Intelligent rehabilitation monitoring system and method

Abstract

The invention provides an intelligent rehabilitation monitoring system and method, and the method comprises the steps: monitoring the health state of a target patient through a sensor, and obtaining health monitoring data of different monitoring items; performing pathological abnormality association analysis on the health feature values extracted from the health monitoring data to obtain the abnormality association degree between the monitoring items on the pathological structure; performing quality mapping evaluation on the health monitoring data of different monitoring items to obtain a quality score of each piece of health monitoring data, and determining the data credibility of each piece of health monitoring data according to the abnormal association degree between all the quality scores and the monitoring items on the pathological structure; when the data credibility of the health monitoring data is smaller than a preset credibility threshold value, the health monitoring data of the monitoring items are reconstructed according to the abnormal association degree of the monitoring items on the pathological structure, and the health index of the target patient in rehabilitation training is generated according to the reconstructed health monitoring data. By adopting the scheme of the invention, the reliability association evaluation of the health monitoring data can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of rehabilitation data processing, and more specifically, to an intelligent rehabilitation monitoring system and method. Background Art

[0002] Rehabilitation data processing plays an important role in health management, especially in supporting personalized rehabilitation training for patients and optimizing treatment effects. With the development of sensor technology, the Internet of Things and artificial intelligence technology, more and more health monitoring data can be collected and analyzed in real time, covering multi-dimensional monitoring items including heart rate, blood oxygen, motion data, electromyography, physiological signals, etc. The traditional rehabilitation process relies on the experience and subjective judgment of medical staff, but with the increase in data volume and the improvement of processing requirements, automated and intelligent data processing methods have gradually become the focus of research. Rehabilitation data processing not only needs to accurately capture the health status, but also needs to conduct in-depth analysis of these data to evaluate the patient's rehabilitation progress and identify potential abnormal changes, and further provide a scientific basis for the adjustment of the rehabilitation plan.

[0003] Existing technologies mainly rely on a single sensor or a simple multi-sensor fusion method in health monitoring of patient rehabilitation training, which usually leads to insufficient integrity and reliability of monitoring data. First, a single sensor is limited by the measurement range, accuracy or environmental interference, resulting in the missing or distorted monitoring data. Secondly, the multi-sensor data fusion process lacks in-depth analysis of pathological characteristics and fails to reveal the potential correlation between different monitoring items, especially under abnormal conditions, and cannot accurately reflect the pathological correlation. In addition, the quality assessment of detection data mostly remains at the basic statistical or rule-based methods, lacks refined correlation assessment based on large-scale training models, and cannot accurately judge the credibility of monitoring data, which easily leads to misjudgment of analysis results. Therefore, how to achieve reliability correlation assessment of health monitoring data has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides an intelligent rehabilitation monitoring system and method, which can realize reliability correlation evaluation of health monitoring data.

[0005] In a first aspect, the present application provides an intelligent rehabilitation monitoring method, comprising the following steps: Monitor the health status of each monitoring item of the target patient during rehabilitation training through different types of sensors, and then obtain the health monitoring data corresponding to each monitoring item; Extract the health feature value of the corresponding monitoring item from each health monitoring data, and then perform pathological abnormality correlation analysis on all health feature values ​​to obtain the abnormal correlation degree between each monitoring item and other monitoring items in the pathological structure; Based on the pre-trained data quality assessment model, the quality mapping assessment of the health monitoring data corresponding to each monitoring item is performed to obtain the quality score of the health monitoring data corresponding to each monitoring item, and then the data credibility of the health monitoring data corresponding to each monitoring item is determined by the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; When the data credibility of the health monitoring data corresponding to the monitoring item is less than the preset credibility threshold, the health monitoring data corresponding to the monitoring item is reconstructed according to the abnormal correlation in the pathological structure between the monitoring item and other monitoring items, and the health index of the target patient in rehabilitation training is generated based on the reconstructed health monitoring data.

[0006] Preferably, extracting the health characteristic value of the corresponding monitoring item from each health monitoring data specifically includes: Normalizing each health monitoring data to obtain each normalized health monitoring data; The health characteristic values ​​of the corresponding monitoring items are extracted from each health monitoring data after normalization.

[0007] Preferably, pathological abnormality correlation analysis is performed on all health feature values ​​to obtain abnormal correlation between each monitoring item and other monitoring items in pathological structure, specifically including: Select a monitoring item as the selected monitoring item; Perform explicit correlation analysis on the health characteristic values ​​between the selected monitoring items and other monitoring items to obtain the explicit correlation degree of the pathology; Conduct implicit correlation analysis on the pathological structure of the health characteristic values ​​between the selected monitoring items and other monitoring items to obtain the implicit correlation degree of the pathology; The abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure is determined by the explicit correlation and the implicit correlation, and the abnormal correlation between the remaining monitoring items and other monitoring items in the pathological structure is further determined.

[0008] Preferably, the quality mapping evaluation is performed on the health monitoring data corresponding to each monitoring item based on the pre-trained data quality evaluation model to obtain the quality score of the health monitoring data corresponding to each monitoring item, specifically including: For each monitoring item, extract the time domain characteristics and frequency domain characteristics of the health monitoring data corresponding to the monitoring item; Using the time domain features and the frequency domain features as input parameters of the data quality assessment model; The quality score of the health monitoring data corresponding to the monitoring item is obtained through the data quality assessment model mapping, and then the quality score of the health monitoring data corresponding to each monitoring item is obtained.

[0009] Preferably, the time domain features specifically include: mean, standard deviation and root mean square.

[0010] Preferably, determining the data credibility of the health monitoring data corresponding to each monitoring item based on the quality scores of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure specifically includes: Select a monitoring item as the selected monitoring item; Determine the credible compensation coefficient of the health monitoring data corresponding to the selected monitoring item according to the abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure; The quality score of the health monitoring data corresponding to the selected monitoring item is compensated by the credible compensation coefficient to obtain the data credibility of the health monitoring data corresponding to the selected monitoring item; Continue to determine the data credibility of the health monitoring data corresponding to the remaining monitoring items.

[0011] Preferably, the monitoring items specifically include range of joint motion, gait, heart rate, body temperature and muscle fatigue.

[0012] In a second aspect, the present application provides an intelligent rehabilitation monitoring system, comprising: The monitoring module is used to monitor the health status of each monitoring item of the target patient during rehabilitation training through different types of sensors, and then obtain the health monitoring data corresponding to each monitoring item; A processing module is used to extract the health characteristic value of the corresponding monitoring item from each health monitoring data, and then perform pathological abnormality correlation analysis on all health characteristic values ​​to obtain the abnormal correlation degree between each monitoring item and other monitoring items in the pathological structure; The processing module is further used to perform quality mapping evaluation on the health monitoring data corresponding to each monitoring item based on the pre-trained data quality evaluation model to obtain the quality score of the health monitoring data corresponding to each monitoring item, and then determine the data credibility of the health monitoring data corresponding to each monitoring item based on the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; The execution module is used to reconstruct the health monitoring data corresponding to the monitoring item according to the abnormal correlation in the pathological structure between the monitoring item and other monitoring items when the data credibility of the health monitoring data corresponding to the monitoring item is less than a preset credibility threshold, and generate the health index of the target patient in rehabilitation training based on the reconstructed health monitoring data.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned intelligent rehabilitation monitoring method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned intelligent rehabilitation monitoring method is implemented.

[0015] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects: In an embodiment of the present application, different types of sensors are used to monitor the health status of each monitoring item of the target patient during rehabilitation training, and then health monitoring data corresponding to each monitoring item is obtained; health feature values ​​corresponding to the monitoring items are extracted from each health monitoring data, and then pathological abnormality correlation analysis is performed on all health feature values ​​to obtain the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; quality mapping evaluation is performed on the health monitoring data corresponding to each monitoring item based on a pre-trained data quality assessment model to obtain a quality score of the health monitoring data corresponding to each monitoring item, and then the data credibility of the health monitoring data corresponding to each monitoring item is determined by the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; when the data credibility of the health monitoring data corresponding to the monitoring item is less than a preset credibility threshold, the health monitoring data corresponding to the monitoring item is reconstructed according to the abnormal correlation between the monitoring item and other monitoring items in the pathological structure, and the health index of the target patient in rehabilitation training is generated based on the reconstructed health monitoring data.

[0016] It can be seen that the present application determines the data credibility of the health monitoring data corresponding to the monitoring item through the quality score of the health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure, and then determines the reliability of the health monitoring data according to the data credibility; first, the health characteristic value of each monitoring item is extracted and the abnormal correlation analysis of the pathological structure is performed to obtain the abnormal correlation between each monitoring item and other monitoring items in the pathological structure. The abnormal correlation can effectively discover the abnormal coordination phenomenon between multiple monitoring items, which can help the system identify whether there is abnormality or distortion in the monitoring data, and can provide an important reference basis for the subsequent data credibility evaluation; secondly, the health monitoring data is evaluated from the aspects of data integrity, stability and accuracy through the pre-trained data quality evaluation model. Refined evaluation in multiple dimensions can improve the accuracy of the quality score of health monitoring data; then, the data credibility of the health monitoring data corresponding to the monitoring item is determined by the quality score of the health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure, and the data credibility of the health monitoring data is obtained by combining the quality score of the health monitoring data with the abnormal correlation, which provides a clear basis for identifying low-quality data and realizes a dynamic evaluation of the overall data reliability; finally, when the data credibility is lower than the preset credibility threshold, reconstructing the health monitoring data can compensate for the impact of insufficient quality of the health monitoring data on the rehabilitation analysis results, thereby generating a highly reliable health index; in summary, the present application scheme can realize the reliability association evaluation of health monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flow chart of an intelligent rehabilitation monitoring method according to some embodiments of the present application; Figure 2 is a schematic diagram of a process of processing rehabilitation monitoring data according to some embodiments of the present application; Figure 3 is a schematic diagram of a process for determining a quality score of health monitoring data according to some embodiments of the present application; Figure 4 is a schematic diagram of the structure of an intelligent rehabilitation monitoring system according to some embodiments of the present application; Figure 5 It is a structural schematic diagram of a computer device for implementing an intelligent rehabilitation monitoring method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0019] refer to Figure 1, which is an exemplary flow chart of an intelligent rehabilitation monitoring method according to some embodiments of the present application. The intelligent rehabilitation monitoring method 100 mainly includes the following steps: In step 101, the health status of each monitoring item of the target patient during rehabilitation training is monitored by different types of sensors, and then the health monitoring data corresponding to each monitoring item is obtained.

[0020] It should be noted that the reference Figure 2 As shown, this figure is a flow chart of rehabilitation monitoring data processing in some embodiments of the present application, which is used to monitor and analyze the health status of target patients during rehabilitation training. The flow chart of rehabilitation monitoring data processing can be divided into the following steps: Monitoring object: This is the starting point of the entire process, indicating the target patient to be monitored; Sensor 1 to Sensor n: These are different types of sensors used to monitor the target patient, and each sensor is responsible for monitoring different health indicators; Feature extraction: Extract key features from the data collected by each sensor; Association identification: Obtain the credibility of the sensor monitoring data by performing association analysis on all the extracted key features; Data reconstruction: Reconstruct the sensor monitoring data based on the obtained credibility.

[0021] It should also be noted that during the rehabilitation monitoring process, the patient's motor function and physiological indicators are usually monitored. In the present application, the monitoring items specifically include joint range of motion, gait, heart rate, body temperature and muscle fatigue; in specific implementation, an angle sensor can be used to monitor the range of motion of the patient's joints to obtain health monitoring data corresponding to the range of joint motion; an accelerometer can be used to monitor the patient's walking gait information to obtain health monitoring data corresponding to the gait; a heart rate monitor can be used to monitor the patient's heart rate to obtain health monitoring data corresponding to the heart rate; a temperature sensor can be used to monitor the patient's body temperature to obtain health monitoring data corresponding to the body temperature; an electromyography sensor can be used to monitor the patient's muscle strength to obtain health monitoring data corresponding to muscle fatigue; in addition, it should be noted that the health monitoring data in the present application refers to a data set, which contains all monitoring data collected by the corresponding sensor.

[0022] In step 102, health feature values ​​corresponding to monitoring items are extracted from various health monitoring data, and then pathological abnormality correlation analysis is performed on all health feature values ​​to obtain abnormal correlation degrees between each monitoring item and other monitoring items in pathological structure.

[0023] In some embodiments, extracting the health feature value of the corresponding monitoring item from each health monitoring data can be achieved by using the following steps: Normalizing each health monitoring data to obtain each normalized health monitoring data; The health characteristic values ​​of the corresponding monitoring items are extracted from each health monitoring data after normalization.

[0024] It should be noted that the monitoring items in the present application specifically include joint motion range, gait, heart rate, body temperature and muscle fatigue, among which the health characteristic value of the joint motion range is the maximum interval value of the joint motion range, the health characteristic value of gait is the gait cycle, the health characteristic value of heart rate is the standard deviation of heart rate fluctuations, the health characteristic value of body temperature is the body temperature change rate, and the health characteristic value of muscle fatigue is the degree of muscle fatigue.

[0025] In specific implementation, each health monitoring data is normalized, and each normalized health monitoring data can be obtained in the following manner, namely: each health monitoring data can be normalized by using the existing minimum-maximum normalization to obtain the normalized health monitoring data; the health characteristic value of the corresponding monitoring item can be extracted from each normalized health monitoring data in the following manner, namely: the interval length value between the maximum value and the minimum value in the health monitoring data corresponding to the joint motion range after normalization can be used as the health characteristic value of the joint motion range (i.e., the maximum interval value of the joint motion range); the periodic characteristic value of the health monitoring data corresponding to the normalization process of gait can be used as the health characteristic value of gait (i.e., the gait cycle); the standard deviation of the health monitoring data corresponding to the normalization process of heart rate can be used as the health characteristic value of heart rate (i.e., the standard deviation of heart rate fluctuation); the change rate of the health monitoring data corresponding to the normalization process of body temperature can be used as the health characteristic value of body temperature (i.e., the body temperature change rate); the attenuation rate of the health monitoring data corresponding to the normalization process of muscle fatigue can be used as the health characteristic value of muscle fatigue (i.e., the degree of muscle fatigue).

[0026] In some embodiments, pathological abnormality correlation analysis is performed on all health feature values ​​to obtain abnormal correlation between each monitoring item and other monitoring items in pathological structure, which can be achieved by the following steps: Select a monitoring item as the selected monitoring item; Perform explicit correlation analysis on the health characteristic values ​​between the selected monitoring items and other monitoring items to obtain the explicit correlation degree of the pathology; Conduct implicit correlation analysis on the pathological structure of the health characteristic values ​​between the selected monitoring items and other monitoring items to obtain the implicit correlation degree of the pathology; The abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure is determined by the explicit correlation and the implicit correlation, and the abnormal correlation between the remaining monitoring items and other monitoring items in the pathological structure is further determined.

[0027] It should be noted that the explicit correlation in this application is an indicator that measures the strength of the explicit relationship between monitoring items shown by statistical analysis, and can be used to quantify the intuitive correlation between monitoring items; the implicit correlation in this application is an indicator that measures the strength of the potential correlation between monitoring items, and the implicit correlation can reveal the deep-level correlation between monitoring items that cannot be directly observed; in addition, the abnormal correlation is an indicator that measures the degree to which the pathological structure between monitoring items deviates from the normal correlation pattern.

[0028] In the specific implementation, a monitoring item is selected as the selected monitoring item; an explicit correlation analysis is performed on the health feature values ​​between the selected monitoring item and other monitoring items, and the explicit correlation degree of the pathology can be obtained. This can be achieved in the following way, namely: the Pearson correlation coefficient of the health feature values ​​between the selected monitoring item and each other monitoring item can be calculated first, and then the average value of all Pearson correlation coefficients is used as the explicit correlation degree of the pathology between the selected monitoring item and other monitoring items; an implicit correlation analysis on the pathological structure is performed on the health feature values ​​between the selected monitoring item and other monitoring items, and the implicit correlation degree of the pathology can be obtained. This can be achieved in the following way, namely: an implicit correlation model is initialized, and the implicit correlation model can be obtained by training based on the autoencoder, and the health feature values ​​between the selected monitoring item and other monitoring items are mapped to the implicit space through the implicit correlation model, wherein the implicit space can be used to extract the implicit correlation between features. Relationship, the implicit space based on the autoencoder can be used in the present application, that is, the encoder part of the autoencoder model is used to map the original health feature value to a low-dimensional implicit representation space, and the implicit variables between each health feature value and other health feature values ​​are extracted through the bottleneck layer, and then the Euclidean distance between each implicit variable is calculated, and the calculated Euclidean distance is used as the implicit correlation degree of the pathology between the selected monitoring item and other monitoring items; determining the abnormal correlation degree between the selected monitoring item and other monitoring items in the pathological structure through the explicit correlation degree and the implicit correlation degree can be achieved in the following way, that is: the inverse of the sum of the explicit correlation degree and the implicit correlation degree can be used as the abnormal correlation degree between the selected monitoring item and other monitoring items in the pathological structure, and the above steps can be repeated to continue to obtain the abnormal correlation degree between the remaining monitoring items and other monitoring items in the pathological structure.

[0029] It should be noted that the specific training process of the implicit association model in the present application is as follows: extract the health feature values ​​of all monitoring items from historical data, and construct a training data set to ensure that the data is complete and covers the characteristics of each monitoring item. The existing autoencoder can be selected as the architecture of the implicit association model, and the initialization of the implicit association model is completed. The training data set is then input into the implicit association model, and forward propagation is performed through the implicit association model. During the encoding process of the forward propagation, each input health feature value is mapped through the encoder network to obtain a low-dimensional implicit variable, and then the Euclidean distance between all implicit variables is calculated, and the Euclidean distance is marked as the output of the implicit association model. During the decoding process of the forward propagation, the implicit variable passes through the decoder network to obtain the reconstructed health feature value, and the loss is further calculated by the reconstructed health feature value and the original health feature value, and the loss is back-propagated and optimized until the loss value reaches the target threshold, then the optimization training of the model is stopped, and the trained implicit association model is verified.

[0030] In step 103, a quality mapping assessment is performed on the health monitoring data corresponding to each monitoring item based on a pre-trained data quality assessment model to obtain a quality score of the health monitoring data corresponding to each monitoring item, and then the data credibility of the health monitoring data corresponding to each monitoring item is determined by the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure.

[0031] It should be noted that the pre-trained data quality assessment model in this application is a model for quantifying the reliability and integrity of health monitoring data; it should also be noted that the training steps of the data quality assessment model in this application are as follows: the first step, data preparation: collect multidimensional data sets, including health monitoring data and their corresponding quality labels (such as data integrity indicators derived from historical analysis); the second step, data annotation: grade the data quality (such as high, medium, and low quality) through expert scoring or rule definition, and construct annotated data for supervised training; the third step, model selection: deep learning models (such as multi-layer perceptrons or Transformer-based models) can be selected to adapt to multidimensional health The complex characteristics of health monitoring data; the fourth step, model training: first normalize the health monitoring data to ensure that data of different dimensions have the same scale, then pass the input data through each layer of the model, gradually extract high-dimensional features, and predict the data quality score, and further use the mean square error to calculate the error (i.e., loss) between the predicted quality score and the true score, and finally optimize the model parameters through gradient descent to gradually reduce the loss function value; the fifth step, model verification: evaluate the model performance on the verification set, and use indicators such as mean square error and accuracy to judge the predictive ability of the model; the sixth step, model optimization: adjust hyperparameters such as learning rate, number of network layers, and number of nodes to improve the generalization ability of the model.

[0032] In some embodiments, reference Figure 3 As shown, this figure is a schematic diagram of the process of determining the quality score of health monitoring data in some embodiments of the present application. In this embodiment, the quality mapping evaluation of the health monitoring data corresponding to each monitoring item is performed based on the pre-trained data quality evaluation model, and the quality score of the health monitoring data corresponding to each monitoring item can be obtained by the following steps: In step 1031, for each monitoring item, time domain features and frequency domain features of health monitoring data corresponding to the monitoring item are extracted; In step 1032, the time domain features and the frequency domain features are used as input parameters of the data quality assessment model; In step 1033, the quality score of the health monitoring data corresponding to the monitoring item is obtained through the data quality assessment model mapping, and then the quality score of the health monitoring data corresponding to each monitoring item is obtained.

[0033] It should be noted that the time domain features in the present application specifically include: mean, standard deviation and root mean square; the frequency domain features in the present application specifically include: spectral energy and frequency band power distribution. In other embodiments, the time domain features and frequency domain features may also include others, which are not specifically limited here.

[0034] In specific implementation, for each monitoring item, the time domain characteristics and frequency domain characteristics of the health monitoring data corresponding to the monitoring item can be extracted by the following method, namely: obtain the health monitoring data corresponding to the monitoring item, and extract the mean, standard deviation and root mean square of the health monitoring data, and use the extracted mean, standard deviation and root mean square as the time domain characteristics of the health monitoring data corresponding to the monitoring item, and then perform Fourier transform on the health monitoring data to obtain spectrum information, and then extract the spectrum energy and frequency band power distribution in the frequency domain from the spectrum information, and use the extracted spectrum energy and frequency band power distribution as the frequency domain characteristics of the health monitoring data corresponding to the monitoring item; the time domain characteristics and the frequency domain characteristics can be used as input parameters of the data quality assessment model in the following way. The implementation is as follows: inputting time domain features and frequency domain features into a data quality assessment model, performing feature mapping through the training weights of the data quality assessment model, and converting the input features into quality scores. The data quality assessment model can be trained on a historical data set through a supervised learning method to learn the relationship between different features and data quality; the quality scores of the health monitoring data corresponding to the monitoring items are obtained through mapping the data quality assessment model, and then the quality scores of the health monitoring data corresponding to each monitoring item are obtained. The following method can be used to implement the method, namely: the output value of the data quality assessment model can be used as the quality score of the health monitoring data corresponding to the monitoring item. Through the above method, the quality scores of the health monitoring data corresponding to each monitoring item can be obtained.

[0035] It should be noted that the quality score in this application is a quantitative evaluation indicator to measure the completeness and accuracy of health monitoring data.

[0036] In some embodiments, determining the data credibility of the health monitoring data corresponding to each monitoring item based on the quality scores of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure can be achieved by the following steps: Select a monitoring item as the selected monitoring item; Determine the credible compensation coefficient of the health monitoring data corresponding to the selected monitoring item according to the abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure; The quality score of the health monitoring data corresponding to the selected monitoring item is compensated by the credible compensation coefficient to obtain the data credibility of the health monitoring data corresponding to the selected monitoring item; Continue to determine the data credibility of the health monitoring data corresponding to the remaining monitoring items.

[0037] It should be noted that the trust compensation coefficient in this application is a compensation coefficient for dynamically adjusting the credibility of the health monitoring data quality; the data credibility in this application is a comprehensive indicator to measure the reliability of health monitoring data, which is used to evaluate the degree of trust in health monitoring data in health index analysis.

[0038] In specific implementation, determining the credible compensation coefficient of the health monitoring data corresponding to the selected monitoring item according to the abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure can be achieved in the following manner, namely: the natural exponential function value of the inverse of the abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure can be used as the credible compensation coefficient of the health monitoring data corresponding to the selected monitoring item, the larger the abnormal correlation, the weaker the correlation between the health monitoring data of each monitoring item, and the lower the reliability of the health monitoring data corresponding to the monitoring item, and a larger adjustment compensation is required; compensating the quality score of the health monitoring data corresponding to the selected monitoring item by the credible compensation coefficient, and obtaining the data credibility of the health monitoring data corresponding to the selected monitoring item can be achieved in the following manner, namely: the product of the credible compensation coefficient and the quality score of the health monitoring data corresponding to the selected monitoring item can be used as the data credibility of the health monitoring data corresponding to the selected monitoring item; finally, repeatedly selecting the remaining monitoring items as the selected monitoring items can continue to determine the data credibility of the health monitoring data corresponding to the remaining monitoring items.

[0039] In step 104, when the data credibility of the health monitoring data corresponding to the monitoring item is less than a preset credibility threshold, the health monitoring data corresponding to the monitoring item is reconstructed according to the abnormal correlation between the monitoring item and other monitoring items in the pathological structure, and the health index of the target patient in rehabilitation training is generated based on the reconstructed health monitoring data.

[0040] It should be noted that when the data credibility of the health monitoring data corresponding to the monitoring item is less than the preset credibility threshold, it means that the health monitoring data has significant problems in reliability and integrity, and may be affected by noise interference, data loss, incorrect collection or abnormal conditions, which results in the health monitoring data being unable to be directly used for subsequent health index analysis and evaluation. It is necessary to improve its credibility through data reconstruction to ensure the accuracy of the overall monitoring system; in addition, the preset credibility threshold in the embodiment of the present application refers to a preset credibility threshold, and its credibility threshold can be dynamically set according to historical data analysis results, domain expert experience or system error tolerance, generally taking the lower limit of the statistical distribution of data credibility or a specific percentile as a benchmark.

[0041] In some embodiments, reconstructing the health monitoring data corresponding to the monitoring item according to the abnormal correlation between the monitoring item and other monitoring items in the pathological structure can be achieved by using the following steps: Obtain the data credibility of the health monitoring data corresponding to the monitoring items; The constraint parameters in the pre-trained data reconstruction model are updated by the data credibility of the health monitoring data corresponding to the monitoring item and the abnormal correlation between the pathological structure of the monitoring item and other monitoring items; The health monitoring data corresponding to the monitoring items are reconstructed according to the data reconstruction model after parameter update to obtain the reconstructed health monitoring data.

[0042] In specific implementation, updating the constraint parameters in the pre-trained data reconstruction model through the data credibility of the health monitoring data corresponding to the monitoring item and the abnormal correlation degree in the pathological structure between the monitoring item and other monitoring items can be achieved in the following manner, namely: combining the abnormal correlation degree in the pathological structure between the monitoring item and other monitoring items, the weighted value between the data credibility and the abnormal correlation degree of the health monitoring data corresponding to the monitoring item is used to update the constraint parameters of the data reconstruction model, and the parameter updating process can use the gradient descent algorithm in supervised learning to optimize the parameters in the model to ensure that it better adapts to the actual situation. In the training stage, the data credibility and the abnormal correlation degree are mapped to the prior constraints and conditional inputs of the generated model to ensure that the model can capture the deep-level correlation structure between the monitoring items; the health monitoring data corresponding to the monitoring item is reconstructed according to the data reconstruction model after the parameter update, and the reconstructed health monitoring data can be achieved in the following manner, namely: applying the updated data reconstruction model to the current monitoring item, and the health monitoring data corresponding to the current monitoring item (the original health monitoring data) is updated in the data reconstruction model, and the health monitoring data is reconstructed through the data reconstruction model to obtain the reconstructed health monitoring data.

[0043] It should be noted that in this application, generating the health index of the target patient in rehabilitation training based on the reconstructed health monitoring data means calculating the health index of the target patient based on the reconstructed health monitoring data, and the health index can be used to evaluate the rehabilitation training effect of the target patient; in specific implementation, all the reconstructed health monitoring data are obtained, and then the key health feature values ​​of the monitoring items in all the reconstructed health monitoring data are extracted, such as heart rate and electromyographic activity intensity, and weights are assigned to different feature values. These weights can be determined through expert experience or data-driven regression analysis, and then a weighted comprehensive scoring method or a machine learning model (such as a support vector machine, a random forest) is used to fuse and calculate the corresponding feature values ​​of all monitoring items to generate a comprehensive health index. Finally, the health index is normalized (for example, mapped to a range of 0-10) to intuitively represent the overall health status of the target patient in rehabilitation training and provide support for subsequent personalized rehabilitation guidance.

[0044] It should also be noted that when the data credibility of the health monitoring data corresponding to the monitoring item is greater than or equal to the preset credibility threshold, the health index of the target patient in rehabilitation training is generated directly based on the health monitoring data corresponding to the monitoring item.

[0045] On the other hand, in some embodiments, the present application provides an intelligent rehabilitation monitoring system, referring to Figure 4 , which is a schematic diagram of the structure of an intelligent rehabilitation monitoring system according to some embodiments of the present application, the intelligent rehabilitation monitoring system 400 includes: a monitoring module 401, a processing module 402 and an execution module 403, which are described as follows: Monitoring module 401, in this application, the monitoring module 401 is mainly used to monitor the health status of each monitoring item of the target patient during rehabilitation training through different types of sensors, and then obtain health monitoring data corresponding to each monitoring item; Processing module 402, in the present application, is used to extract health feature values ​​of corresponding monitoring items from various health monitoring data, and then perform pathological abnormality correlation analysis on all health feature values ​​to obtain abnormal correlation between each monitoring item and other monitoring items in pathological structure; The processing module 402 in the present application is also used to perform quality mapping evaluation on the health monitoring data corresponding to each monitoring item based on the pre-trained data quality evaluation model to obtain the quality score of the health monitoring data corresponding to each monitoring item, and then determine the data credibility of the health monitoring data corresponding to each monitoring item according to the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; Execution module 403. In the present application, execution module 403 is mainly used to reconstruct the health monitoring data corresponding to the monitoring item according to the abnormal correlation in the pathological structure between the monitoring item and other monitoring items when the data credibility of the health monitoring data corresponding to the monitoring item is less than a preset credibility threshold, and generate a health index of the target patient in rehabilitation training based on the reconstructed health monitoring data.

[0046] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned intelligent rehabilitation monitoring method.

[0047] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for implementing an intelligent rehabilitation monitoring method according to some embodiments of the present application. The intelligent rehabilitation monitoring method in the above embodiment can be Figure 5 The computer device 500 shown in the figure is implemented, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503 and at least one communication interface 504.

[0048] The processor 501 may be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0049] The communication bus 502 may be used to transmit information between the above-mentioned components.

[0050] The memory 503 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 503 may exist independently and be connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0051] The memory 503 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 501. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The intelligent rehabilitation monitoring method in the above embodiment can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0052] The communication interface 504 uses any transceiver or other device for communicating 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 an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0054] The above-mentioned computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0055] In addition, the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned intelligent rehabilitation monitoring method is implemented.

[0056] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0057] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent rehabilitation monitoring method, characterized in that: The steps include: Monitor the health status of each monitoring item of the target patient during rehabilitation training through different types of sensors, and then obtain the health monitoring data corresponding to each monitoring item; Extract the health feature value of the corresponding monitoring item from each health monitoring data, and then perform pathological abnormality correlation analysis on all health feature values ​​to obtain the abnormal correlation degree between each monitoring item and other monitoring items in the pathological structure; Based on the pre-trained data quality assessment model, the quality mapping assessment of the health monitoring data corresponding to each monitoring item is performed to obtain the quality score of the health monitoring data corresponding to each monitoring item, and then the data credibility of the health monitoring data corresponding to each monitoring item is determined by the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; When the data credibility of the health monitoring data corresponding to the monitoring item is less than the preset credibility threshold, the health monitoring data corresponding to the monitoring item is reconstructed according to the abnormal correlation in the pathological structure between the monitoring item and other monitoring items, and the health index of the target patient in rehabilitation training is generated based on the reconstructed health monitoring data.

2. The intelligent rehabilitation monitoring method according to claim 1, characterized in that: Extracting the health characteristic values ​​of the corresponding monitoring items from each health monitoring data specifically includes: Normalizing each health monitoring data to obtain each normalized health monitoring data; The health characteristic values ​​of the corresponding monitoring items are extracted from each health monitoring data after normalization.

3. The intelligent rehabilitation monitoring method according to claim 1, characterized in that: Perform pathological abnormality correlation analysis on all health feature values ​​to obtain the abnormal correlation between each monitoring item and other monitoring items in the pathological structure, including: Select a monitoring item as the selected monitoring item; Perform explicit correlation analysis on the health characteristic values ​​between the selected monitoring items and other monitoring items to obtain the explicit correlation degree of the pathology; Conduct implicit correlation analysis on the pathological structure of the health characteristic values ​​between the selected monitoring items and other monitoring items to obtain the implicit correlation degree of the pathology; The abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure is determined by the explicit correlation and the implicit correlation, and the abnormal correlation between the remaining monitoring items and other monitoring items in the pathological structure is further determined.

4. The intelligent rehabilitation monitoring method according to claim 1, characterized in that: Based on the pre-trained data quality assessment model, the quality mapping assessment of the health monitoring data corresponding to each monitoring item is performed to obtain the quality scores of the health monitoring data corresponding to each monitoring item, including: For each monitoring item, extract the time domain characteristics and frequency domain characteristics of the health monitoring data corresponding to the monitoring item; Using the time domain features and the frequency domain features as input parameters of the data quality assessment model; The quality score of the health monitoring data corresponding to the monitoring item is obtained through the data quality assessment model mapping, and then the quality score of the health monitoring data corresponding to each monitoring item is obtained.

5. The intelligent rehabilitation monitoring method according to claim 4, characterized in that: The time domain features specifically include: mean, standard deviation and root mean square.

6. The intelligent rehabilitation monitoring method according to claim 1, characterized in that: Determining the data credibility of the health monitoring data corresponding to each monitoring item based on the quality scores of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure specifically includes: Select a monitoring item as the selected monitoring item; Determine the credible compensation coefficient of the health monitoring data corresponding to the selected monitoring item according to the abnormal correlation between the selected monitoring item and other monitoring items in the pathological structure; The quality score of the health monitoring data corresponding to the selected monitoring item is compensated by the credible compensation coefficient to obtain the data credibility of the health monitoring data corresponding to the selected monitoring item; Continue to determine the data credibility of the health monitoring data corresponding to the remaining monitoring items.

7. The intelligent rehabilitation monitoring method according to claim 1, characterized in that: The monitoring items specifically include range of joint motion, gait, heart rate, body temperature and muscle fatigue.

8. An intelligent rehabilitation monitoring system, characterized in that: include: The monitoring module is used to monitor the health status of each monitoring item of the target patient during rehabilitation training through different types of sensors, and then obtain the health monitoring data corresponding to each monitoring item; A processing module is used to extract the health characteristic value of the corresponding monitoring item from each health monitoring data, and then perform pathological abnormality correlation analysis on all health characteristic values ​​to obtain the abnormal correlation degree between each monitoring item and other monitoring items in the pathological structure; The processing module is further used to perform quality mapping evaluation on the health monitoring data corresponding to each monitoring item based on the pre-trained data quality evaluation model to obtain the quality score of the health monitoring data corresponding to each monitoring item, and then determine the data credibility of the health monitoring data corresponding to each monitoring item based on the quality score of all health monitoring data and the abnormal correlation between each monitoring item and other monitoring items in the pathological structure; The execution module is used to reconstruct the health monitoring data corresponding to the monitoring item according to the abnormal correlation in the pathological structure between the monitoring item and other monitoring items when the data credibility of the health monitoring data corresponding to the monitoring item is less than a preset credibility threshold, and generate the health index of the target patient in rehabilitation training based on the reconstructed health monitoring data.

9. A computer device, comprising a memory and a processor, wherein the memory stores a code, characterized in that: The processor is configured to obtain the code and execute the intelligent rehabilitation monitoring method according to 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 a processor, the intelligent rehabilitation monitoring method according to any one of claims 1 to 7 is implemented.

Images