for performing time alignment processing on the work data, the sleep data and the exercise data to obtain target health data; A health feature extraction module, used to perform feature extraction processing on the target health data to obtain a health feature sequence; The health status detection module is used to perform health status detection on the health feature sequence through a health status detection model based on the Transformer architecture to obtain the health status index of the user.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the health status detection method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the health status detection method according to any one of claims 1 to 7 is implemented.

Images