Abdominal respiration training method and system based on intelligent guiding algorithm

Abstract

The invention relates to the technical field of intelligent medical rehabilitation, in particular to an abdominal respiration training method and system based on an intelligent guiding algorithm. The continuous real-time breathing data sequence contains at least one of a breathing frequency periodic change characteristic, a breathing amplitude dynamic change characteristic and a breathing phase accurate definition characteristic which change along with time; comprehensively evaluating the continuous real-time breathing data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive breathing health index for evaluating the breathing state of the user; according to the comprehensive respiratory health index, based on inference rules in a knowledge base, a customized respiratory training plan is generated through an intelligent decision engine combining a deep learning algorithm and a biological feedback mechanism. By implementing the method, potential information in the respiration data can be deeply mined, so that a more scientific and personalized respiration training plan is generated.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical rehabilitation technology, and in particular to an abdominal breathing training method and system based on an intelligent guidance algorithm. Background Art

[0002] Abdominal breathing is an important breathing method. It allows the diaphragm to move up and down, thereby increasing the range of motion of the diaphragm and directly affecting the ventilation of the lungs. Through abdominal breathing, the chest can be expanded to the maximum extent, the alveoli in the lower part of the lungs can be expanded and contracted, allowing more oxygen to enter the lungs, thereby improving cardiopulmonary function. In addition, abdominal breathing can also reduce lung infections, improve the function of abdominal organs, and lower blood pressure by reducing abdominal pressure, which is very beneficial for patients with hypertension. Therefore, abdominal breathing is widely used in health care and treatment of diseases in various parts of the human body.

[0003] Although the benefits of abdominal breathing are widely recognized, in practical applications, users often find it difficult to correctly master the skills of abdominal breathing and are unable to perform abdominal breathing training continuously and effectively. At present, traditional abdominal breathing training methods mostly rely on the guidance of coaches or doctors, lacking personalized feedback and real-time adjustments. At the same time, it is difficult for users to persist in training without guidance, resulting in poor training results. In order to solve the above problems, relevant researchers have begun to explore abdominal breathing training methods based on intelligent technology, aiming to monitor the user's breathing status in real time through intelligent algorithms and sensor technology, and provide personalized feedback and guidance based on this, so as to help users perform abdominal breathing training correctly and effectively. Although these intelligent methods can improve the pertinence and effectiveness of training, enhance user participation and training experience, they still have some limitations, such as difficulty in accurately and comprehensively obtaining and analyzing users' continuous real-time breathing data, and generating scientific and personalized breathing training plans based on these data. Therefore, it is necessary to improve the intelligent algorithm to deeply mine the potential information in the breathing data, so as to generate a more scientific and personalized breathing training plan. Summary of the invention

[0004] In order to deeply mine the potential information in the breathing data and generate a more scientific and personalized breathing training plan, the purpose of the present invention is to provide an abdominal breathing training method and system based on an intelligent guidance algorithm. The technical solutions adopted are as follows:

[0005] In a first aspect, the present application discloses an abdominal breathing training method based on an intelligent guidance algorithm, the method comprising:

[0006] S1. Acquire a continuous real-time respiratory data sequence of a user, wherein the continuous real-time respiratory data sequence includes at least one of a periodic change characteristic of respiratory frequency that changes over time, a dynamic change characteristic of respiratory amplitude, and an accurate definition characteristic of respiratory phase;

[0007] S2, comprehensively evaluating the continuous real-time respiratory data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory status;

[0008] S3. According to the comprehensive respiratory health index, based on the inference rules in the knowledge base, a customized respiratory training plan is generated by an intelligent decision-making engine that combines a deep learning algorithm with a biofeedback mechanism.

[0009] Further, in step S2, the continuous real-time respiratory data sequence is comprehensively evaluated according to the statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory state, including:

[0010] S21, performing a preprocessing operation on the continuous real-time respiratory data sequence to obtain a preprocessed respiratory data sequence, wherein the preprocessing operation includes at least one of denoising, smoothing, and outlier detection and correction;

[0011] S22, extracting multi-dimensional respiratory features from the pre-processed respiratory data sequence based on respiratory waveform analysis;

[0012] S23, obtaining historical breathing data of the user, and setting a dynamic threshold based on statistics based on the historical breathing data;

[0013] S24, comparing and analyzing the multi-dimensional respiratory characteristics with the set dynamic threshold, and calculating the corresponding comprehensive respiratory health index by weighted average method based on the analysis results.

[0014] Furthermore, in step S23, the dynamic threshold setting based on statistics is performed based on the historical breathing data of the user and the multi-dimensional breathing characteristics, including:

[0015] S231, determining key statistical indicators related to the multi-dimensional respiratory characteristics from historical respiratory data;

[0016] S232, performing feature importance analysis on the multi-dimensional respiratory features based on the key statistical indicators to determine the contribution of each respiratory feature to the description of the respiratory state;

[0017] S233. Based on the feature importance analysis results, individual differences of users, and real-time breathing data, an adaptive statistical filtering algorithm is used to dynamically optimize and adjust the threshold of the multi-dimensional breathing feature in real time.

[0018] Further, in step S232, the feature importance analysis of the multi-dimensional respiratory features based on the key statistical indicators includes:

[0019] S2321. For each dimension of the respiratory feature, calculate the feature contribution based on the corresponding key statistical indicators to obtain a feature contribution score;

[0020] S2322. Based on the feature contribution scores, the importance of multi-dimensional respiratory features is analyzed through quantitative evaluation.

[0021] Furthermore, in step S233, the threshold of the multi-dimensional respiratory feature is dynamically optimized and adjusted in real time using an adaptive statistical filtering algorithm based on the feature importance analysis results, individual differences of users, and real-time respiratory data, including:

[0022] S2331, obtaining basic information of the user, and setting an initial threshold through a heuristic algorithm based on the basic information and feature importance analysis results;

[0023] S2332, applying an adaptive statistical filtering algorithm, combining the user's real-time breathing data and individual differences to accurately estimate breathing characteristics and detect anomalies;

[0024] S2333. Adjust the initial threshold based on the accurate estimation of real-time respiratory characteristics and the abnormal detection results to achieve dynamic optimization of the threshold.

[0025] Further, in step S24, the multi-dimensional respiratory characteristics are compared and analyzed with the set dynamic threshold, and according to the analysis results, the corresponding comprehensive respiratory health index is calculated by weighted average method, including:

[0026] S241, comparing the real-time data of the multi-dimensional respiratory features with the set dynamic thresholds respectively, to obtain the deviation value of each respiratory feature relative to the dynamic threshold;

[0027] S242, assigning a corresponding respiratory health score to each respiratory feature through a linear conversion function based on the deviation value of each respiratory feature relative to the dynamic threshold;

[0028] S243, based on the feature importance analysis result, assigning a corresponding importance weight to each respiratory feature through a weight mapping table;

[0029] S244, performing weighted average calculation based on the respiratory health score of each respiratory characteristic and the corresponding importance weight to obtain a corresponding comprehensive respiratory health index.

[0030] Furthermore, in step S242, the formula of the linear conversion function includes:

[0031]

[0032] Among them, score i represents the respiratory health score of the i-th respiratory feature, p i represents the deviation of the i-th respiratory feature relative to the dynamic threshold, thresh max Indicates the maximum value of the dynamic threshold preset, thresh min It represents the minimum value preset by the dynamic threshold, and e represents the exponential parameter of the linear conversion, which is used to adjust the sensitivity of the score to the deviation value.

[0033] Further, in step S3, according to the comprehensive respiratory health index, based on the inference rules in the knowledge base, a customized respiratory training plan is generated by an intelligent decision engine combining a deep learning algorithm and a biofeedback mechanism, including:

[0034] S31, comparing the comprehensive respiratory health index with the threshold ranges corresponding to the health standards and pathological characteristics in the knowledge base to obtain corresponding comparison results;

[0035] S32, based on the comparison result, logical judgment is performed through the inference rules in the knowledge base to preliminarily determine the breathing training direction and goal required by the user;

[0036] S33, based on the user's personal information and the initially determined breathing training direction and goal, applying a deep learning algorithm to generate a personalized training plan to obtain customized breathing training suggestions;

[0037] S34. Integrate the biofeedback mechanism to optimize the breathing training plan by real-time monitoring of the user's physiological reactions and adjusting the training plan, so that the training plan is more in line with the user's actual needs and physiological state.

[0038] In a second aspect, the present application discloses an abdominal breathing training system based on an intelligent guidance algorithm, the system comprising a real-time breathing data acquisition module, a comprehensive breathing state evaluation module, and a breathing training plan generation module, wherein:

[0039] The real-time respiratory data acquisition module is used to acquire a continuous real-time respiratory data sequence of the user, wherein the continuous real-time respiratory data sequence includes at least one of a periodic change characteristic of respiratory frequency over time, a dynamic change characteristic of respiratory amplitude, and an accurate definition characteristic of respiratory phase;

[0040] The respiratory state comprehensive evaluation module is used to comprehensively evaluate the continuous real-time respiratory data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory state;

[0041] The breathing training plan generation module is used to generate a customized breathing training plan according to the comprehensive respiratory health index, based on the inference rules in the knowledge base, through an intelligent decision-making engine that combines a deep learning algorithm with a biofeedback mechanism.

[0042] In a third aspect, the present application discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the abdominal breathing training method based on an intelligent guidance algorithm.

[0043] The present invention has the following beneficial effects:

[0044] 1) Based on the statistical analysis algorithm of dynamic threshold setting and multi-dimensional feature fusion, a personalized comprehensive respiratory health index is generated according to the individual differences of users and real-time respiratory data while comprehensively considering multiple respiratory features, which helps to understand the user's respiratory health status more accurately;

[0045] 2) Based on the inference rules and deep learning algorithms in the knowledge base, the intelligent decision-making engine enables the generation of customized breathing training plans based on the comprehensive respiratory health index. These plans can be optimized according to the user's current breathing condition, thereby ensuring the scientificity and effectiveness of the training plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 A method flow chart of an abdominal breathing training method based on an intelligent guidance algorithm provided by an embodiment of the present invention;

[0048] Figure 2A system structure diagram of an abdominal breathing training system based on an intelligent guidance algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the abdominal breathing training method and system based on the intelligent guidance algorithm proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0050] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0051] The specific scheme of an abdominal breathing training method and system based on an intelligent guidance algorithm provided by the present invention is described in detail below in conjunction with the accompanying drawings.

[0052] See also Figure 1 , which shows a method flow chart of an abdominal breathing training method based on an intelligent guidance algorithm provided by an embodiment of the present invention, the method comprising:

[0053] Step S1, obtaining a continuous real-time respiratory data sequence of the user, wherein the continuous real-time respiratory data sequence includes at least one of a periodic change characteristic of respiratory frequency over time, a dynamic change characteristic of respiratory amplitude, and a precise definition characteristic of respiratory phase.

[0054] Step S2, comprehensively evaluating the continuous real-time respiratory data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory status.

[0055] Step S3, according to the comprehensive respiratory health index, based on the inference rules in the knowledge base, a customized respiratory training plan is generated by an intelligent decision engine that combines a deep learning algorithm with a biofeedback mechanism.

[0056] From the above, it can be seen that the abdominal breathing training method based on the intelligent guidance algorithm disclosed in the present application, on the one hand, can generate a personalized comprehensive respiratory health index based on the user's individual differences and real-time respiratory data while comprehensively considering multiple respiratory characteristics based on the statistical analysis algorithm of dynamic threshold setting and multi-dimensional feature fusion, which helps to understand the user's respiratory health status more accurately; on the other hand, it can also generate customized breathing training plans based on the comprehensive respiratory health index based on the inference rules and deep learning algorithms in the knowledge base, using the intelligent decision-making engine. These plans can be optimized according to the user's current respiratory condition, thereby ensuring the scientificity and effectiveness of the training plan.

[0057] In one embodiment, in step S2, the continuous real-time respiratory data sequence is comprehensively evaluated according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory state, including:

[0058] Step S21, performing a preprocessing operation on the continuous real-time respiratory data sequence to obtain a preprocessed respiratory data sequence, wherein the preprocessing operation includes at least one of denoising, smoothing, and outlier detection and correction.

[0059] Specifically, the denoising process here includes setting a threshold value, treating data above the threshold as noise and removing it; the smoothing process includes calculating the average value of the data within a certain window to smooth the data and reduce fluctuations; outlier detection and correction includes using statistical methods to identify outliers that are significantly deviated from the normal data range, and then using mean replacement, median replacement and other methods (determined by the nature of the outliers and the characteristics of the data distribution) to correct the detected outliers.

[0060] Step S22: extracting multi-dimensional respiratory features from the pre-processed respiratory data sequence based on respiratory waveform analysis.

[0061] Specifically, these features include respiratory frequency, respiratory amplitude, respiratory cycle variability, and respiratory phase relationship. Among them, the respiratory frequency is based on the periodic change characteristics of the respiratory frequency, and is calculated by calculating the number of peaks of the respiratory waveform per unit time; the respiratory amplitude is based on the dynamic change characteristics of the respiratory amplitude, and is calculated by measuring the difference between the maximum and minimum values ​​of each cycle of the respiratory waveform; the respiratory cycle variability is based on the periodic change characteristics of the respiratory frequency, and is calculated by calculating the standard deviation of the length of continuous respiratory cycles; the respiratory phase relationship is based on the precise definition of the respiratory phase, and is calculated by analyzing the time ratio of the inhalation segment and the exhalation segment.

[0062] Step S23, obtaining the user's historical breathing data, and performing statistical dynamic threshold setting based on the historical breathing data.

[0063] Specifically, this application analyzes historical respiratory data to determine key statistical indicators, and then evaluates the importance of multi-dimensional respiratory features based on the key statistical indicators. Then, based on the results of feature importance analysis, combined with individual differences of users and acquired real-time respiratory data, an adaptive statistical filtering algorithm is used to dynamically adjust the threshold of respiratory features to achieve accurate and personalized respiratory status monitoring.

[0064] Step S24, comparing and analyzing the multi-dimensional respiratory characteristics with the set dynamic threshold, and calculating the corresponding comprehensive respiratory health index by weighted average method based on the analysis results.

[0065] Specifically, the present application calculates the deviation value by comparing the real-time data of multi-dimensional respiratory features with the dynamic threshold. Afterwards, a linear conversion function is used to assign a corresponding health score to each respiratory feature according to the size and direction of the deviation value. Subsequently, based on the results of the feature importance analysis, a specific weight value is preset for each respiratory feature through a weight mapping table, wherein the weight mapping table reflects the relative importance of different respiratory features in describing the respiratory health status. Finally, the present application performs a weighted average calculation based on the health score and the assigned weights to calculate the corresponding comprehensive respiratory health index.

[0066] In the above embodiment, by obtaining the user's historical breathing data and setting a dynamic threshold based on statistics, it is possible to tailor a suitable respiratory health standard for the user, thereby improving the accuracy and pertinence of the health assessment. In addition, the multi-dimensional respiratory characteristics are compared and analyzed with the set dynamic threshold, and the corresponding comprehensive respiratory health index is calculated by the weighted average method based on the analysis results. Among them, this index can more comprehensively and intuitively reflect the user's respiratory health status, and provide an important reference for the formulation of subsequent respiratory training plans.

[0067] In one embodiment, in step S23, the dynamic threshold setting based on statistics based on the historical breathing data of the user and the multi-dimensional breathing characteristics includes:

[0068] Step S231, determining key statistical indicators related to the multi-dimensional respiratory features from historical respiratory data.

[0069] Specifically, these key statistical indicators include: the mean and standard deviation of respiratory frequency to measure the stability of respiratory rate; the mean, maximum and minimum values ​​of respiratory amplitude to measure the normal range of respiratory amplitude; the coefficient of variation of the time interval between adjacent respiratory cycles to measure the regularity of respiratory rhythm; the ratio of inhalation to exhalation time, and the duration and variability of the inhalation phase and the exhalation phase to measure whether the relationship between respiratory phases is normal.

[0070] Step S232: performing feature importance analysis on the multi-dimensional respiratory features based on the key statistical indicators to determine the contribution of each respiratory feature to the description of the respiratory state.

[0071] Specifically, this application will use key statistical indicators to quantitatively calculate the feature contribution and derive the feature contribution score. Then, based on these scores, the importance of multi-dimensional respiratory features will be deeply analyzed through quantitative evaluation, thus providing a key basis for the subsequent dynamic threshold setting.

[0072] Step S233, based on the feature importance analysis results, individual differences of users, and real-time breathing data, an adaptive statistical filtering algorithm is used to dynamically optimize and adjust in real time the threshold of the multi-dimensional breathing feature.

[0073] Specifically, this application will first use a heuristic algorithm to set a personalized initial breathing threshold based on the user's basic information and feature importance. Then, an adaptive statistical filtering algorithm is applied, and combined with real-time respiratory data and individual differences, respiratory characteristics are accurately estimated and abnormality detection is performed. Finally, the threshold is dynamically adjusted based on the real-time detection results, thereby achieving continuous optimization and precision of respiratory health monitoring.

[0074] In the above embodiment, by mining key statistical indicators from historical data, accurately analyzing the importance of multi-dimensional respiratory characteristics, and combining individual differences of users and real-time respiratory data, an adaptive statistical filtering algorithm is used to dynamically optimize and adjust the respiratory characteristic threshold in real time, thereby significantly improving the accuracy and personalization level of respiratory health monitoring.

[0075] In one embodiment, in step S232, performing feature importance analysis on the multi-dimensional respiratory features based on the key statistical indicators includes:

[0076] Step S2321, for each dimension of the respiratory feature, calculate the feature contribution based on the corresponding key statistical indicators to obtain a feature contribution score.

[0077] Specifically, for each dimension of respiratory features, this application will perform weighted calculation based on its corresponding key statistical indicators to obtain the corresponding feature contribution score.

[0078] In one of the embodiments, before performing the weighted calculation, taking into account that different statistical indicators may have different dimensions and value ranges, in order to eliminate the impact of such differences on the weighted calculation, the present application standardizes each key statistical indicator to convert each indicator to the same dimension and value range, thereby ensuring the accuracy of the weighted calculation results.

[0079] In one embodiment, during the weighted calculation process, the present application assigns a corresponding weight value to each key statistical indicator based on the correlation analysis results of each key statistical indicator relative to the overall respiratory state to reflect the relative contribution of each indicator in the description of the overall respiratory state. Specifically, the present application calculates the degree of association between each statistical indicator and the overall respiratory state, and quantifies the strength of the linear relationship between each statistical indicator and the overall respiratory state by performing a correlation analysis.

[0080] In one embodiment, the present application uses the Pearson correlation coefficient as a quantitative indicator to calculate the correlation coefficient between each key statistical indicator and the overall respiratory state; then, according to the value of the correlation coefficient, a linear mapping function is used to assign a corresponding weight value. Specifically, the linear mapping function formula here is as follows:

[0081]

[0082] Among them, W i is the weight value assigned to the i-th indicator, W max is the preset maximum weight value, r max With r min are the preset maximum and minimum correlation coefficient values, r i is the correlation coefficient of the ith indicator relative to the overall respiratory state.

[0083] Step S2322: Based on the feature contribution scores, importance analysis of multi-dimensional respiratory features is performed through quantitative evaluation.

[0084] Specifically, for each dimension of the breathing feature, this application will divide the corresponding feature contribution score by the sum of all feature contribution scores to obtain the relative importance ratio of the feature. Subsequently, this application will set dynamic thresholds based on the ratio value, individual differences of users, and real-time breathing data.

[0085] In the above embodiment, based on the feature contribution score, the importance analysis of multi-dimensional respiratory features is performed through quantitative evaluation, which can clarify the relative importance of each feature in the overall respiratory feature set. This helps to understand the description of the respiratory state by different features, thereby providing a basis for subsequent analysis and decision-making.

[0086] In one embodiment, in step S233, the adaptive statistical filtering algorithm is used to dynamically optimize and adjust the threshold of the multi-dimensional respiratory feature in real time according to the feature importance analysis result, the individual differences of the users, and the real-time respiratory data, including:

[0087] Step S2331, obtaining basic information of the user, and setting an initial threshold through a heuristic algorithm based on the basic information and the result of feature importance analysis.

[0088] Specifically, this application will obtain the user's basic information, such as age, gender, height, health status, etc. Furthermore, this application will set an initial threshold set based on the user's basic information and feature importance analysis results. Afterwards, during the operation of the heuristic algorithm, the threshold set will be continuously adjusted through a series of iterative steps to optimize the recognition rate of different respiratory abnormalities. Finally, when the algorithm stops iterating, the final initial threshold setting is output.

[0089] Step S2332, applying an adaptive statistical filtering algorithm, combining the user's real-time breathing data and individual differences to accurately estimate breathing characteristics and detect anomalies.

[0090] Specifically, during the application of the adaptive statistical filtering algorithm, the filter parameters will be dynamically adjusted according to the historical breathing data and the individual differences of the user to achieve accurate estimation of the breathing characteristics. In this process, the algorithm will use the historical breathing data to establish an initial filtering model, and update the model parameters in real time with the arrival of new data to adapt to the changes in the user's breathing characteristics. After multiple iterations and optimizations, the optimized target filtering model is obtained.

[0091] Furthermore, after obtaining the target filtering model, the present application will input the acquired real-time breathing data into the model to accurately estimate the breathing characteristics, and further compare the estimated breathing characteristic value with the preset normal range. When it is determined that it exceeds the preset normal range, the detection result of abnormal breathing is output.

[0092] Step S2333, adjusting the initial threshold based on the accurate estimation of the real-time breathing characteristics and the abnormality detection result to achieve dynamic optimization of the threshold.

[0093] Specifically, the application will analyze the frequency of the continuous occurrence of abnormal detection results, that is, count the number of times the abnormal detection results appear continuously within a certain period of time. When it is determined that the frequency value is higher than the preset threshold, it is considered that there is a problem with the current initial threshold setting, and the initial threshold will be adjusted according to the preset step size and adjustment direction (increase or decrease).

[0094] In the above embodiments, the accuracy and adaptability of respiratory health monitoring are improved through personalized respiratory threshold setting, accurate respiratory feature estimation and anomaly detection, and dynamic optimization of thresholds based on real-time results.

[0095] In one embodiment, in step S24, the multi-dimensional respiratory characteristics are compared and analyzed with the set dynamic threshold, and according to the analysis results, the corresponding comprehensive respiratory health index is calculated by weighted average method, including:

[0096] Step S241 , comparing the real-time data of the multi-dimensional respiratory features with the set dynamic thresholds respectively, to obtain the deviation value of each respiratory feature relative to the dynamic threshold.

[0097] Specifically, the present application will determine the value of each respiratory feature based on the acquired real-time data, and determine the deviation value of each respiratory feature relative to the dynamic threshold by difference calculation. It should be noted that the implementation of this step is intended to quantify the deviation between the respiratory feature and the health standard.

[0098] Step S242, based on the deviation value of each respiratory feature relative to the dynamic threshold, a corresponding respiratory health score is assigned to each respiratory feature through a linear conversion function.

[0099] Specifically, the present application will map the deviation value of each respiratory feature relative to the dynamic threshold to a respiratory health score value based on a pre-set linear conversion function, wherein a higher score means a better respiratory health status.

[0100] Step S243: Based on the feature importance analysis result, a corresponding importance weight is assigned to each respiratory feature through a weight mapping table.

[0101] Specifically, the present application constructs a weight mapping table based on expert experience, which maps the importance score of each respiratory feature to the corresponding importance weight. Specifically, during the construction process, the present application assigns a corresponding importance score interval to each respiratory feature based on the expert's evaluation results. Subsequently, a corresponding weight mapping table is constructed based on these importance score intervals by linear mapping, which can map the specific score values ​​falling into the importance score interval to the corresponding importance weight. In this way, no matter which interval the importance score of the respiratory feature falls into, the corresponding weight value can be accurately found.

[0102] Step S244, performing weighted average calculation based on the respiratory health score of each respiratory feature and the corresponding importance weight to obtain a corresponding comprehensive respiratory health index.

[0103] Specifically, this application will multiply the respiratory health score of each respiratory feature by its corresponding importance weight to obtain a corresponding weighted score. Afterwards, all weighted scores are summed up to obtain the total weighted score. Finally, this weighted score sum is divided by the sum of all importance weights to obtain the corresponding comprehensive respiratory health index. It should be noted that this index comprehensively reflects the numerical value of the user's respiratory health status, and is a comprehensive evaluation result that comprehensively considers multi-dimensional respiratory features and their importance weights.

[0104] In the above embodiment, accurate monitoring and evaluation of the user's respiratory health status is achieved by comparing multi-dimensional respiratory characteristics with dynamic thresholds in real time, using linear conversion functions to assign health scores, and assigning weights based on feature importance, and finally calculating a comprehensive respiratory health index.

[0105] In one embodiment, in step S242, the formula of the linear conversion function includes:

[0106]

[0107] Among them, score i represents the respiratory health score of the i-th respiratory feature, p i represents the deviation of the i-th respiratory feature relative to the dynamic threshold, thresh max Indicates the maximum value of the dynamic threshold preset, thresh min It represents the minimum value preset by the dynamic threshold, and e represents the exponential parameter of the linear conversion, which is used to adjust the sensitivity of the score to the deviation value.

[0108] In one embodiment, in step S3, generating a customized breathing training plan based on the comprehensive respiratory health index and the inference rules in the knowledge base through an intelligent decision engine combining a deep learning algorithm and a biofeedback mechanism includes:

[0109] Step S31, comparing the comprehensive respiratory health index with the health standards and threshold ranges corresponding to pathological characteristics in the knowledge base to obtain corresponding comparison results.

[0110] Specifically, this application will compare each indicator in the comprehensive respiratory health index with the threshold range of health standards and pathological characteristics preset in the knowledge base one by one. Through the comparison, it is determined whether each respiratory health indicator of the user is within the normal range, close to the abnormal range, or has exceeded the normal range, and a comprehensive comparison result is obtained based on this.

[0111] Step S32, based on the comparison result, logical judgment is performed through the inference rules in the knowledge base to preliminarily determine the breathing training direction and goal required by the user.

[0112] Specifically, this application will extract corresponding reasoning rules from the knowledge base based on the evaluation of the user's respiratory health status and possible pathological problems in the comparison results. These reasoning rules are formulated based on medical research and expert opinions, and are used to guide how to determine the appropriate breathing training direction and goals based on the user's respiratory health status. For example, if the comparison results show that the user's breathing rate is high, the reasoning rules will recommend taking breathing training directions that can reduce the breathing rate, such as abdominal breathing, deep breathing, etc. By applying these reasoning rules, logical judgments can be made to preliminarily determine the breathing training direction and goals required by the user.

[0113] Step S33, based on the user's personal information and the initially determined breathing training direction and goal, a deep learning algorithm is applied to generate a personalized training plan to obtain customized breathing training suggestions.

[0114] Specifically, this application will form a corresponding user profile based on the user's personal information and the initially determined breathing training direction and goals. After that, the user profile will be analyzed and processed using a deep learning algorithm, and a personalized training plan will be generated based on the output of the algorithm.

[0115] Step S34, integrating the biofeedback mechanism, optimizing the breathing training plan by real-time monitoring of the user's physiological response and adjusting the training plan, so that the training plan is more in line with the user's actual needs and physiological state.

[0116] Specifically, this application will integrate the biofeedback mechanism, use wearable devices to monitor the user's physiological reactions in real time, evaluate the training effect and risk based on the real-time monitored data and the preset algorithm threshold, and dynamically adjust the training plan accordingly. Subsequently, the adjusted training plan will be fed back to the user in a timely manner, and the user will confirm or propose modification suggestions based on their own situation and experience, so that the training plan is more in line with the user's actual needs and physiological state.

[0117] In the above embodiment, breathing training suggestions are formulated by comprehensively evaluating the user's respiratory health index and combining the knowledge base with the deep learning algorithm, and the breathing training suggestions are optimized and adjusted through real-time monitoring to ensure that the breathing training suggestions meet the user's actual needs and fit their physiological state, thereby ensuring that the training effect can be effectively improved.

[0118] Please refer to Figure 2 The present application discloses an abdominal breathing training system based on an intelligent guidance algorithm, the system comprising a real-time breathing data acquisition module, a comprehensive breathing state evaluation module, and a breathing training plan generation module, wherein:

[0119] The real-time respiratory data acquisition module is used to acquire the user's continuous real-time respiratory data sequence, and the continuous real-time respiratory data sequence covers at least one of the periodic change characteristics of respiratory frequency over time, the dynamic change characteristics of respiratory amplitude, and the precise definition characteristics of respiratory phase.

[0120] The respiratory status comprehensive evaluation module is used to comprehensively evaluate the continuous real-time respiratory data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory status.

[0121] The breathing training plan generation module is used to generate a customized breathing training plan according to the comprehensive respiratory health index, based on the inference rules in the knowledge base, through an intelligent decision-making engine that combines a deep learning algorithm with a biofeedback mechanism.

[0122] In one embodiment, the modules are also used to implement the abdominal breathing training method based on the intelligent guidance algorithm as described in any of the aforementioned method embodiments, which is not limited in this application.

[0123] From the above, it can be seen that the abdominal breathing training system based on the intelligent guidance algorithm disclosed in the present application, on the one hand, can generate a personalized comprehensive respiratory health index based on the user's individual differences and real-time respiratory data, based on the statistical analysis algorithm of dynamic threshold setting and multi-dimensional feature fusion, while comprehensively considering multiple respiratory characteristics, which helps to understand the user's respiratory health status more accurately; on the other hand, it can also generate customized breathing training plans based on the comprehensive respiratory health index based on the inference rules and deep learning algorithms in the knowledge base, using the intelligent decision-making engine. These plans can be optimized according to the user's current respiratory condition, thereby ensuring the scientificity and effectiveness of the training plan.

[0124] The present application discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the abdominal breathing training method based on an intelligent guidance algorithm.

[0125] From the above, it can be seen that a computer-readable storage medium disclosed in the present application, on the one hand, can generate a personalized comprehensive respiratory health index based on the statistical analysis algorithm of dynamic threshold setting and multi-dimensional feature fusion, taking into account multiple respiratory characteristics, according to the user's individual differences and real-time respiratory data, which helps to understand the user's respiratory health status more accurately; on the other hand, it can also generate customized respiratory training plans based on the comprehensive respiratory health index based on the inference rules and deep learning algorithms in the knowledge base, using an intelligent decision-making engine. These plans can be optimized according to the user's current respiratory condition, thereby ensuring the scientificity and effectiveness of the training plan.

[0126] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0127] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An abdominal breathing training method based on an intelligent guidance algorithm, characterized in that: The method comprises: S1. Acquire a continuous real-time respiratory data sequence of a user, wherein the continuous real-time respiratory data sequence includes at least one of a periodic change characteristic of respiratory frequency that changes over time, a dynamic change characteristic of respiratory amplitude, and an accurate definition characteristic of respiratory phase; S2, comprehensively evaluating the continuous real-time respiratory data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory status; S3. According to the comprehensive respiratory health index, based on the inference rules in the knowledge base, a customized respiratory training plan is generated by an intelligent decision-making engine that combines a deep learning algorithm with a biofeedback mechanism.

2. The method according to claim 1, characterized in that In step S2, the continuous real-time respiratory data sequence is comprehensively evaluated according to the statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory state, including: S21, performing a preprocessing operation on the continuous real-time respiratory data sequence to obtain a preprocessed respiratory data sequence, wherein the preprocessing operation includes at least one of denoising, smoothing, and outlier detection and correction; S22, extracting multi-dimensional respiratory features from the pre-processed respiratory data sequence based on respiratory waveform analysis; S23, obtaining historical breathing data of the user, and setting a dynamic threshold based on statistics based on the historical breathing data; S24, comparing and analyzing the multi-dimensional respiratory characteristics with the set dynamic threshold, and calculating the corresponding comprehensive respiratory health index by weighted average method based on the analysis results.

3. The method according to claim 2, characterized in that In step S23, the dynamic threshold setting based on statistics is performed based on the historical breathing data of the user and the multi-dimensional breathing characteristics, including: S231, determining key statistical indicators related to the multi-dimensional respiratory characteristics from historical respiratory data; S232, performing feature importance analysis on the multi-dimensional respiratory features based on the key statistical indicators to determine the contribution of each respiratory feature to the description of the respiratory state; S233. Based on the feature importance analysis results, individual differences of users, and real-time breathing data, an adaptive statistical filtering algorithm is used to dynamically optimize and adjust the threshold of the multi-dimensional breathing feature in real time.

4. The method according to claim 3, characterized in that In step S232, the feature importance analysis of the multi-dimensional respiratory features based on the key statistical indicators includes: S2321. For each dimension of the respiratory feature, calculate the feature contribution based on the corresponding key statistical indicators to obtain a feature contribution score; S2322. Based on the feature contribution scores, the importance of multi-dimensional respiratory features is analyzed through quantitative evaluation.

5. The method according to claim 3, characterized in that: In step S233, the threshold of the multi-dimensional respiratory feature is dynamically optimized and adjusted in real time using an adaptive statistical filtering algorithm based on the feature importance analysis results, individual differences of users, and real-time respiratory data, including: S2331, obtaining basic information of the user, and setting an initial threshold through a heuristic algorithm based on the basic information and feature importance analysis results; S2332, applying an adaptive statistical filtering algorithm, combining the user's real-time breathing data and individual differences to accurately estimate breathing characteristics and detect anomalies; S2333. Adjust the initial threshold based on the accurate estimation of real-time respiratory characteristics and abnormal detection results to achieve dynamic optimization of the threshold.

6. The method according to claim 2, characterized in that In step S24, the multi-dimensional respiratory characteristics are compared and analyzed with the set dynamic threshold, and according to the analysis results, the corresponding comprehensive respiratory health index is calculated by weighted average method, including: S241, comparing the real-time data of the multi-dimensional respiratory features with the set dynamic thresholds respectively, to obtain the deviation value of each respiratory feature relative to the dynamic threshold; S242, assigning a corresponding respiratory health score to each respiratory feature through a linear conversion function based on the deviation value of each respiratory feature relative to the dynamic threshold; S243, based on the feature importance analysis result, assigning a corresponding importance weight to each respiratory feature through a weight mapping table; S244, performing weighted average calculation based on the respiratory health score of each respiratory characteristic and the corresponding importance weight to obtain a corresponding comprehensive respiratory health index.

7. The method according to claim 6, characterized in that In step S242, the formula of the linear conversion function includes: Among them, score i represents the respiratory health score of the i-th respiratory feature, p i represents the deviation of the i-th respiratory feature relative to the dynamic threshold, thresh max Indicates the maximum value of the dynamic threshold preset, thresh min It represents the minimum value preset by the dynamic threshold, and e represents the exponential parameter of the linear conversion, which is used to adjust the sensitivity of the score to the deviation value.

8. The method according to claim 1, characterized in that In step S3, according to the comprehensive respiratory health index, based on the inference rules in the knowledge base, a customized respiratory training plan is generated by an intelligent decision engine combining a deep learning algorithm and a biofeedback mechanism, including: S31, comparing the comprehensive respiratory health index with the threshold ranges corresponding to the health standards and pathological characteristics in the knowledge base to obtain corresponding comparison results; S32, based on the comparison result, logical judgment is performed through the inference rules in the knowledge base to preliminarily determine the breathing training direction and goal required by the user; S33, based on the user's personal information and the initially determined breathing training direction and goal, applying a deep learning algorithm to generate a personalized training plan to obtain customized breathing training suggestions; S34. Integrate the biofeedback mechanism to optimize the breathing training plan by real-time monitoring of the user's physiological reactions and adjusting the training plan, so that the training plan is more in line with the user's actual needs and physiological state.

9. An abdominal breathing training system based on intelligent guidance algorithm, characterized in that: The system includes a real-time respiratory data acquisition module, a respiratory state comprehensive evaluation module, and a respiratory training plan generation module, wherein: The real-time respiratory data acquisition module is used to acquire a continuous real-time respiratory data sequence of the user, wherein the continuous real-time respiratory data sequence includes at least one of a periodic change characteristic of respiratory frequency over time, a dynamic change characteristic of respiratory amplitude, and an accurate definition characteristic of respiratory phase; The respiratory state comprehensive evaluation module is used to comprehensively evaluate the continuous real-time respiratory data sequence according to a statistical analysis algorithm based on dynamic threshold setting and multi-dimensional feature fusion to obtain a comprehensive respiratory health index for evaluating the user's respiratory state; The breathing training plan generation module is used to generate a customized breathing training plan according to the comprehensive respiratory health index, based on the inference rules in the knowledge base, through an intelligent decision-making engine that combines a deep learning algorithm with a biofeedback mechanism.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, the abdominal breathing training method based on the intelligent guidance algorithm is implemented.

Images