Multi-exercise health index measuring system and method based on intelligent wearing

Abstract

The invention provides a multi-exercise health index measurement system and method based on intelligent wearing. The method comprises the following steps: acquiring a multi-modal data stream during exercise of a target user; determining the motion type of the target user based on the human body biomechanical characteristics in different motion modes in combination with the multi-modal data stream, and identifying according to the motion type and the multi-modal data stream to obtain the motion characteristics of the target user in different time scales; distributing a feature influence coefficient of the motion feature under each time scale according to the association strength of the motion feature under each time scale and the key health index of the target user; determining the attenuation fluctuation of the motion function of the target user in the motion process according to all the feature influence coefficients and the motion features under each time scale; and based on the attenuation fluctuation and the health baseline during the motion of the target user, obtaining a measurement result of the motion health index of the target user. According to the scheme, on the basis of attenuation fluctuation, the exercise modes of the user can be distinguished, and the exercise health indexes can be accurately measured.

Description

Technical Field

[0001] This application relates to the field of behavioral feature recognition technology, and more specifically, to a multi-sports health indicator measurement system and method based on smart wearable devices. Background Technology

[0002] Behavioral feature recognition, as an important foundation for human-computer interaction and health monitoring, is evolving from single action detection to understanding complex behavioral patterns. Although traditional methods based on inertial sensors or video analysis are effective in specific scenarios, they face bottlenecks such as large interference from dynamic environments, strong individual behavioral differences, and semantic fragmentation of continuous behaviors. Especially in the field of wearable devices, existing technologies mostly rely on threshold segmentation or template matching, which makes it difficult to analyze the temporal correlation and contextual dependence in motion sequences, resulting in a decrease in cross-scenario behavior classification accuracy. With the development of multimodal sensor fusion and edge computing, the accuracy of behavioral feature recognition is becoming higher and higher, making the effect of health measurement more and more obvious.

[0003] In existing behavioral feature recognition, behavioral feature recognition mainly determines an individual's behavioral state by analyzing their behavioral patterns. First, sensors are used to collect behavioral data and extract key features. Then, machine learning or deep learning algorithms are used to model and classify these key features to achieve the recognition of specific behaviors. However, in behavioral feature recognition for measuring multiple sports health indicators, traditional health monitoring devices often use fixed algorithms to process sports data, which cannot distinguish the biomechanical differences of different sports modes (such as running / swimming / cycling). This makes the measured health indicators inconsistent with individual physiological characteristics, resulting in low accuracy of health indicator measurement results. Therefore, how to distinguish users' sports modes and accurately measure sports health indicators has become a challenge for the industry. Summary of the Invention

[0004] This application provides a multi-sports health indicator measurement system and method based on smart wearables, which can distinguish the user's sports mode and accurately measure sports health indicators.

[0005] In a first aspect, this application provides a method for measuring multiple sports health indicators based on smart wearables, comprising the following steps: Acquire multimodal data streams of physiological and motion parameters of the target user during exercise; Based on the biomechanical characteristics of the human body under different movement modes, combined with the multimodal data stream, the movement mode of the target user is dynamically identified to obtain the movement type of the target user. Based on the movement type, the multimodal data stream is used to identify features at multiple time scales to obtain the movement features of the target user at different time scales. Determine the correlation strength between motion features and key health indicators of the target user at each time scale, and then assign the feature influence coefficient of the motion features at each time scale based on all the correlation strengths; The decline and fluctuation of the target user's motor function during exercise are determined based on all the characteristic influence coefficients and the motion characteristics at various time scales. The measurement results of the target user's exercise health index are obtained based on the attenuation fluctuations and the target user's health baseline during exercise.

[0006] In some embodiments, the movement patterns of a target user are dynamically identified based on the biomechanical characteristics of the human body under different movement modes, combined with the multimodal data stream, to obtain the movement type of the target user, specifically including: Identify multiple modal features of the target user's movement from the multimodal data stream; All modal features are fused to obtain the fused modal features of the target user during movement; Extracting human biomechanical characteristics under different motion patterns; Select a motion mode as the selected motion mode, and perform correlation analysis between the human biomechanical characteristics of the selected motion mode and the fused modal characteristics to obtain the motion correlation degree of the selected motion mode. Continue to determine the motion correlation of the remaining motion patterns; Based on the correlation between various sports, the sports types of the target user are selected from all sports modes.

[0007] In some embodiments, performing feature recognition on the multimodal data stream at multiple time scales based on the motion type to obtain the motion features of the target user at different time scales specifically includes: The time scale level for feature recognition of target users is set according to the type of movement. The preprocessed multimodal data stream is divided according to the time scale hierarchy to obtain data subsets at multiple time scales; Select a time scale as the selected time scale, and perform multimodal feature recognition on a subset of data under the selected time scale to obtain multiple modal recognition features; Determine the motion characteristics of the target user at a selected time scale based on all modal recognition features; Continue to determine the motion characteristics of the target user over the remaining time scale.

[0008] In some embodiments, determining the correlation strength between motion characteristics at each time scale and key health indicators of the target user specifically includes: Obtain key health metrics for target users; Correlation analysis was performed between the motion characteristics at various time scales and the key health indicators to obtain the correlation strength between the motion characteristics and the health indicators at each time scale.

[0009] In some embodiments, assigning the feature influence coefficient of the motion feature at each time scale based on all correlation strengths specifically includes: Determine the contribution of the target user's motion characteristics at each time scale; Normalize all association strengths to obtain a normalized set of association strengths; Calculate the feature weights of motion features at each time scale based on the set of correlation strengths; The feature influence coefficient of the motion feature at each time scale is determined by the feature weight and the contribution of the motion feature at each time scale.

[0010] In some embodiments, determining the attenuation fluctuation of the target user's motor function during exercise based on all characteristic influence coefficients and motion characteristics at various time scales specifically includes: By performing a weighted fusion analysis of motion characteristics at various time scales and their corresponding feature influence coefficients, all motion function indices of the target user during the exercise process are obtained. Based on all the motor function indices, determine the decline and fluctuation of the target user's motor function during exercise.

[0011] In some embodiments, the multimodal data stream includes physiological parameter data and motion parameter data.

[0012] Secondly, this application provides a multi-sports health indicator measurement system based on smart wearable devices, comprising: The acquisition module is used to acquire multimodal data streams of physiological and motion parameters of the target user during exercise; The processing module is used to dynamically identify the target user's movement pattern based on the human biomechanical characteristics under different movement modes and the multimodal data stream, to obtain the target user's movement type, and to perform feature identification on the multimodal data stream at multiple time scales based on the movement type, to obtain the target user's movement characteristics at different time scales. The processing module is also used to determine the correlation strength between motion features and key health indicators of the target user at each time scale, and then assign the feature influence coefficient of the motion features at each time scale based on all the correlation strengths. The processing module is also used to determine the attenuation fluctuation of the target user's motor function during exercise based on all the feature influence coefficients and the motion characteristics at each time scale; The execution module is used to extract the measurement results of the target user's exercise health indicators based on the attenuation fluctuations and the target user's health baseline during exercise.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for measuring multiple sports health indicators based on smart wearables.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for measuring multiple sports health indicators based on smart wearables.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The multi-sports health index measurement system and method based on smart wearables provided in this application first acquires multimodal data streams of physiological and exercise parameters of the target user during exercise; secondly, based on the human biomechanical characteristics under different exercise modes and combined with the multimodal data streams, the exercise mode of the target user is dynamically identified to obtain the exercise type of the target user, and feature identification at multiple time scales is performed on the multimodal data streams according to the exercise type to obtain the exercise characteristics of the target user at different time scales; further, the correlation strength between the exercise characteristics at each time scale and the key health indicators of the target user is determined, and then the feature influence coefficient of the exercise characteristics at each time scale is assigned according to all the correlation strengths; then, the decay fluctuation of the target user's motor function during exercise is determined according to all the feature influence coefficients and the exercise characteristics at each time scale; finally, the measurement results of the target user's sports health index are extracted based on the decay fluctuations and the target user's health baseline during exercise.

[0016] Therefore, this application can distinguish users' exercise patterns and accurately measure exercise health indicators. First, it acquires multimodal data streams of physiological and exercise parameters of the target user during exercise to comprehensively reflect the user's physiological state and dynamic behavior during exercise, thus providing a rich information foundation for accurate measurement of health indicators. Second, based on the biomechanical characteristics of the human body under different exercise patterns, combined with the multimodal data stream, it dynamically identifies the target user's exercise type to conduct targeted health indicator assessments. Furthermore, based on the exercise type, it performs feature identification of the multimodal data stream at multiple time scales to obtain the target user's exercise characteristics at different time scales, accurately depicting the target user's exercise state at different time scales. This provides data support for personalized exercise monitoring and health assessment, thereby avoiding the problems caused by biomechanical differences between different exercise patterns. The method addresses the following: First, it assesses the error in evaluating exercise health indicators. Second, it assigns a feature influence coefficient for each time scale based on the correlation strength between exercise characteristics and key health indicators of the target user. This accurately assesses the contribution of exercise characteristics to health status at different time scales, ensuring more precise and personalized health assessments. Third, it determines the decay fluctuations in the target user's motor function during exercise based on all feature influence coefficients and exercise characteristics at each time scale. This effectively identifies the target user's energy consumption patterns and fatigue occurrence, providing a basis for measuring the target user's exercise health indicators. Finally, it extracts the target user's exercise health indicators based on decay fluctuations and the target user's health baseline during exercise. In summary, the technical solution provided in this application can distinguish users' exercise patterns and accurately measure exercise health indicators. Attached Figure Description

[0017] Figure 1 This is an exemplary flowchart of a multi-sports health index measurement method based on smart wearables, according to some embodiments of this application; Figure 2 This is an exemplary flowchart illustrating the determination of motion features according to some embodiments of this application; Figure 3 This is an exemplary flowchart illustrating the determination of feature influence coefficients according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a multi-sports health index measurement system based on smart wearables, according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device that implements a multi-sports health index measurement method based on smart wearables, according to some embodiments of this application. Detailed Implementation

[0018] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] refer to Figure 1 The figure is an exemplary flowchart of a multi-sports health index measurement method based on smart wearables according to some embodiments of this application. The multi-sports health index measurement method 100 based on smart wearables mainly includes the following steps: In step 101, a multimodal data stream of physiological and motion parameters of the target user during exercise is acquired.

[0020] In practice, the multimodal data stream of physiological and exercise parameters of the target user during exercise is obtained through the data processing terminal of the smart wearable multi-sports health index measurement device configured on the target user. The multimodal data stream includes physiological parameter data and exercise parameter data.

[0021] It should be noted that the intelligent wearable multi-sports health indicator measurement device includes at least one physiological parameter sensor unit with the ability to collect physiological parameters and at least one motion parameter sensor unit with the ability to collect motion parameters. The physiological parameter sensor unit includes a heart rate sensor, a blood oxygen sensor, a respiratory rate sensor, etc., and the motion parameter sensor unit includes an accelerometer, a gyroscope, a GPS module, a cadence sensor, etc. During the target user's exercise, each sensor unit collects physiological parameters and motion parameters in real time according to a predetermined sampling rate, generating a raw data stream. The raw data collected by each sensor unit is transmitted to the data processing terminal in real time through a built-in wireless communication module (such as Bluetooth or Wi-Fi), with timestamp information attached to achieve data synchronization. The data processing terminal integrates the received data from each sensor to form a multimodal data stream containing physiological parameters and motion parameters, providing basic data support for subsequent data preprocessing, feature extraction, and dynamic recognition of exercise patterns.

[0022] It should also be noted that, in this application, physiological parameters refer to indicators reflecting the user's physiological state, such as heart rate, blood oxygen, and respiratory rate; in this application, motion parameters refer to indicators reflecting the user's dynamic physical characteristics during exercise, such as acceleration, angular velocity, cadence, and displacement; in this application, multimodal data stream refers to an information set that simultaneously contains physiological parameter data and motion parameter data. The multimodal data stream simultaneously collects physiological parameters (such as heart rate, blood oxygen, and respiratory rate) and motion parameters (such as acceleration, angular velocity, cadence, and displacement), which can comprehensively reflect the user's physiological state and dynamic behavior during exercise from multiple dimensions, thereby providing a rich information foundation for the accurate measurement of health indicators.

[0023] In step 102, the target user's movement pattern is dynamically identified based on the human biomechanical characteristics under different movement modes and the multimodal data stream to obtain the target user's movement type. Based on the movement type, the multimodal data stream is used to identify features at multiple time scales to obtain the target user's movement features at different time scales.

[0024] In some embodiments, the movement pattern of a target user can be dynamically identified based on the biomechanical characteristics of the human body under different movement modes, combined with the multimodal data stream, to obtain the movement type of the target user. This can be achieved through the following steps: Identify multiple modal features of the target user's movement from the multimodal data stream; All modal features are fused to obtain the fused modal features of the target user during movement; Extracting human biomechanical characteristics under different motion patterns; Select a motion mode as the selected motion mode, and perform correlation analysis between the human biomechanical characteristics of the selected motion mode and the fused modal characteristics to obtain the motion correlation degree of the selected motion mode. Continue to determine the motion correlation of the remaining motion patterns; Based on the correlation between various sports, the sports types of the target user are selected from all sports modes.

[0025] In specific implementation, firstly, multiple modal features of the target user during movement are identified from the multimodal data stream. Specifically, physiological parameter data and motion parameter data are extracted from the multimodal data stream. The mean of all physiological parameter values ​​in the physiological parameter data is calculated, and the result is used as the physiological modal feature of the target user during movement. Similarly, the mean of all motion parameter data in the motion parameter data is calculated, and the result is used as the motion modal feature of the target user during movement. Then, the physiological modal features and the motion modal features are combined to obtain multiple modal features of the target user during movement. The physiological parameter data represents multiple... The set of physiological parameter values ​​is represented by respiratory rate in this embodiment. The motion parameter data represents a set containing multiple motion parameter values, represented by acceleration in this embodiment. Furthermore, other types of parameter values ​​can be used to represent the physiological parameter values ​​and motion characteristic parameter values ​​in other embodiments, which are not limited here. Next, all modal features are fused to obtain the fused modal features of the target user during exercise. That is, all modal features are weighted and fused to obtain the fused modal features of the target user during exercise. The weight of each modal feature can be set from 0 to [missing value] according to its influence on exercise health measurement. The range is not limited to 1; further, human biomechanical features under different exercise modes are extracted from the smart wearable multi-sports health monitoring database. These biomechanical features are extracted using a dynamic time warping method based on historical multimodal data streams, which will not be elaborated further here. The different exercise modes include running, cycling, swimming, etc.; even further, one exercise mode is selected as the chosen exercise mode, and the human biomechanical features under the chosen exercise mode are correlated with the fused modal features to obtain the exercise correlation degree of the chosen exercise mode. That is: selecting one exercise mode as the chosen exercise mode, and using cosine similarity to analyze the human biomechanical features under the chosen exercise mode... The motion correlation degree of the selected motion mode is calculated by combining the features with the fused modal features. In addition, in other embodiments, other correlation calculation methods can be used to calculate the motion correlation degree of the selected motion mode, which is not limited here. Then, the motion correlation degree of the remaining motion modes is determined by the method of "conducting correlation analysis between the human biomechanical features in the selected motion mode and the fused modal features to obtain the motion correlation degree of the selected motion mode". Finally, the motion type of the target user is selected from all motion modes according to each motion correlation degree, that is, the motion mode corresponding to the largest motion correlation degree is selected as the motion type of the target user.

[0026] It should be noted that in this embodiment, modal features represent feature data of different dimensions, and the modal features reflect the feature states of information of various dimensions during the target user's movement; in this embodiment, fused modal features represent the feature values ​​after the fusion of multiple modal features; in this embodiment, the smart wearable multi-movement health monitoring database refers to a comprehensive data platform, mainly used to store, manage and analyze multimodal data collected by smart wearable devices during exercise and health monitoring. This database not only includes raw physiological parameter data (such as heart rate, blood oxygen saturation, respiratory rate, etc.) and motion parameter data (such as acceleration, angular velocity, cadence, displacement, etc.), but also feature data after preprocessing, feature extraction and fusion, as well as the user's health... Metadata such as baseline, exercise history, and environmental information; in this embodiment, human biomechanical characteristics represent the mechanically related movement characteristics during human movement; in this embodiment, the movement correlation degree represents the degree of correlation between the current movement characteristics and the movement characteristics under different movement modes. The movement correlation degree reflects the similarity between the target user's current movement state and each movement mode. By determining the movement correlation degree, the movement mode that best matches the target user's actual movement type can be effectively selected. In this application, the movement type represents the current movement mode of the target user. By determining the movement type, targeted health indicator assessments can be performed on the target user, thereby avoiding the error in the assessment of exercise health indicators caused by the biomechanical differences of different movement modes.

[0027] In some embodiments, reference Figure 2 As shown in the figure, this is an exemplary flowchart of determining motion features according to some embodiments of this application. In this embodiment, feature recognition of the multimodal data stream at multiple time scales based on the motion type to obtain the motion features of the target user at different time scales can be achieved by the following steps: First, in step 1021, the time scale level for feature recognition of the target user is set according to the type of motion; Secondly, in step 1022, the preprocessed multimodal data stream is divided according to the time scale hierarchy to obtain data subsets at multiple time scales; Furthermore, in step 1023, a time scale is selected as the selected time scale, and multimodal feature recognition is performed on the data subset under the selected time scale to obtain multiple modal recognition features; Then, in step 1024, the motion characteristics of the target user at the selected time scale are determined based on all modal recognition features; Finally, in step 1025, the motion characteristics of the target user over the remaining time scale are determined.

[0028] In specific implementation, firstly, the time scale hierarchy for feature recognition of the target user is set according to the motion type, that is, the time scale hierarchy for feature recognition of the target user is divided into short time scales and long time scales. Furthermore, in other embodiments, it can be divided into other time scale hierarchies, which are not limited here. Secondly, the preprocessed multimodal data stream is segmented according to the time scale hierarchy to obtain data subsets at multiple time scales, that is, the preprocessed multimodal data stream is segmented according to the time scale hierarchy through a preset time window to obtain data subsets at multiple time scales, for example, the multimodal data stream is divided into data subsets at short time scales and data subsets at long time scales. Further, a time scale is selected as the selected time scale, and multimodal feature recognition is performed on the data subset at the selected time scale to obtain multiple modal recognition features, that is, a time scale is selected as the selected time scale, and the corresponding feature extraction algorithm is used to extract features from different modal data in the data subset at the selected time scale to obtain multiple modal recognition features. The modal recognition features include physiological modal data and motion modal data. For example, the motion modal data in the data subset is used to extract features by calculating the standard deviation, and the standard deviation result is used as the modal recognition feature corresponding to the motion modal data. The motion modal data uses acceleration data. The physiological modal data in the data subset can be extracted by Fourier transform to extract frequency domain data, and the mean of the frequency domain data is used as the modal recognition feature corresponding to the physiological modal data, thus obtaining multiple modal recognition features. Then, the motion features of the target user at the selected time scale are determined based on all modal recognition features. That is, all modal recognition features are weighted and fused to obtain the motion features of the target user at the selected time scale. The weight of each modal recognition feature can be set between 0 and 1 according to the influence of the corresponding modality on the motion feature, which is not limited here. Finally, the motion features of the target user at the remaining time scales are determined by the method of "determining the motion features of the target user at the selected time scale based on all modal recognition features".

[0029] It should be noted that, in this embodiment, the time scale level refers to the level of the time dimension. Specifically, in this embodiment, the time scale level refers to the dimension obtained by dividing the time series into multiple levels in order to analyze the movement characteristics of the target user during the exercise and health monitoring process. Different time scale levels correspond to different granularities of exercise analysis. In this embodiment, the data subset represents a data segment in the multimodal data stream. Specifically, in this embodiment, the data subset represents a data segment divided from the multimodal data stream at different time scale levels. These data subsets contain the movement and physiological parameter information of the target user in a specific time dimension, which are used for further feature extraction and analysis. In this embodiment, the modality recognition feature represents the feature that can characterize the movement state of the target user under different data modalities. In this application, the movement feature represents the parameter that describes the state of the target user during the movement process. By determining the movement feature, the movement state of the target user at different time scales can be accurately characterized, providing data support for personalized exercise monitoring and health assessment.

[0030] In step 103, the correlation strength between motion features and key health indicators of the target user at each time scale is determined, and then the feature influence coefficient of the motion features at each time scale is assigned based on all the correlation strengths.

[0031] In some embodiments, determining the correlation strength between motion characteristics and key health indicators of the target user at each time scale can be achieved through the following steps: Obtain key health metrics for target users; Correlation analysis was performed between the motion characteristics at various time scales and the key health indicators to obtain the correlation strength between the motion characteristics and the health indicators at each time scale.

[0032] In specific implementation, firstly, the key health indicators of the target user are obtained, that is, the key health indicators of the target user are obtained through the target user's exercise type. For example, when the target user's exercise type is running, heart rate is used as the key health indicator; when the target user's exercise type is swimming, blood pressure is used as the key health indicator. There are no restrictions on other exercise types. Then, the correlation analysis between the exercise characteristics at each time scale and the key health indicators is performed to obtain the correlation strength between the exercise characteristics and the health indicators at each time scale. That is, the correlation strength between the exercise characteristics and the health indicators at each time scale is calculated using the Pearson correlation coefficient. This will not be elaborated here. In addition, in other embodiments, the correlation strength between the exercise characteristics and the health indicators at each time scale can also be calculated using the following steps, which are not limited here.

[0033] It should be noted that, in this embodiment, the key health indicators refer to the exercise parameters used to assess the health status of the target user. These key health indicators are usually highly correlated with the user's health and can reflect the changing trend of their physical condition. In this application, the correlation strength represents the degree of correlation between the target user's exercise characteristics and their key health indicators at different time scales. The correlation strength reflects the degree of influence of specific exercise characteristics on health status. By determining the correlation strength, the relationship between exercise characteristics and health indicators can be effectively identified, thereby optimizing the health assessment model to select features closely related to health status and improve the accuracy of exercise health assessment.

[0034] In some embodiments, reference Figure 3 As shown in the figure, this is an exemplary flowchart of determining feature influence coefficients according to some embodiments of this application. In this embodiment, the allocation of feature influence coefficients for motion features at each time scale based on all correlation strengths can be achieved by the following steps: First, in step 1031, the contribution of the target user's motion features at each time scale is determined; Secondly, in step 1032, all correlation strengths are normalized to obtain a normalized set of correlation strengths; Then, in step 1033, the feature weights of the motion features at each time scale are calculated based on the set of correlation strengths; Finally, in step 1034, the feature influence coefficient of the motion feature at each time scale is determined by the feature weight and motion feature contribution degree corresponding to each time scale.

[0035] In specific implementation, firstly, the contribution of the target user's motion characteristics at each time scale is determined. This contribution can be set between 0 and 1 by assessing the impact of motion characteristics on health indicators at each time scale. For example, the contribution of the target user's motion characteristics at a short time scale is set to 0.43, and at a long time scale to 0.57. In other embodiments, the contribution can be determined according to actual needs. Secondly, all correlation strengths are normalized to obtain a normalized set of correlation strengths. This is achieved by using minimum-maximum normalization to normalize each correlation strength and combining all normalized correlation strengths. Then, the motion characteristics contribution at each time scale is calculated based on this set of correlation strengths. The feature weights of the motion features are calculated as follows: the mean of all association strengths in the association strength set is calculated, and for each time scale, the quotient of the association strength value at the time scale and the mean is used as the feature weight of the motion feature at the time scale, thus obtaining the feature weight of the motion feature at each time scale. In addition, in other embodiments, other calculation methods can be used to calculate the feature weights of the motion features at each time scale, which are not limited here. Finally, the feature influence coefficient of the motion feature at each time scale is determined by the feature weights and the contribution of the motion feature corresponding to each time scale, that is: for each time scale, the feature weights and the contribution of the motion feature corresponding to the time scale are multiplied, and the result of the multiplication is used as the feature influence coefficient of the motion feature at the time scale, thus obtaining the feature influence coefficient of the motion feature at each time scale.

[0036] It should be noted that, in this embodiment, the contribution degree of the motion feature represents the degree of influence of the motion feature on the change of the target health indicator. The contribution degree of the motion feature reflects the intensity of the influence of the motion feature on the health indicator during exercise. In this embodiment, the association strength set represents the combination of multiple association strengths. In this embodiment, the feature weight represents the weight value assigned to the motion feature. The feature weight reflects the importance or influence of different motion features on the target health indicator at multiple time scales. In this application, the feature influence coefficient represents the comprehensive influence of the motion feature on the target user's health indicator. The feature influence coefficient quantifies the role of each motion feature in the change of health indicator at each time scale. By calculating the feature influence coefficient, the contribution of the motion feature to the health status at different time scales can be accurately assessed, thereby ensuring that the health assessment is more accurate and personalized. A motion feature with a larger feature influence coefficient indicates that it has a greater impact on the change of health indicator, and vice versa. Therefore, the feature influence coefficient can provide a basis for individualized health indicator measurement.

[0037] In step 104, the decay fluctuation of the target user's motor function during exercise is determined based on all the characteristic influence coefficients and the motion characteristics at each time scale.

[0038] In some embodiments, determining the attenuation fluctuation of a target user's motor function during exercise based on all characteristic influence coefficients and motion characteristics at various time scales can be achieved through the following steps: By performing a weighted fusion analysis of motion characteristics at various time scales and their corresponding feature influence coefficients, all motion function indices of the target user during the exercise process are obtained. Based on all the motor function indices, determine the decline and fluctuation of the target user's motor function during exercise.

[0039] In practice, firstly, the motion characteristics at each time scale are weighted and fused with the corresponding feature influence coefficients to obtain all the motion function indices of the target user during the exercise process. That is, for each time scale, the motion characteristics and feature influence coefficients corresponding to the time scale are multiplied, and the result of the multiplication calculation is used as the motion function index of the target user during the exercise process, thus obtaining all the motion function indices of the target user during the exercise process. Then, based on all the motion function indices, the decay fluctuation of the target user's motion function during the exercise process is determined. That is, the standard deviation of all the motion function indices is calculated, and the result of the standard deviation calculation is used as the decay fluctuation of the target user's motion function during the exercise process.

[0040] It should be noted that, in this embodiment, the motor function index represents a quantitative indicator of the target user's performance during exercise. The motor function index characterizes the user's physiological function during exercise, and its level reflects the target user's physical condition, exercise intensity, endurance level, and potential fatigue. In this embodiment, the motor function change curve represents a graphical representation of the target user's motor function index changing over time. The curve shows the fluctuation trend, fluctuations, and decay or increase of the motor function index during exercise. In this application, decay fluctuations represent the... The degree of fluctuation in motor function over time is considered. The greater the decay fluctuation, the greater the degree of fluctuation in motor function over time; the smaller the decay fluctuation, the smaller the degree of fluctuation in motor function over time. The decay fluctuation is not just a single linear decline, but includes periodic fluctuations. That is, at certain times, motor function may briefly recover or experience slight fluctuations (such as slight fatigue recovery or consumption fluctuations), but the overall trend is a gradual decline. By determining the decay fluctuation, the pattern of physical consumption and the occurrence of fatigue during exercise can be effectively judged, thus providing a basis for measuring the target user's sports health indicators.

[0041] In step 105, the measurement results of the target user's exercise health index are obtained based on the attenuation fluctuation and the target user's health baseline during exercise.

[0042] In some embodiments, the measurement results of the target user's exercise health index extracted based on the attenuation fluctuation and the target user's health baseline during exercise can be obtained by the following steps: Obtain the target user's health baseline during exercise; The attenuation fluctuations are mapped and compared with the health baseline to obtain the measurement results of the target user's exercise health indicators.

[0043] In specific implementation, firstly, the target user's health baseline is obtained from the smart wearable multi-sports health monitoring database. This health baseline is obtained by evaluating the historical decline fluctuations of motor function through an existing health assessment model. The range formed by the minimum and maximum decline fluctuations of historical motor function is used as the health baseline, which will not be elaborated here. Then, the decline fluctuations are mapped and compared with the health baseline to obtain the measurement results of the target user's sports health indicators. That is, the lower limit and upper limit of the health baseline are extracted, the difference between the decline fluctuations and the upper limit is calculated to obtain the first difference calculation result, the difference between the upper limit and the lower limit is calculated to obtain the second difference calculation result, and the quotient of the first difference calculation result and the second difference calculation result is used as the measurement result of the target user's sports health indicators.

[0044] It should be noted that in this application, the health baseline represents the benchmark range for assessing exercise health. By determining the health baseline, a reference standard can be effectively provided for an individual's exercise health status, thereby determining the measurement results of the user's exercise health.

[0045] In another aspect, in some embodiments, this application provides a multi-sports health indicator measurement system based on smart wearables, with reference to... Figure 4 The figure is a schematic diagram of the structure of a multi-sports health index measurement system based on smart wearables according to some embodiments of this application. The multi-sports health index measurement system 200 based on smart wearables includes: an acquisition module 201, a processing module 202, and an execution module 203, which are described below: The acquisition module 201 in this application is mainly used to acquire multimodal data streams of physiological parameters and motion parameters of the target user during exercise; Processing module 202, in this application, is mainly used to dynamically identify the movement pattern of the target user based on the human biomechanical characteristics under different movement modes and the multimodal data stream, to obtain the movement type of the target user, and to perform feature identification of the multimodal data stream at multiple time scales according to the movement type, to obtain the movement characteristics of the target user at different time scales; The processing module 202 is also used to determine the correlation strength between motion features and key health indicators of the target user at each time scale, and then assign the feature influence coefficient of the motion features at each time scale based on all the correlation strengths. In addition, the processing module 202 is also used to determine the attenuation fluctuation of the target user's motor function during exercise based on all the feature influence coefficients and the motion characteristics at each time scale; The execution module 203 in this application is mainly used to extract the measurement results of the target user's exercise health index based on the attenuation fluctuation and the target user's health baseline during exercise.

[0046] In addition, this application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described method for measuring multiple sports health indicators based on smart wearables.

[0047] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device implementing a multi-sports health index measurement method based on smart wearables, according to some embodiments of this application. The multi-sports health index measurement method based on smart wearables in the above embodiments can be... Figure 5 The computer device shown is used to implement this, and the computer device 300 includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0048] The processor 301 can be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more devices used to control the execution of the multi-sports health index measurement method based on smart wearables in this application.

[0049] The communication bus 302 can be used to transmit information between the aforementioned components.

[0050] The memory 303 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CDROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via a communication bus 302. The memory 303 may also be integrated with the processor 301.

[0051] The memory 303 stores program code for executing the solution of this application, and its execution is controlled by the processor 301. The processor 301 executes the program code stored in the memory 303. The program code may include one or more software modules. In the above embodiments, the determination of the multi-sports health index measurement method based on smart wearables can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0052] Communication interface 304 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0053] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single CPU) processor or a multi-core (multi CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0054] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0055] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for measuring multiple sports health indicators based on smart wearables.

[0056] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0057] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for measuring multiple sports health indicators based on smart wearables, characterized in that, Includes the following steps: Acquire multimodal data streams of physiological and motion parameters of the target user during exercise; Based on the biomechanical characteristics of the human body under different movement modes, combined with the multimodal data stream, the movement mode of the target user is dynamically identified to obtain the movement type of the target user. Based on the movement type, the multimodal data stream is used to identify features at multiple time scales to obtain the movement features of the target user at different time scales. Determine the correlation strength between motion features and key health indicators of the target user at each time scale, and then assign the feature influence coefficient of the motion features at each time scale based on all the correlation strengths; The decline and fluctuation of the target user's motor function during exercise are determined based on all the characteristic influence coefficients and the motion characteristics at various time scales. The measurement results of the target user's exercise health index are obtained based on the attenuation fluctuations and the target user's health baseline during exercise.

2. The method as described in claim 1, characterized in that, Based on the biomechanical characteristics of the human body under different movement modes, combined with the multimodal data stream, the movement modes of the target user are dynamically identified, and the specific movement types of the target user include: Identify multiple modal features of the target user's movement from the multimodal data stream; All modal features are fused to obtain the fused modal features of the target user during movement; Extracting human biomechanical characteristics under different motion patterns; Select a motion mode as the selected motion mode, and perform correlation analysis between the human biomechanical characteristics of the selected motion mode and the fused modal characteristics to obtain the motion correlation degree of the selected motion mode. Continue to determine the motion correlation of the remaining motion patterns; Based on the correlation between various sports, the sports types of the target user are selected from all sports modes.

3. The method as described in claim 1, characterized in that, Based on the motion type, feature recognition is performed on the multimodal data stream at multiple time scales to obtain the motion features of the target user at different time scales, specifically including: The time scale level for feature recognition of target users is set according to the type of movement. The preprocessed multimodal data stream is divided according to the time scale hierarchy to obtain data subsets at multiple time scales; Select a time scale as the selected time scale, and perform multimodal feature recognition on a subset of data under the selected time scale to obtain multiple modal recognition features; Determine the motion characteristics of the target user at a selected time scale based on all modal recognition features; Continue to determine the motion characteristics of the target user over the remaining time scale.

4. The method as described in claim 1, characterized in that, Determining the correlation strength between motion characteristics and key health indicators of target users at each time scale specifically includes: Obtain key health metrics for target users; Correlation analysis was performed between the motion characteristics at various time scales and the key health indicators to obtain the correlation strength between the motion characteristics and the health indicators at each time scale.

5. The method as described in claim 1, characterized in that, The feature influence coefficients for motion features at each time scale are assigned based on all correlation strengths, specifically including: Determine the contribution of the target user's motion characteristics at each time scale; Normalize all association strengths to obtain a normalized set of association strengths; Calculate the feature weights of motion features at each time scale based on the set of correlation strengths; The feature influence coefficient of the motion feature at each time scale is determined by the feature weight and the contribution of the motion feature at each time scale.

6. The method as described in claim 1, characterized in that, Based on all the characteristic influence coefficients and motion characteristics at various time scales, the specific factors determining the decline and fluctuation of the target user's motor function during exercise include: By performing a weighted fusion analysis of motion characteristics at various time scales and their corresponding feature influence coefficients, all motion function indices of the target user during the exercise process are obtained. Based on all the motor function indices, determine the decline and fluctuation of the target user's motor function during exercise.

7. The method as described in claim 1, characterized in that, The multimodal data stream includes physiological parameter data and motion parameter data.

8. A multi-sports health indicator measurement system based on smart wearable devices, characterized in that, include: The acquisition module is used to acquire multimodal data streams of physiological and motion parameters of the target user during exercise; The processing module is used to dynamically identify the target user's movement pattern based on the human biomechanical characteristics under different movement modes and the multimodal data stream, to obtain the target user's movement type, and to perform feature identification on the multimodal data stream at multiple time scales based on the movement type, to obtain the target user's movement characteristics at different time scales. The processing module is also used to determine the correlation strength between motion features and key health indicators of the target user at each time scale, and then assign the feature influence coefficient of the motion features at each time scale based on all the correlation strengths. The processing module is also used to determine the attenuation fluctuation of the target user's motor function during exercise based on all the feature influence coefficients and the motion characteristics at each time scale; The execution module is used to extract the measurement results of the target user's exercise health indicators based on the attenuation fluctuations and the target user's health baseline during exercise.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the multi-sports health indicator measurement method based on smart wearables as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the multi-sports health index measurement method based on smart wearables as described in any one of claims 1 to 7.

Images