Personalized user parameter monitoring method, smart bracelet and medium

Abstract

The invention discloses a personalized user parameter monitoring method, a smart bracelet and a medium. The method comprises the following steps: when a user does not wear the smart bracelet, acquiring an emotional state of the user; when the emotional state is abnormal or the user belongs to a special user group, if the smart bracelet is located in a preset position range, acquiring video data, and identifying a real-time health state and a real-time emotional state of the user; when the user cannot be recognized in the video data and the real-time emotional state is abnormal, obtaining target sound data, and determining the current behavior of the user; calculating a real-time similarity; when the real-time similarities within the first preset duration are all lower than a preset similarity threshold value or abnormal sound data exist in the real-time sound data, determining a prompting mode; when the user is prompted according to the prompting mode, if normal reply of the user is detected, determining a target health state of the user; and when the target health state is abnormal, sending early warning information. According to the invention, the efficiency of finding the health abnormality of the user when the user living alone does not wear the smart bracelet is improved.

Description

Technical Field

[0001] This application relates to the field of health monitoring, and particularly to a personalized user parameter monitoring method, a smart bracelet, and a medium. Background Art

[0002] In today's digital age, people's attention to health management has been increasing day by day. Smart wearable devices, especially smart bracelets, play a crucial role in the field of health monitoring because they can monitor multiple physiological parameters of users in real time. Smart bracelets are not only convenient to carry but also can provide users with instant health data feedback, and have become a powerful assistant for the daily health management of many people.

[0003] Currently, the existing smart bracelets on the market mainly collect users' physiological data in real time through various sensors that contact the user's wrist skin, such as heart rate sensors, blood oxygen sensors, etc. These sensors can relatively accurately monitor various physiological indicators of users in the normal wearing state and transmit the data to associated mobile devices or the cloud through wireless communication technologies such as Bluetooth, facilitating users to view and analyze.

[0004] However, when users do not wear smart bracelets for various reasons, the bracelets cannot continue to monitor the health status of users. For solitary users, especially those with abnormal emotions or belonging to special groups, during the period when they do not wear the bracelets, once a sudden physical condition or a dangerous situation occurs, if the bracelets cannot detect and issue an alarm in time, it will cause users to be unable to obtain external help in the first time, which may result in serious consequences. Summary of the Invention

[0005] This application provides a personalized user parameter monitoring method, a smart bracelet, and a medium, which can improve the efficiency of detecting users' health abnormalities when solitary users do not wear smart bracelets.

[0006] In a first aspect, the present application provides a personalized user parameter monitoring method, the method comprising: when a target user is not wearing a smart bracelet, obtaining the emotional state of the target user when taking off the smart bracelet; when the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is within a preset position range, obtaining real-time video data collected by a preset image acquisition device; based on the real-time video data, identifying the real-time health status and real-time emotional state of the target user; when the target user cannot be identified in the real-time video data and the real-time emotional state is abnormal, obtaining target sound data of the target position where the target user is located; determining the current behavior of the target user based on the target sound data, the target position and the historical behavior record; and calculating the positive correlation coefficient between the current behavior and the target user. The real-time similarity between the normal sound data and the real-time sound data of the target location; when the real-time similarity at each time point within the first preset time period is lower than the preset similarity threshold or there is abnormal sound data in the preset abnormal sound set in the real-time sound data, determine the prompt method of the abnormal confirmation information according to the mobile terminal usage status of the target user and the personal information of the target user; when the target device is controlled to prompt the target user to reply to the abnormal confirmation information according to the prompt method, if a normal reply of the target user is detected within the second preset time period, determine the target health status of the target user according to the normal reply and the reply method, and the normal reply is a reply indicating that there is no abnormality in the physical state of the target user; when the target health status is abnormal, send an early warning message to the preset contact.

[0007] By adopting the above technical solution, when the user is not wearing a smart bracelet, the user's health abnormalities can be discovered in time by combining multi-source data such as image acquisition equipment, sound acquisition equipment, and user mobile terminals. When the user is in an abnormal mood or belongs to a special user group, the video data and sound data of the user's area are actively obtained, and combined with historical behavior records for comprehensive analysis, so as to accurately judge the user's current health status and behavior pattern. Once an abnormality is found, the appropriate prompt method can be selected according to the user's specific situation, and the user can be reminded to confirm his or her own status in time, and an early warning message can be quickly sent to the preset contact when the user fails to respond normally, which improves the efficiency of discovering user health abnormalities when the single user is not wearing a smart bracelet, effectively makes up for the monitoring gap of traditional smart bracelets when the user is not wearing them, and reduces the health risks and potential hazards caused by the user not wearing the bracelet.

[0008] In some embodiments in combination with some embodiments of the first aspect, when the target device is controlled to prompt the target user to reply to the abnormal confirmation information in the prompt manner, if a normal reply from the target user is detected within the second preset duration, the target health status of the target user is determined according to the normal reply and the reply manner, which specifically includes: when the display device is controlled to display the abnormal confirmation information to the target user in the display manner, if a normal reply from the target user is detected within the second preset duration, the reply manner of the normal reply is obtained; if the reply manner belongs to the first manner set, the target health status of the target user is determined according to the voice feature data of the voice reply information and a preset voice feature library, where the first manner set is to indicate that the physical state is normal through voice; if the reply manner belongs to the second manner set, the action operation data of the target user is obtained, where the second manner set is to indicate that the physical state is normal through an action operation of operating a mobile terminal or pressing a fixed button; the target health status is determined according to the action operation data and the historical action operation data of the target user.

[0009] By adopting the above technical solutions, diverse feedback channels are provided for users, fully considering the convenience and feasibility of users in different scenarios. Through the classification processing of the reply manners, whether it is a voice reply or an action operation reply, the target health status of the user can be accurately judged by means of specific data matching and analysis. This not only improves the accuracy of health judgment but also can effectively cope with various complex situations, ensuring that when the user is not wearing a smart bracelet and is suspected of having a health abnormality, the true state of the user can be quickly and reliably confirmed, the abnormality can be promptly excluded or the warning mechanism can be activated when necessary, further enhancing the ability to protect the health and safety of the solitary and special user groups and reducing the situations of misjudgment and omission.

[0010] In some embodiments in combination with some embodiments of the first aspect, when the target health status is determined according to the action operation data and the historical action operation data of the target user, it specifically includes: determining whether there is an abnormality in the action operation of the target user according to the action operation data and the historical action operation data of the target user; if not, the first voice data within the preset third duration before the prompt time point and the second voice data within the preset third duration after the prompt time point are obtained, where the prompt time point is the time point for prompting the target user to reply to the abnormal confirmation information; the first real-time state of the sound collection device at the target location before the prompt time point and the second real-time state after the prompt time point are determined according to the first voice data and the second voice data; if the first real-time state of the sound collection device is normal and the second real-time state is abnormal, it is determined that the target health status is abnormal.

[0011] By adopting the above technical solution, through comprehensive analysis of action operation data and historical data, it is possible to preliminarily determine whether there are abnormalities in the user's action operations, thereby providing a preliminary basis for the assessment of the health status. On this basis, by combining the first sound data and the second sound data before and after the prompt time point, the real-time state change of the sound collection device is further analyzed. If the state of the sound collection device changes from normal to abnormal after the prompt time point, it indicates that there may be sudden situations or artificially caused abnormalities in the sound data, and thus it can be inferred that the user's health status may be abnormal. This multi-dimensional and multi-stage analysis method can effectively avoid misjudgments of single data or single time points, ensure that the system accurately identifies the user's health status in a complex and changeable environment, timely discovers potential health risks, further improves the intelligence level and security guarantee ability of the user parameter monitoring system, and provides more comprehensive and accurate health monitoring services for users.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after the step of sending a warning message to a preset contact person when the target health status is abnormal, the method further includes: when it is detected that the target user makes a call, obtaining the call data of the target user; according to the call data, determining the emotional value of the target user during the call; when the emotional value of the target user at the end of the call exceeds the emotional value at the start of the call, according to the call data, determining the reason for the deepening of the emotional fluctuation of the target user, where the reason for the deepening of the emotional fluctuation includes the call reason and the self reason; according to the reason for the deepening of the emotional fluctuation, executing corresponding soothing strategies, where the soothing strategies include a first soothing strategy and a second soothing strategy, the first soothing strategy is to control the target terminal device to send a call request to any contact person in the target contact person set, and the second soothing strategy is to, when it is detected that the target terminal device receives a call request from a contact person other than the contact persons in the target contact person set, reject the call request and send a preset reply text message to the contact person who sent the call request.

[0013] By adopting the above technical solution, by obtaining and analyzing the user's call data, identifying the emotional changes of the user during the call, and when the user's emotional fluctuation deepens, executing corresponding soothing strategies according to the reasons for the emotional fluctuation (such as emotional changes caused by the call content or the user's own emotional problems), such as actively contacting the contact persons in the target contact person set to provide timely emotional support for the user; or automatically filtering out irrelevant call interferences to avoid negative factors from exacerbating the user's bad emotions. This active intervention and soothing mechanism can effectively relieve the user's emotional stress and prevent emotional problems from deteriorating further, thereby providing more comprehensive and considerate health management services for users and further improving the intelligence level and user experience of the system.

[0014] In combination with some embodiments of the first aspect, in some embodiments, according to the reason for the deepening of the emotional fluctuation, corresponding soothing strategies are executed, specifically including: obtaining the historical call records of the preset contact; determining a set of target contacts according to the historical call records, and the target user can make a call with any contact in the set of target contacts to reduce the degree of emotional fluctuation; if the reason for the deepening of the fluctuation is a call reason, execute the first soothing strategy; if the reason for the deepening of the fluctuation is a self reason, execute the second soothing strategy.

[0015] With the above technical solutions, by analyzing the historical call records of the preset contact, the set of target contacts that can effectively reduce the emotional fluctuation of the target user is accurately screened out, laying a solid foundation for subsequent soothing actions. When the reason for the deepening of the emotional fluctuation is clear, corresponding strategies are quickly matched according to different causes. If the call has an adverse effect, the first soothing strategy is used to quickly connect the call of the contact who can calm the emotion to help the user reverse the emotion; if it is due to self reasons, the second soothing strategy is used to reduce external interference, allowing the user to adjust the state in a relatively quiet environment, improving the efficiency and effect of emotional soothing.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of obtaining the emotional state of the target user when the target user removes the smart bracelet when the target user is not wearing the smart bracelet, the method further includes: in the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is outside the preset position range, obtaining the number of users within the preset range of the target user; when the number of users is less than the preset number threshold, obtaining the current position and moving direction of the target user; predicting the set of positions where the target user will be located after moving a preset distance according to the current position and the moving direction; obtaining the danger coefficient of each position in the set of positions; when there is a dangerous position in the set of positions whose danger coefficient exceeds the preset danger threshold, prompting the target user to avoid the dangerous position through a preset prompting method.

[0017] With the above technical solutions, when the user is not wearing the smart bracelet and the emotional state is abnormal or belongs to a special user group, by real-time monitoring the user's position and moving direction, predicting the possible dangerous positions that the user may reach, and timely prompting the user to avoid the dangerous area, the potential safety risks caused by the user's abnormal emotion are reduced.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of prompting the target user to avoid the dangerous position through a preset prompting method when there is a dangerous position in the set of positions whose danger coefficient exceeds the preset danger threshold, the method further includes: when detecting that the real-time position of the target user is the dangerous position, sending the real-time position of the target user to the preset contact.

[0019] With the above technical solution, by continuously monitoring the user's location and comparing it with the dangerous locations, the warning mechanism can be triggered immediately when the user enters a high-risk area, and the real-time location information can be sent to the preset contacts so that the contacts can take actions promptly. This real-time location sharing and warning function can not only effectively reduce the safety risks of users in dangerous areas, but also provide additional safety guarantees for users.

[0020] Combined with some embodiments of the first aspect, in some embodiments, after the step of sending a warning message to the preset contacts when the target health status is abnormal, the method further includes: obtaining the historical health records of the target user within a preset time period; determining the number of times of emotional abnormality of the target user within the preset time period according to the historical health records; when the number of times of emotional abnormality exceeds a preset number threshold, setting the operating mode of the smart bracelet to a preset incentive mode, and sending the historical health records to medical experts; updating the health monitoring strategy of the target user according to the received feedback information of the medical experts.

[0021] With the above technical solution, when the number of times of emotional abnormality of the user exceeds the preset threshold, the operating mode of the smart bracelet is switched to the preset incentive mode to help the user better manage emotions. At the same time, the historical health records are sent to medical experts to obtain professional opinions and suggestions. According to the feedback information of the medical experts, the health monitoring strategy of the target user can be updated in a timely manner, so as to more accurately meet the health management needs of the user. This dynamic adjustment mechanism based on historical data and professional feedback can not only provide personalized health management services for users, but also take more effective intervention measures in a timely manner when the user's emotions or health conditions frequently show abnormalities, further improving the intelligent level and health management effect of the system, and providing more comprehensive and scientific health support for users.

[0022] In a second aspect, an embodiment of the present application provides a smart bracelet, including: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the smart bracelet to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0023] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, including instructions, when the above instructions run on the smart bracelet, enabling the smart bracelet to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0024] Understandably, both the smart bracelet provided by the second aspect and the storage medium provided by the third aspect are used to execute the method provided by this application. Therefore, the beneficial effects they can achieve can refer to the beneficial effects in the corresponding method, which will not be elaborated here.

[0025] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. When this application monitors that the user is not wearing the bracelet and the emotional state is abnormal or the user belongs to a special group, it monitors the user's physical state in real time through video data and sound data. Once physical abnormalities are found, it selects an appropriate prompting method according to the specific situation of the user, timely reminds the user to confirm their own state, and quickly sends a warning message to the preset contact when the user fails to reply normally, improving the efficiency of detecting the user's health abnormalities when a solitary user is not wearing the smart bracelet, effectively making up for the monitoring gap of traditional smart bracelets when the user is not wearing them, and reducing the health risks and potential hazards caused by the user not wearing the bracelet.

[0026] 2. This application analyzes the historical call records of the preset contacts, screens out the set of contacts that can effectively reduce the user's emotional fluctuations, laying a foundation for subsequent soothing of the user's emotions. When the reason for the deepening of the user's emotional fluctuations is a call reason, it uses the first soothing strategy to timely connect to the call of the contact who can calm the emotions to help the user reverse the emotions; if it is due to their own reasons, it uses the second soothing strategy to reduce external interference and let the user adjust their state in a relatively quiet environment, improving the efficiency and effect of emotional soothing.

[0027] 3. This application predicts the dangerous positions that the user may reach by monitoring the user's location and moving direction in real time, and timely prompts the user to avoid dangerous areas, reducing the potential safety risks caused by the user's abnormal emotions. At the same time, by continuously monitoring the user's location and comparing it with the dangerous positions, it can immediately trigger the warning mechanism when the user enters the high-risk area and send the real-time location information to the preset contact, reducing the safety risks of the user in the dangerous area. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a structural schematic diagram of a system architecture to which the personalized user parameter monitoring method in the embodiments of this application can be applied; Figure 2 is a flowchart of the personalized user parameter monitoring method in the embodiments of this application; Figure 3 is another flowchart of the personalized user parameter monitoring method in the embodiments of this application; Figure 4 is an exemplary hardware structural schematic diagram of the smart bracelet in the embodiments of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] The terms used in the following embodiments of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. As used in the specification and appended claims of this application, the singular forms "a", "an", "the", "above-mentioned", "said", and "this" are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in this application refers to any or all possible combinations including one or more of the listed items.

[0030] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as implying or suggesting relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of this application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0031] Figure 1 It is a schematic structural diagram of a system architecture to which the personalized user parameter monitoring method in the embodiments of this application can be applied.

[0032] Please refer to Figure 1 , the user parameter monitoring system includes an image acquisition device, a sound acquisition device, a sound playback device, a terminal device, and a smart bracelet.

[0033] As the core component of the system, the smart bracelet is used to analyze and process the data collected by devices such as the image acquisition device, the sound playback device, and the terminal device, and send control instructions to devices such as the sound playback device and the terminal device. The image acquisition device is used to acquire the image data of the user's area and transmit the acquired image data to the smart bracelet. The sound acquisition device is used to acquire the sound data of the user's area and transmit the acquired sound data to the smart bracelet. The sound playback device is used to receive the control instructions sent by the smart bracelet and play the corresponding information according to the control instructions. The terminal device is used to acquire the user's call information, transmit the acquired call information to the smart bracelet, and receive the control instructions of the smart bracelet and perform relevant operations according to the control instructions.

[0034] Among them, the terminal device includes devices such as mobile phones and tablets that can perform telephone communication.

[0035] Through the above system architecture, the user parameter monitoring system can determine the user's health status through the image data and sound data of the user's area, and can also monitor the user's health status when the user is not wearing the smart bracelet, avoiding the interruption of health monitoring or data loss caused by the user not wearing the smart bracelet.

[0036] In the related art, existing smart bracelets mainly collect users' physiological data in real time through various sensors that come into contact with the user's wrist skin, such as heart rate sensors, blood oxygen sensors, etc. These sensors can relatively accurately monitor various physiological indicators of users in the normal wearing state, and transmit the data to the associated mobile device or cloud through wireless communication technologies such as Bluetooth, facilitating users to view and analyze. However, when the user does not wear the smart bracelet for various reasons, the bracelet cannot continue to monitor the user's health status. For solitary users, especially those with abnormal emotions or belonging to special groups, during the period when the bracelet is not worn, once a sudden physical condition or a dangerous situation occurs, if the bracelet cannot detect and issue an alarm in time, it will cause the user to be unable to obtain external help in the first time, which may result in serious consequences.

[0037] By adopting the personalized user parameter monitoring method in the embodiment of the present application, when it is detected that the user does not wear the bracelet and the emotional state is abnormal or the user belongs to a special group, the physical state of the user is monitored in real time through video data and sound data. When it is detected that the physical state of the user is abnormal, an alarm is issued in time, improving the efficiency of detecting the user's health abnormality when the solitary user does not wear the smart bracelet.

[0038] The following will be combined with Figure 2 to illustrate the method of the embodiment of the present application.

[0039] Please refer to Figure 2 , which is a schematic flowchart of a personalized user parameter monitoring method in the embodiment of the present application.

[0040] S201. When the target user does not wear the smart bracelet, obtain the emotional state of the target user when the smart bracelet is taken off.

[0041] Specifically, the wearing state of the bracelet and the target user is monitored in real time through sensors (such as acceleration sensors, Hall sensors, etc.) built in the smart bracelet. When the acceleration sensor detects that there is no regular acceleration change generated by human movement for a period of time, and the Hall sensor detects that the distance between the magnet and the sensor exceeds a preset threshold (this threshold can be set according to the actual wearing situation and is used to judge whether the bracelet leaves the human body), it is determined that the target user has taken off the smart bracelet.

[0042] When the target user takes off the smart bracelet (does not wear the smart bracelet), obtain the body parameters monitored by the smart bracelet within a preset period of time before the target user takes off the smart bracelet.

[0043] Then, a machine learning model or a preset multi-parameter fusion rule is used to determine the user's emotional state. For the machine learning model, an emotion recognition model based on support vector machine (SVM) or neural network can be trained. Parameters such as heart rate, skin conductance response, and body temperature are used as input features. After training with a large amount of labeled data, the model can accurately predict the user's emotional state based on the input body parameters, such as normal (including happy, calm, etc.) and abnormal (including sad, angry, etc.). For the multi-parameter fusion rule, first obtain the preset weights of different body parameters such as heart rate, skin conductance response, and body temperature, and the emotion index values corresponding to different change ranges of each parameter. Then, the emotion index values of each parameter are weighted and summed according to the set weights to obtain a comprehensive emotion index. Finally, according to the preset comprehensive emotion index threshold range, the user's emotional state is judged. If the comprehensive emotion index is in a lower range, it is determined that the user's emotion is normal; if it is in a higher range, it is determined that the user's emotion is abnormal.

[0044] S202. In the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is within the preset position range, obtain the real-time video data collected by the preset image acquisition device.

[0045] Specifically, in the case where the emotional state is abnormal or the identity identifier of the target user belongs to the identity identifier of a preset special user group (such as the elderly, disabled, etc.), obtain the position information collected by the built-in positioning device of the smart bracelet. Compare the position information with the preset position range to determine whether the smart bracelet is within the preset position range. The preset position range can be a specific geographical area (such as the user's residential community, a specific area in the home, etc.), which is defined by the longitude and latitude coordinate range.

[0046] When the smart bracelet is within the preset position range, send an instruction to the preset image acquisition device through the network. The preset image acquisition device can be a smart camera installed in the user's home, a surveillance camera in the community, etc. After receiving the instruction, the preset image acquisition device starts to collect real-time video data and transmits the collected real-time video data to the smart bracelet through the network.

[0047] S203. According to the real-time video data, identify the real-time health state and real-time emotional state of the target user.

[0048] Specifically, first use a target detection algorithm (such as the YOLO series algorithm, Faster R-CNN algorithm, etc.) to process the real-time video frames to identify the human targets in the video. After identifying the human targets, use a face recognition algorithm (such as a face recognition model based on deep learning, etc.) and a human feature matching algorithm to determine whether the human target is the target user through the pre-stored appearance features (such as facial features, body shape features, etc.) of the target user.

[0049] When the target user is recognized, a target tracking algorithm (such as KCF (Kernelized Correlation Filters), etc.) is used to continuously track the position of the target user in the video sequence to ensure that relevant information of the target user can be accurately obtained in subsequent processing.

[0050] For real-time health status recognition, the position information of each joint point of the target user's body is extracted through human pose estimation technologies (such as OpenPose, HRNet, etc.) to judge the user's posture (such as standing, sitting, lying down, etc.) and actions (such as walking, running, falling, etc.). Abnormal postures or actions imply that the user may have health problems, such as the action of suddenly falling. When there are preset dangerous actions in the user's posture and actions, and the dangerous actions last for a preset duration, the real-time health status of the user is determined to be abnormal. At the same time, some physiological characteristics of the target user are extracted from the video using computer vision technology. The video frame is input into a pre-trained deep learning-based physiological feature recognition model to obtain the physiological feature parameters output by the model. These physiological feature parameters are compared with the normal physiological index range to determine the real-time health status of the user.

[0051] Among them, the physiological feature recognition model can be constructed using a convolutional neural network (CNN) architecture. In the model training stage, a large number of video frame data with different physiological feature states standardized are required to train the model. During the training process, the parameters of the model are continuously adjusted to minimize the error between the prediction result and the annotation value. When the video frame is input into the trained physiological feature recognition model, the model first performs data preprocessing on the video frame, adjusts the size to a fixed size adapted to the model, normalizes the pixel values at the same time, and ensures that the number of channels meets the requirements. Then it enters the convolutional layer, and multiple convolutional kernels slide for convolution to extract local features such as edges and textures. The ReLU activation function is introduced for non-linearity, and then the dimension is reduced through the pooling layer for downsampling. Subsequently, the features extracted at different levels are fused in deeper layers and enter the fully connected layer. The fully connected layer multiplies the fused feature vector by the weight matrix learned during the model training process, adds the bias, and after a series of complex operations, finally outputs the values corresponding to various physiological feature parameters. These values are the prediction results of the model for the physiological features of the target user in the input video frame.

[0052] For real-time emotion state recognition, a deep learning-based facial expression recognition model (such as a MobileNet-based expression recognition model) is used to analyze the facial expressions of the target user. The model extracts features and classifies the facial images in the video frames to determine whether the user's expression belongs to basic emotion categories such as happy, sad, angry, surprised, etc. At the same time, the emotional state is judged by analyzing information such as the amplitude, speed, and posture of the target user's body movements. The body language recognition model that has been pre-trained is used to analyze and classify the body movements in the video. Finally, a weighted sum is performed according to the preset weights of the facial expressions and body language and the preset scores corresponding to their respective analysis results to obtain a comprehensive emotion score. According to the preset range of emotion score thresholds, the real-time emotion state of the user is determined to be normal or abnormal.

[0053] Among them, the facial expression recognition model is constructed based on deep learning technology. Taking the MobileNet-based expression recognition model as an example, in the model training stage, a large number of facial image data with different expression labels are required to train the model. The facial image of the target user is input into the trained model, and the model will process the image according to the trained parameters and structure. After a series of convolutional, activation, pooling, and fully connected operations, the probability scores of each emotion category are output, and the category with the highest score is the emotion state corresponding to the facial expression of the target user judged by the model.

[0054] The body language recognition model is constructed based on deep learning or traditional machine learning technology. In the model training stage, a large number of video data labeled with different body movements and their corresponding emotion or state labels are required to train the model. The video data of the target user is input into the model, and the model will analyze the body movements in the video according to the trained parameters and structure. After a series of operations such as feature extraction, feature fusion, and classification, the emotion or state category corresponding to the body movement is output.

[0055] S204. When the target user cannot be recognized in the real-time video data and the real-time emotion state is abnormal, obtain the target sound data at the target location where the target user is located.

[0056] Specifically, a target detection algorithm is used to identify the human target in the video. When the human target is not the target user (that is, the target user cannot be recognized in the real-time video data) and the real-time emotion state of the target user is abnormal, obtain the positioning information of the target user's mobile terminal (such as a mobile phone), and at the same time extract the first video data when the target user was last detected.

[0057] Then, use the object detection algorithm to process the first video data, and continuously track the position change of the target user in the video sequence in combination with the object tracking algorithm (such as KCF, CSRT, etc.), and determine the walking direction of the target user according to the position change. At the same time, according to the direction correspondence table of each area in the pre-stored preset position range, obtain the direction correspondence table corresponding to the area where the target user is located in the first video data. Among them, the direction correspondence table specifies the other areas that will be reached when walking in different directions from this area.

[0058] Next, determine the position area where the user is currently located according to the walking direction of the target user and the direction correspondence table corresponding to the walking direction. Compare the mobile phone positioning information with the position area range determined according to the video. If the position shown by the mobile phone positioning is within the position area range determined according to the video, determine the position shown by the mobile phone positioning as the user's location; if the position shown by the mobile phone positioning is not within the position area range determined according to the video, determine the position area determined according to the walking direction in the video as the user's location.

[0059] Finally, send an instruction to the preset sound collection device (such as a smart speaker, microphone array, etc. installed in this position area) through the network. After receiving the instruction, the sound collection device starts to collect the target sound data at the target position where the target user is located, and transmits the collected target sound data to the smart bracelet through the network.

[0060] S205. Determine the current behavior of the target user according to the target sound data, target position, and historical behavior record.

[0061] Specifically, first preprocess the received target sound data, including operations such as removing noise and performing audio normalization, etc., to improve the quality and comparability of the sound data.

[0062] Then, extract the historical sound data of each behavior corresponding to the target position from the historical behavior record of the target user. When these historical sound data were recorded, they were classified and stored according to different behavior types, and each behavior has corresponding historical sound data samples.

[0063] Next, use a voice matching algorithm (such as the dynamic time warping (DTW) algorithm, a deep learning-based audio similarity matching model, etc.) to match the target voice data with the historical voice data of each behavior extracted. For the dynamic time warping (DTW) algorithm, it can calculate the similarity between the target voice data and the historical voice data on the time axis, find the best matching path between the two, and thus obtain their similarity degree. For a deep learning-based audio similarity matching model, it usually first extracts features from the audio data, such as using Mel-frequency cepstral coefficients (MFCC) and other features to represent the audio, and then inputs the extracted features into the trained model, and the model will output the similarity scores between the target voice data and each historical voice data.

[0064] Finally, according to the similarity scores obtained by the matching algorithm, select the historical behavior with the highest similarity score as the current behavior of the target user.

[0065] S206. Calculate the real-time similarity between the normal voice data corresponding to the current behavior and the real-time voice data at the target location.

[0066] Specifically, first preprocess the real-time voice data. Then, extract the normal voice data corresponding to the current behavior from the historical behavior records of the target user. Next, use a voice matching algorithm to match the normal voice data with the real-time voice data to obtain the real-time similarity between the normal voice data and the real-time voice data.

[0067] S207. When the real-time similarity at each time point within the first preset duration is lower than the preset similarity threshold or there is abnormal voice data in the preset abnormal voice set in the real-time voice data, determine the prompting method of the abnormal confirmation information according to the usage status of the target user's mobile terminal and the personal information of the target user.

[0068] Specifically, use a voice matching algorithm to match the real-time voice data with each abnormal voice data in the preset abnormal voice set to obtain the first similarity between the real-time voice data and each abnormal voice data. When there is a first similarity greater than the preset threshold, determine that there is abnormal voice data in the real-time voice data.

[0069] When the real-time similarity at each time point within the first preset duration is lower than the preset similarity threshold or there is abnormal voice data in the preset abnormal voice set in the real-time voice data, if the positioning location of the mobile terminal is within the location area of the target user, obtain the usage status of the mobile terminal. Through the communication connection established with the mobile terminal, send a request instruction for obtaining the usage status to the mobile terminal. After receiving the instruction, the mobile terminal detects its own usage status and sends the usage status to the smart bracelet.

[0070] If the mobile terminal is in the foreground with an application running and the mobile phone screen is in the lit state, it indicates that the user is using the mobile phone. If the personal information of the target user shows that the user can hear sounds, determine that the prompting method is to play a specific abnormal prompt sound through the audio playback system of the mobile terminal, and at the same time display the abnormal confirmation information in a prominent manner (such as pop-up windows, flashing prompts, etc.) on the interface of the application currently running on the mobile terminal or in the system notification bar. If the personal information of the target user shows that the user cannot hear sounds but can see, determine that the prompting method is to display the abnormal confirmation information on the mobile terminal and control the indoor lights or control the preset warning devices (such as variable signboards, etc.) in the area where the user is located to display the confirmation information. If the personal information of the target user shows that the user cannot hear sounds and cannot see, determine that the prompting method is to send the confirmation information to the preset contacts.

[0071] If the mobile terminal is in the non-running state and the mobile phone screen is in the black screen state or the positioning location of the mobile terminal is not within the location area of the target user, it indicates that the user is not using the mobile phone. If the personal information of the target user shows that the user can hear sounds, determine that the prompting method is to play a specific abnormal prompt sound through the sound playback device corresponding to the area where the user is located. If the personal information of the target user shows that the user cannot hear sounds but can see, determine that the prompting method is to control the indoor lights or control the preset warning devices (such as variable signboards, etc.) in the area where the user is located to display the confirmation information. If the personal information of the target user shows that the user cannot hear sounds and cannot see, determine that the prompting method is to send the confirmation information to the preset contacts.

[0072] S208. In the case of controlling the target device to prompt the target user to reply with the abnormal confirmation information according to the prompting method, if the normal reply of the target user is detected within the second preset duration, determine the target health status of the target user according to the normal reply and the reply method.

[0073] Among them, the normal reply is a reply indicating that there is no abnormality in the physical state of the target user.

[0074] Specifically, according to the prompting method, confirm the target device and its corresponding control instruction. Send the corresponding control instruction to the target device, and record the current time point as the prompting time point. After receiving the instruction, the target device prompts the target user to reply with the abnormal confirmation information according to the prompting method.

[0075] Start timing from the time point when the control instruction is sent, and determine whether the target user has made a normal response within the second preset duration. Within the second preset duration, by continuously monitoring the information fed back by the mobile terminal and the sound data collected by the voice acquisition device, determine whether the target user has made a normal response. If the voice of the target user is detected in the sound data, use speech recognition technology to convert the speech content into text form. Then, compare the converted text with the text content corresponding to the specific normal response sound data in the preset template. Through the keyword matching algorithm, extract the keywords in the text and compare them one by one with the keywords in the normal response template. If the keyword matching degree reaches the preset threshold and the semantics are coherent and reasonable, and it can clearly express that the physical state of the target user is normal, it is determined as a normal response, and the response method is determined to be a voice response. When receiving the information fed back by the mobile terminal, analyze the feedback content. If the feedback content is text information, perform natural language processing on the text. Then, compare the processed text information with the preset normal response template to determine whether it meets the semantic and format requirements of the normal response. If it meets the requirements, it is determined as a normal response, and the response method is determined to be an action response. If the feedback content is an operation instruction (such as clicking a specific button, swiping the screen, etc.), according to the preset mapping relationship between mobile terminal operations and responses, determine whether the response content corresponding to the operation indicates that the physical state of the target user is normal. If it meets the requirements, it is determined as a normal response, and the response method is determined to be an action response.

[0076] When a normal response from the target user is detected within the second preset duration, determine the target health status of the target user according to the normal response and the response method.

[0077] If the response method of the target user is to indicate a normal physical state through a voice response or a sound of a specific frequency, first process the voice response information of the target user to extract the voice feature data. Compare the extracted voice feature data with the preset voice feature library. The preset voice feature library is obtained through training and analysis based on the voice data of a large number of healthy individuals and individuals with different health conditions, and it contains voice feature templates corresponding to various health states. If the matching degree of the voice feature data with the voice feature template representing a normal health state exceeds the preset threshold, determine the target health status as normal. If the matching degree of the voice feature data with the voice feature template of an abnormal health state exceeds the preset threshold, determine the target health status as abnormal.

[0078] If the response method of the target user indicates normal physical status through actions such as operating a mobile terminal or pressing a fixed button, first communicate with the device operated by the user to obtain the operation data of the device. The operation data includes, but is not limited to, information such as the speed, strength, accuracy, number of operations, and time interval of operations. For example, if the device is a mobile terminal, when the mobile terminal detects that the user performs a response operation on a preset response application or pop-up window, etc., it will rely on various sensors and system mechanisms integrated inside it to collect operation data. When the smart bracelet sends corresponding control instructions to the mobile terminal, the mobile terminal will send the collected operation data to the smart bracelet; if the target device is a fixed button, the fixed button collects operation data through built-in sensors and sends the operation data to the smart bracelet. Then, compare and analyze the obtained operation data with the historical operation data of the target user, and calculate the similarity or difference degree between the current operation data and the historical operation data. The historical operation data is the operation habits and patterns of the target user in a normal state recorded and stored by the electronic device in the past period of time. If the similarity between the current operation data and the historical operation data exceeds the preset threshold, that is, the action operation of the target user is normal, and each operation index (such as operation speed, strength, etc.) is similar to the corresponding index in the historical data and within a reasonable fluctuation range, determine that the target health status is normal. If the similarity is lower than the preset threshold, that is, the action operation of the target user is abnormal, determine that the target health status is abnormal.

[0079] S209. When the target health status is abnormal, send a warning message to a preset contact.

[0080] When the target health status is abnormal or the real-time health status of the target user monitored through video data is abnormal, send the preset warning message to the preset contact through the built-in communication module.

[0081] In the embodiments of the present application, when it is monitored that the user is not wearing the bracelet and the emotional state is abnormal or the user belongs to a special population, the physical state of the user is monitored in real time through video data and sound data. When it is monitored that the physical state of the user is abnormal, select a suitable prompt method according to the specific situation of the user, timely remind the user to confirm their own state, and quickly send a warning message to the preset contact when the user fails to reply normally, improving the efficiency of discovering the health abnormality of the user when the solitary user is not wearing the smart bracelet, effectively making up for the monitoring gap of the traditional smart bracelet when the user is not wearing it, and reducing the health risks and potential hazards caused by the user not wearing the bracelet.

[0082] The following combines Figure 3 to further illustrate the method of the embodiments of the present application.

[0083] Please refer to Figure 3, which is another schematic flowchart of the personalized user parameter monitoring method in the embodiments of the present application.

[0084] S301. When the target user is not wearing the smart bracelet, obtain the emotional state of the target user when taking off the smart bracelet.

[0085] S302. In the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is within the preset position range, obtain the real-time video data collected by the preset image acquisition device.

[0086] S303. According to the real-time video data, identify the real-time health state and real-time emotional state of the target user.

[0087] S304. When the target user cannot be recognized in the real-time video data and the real-time emotional state is abnormal, obtain the target sound data at the target position where the target user is located.

[0088] S305. Determine the current behavior of the target user.

[0089] S306. Calculate the real-time similarity between the normal sound data corresponding to the current behavior and the real-time sound data at the target position.

[0090] S307. When the real-time similarity at each time point within the first preset duration is lower than the preset similarity threshold or there is abnormal sound data in the preset abnormal sound set in the target sound data, determine the prompting method of the abnormal confirmation information.

[0091] S308. In the case where the control display device displays the abnormal confirmation information to the target user according to the display method, if a normal response of the target user is detected within the second preset duration, obtain the response method of the normal response.

[0092] S309. If the response method belongs to the first method set, determine the target health state of the target user.

[0093] If the response method belongs to the first method set, determine the target health state of the target user according to the sound feature data of the voice response information and the preset sound feature library.

[0094] Among them, the first method set is to indicate that the physical state is normal through voice.

[0095] S310. If the response method belongs to the second method set, obtain the action operation data of the target user.

[0096] Among them, the second method set is to indicate that the physical state is normal through the action operation of operating the mobile terminal or pressing the fixed button.

[0097] S311. Determine whether there is an abnormality in the action operation of the target user.

[0098] Determine whether there is an abnormality in the action operation of the target user according to the action operation data and the historical action operation data of the target user.

[0099] If so, execute the steps of S312; if not, execute the steps of S313.

[0100] S312. Determine that the target health status is abnormal.

[0101] Steps S301 - S312 are similar to Figure 2 Steps S201 - S208 in the illustrated embodiment, and reference may be made to the descriptions in steps S201 - S208, which will not be elaborated here.

[0102] S313. Obtain first sound data within a preset third time period before the prompt time point and second sound data within a preset third time period after the prompt time point.

[0103] Among them, the prompt time point is the time point for prompting the target user to reply with an abnormality confirmation message.

[0104] Specifically, obtain the prompt time point. Calculate the target time point obtained by adding the third time period to the prompt time point. When the current time point reaches the target time point, obtain the historical sound data set. The sound data in the historical sound data set is sorted in chronological order, and each time point corresponds to a segment of sound data. Extract the first sound data within a preset third time period before the prompt time point and the second sound data within a preset third time period after the prompt time point from the historical sound data set.

[0105] S314. Determine the first real - time state of the sound collection device at the target location before the prompt time point and the second real - time state after the prompt time point.

[0106] According to the first sound data and the second sound data, determine the first real - time state of the sound collection device at the target location before the prompt time point and the second real - time state after the prompt time point.

[0107] Specifically, first determine that the first real - time state of the sound collection device at the target location before the prompt time point is normal.

[0108] Then, when the voice data at each time point in the first voice data is not empty, determine whether the voice data corresponding to the current time point in the second voice data is empty. If so, it means that the target user has turned off or damaged the voice collection device after receiving the confirmation message, resulting in the inability to obtain the voice data collected by the voice collection device, and determine that the second real-time state is abnormal; if not, compare the first voice data with the second voice data. Calculate the average volume at each time point according to the first voice data and the second voice data. Judge the change of the average volume within a preset time period according to the average volume at each time point. If the volume change value within a certain target time period in the time period corresponding to the second voice data exceeds the preset volume reduction threshold, determine whether the average volume at each time point within a period of time after the target time period is within the preset error range compared with the average volume corresponding to the target time period. If so, it means that there is an unreasonable operation of covering the device with a covering object by someone, resulting in a significant reduction in the volume collected by the device, and determine that the second real-time state is abnormal. If not, determine that the second real-time state is normal.

[0109] S315. If the first real-time state of the voice collection device is normal and the second real-time state is abnormal, determine that the target health state is abnormal.

[0110] If the first real-time state of the voice collection device is normal and the second real-time state is abnormal, determine that the target health state is abnormal; if the first real-time state and the second real-time state of the voice collection device are the same, determine that the target health state is normal. At the same time, if there is abnormal voice data in the preset abnormal voice set in the second voice data, determine that the target health state is abnormal.

[0111] S316. When the target health state is abnormal, send a warning message to the preset contact.

[0112] Step S316 is similar to Figure 2 Step S209 in the embodiment shown, and reference can be made to the description in step S209, which will not be elaborated here.

[0113] S317. When it is detected that the target user is making a call, obtain the call data of the target user.

[0114] Specifically, communicate with the mobile terminal of the target user to obtain the real-time call status (including not in a call, in a call, etc.) fed back by the mobile terminal. When the call status is in a call (that is, it is detected that the target user is making a call), obtain the real-time call data transmitted by the mobile terminal in real time, or when it is monitored that the call status of the target user changes from in a call to not in a call, obtain the call data during the call (including audio data, call duration, call time, etc.) transmitted by the mobile terminal.

[0115] If the call data of the target user cannot be obtained from the mobile terminal, the voice data of the target user during the call can be collected through a preset voice collection module as the call data of the target user.

[0116] S318. Determine the emotional value of the target user during the call.

[0117] Determine the emotional value of the target user during the call according to the call data.

[0118] Specifically, from the obtained call data, using voice recognition technology, by analyzing voice characteristics such as the timbre, pitch, and speaking speed of different speakers, the voice data of the target user is separated from the call audio to obtain multiple discrete voice data, and each voice data corresponds to a time period.

[0119] Then, perform acoustic feature analysis on the voice data of the target user in each time period, such as analyzing features such as pitch, volume, speaking speed, and formants. These feature changes can reflect the user's emotional state. Use. The acoustic features extracted in each time period are sequentially input into a pre-trained emotion recognition model based on acoustic features, and the model outputs the first emotion category (such as positive, negative, neutral, etc.) of the target user in each time period. Then, obtain the preset first emotion score corresponding to the first emotion category.

[0120] Among them, the emotion recognition model is constructed based on a deep learning architecture (such as a recurrent neural network, a convolutional neural network, etc.). In model training, a large amount of voice data marked with different emotional states is used to train the model. Taking the recurrent neural network as an example, when the acoustic features are input into the model, the acoustic features first enter the input layer and then are transmitted to the hidden layer. The neurons in the hidden layer, based on the current input and the previous state, perform non-linear transformation through an activation function to extract and update emotion features, and their states are finally transmitted to the output layer. The output layer calculates the probability distribution of each emotion category through weighted summation and the Softmax function, and the one with the highest probability is the output first emotion category.

[0121] Next, use speech recognition technology to convert the voice data of the target user into text content, perform natural language processing on the converted text, adopt a sentiment analysis method based on a dictionary, construct a dictionary containing a large number of sentiment words, assign sentiment polarity (positive, negative, or neutral), scores, and weights to each word, count the scores and weights of each sentiment word in the text, and calculate the sentiment score of the text as the second emotion category score.

[0122] Finally, comprehensively calculate the first emotion score and the second emotion score to obtain the emotional value of the target user in each time period during the call. For example, use the weighted average method to calculate the emotional value at each time point.

[0123] S319. When the emotional value of the target user at the end of the call exceeds the emotional value at the start of the call, determine the reason for the deepening of the emotional fluctuation of the target user according to the call data.

[0124] Among them, the reasons for the deepening of emotional fluctuations include call reasons and personal reasons.

[0125] Specifically, when the emotional value at the time point corresponding to the end time of the call of the target user exceeds the emotional value at the time point corresponding to the start of the call, obtain the call user information of the call with the target user. Then, compare the call user information with the information of each contact in the preset contacts.

[0126] If the call user does not belong to the preset contacts, determine that the reason for the deepening of the emotional fluctuation of the target user is the call reason.

[0127] If the call user belongs to the preset contacts, query the historical call records of the target user and the call user from the local database or the associated user information storage system. Among them, the historical call records include the emotional values of the target user at each time period during the call and the reasons for the deepening of emotional fluctuations. Traverse the historical call records, count the total number of calls between the target user and the call user, and the total number of targets whose reason for the deepening of emotional fluctuations is the call reason. Calculate the ratio of the total number of targets to the total number of calls.

[0128] If the ratio is greater than the preset ratio threshold, determine that the reason for the deepening of the emotional fluctuation of the target user is the call reason.

[0129] If the ratio is less than or equal to the preset ratio threshold, convert the audio data during the call into text through speech recognition technology. Then, preprocess the converted text, including removing special characters, stop words, etc. from the text, and then perform word segmentation on the text to split the long text into individual words or phrases.

[0130] For the text sentiment score, perform natural language processing on the preprocessed text. Adopt a sentiment analysis method based on a dictionary, construct a dictionary containing a large number of sentiment words, assign sentiment polarity (positive, negative or neutral), score and weight to each word, count the scores and weights of each sentiment word in the text, and calculate the text sentiment score through weighted calculation.

[0131] For the tone intensity of both parties, judge the intensity of the tone by counting the number and usage frequency of interjections and modal particles in the text, as well as factors such as the emotional intensity of the words. According to the number and usage frequency of interjections and modal particles, calculate the tone intensity score according to the preset calculation rules.

[0132] Regarding the attitudes of both parties, analyze whether there are words in the preset vocabulary set corresponding to attitudes such as accusation, criticism, support, and encouragement in the text, and obtain the preset attitude score corresponding to the number of words.

[0133] Based on the above text sentiment score, tone intensity score, and attitude score, conduct a quantitative evaluation. Different indicators and weights can be set, and the final evaluation score can be obtained through weighted calculation.

[0134] Judge whether there is an increase in emotional fluctuations caused by the call content according to the quantitative evaluation score. If the evaluation score is greater than the preset evaluation threshold, it is determined that the reason for the increase in the target user's emotional fluctuations is the call; if the evaluation score is less than or equal to the preset evaluation threshold, it is determined that the reason for the increase in emotional fluctuations is the user's own reason.

[0135] S320. Obtain the historical call records of the preset contacts.

[0136] Query the historical call records of the target user and each user in the preset contacts from the local database or the associated user information storage system.

[0137] S321. Determine the target contact set according to the historical call records.

[0138] Among them, the call between the target user and any contact in the target contact set can reduce the degree of emotional fluctuations.

[0139] Specifically, traverse the historical call records of each contact in the preset contacts, count the total number of calls between the target user and each contact, and the total number of target calls. Among them, the target call refers to the call in which the emotional value of the user at the time point corresponding to the end of the call is less than the emotional value at the time point corresponding to the start of the call, and the reason for the increase in emotional fluctuations is the call.

[0140] Divide the total number of target calls by the total number of calls to obtain the soothing effect value of each contact on the target user. Select the target contacts corresponding to the target soothing effect values whose soothing effect values exceed the preset effect threshold and store them in the target contact set.

[0141] S322. If the reason for the increase in fluctuations is the call, execute the first soothing strategy.

[0142] If the reason for the increase in fluctuations is the call, generate corresponding control instructions according to the target contact set, and send the control instructions to the target mobile terminal for the call operation of the target user. After receiving the control instructions, the target mobile terminal sends a call request to any contact in the target contact set.

[0143] S323. If the reason for the increase in fluctuations is the user's own reason, execute the second soothing strategy.

[0144] Specifically, if the reason for the deepening of the fluctuation is its own reason, establish a communication connection with the target mobile terminal and monitor the call requests received by the target mobile terminal in real time. When it is detected that the target mobile terminal receives a call request, obtain the call contact information of the sender of the call request, and compare the call contact information with the contacts in the target contact set. If the call contact does not belong to the target contact set, send a control instruction including rejecting the communication and sending a preset reply text message to the target mobile terminal. After receiving the instruction, the target mobile terminal rejects the communication request and sends a preset reply text message to the call contact.

[0145] S324. Obtain the historical health records of the target user within a preset time period.

[0146] Query the historical health records of the target user within a preset time period before the current time point from the local database or the associated user information storage system. Among them, the historical health records record data such as the physical parameters, physical states, and emotional states of the target user in each sub-time period within the preset time period.

[0147] S325. Determine the number of abnormal emotions of the target user within a preset time period.

[0148] Traverse the historical health records and count the total number of sub-time periods with abnormal emotional states of the target user within the preset time period as the number of abnormal emotions of the target user within the preset time period.

[0149] S326. When the number of abnormal emotions exceeds the preset number threshold, set the operating mode of the smart bracelet to the preset incentive mode.

[0150] When the number of emotional fluctuations exceeds the preset number threshold, obtain the preset incentive mode and set the operating mode of the smart bracelet to the incentive mode.

[0151] Among them, the incentive mode includes continuously scrolling positive encouragement statements on the bracelet screen, enabling higher-frequency health monitoring, and increasing the number of times to remind the user to carry out various activities beneficial to physical and mental health, etc., to help the user promptly detect their own emotional states. At the same time, relieve the user's negative emotions with positive psychological suggestions.

[0152] S327. Send the historical health records to a medical expert.

[0153] Send the historical health records to a medical expert through the built-in communication module.

[0154] S328. Update the health monitoring strategy of the target user.

[0155] Update the health monitoring strategy of the target user according to the feedback information received from the medical expert.

[0156] Specifically, obtain feedback information from medical experts. The feedback information includes physical monitoring suggestions put forward by medical experts for the current health status of the target user. The physical monitoring suggestions include the physiological indicators to be monitored and their corresponding monitoring frequencies. According to the physical monitoring suggestions in the feedback information, adjust the monitored physiological indicators and their corresponding monitoring frequencies in the health monitoring strategy of the target user. When the target user wears the smart bracelet, collect various physiological indicators of the target user according to the health monitoring strategy to monitor the health status of the target user.

[0157] S329. In the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is outside the preset position range, obtain the number of users within the preset range of the target user.

[0158] Specifically, in the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is outside the preset position range, determine whether the smart bracelet is in a moving state according to the real-time position of the smart bracelet.

[0159] If the smart bracelet is in an unmoved state, obtain the real-time position of the mobile terminal of the target user. When the real-time position of the mobile terminal is not within the preset range of the real-time position of the smart bracelet, send a warning message to the preset contact through the built-in communication module.

[0160] If the smart bracelet is in a moving state, obtain the real-time position information of the smart bracelet through the built-in positioning module of the smart bracelet, and at the same time use this position information as the center to delimit a preset geographical range (such as a circular area with the smart bracelet position as the center and a certain distance as the radius).

[0161] Then, communicate with the surrounding communication base stations to obtain the quantity information of user devices (such as mobile phones, smart watches, etc.) within the preset range of the target user. Among them, the communication base station can count the number of users by identifying the unique identifier (such as IMEI code, etc.) of the user device connected to the base station. Obtain the user quantity information fed back by the base station to get the number of users within the preset range of the target user.

[0162] In some embodiments, the number of users within the preset range of the target user can be obtained by communicating with the surrounding monitoring devices (such as cameras, sensors, etc.). The smart bracelet sends a request to the monitoring device, and the monitoring device uses its own image recognition, sensing and other technologies to identify and count the personnel within the covered area, and sends the statistical result to the smart bracelet. The smart bracelet processes and analyzes the obtained information to get the number of users within the preset range of the target user.

[0163] S330. When the number of users is less than the preset number threshold, obtain the current location and moving direction of the target user.

[0164] Specifically, when the number of users is less than the preset number threshold, obtain the current location and historical location movement data of the target user through the built-in positioning module.

[0165] Then, analyze and process the historical location movement data. Using the time series analysis method, arrange the location information at each time point in chronological order. By calculating the vector relationship between the locations at adjacent time points, obtain the displacement direction of the target user in different time periods.

[0166] Finally, obtain the displacement direction in the previous time period of the current time point as the moving direction of the target user.

[0167] S331. According to the current location and moving direction, predict the set of locations where the target user will be after moving a preset distance.

[0168] Specifically, according to the obtained current location (latitude and longitude coordinates) and moving direction (expressed in angles, such as the angle rotated clockwise from the due north direction) of the target user, combined with the preset moving distance, use the principle of trigonometric functions to calculate the possible location coordinates after moving the preset distance in this moving direction.

[0169] Considering the possible errors and uncertainties in the actual movement process, the smart bracelet will generate multiple possible location points within a certain range centered on the calculated location according to the preset rules. These location points together form a set of locations to simulate the possible deviations in the movement process of the target user.

[0170] S332. Obtain the danger coefficients of each location in the set of locations.

[0171] Search for the danger coefficients corresponding to each location in the set of locations in the preset danger coefficient correspondence table.

[0172] In some embodiments, a danger assessment system including various risk factors and their weights can be established to calculate the danger coefficients of each location in the set of locations.

[0173] First, query the geographic information database to obtain the topographic and geomorphic data of each location in the set of locations. Obtain the basic danger scores corresponding to this topographic and geomorphic data in the preset topographic and geomorphic correspondence table.

[0174] Next, retrieve the traffic information database to understand the traffic conditions of each location in the set of locations. Obtain the traffic danger scores corresponding to this traffic condition in the preset traffic condition correspondence table.

[0175] Then, communicate with the local public security department through the data interface of the local public security department to obtain the public security situation data of each location in the location set. Obtain the public security risk score corresponding to the public security situation data in the preset public security situation correspondence table.

[0176] Finally, perform weighted calculation according to the weights of each pre-set risk factor and their corresponding scores to obtain the risk coefficient of each location in the location set.

[0177] S333. When there are dangerous locations in the location set with a risk coefficient exceeding the preset risk threshold, prompt the target user to avoid the dangerous locations through a preset prompt method.

[0178] Specifically, when there are dangerous locations in the location set with a risk coefficient exceeding the preset risk threshold, prompt the target user to avoid the dangerous locations according to the pre-set prompt method. For example, the prompt method is a mobile phone voice prompt, and the smart bracelet will connect to the target user's mobile phone and send a control instruction. After receiving the instruction, the mobile phone automatically plays the voice prompt content and displays the coordinates of the dangerous location.

[0179] S334. When it is detected that the real-time location of the target user is a dangerous location, send the real-time location of the target user to the preset contact.

[0180] Obtain the real-time location of the target user in real time, compare the real-time location with the dangerous location, and when the real-time location is the same as the dangerous location, send the real-time location and warning information of the target user to the preset contact through the built-in communication module.

[0181] In the embodiment of the present application, by obtaining and analyzing the user's call data, identifying the emotional changes of the user during the call, and when the user's emotional fluctuations deepen, performing corresponding soothing strategies according to the reasons for the emotional fluctuations (such as emotional changes caused by call content or the user's own emotional problems), the efficiency and effect of emotional soothing are improved. At the same time, by real-time monitoring the user's location and moving direction, predicting the possible dangerous locations that the user may reach, and timely prompting the user to avoid the dangerous areas, the potential safety risks caused by the user's abnormal emotions are reduced. When the user enters a high-risk area, the warning mechanism is immediately triggered, and the real-time location information is sent to the preset contact so that the contact can take prompt actions to reduce the safety risks of the user in the dangerous area.

[0182] The personalized user parameter monitoring method in the embodiment of the present application is described above. Below, in combination with the above personalized user parameter monitoring method, the smart bracelet in the embodiment of the present application is described in detail.

[0183] Please refer to Figure 4 , which is an exemplary hardware structure schematic diagram of the smart bracelet in the embodiment of the present application.

[0184] In some embodiments, the smart bracelet 400 includes a computer device, which may be a terminal device. The computer device includes a processor 401, a memory 402, a sensor module 403, a communication module 404, an input device 405, and an output device 406 connected through a system bus. Among them, the processor 401 of the computer device is used to provide computing and control capabilities. The memory 402 of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database is used to store data. The sensor module 403 of the computer device is used to collect the user's body parameters, etc. The communication module 404 of the computer device is used to send information to the user's preset contacts and transmit communication data between devices such as a voice collection device, an image collection device, and a terminal device and the smart bracelet. The input device 405 of the computer device is used to receive instructions directly issued by the user or instructions issued by the user through a terminal device, etc. The output device 406 of the computer device is used to display the user's body parameters, etc. When the computer program is executed by the processor 401, it realizes the personalized user parameter monitoring method in the embodiments of the present application.

[0185] Those skilled in the art can understand that Figure 4 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different component layout.

[0186] In some embodiments of the present application, a computer-readable storage medium is provided, including instructions, which can cause the smart bracelet 400 to execute the personalized user parameter monitoring method in the embodiments of the present application when the instructions run on the smart bracelet 400.

[0187] In some embodiments of the present application, a computer program product is further provided, which can cause the smart bracelet 400 to execute the personalized user parameter monitoring method in the embodiments of the present application when the computer program product runs on the smart bracelet 400.

[0188] As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

[0189] As used in the foregoing embodiments, depending on the context, the term "when" may be construed to mean "if", "after", "in response to determining", or "in response to detecting". Similarly, depending on the context, the phrase "upon determining" or "if (the stated condition or event) is detected" may be construed to mean "if determined", "in response to determining", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0190] In the foregoing embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more integrated available media. The available medium may be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid-state drive), etc.

[0191] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the foregoing embodiments can be completed by relevant hardware instructed by a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it may include the processes of the foregoing method embodiments. The foregoing storage medium includes various media that can store program codes, such as ROM or random access memory RAM, magnetic disks, or optical discs.

Claims

1. A personalized user parameter monitoring method, characterized in that: include: When the target user is not wearing the smart bracelet, obtaining the emotional state of the target user when taking off the smart bracelet; In the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is within a preset position range, real-time video data collected by a preset image acquisition device is obtained; According to the real-time video data, identifying the real-time health status and real-time emotional status of the target user; When the target user cannot be identified in the real-time video data and the real-time emotional state is abnormal, obtaining target sound data of the target position where the target user is located; Determining the current behavior of the target user according to the target sound data, the target location, and the historical behavior record; Calculating the real-time similarity between the normal sound data corresponding to the current behavior and the real-time sound data at the target location; When the real-time similarity at each time point within the first preset time period is lower than a preset similarity threshold or abnormal sound data in a preset abnormal sound set exists in the real-time sound data, determining a prompting method for abnormal confirmation information according to the mobile terminal usage status of the target user and the personal information of the target user; In the case where the target device is controlled to prompt the target user to reply to the abnormality confirmation information in the prompting manner, if a normal reply of the target user is detected within a second preset time period, the target health status of the target user is determined according to the normal reply and the reply manner, wherein the normal reply is a reply indicating that there is no abnormality in the physical state of the target user; When the target health status is abnormal, an early warning message is sent to a preset contact.

2. The method according to claim 1, characterized in that In the case where the target device is controlled to prompt the target user to reply to the abnormal confirmation information in the prompting manner, if a normal reply of the target user is detected within a second preset time period, determining the target health status of the target user according to the normal reply and the reply manner specifically includes: In the case of controlling the display device to display the abnormal confirmation information to the target user in the display mode, if a normal reply of the target user is detected within a second preset time period, obtaining a reply mode of the normal reply; If the reply mode belongs to a first mode set, determining the target health state of the target user according to the sound feature data of the sound reply information and a preset sound feature library, the first mode set being to indicate a normal physical state through sound; If the reply mode belongs to the second mode set, obtaining the action operation data of the target user, wherein the second mode set is an action operation indicating a normal physical state by operating a mobile terminal or pressing a fixed button; The target health status is determined according to the motion operation data and the historical motion operation data of the target user.

3. The method according to claim 2, characterized in that The determining the target health status according to the action operation data and the historical action operation data of the target user specifically includes: Determining whether the action operation of the target user is abnormal according to the action operation data and the historical action operation data of the target user; If not, obtaining first sound data within a preset third time period before the prompt time point and second sound data within a preset third time period after the prompt time point, wherein the prompt time point is a time point for prompting the target user to reply to the abnormality confirmation information; Determine, according to the first sound data and the second sound data, a first real-time state of the sound collection device at the target location before the prompt time point and a second real-time state after the prompt time point; If the first real-time state of the sound collection device is normal and the second real-time state is abnormal, it is determined that the target health state is abnormal.

4. The method according to claim 1, characterized in that: After the step of sending a warning message to a preset contact when the target health status is abnormal, the method further includes: When it is detected that the target user is making a call, acquiring the call data of the target user; Determining the emotion value of the target user during the call according to the call data; When the emotion value of the target user at the end of the call exceeds the emotion value at the start of the call, determining the reason for the deepening of the emotion fluctuation of the target user according to the call data, wherein the reason for the deepening of the emotion fluctuation includes the call reason and the self reason; According to the cause of the deepening of the emotional fluctuation, a corresponding soothing strategy is executed, and the soothing strategy includes a first soothing strategy and a second soothing strategy. The first soothing strategy is to control the target terminal device to send a call request to any contact in the target contact set, and the second soothing strategy is to detect that when the target terminal device receives a call request from a contact other than the target contact set, reject the call request and send a preset reply text message to the contact who sent the call request.

5. The method according to claim 4, characterized in that The corresponding soothing strategies are implemented according to the reasons for the deepening of the emotional fluctuations, including: Obtaining historical call records of the preset contact; Determine a target contact set according to the historical call records, and the target user can talk to any contact in the target contact set to reduce the degree of emotional fluctuation; If the cause of the aggravated fluctuation is due to a call, the first appeasement strategy is implemented; If the cause of the deepening fluctuation is due to personal reasons, implement the second soothing strategy.

6. The method according to claim 1, characterized in that After the step of obtaining the emotional state of the target user when the target user takes off the smart bracelet when the target user is not wearing the smart bracelet, the method further includes: In the case where the emotional state is abnormal or the target user belongs to a preset special user group, if the smart bracelet is outside the preset position range, the number of users within the preset range of the target user is obtained; When the number of users is less than a preset number threshold, obtaining the current location and moving direction of the target user; Predicting, based on the current position and the moving direction, a set of positions where the target user will be located after moving a preset distance; Obtaining a risk factor for each position in the position set; When there is a dangerous location in the location set whose danger factor exceeds a preset danger threshold, the target user is prompted to avoid the dangerous location in a preset prompting manner.

7. The method according to claim 6, characterized in that After the step of prompting the target user to avoid the dangerous location in a preset prompting manner when there is a dangerous location in the location set with a danger coefficient exceeding a preset danger threshold, the method further includes: When it is detected that the real-time location of the target user is the dangerous location, the real-time location of the target user is sent to the preset contact.

8. The method according to claim 1, characterized in that: After the step of sending a warning message to a preset contact when the target health status is abnormal, the method further includes: Obtaining the historical health records of the target user within a preset time period; Determine the number of abnormal emotions of the target user within a preset time period based on the historical health records; When the number of abnormal emotions exceeds a preset number threshold, the operation mode of the smart bracelet is set to a preset incentive mode, and the historical health record is sent to a medical expert; The health monitoring strategy of the target user is updated according to the feedback information received from the medical expert.

9. A smart bracelet, characterized in that: include: one or more processors and memory; The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the smart bracelet to execute the method as described in any one of claims 1-8.

10. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on the smart bracelet, the smart bracelet executes the method as described in any one of claims 1 to 8.

Images