Method and system for measuring exercise heart rate variability
By collecting and analyzing the patient's exercise ECG in the exercise plate test and creating a heart rate analysis model, the problem that the prior art cannot effectively detect heart rate variability under exercise conditions is solved, and a more reliable heart rate variability detection is achieved.
Patent Information
- Application Number
- CN202411443953.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-16
- Publication Date
- 2025-02-18
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing 24-hour dynamic electrocardiogram cannot effectively collect data when the patient performs large or vigorous exercise, resulting in a lack of heart rate variability detection and cannot reflect the heart's load capacity.
Provide a method and system for measuring exercise heart rate variability. By collecting multiple exercise electrocardiograms of patients in exercise plate tests, dividing them into training sets and test sets, creating a heart rate analysis model, training and validating the model to output a patient's heart rate analysis report, Calculate the patient's heart rate variability.
It realizes effective collection and analysis of electrocardiogram data under exercise conditions, and the calculated heart rate variability is more reliable and can more comprehensively reflect the heart's load capacity.
Smart Images
Figure CN119453956A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of medical technology, and in particular to a method and system for measuring heart rate variability during exercise. Background Art
[0002] Heart rate variability (HRV) reflects the activity of the autonomic nervous system and quantitatively evaluates the tension and balance of the cardiac sympathetic and vagal nerves, thereby judging the severity and prevention of cardiovascular diseases. It may be a valuable indicator for predicting sudden cardiac death and arrhythmic events.
[0003] At present, heart rate variability (HRV) is only used in 24-hour dynamic electrocardiogram examinations, and has not yet been used in exercise treadmill electrocardiograms. Since 24-hour dynamic electrocardiograms are mostly performed according to normal daily routines, they generally cannot be performed with large or strenuous exercise, and sweating can easily cause the electrodes to fall off, resulting in missing test data. Moreover, since 24-hour dynamic electrocardiograms cannot be performed with strenuous exercise, their test data lacks exercise heart rate variability and cannot reflect the heart's load capacity. Summary of the invention
[0004] Based on this, it is necessary to provide a method and system for measuring heart rate variability during exercise, as 24-hour dynamic electrocardiogram examinations are mostly conducted according to normal daily routines, and generally large or strenuous exercise is not allowed. Sweating is easy to cause the electrodes to fall off, resulting in missing test data. Moreover, precisely because 24-hour dynamic electrocardiogram examinations cannot involve strenuous exercise, the test data lacks exercise heart rate variability and cannot reflect the heart's load capacity.
[0005] On the one hand, the present application provides a method for measuring heart rate variability during exercise, the method comprising:
[0006] Collect multiple patient exercise electrocardiograms, and divide the multiple patient exercise electrocardiograms into a training set and a test set;
[0007] Creating a heart rate analysis model, inputting a training set into the heart rate analysis model to train the heart rate analysis model, and obtaining a trained heart rate analysis model; the trained heart rate analysis model has the ability to output a patient heart rate analysis report based on an input patient exercise electrocardiogram;
[0008] Input the test set into the trained heart rate analysis model to obtain the patient's heart rate analysis report output by the trained heart rate analysis model; record the patient's heart rate analysis report output by the trained heart rate analysis model as a predicted heart rate report;
[0009] Determine whether the prediction training report is qualified. If the prediction training report is qualified, the heart rate analysis model training is considered completed;
[0010] The step of inputting the training set into the heart rate analysis model to train the heart rate analysis model comprises:
[0011] Based on each patient's exercise electrocardiogram in the training set, corresponding patient motion data is obtained;
[0012] Combined with each patient's motion data, the patient's heart rate detection data is calculated in the time domain and frequency domain respectively;
[0013] The patient's heart rate variability is output based on the patient's heart rate detection data.
[0014] On the other hand, the present application also provides a system for measuring heart rate variability during exercise, the system comprising:
[0015] A collection terminal, which is worn at the patient's collection site;
[0016] A collection module, which is arranged at a collection end, and collects the patient's exercise electrocardiogram through the collection module in the collection end;
[0017] The acquisition module is communicatively connected to the processor, and the processor is used to perform the exercise heart rate variability measurement method as described in any one of the above on the patient's exercise electrocardiogram.
[0018] The present application relates to a method and system for measuring heart rate variability during exercise. The method for measuring heart rate variability during exercise collects multiple patient exercise electrocardiograms (ECGs) and divides them into a training set and a test set, then creates a heart rate analysis model, inputs the training set into the heart rate analysis model to train the heart rate analysis model to obtain the trained heart rate analysis model, then inputs the test set into the trained heart rate analysis model to obtain a patient heart rate analysis report output by the trained heart rate analysis model, and finally determines whether the predicted training report is qualified. If the predicted training report is qualified, it is considered that the training of the heart rate analysis model is complete. The method for measuring heart rate variability during exercise provided by the present application can collect patient exercise ECGs under an exercise treadmill test, and calculate the patient's heart rate variability based on the patient's exercise ECG. Since the collection time required for the exercise treadmill test is shorter and can better reflect the heart's load capacity, the patient's heart rate variability calculated in this way is more reliable. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 A flowchart of a method for measuring heart rate variability during exercise provided in one embodiment of the present application.
[0020] Figure 2 A schematic diagram of the structure of a system for measuring heart rate variability during exercise provided in one embodiment of the present application.
[0021] Figure 3A patient's exercise electrocardiogram collected by a method for measuring exercise heart rate variability provided in one embodiment of the present application.
[0022] Reference numerals:
[0023] 100, acquisition terminal; 200, acquisition module; 300, processor. DETAILED DESCRIPTION
[0024] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0025] The present application provides a method and system for measuring heart rate variability during exercise.
[0026] like Figure 1 As shown, in one embodiment of the present application, a method for measuring heart rate variability during exercise is provided, and the method for measuring heart rate variability during exercise includes:
[0027] S100, collecting multiple patient exercise electrocardiograms, and dividing the multiple patient exercise electrocardiograms into a training set and a test set.
[0028] Specifically, the patient exercise electrocardiogram described in the present application is a 12-lead electrocardiogram collected when the patient performs an exercise treadmill test.
[0029] The reason why the patient's exercise electrocardiogram is collected when the patient is doing an exercise treadmill test is that the HRV during exercise can be detected simultaneously based on the electrocardiogram waveform detected during exercise, thereby collecting more comprehensive heart-related data.
[0030] S200, creating a heart rate analysis model, inputting a training set into the heart rate analysis model to train the heart rate analysis model, and obtaining a trained heart rate analysis model. The trained heart rate analysis model has the ability to output a patient heart rate analysis report based on an input patient exercise electrocardiogram.
[0031] S300, input the test set into the trained heart rate analysis model, obtain the patient's heart rate analysis report output by the trained heart rate analysis model, and record the patient's heart rate analysis report output by the trained heart rate analysis model as the predicted heart rate report.
[0032] S400, judging whether the predicted training report is qualified. If the predicted training report is qualified, it is considered that the heart rate analysis model training is completed.
[0033] The S200 includes:
[0034] S210, acquiring corresponding patient motion data based on each patient motion electrocardiogram in the training set.
[0035] S220, combining each patient's motion data, and calculating the patient's heart rate detection data in the time domain and frequency domain.
[0036] Specifically, the time domain and frequency domain are fundamental properties of signals, so signals can be analyzed in a variety of ways, each providing a different perspective. The fastest way to solve a problem is not necessarily the most obvious way, and the different perspectives used to analyze signals are called domains. The time domain and frequency domain clearly reflect the interaction between signals and interconnects.
[0037] S230: Output the patient's heart rate variability based on the patient's heart rate detection data.
[0038] In this embodiment, multiple patient exercise electrocardiograms are collected and divided into a training set and a test set, and then a heart rate analysis model is created. The training set is input into the heart rate analysis model to train the heart rate analysis model to obtain the trained heart rate analysis model, and then the test set is input into the trained heart rate analysis model to obtain the patient heart rate analysis report output by the trained heart rate analysis model. Finally, it is determined whether the predicted training report is qualified. If the predicted training report is qualified, it is considered that the heart rate analysis model training is completed. The exercise heart rate variability measurement method provided in this application can collect patient exercise electrocardiograms under exercise treadmill tests, and calculate the patient's heart rate variability based on the patient's exercise electrocardiogram. Since the collection time required for the exercise treadmill test is shorter and can better reflect the heart's load capacity, the patient's heart rate variability calculated in this way is more reliable.
[0039] In one embodiment of the present application, the S100 includes:
[0040] S110, creating an electrocardiogram database.
[0041] S120, collecting multiple patient exercise electrocardiograms, and putting all collected patient exercise electrocardiograms into an electrocardiogram database.
[0042] Specifically, for multiple patient exercise ECGs, the data source can be either different patients, with one patient exercise ECG collected from each patient, or the same patient, with multiple patient exercise ECGs collected from the same patient. It can also be a combination of the two, that is, multiple patient exercise ECGs are collected from each patient.
[0043] S130, setting the division ratio between the training set and the test set, and dividing all patient exercise electrocardiograms in the electrocardiogram database into the training set and the test set according to the division ratio.
[0044] Specifically, in order to improve the training effect of the heart rate analysis model and make the output of the heart rate analysis model more reliable, the proportion of the training set should be higher than the proportion of the test set.
[0045] In this embodiment, multiple patient exercise ECGs are collected, and all collected patient exercise ECGs are placed in an ECG database. Then, the division ratio between the training set and the test set is set, and all patient exercise ECGs in the ECG database are divided into the training set and the test set according to the division ratio.
[0046] Since the patient's exercise ECG is collected while the patient is undergoing a treadmill exercise test, and the treadmill exercise test is a test in which the patient exercises on a treadmill (running or brisk walking), the time required to collect the patient's exercise ECG is relatively short, only 5 to 10 minutes. Therefore, in terms of the efficiency of collecting the patient's exercise ECG, the collection method described in the present application is also superior to the 24-hour dynamic ECG test.
[0047] At the same time, since the exercise intensity of the treadmill exercise test is higher, it can better reflect the load capacity of the heart, so that the patient's heart rate variability obtained by the exercise heart rate variability measurement method described in this application is more comprehensive.
[0048] In one embodiment of the present application, the S210 includes:
[0049] S221, randomly select a patient's exercise ECG from the training set.
[0050] S222, based on the electrocardiogram waveform recorded in the selected patient exercise electrocardiogram, the patient exercise data is collected, and a coupling relationship between the patient exercise data and the patient exercise electrocardiogram is established. The patient exercise data includes the patient's target heart rate, the exercise termination heart rate, the percentage of the termination heart rate to the patient's target heart rate, and the heart rate in the recovery period for 1 minute.
[0051] Specifically, the patient's exercise data also includes the heart rate for 2 minutes during the recovery period, the heart rate for 3 minutes during the recovery period, the heart rate for 4 minutes during the recovery period, the heart rate for 5 minutes during the recovery period, the heart rate for 6 minutes during the recovery period, metabolic equivalent, and the degree of depression of electrocardiogram leads during exercise.
[0052] These patient motion data can reflect the heart rate variability of the corresponding patients in various aspects. Heart rate variability refers to the slight difference between successive heartbeats. It is caused by the modulation of the heart's sinoatrial node by the autonomic nervous system, which generally causes the heart beat intervals to differ and fluctuate by tens of milliseconds.
[0053] Heart rate variability (HRV) represents a quantitative measurement that reflects the degree and regularity of heart rate changes by measuring the variability of continuous normal RR interval changes, so as to judge its impact on cardiovascular activity. A decrease in HRV indicates an increase in sympathetic nerve tension, which can reduce the ventricular fibrillation threshold and is an unfavorable factor. An increase in HRV indicates an increase in parasympathetic nerve tension, which increases the ventricular fibrillation threshold and is a protective factor.
[0054] S223, return to randomly select a patient exercise electrocardiogram from the training set until all patient exercise electrocardiograms in the electrocardiogram database are selected, and obtain the patient exercise data corresponding to each patient exercise electrocardiogram.
[0055] In this embodiment, since each patient exercise electrocardiogram includes an electrocardiogram waveform, various patient exercise data can be collected according to the electrocardiogram waveform included in the patient exercise electrocardiogram, and the collected patient exercise data can be compared with the patient exercise electrocardiogram. Figure 1 One-to-one correspondence ensures that the patient motion data contained in each patient motion electrocardiogram in the electrocardiogram database can be reflected.
[0056] In one embodiment of the present application, the S220 includes:
[0057] S221, randomly selecting a patient exercise electrocardiogram and the patient exercise data corresponding to the patient exercise electrocardiogram from the electrocardiogram data.
[0058] S222, calculating the average RR interval of the patient's exercise electrocardiogram using Formula 1.
[0059] RR=T2-T1 formula 1.
[0060] Among them, RR is the average RR interval of the patient's exercise electrocardiogram, T1 is the peak time of the previous QRS wave, and T2 is the peak time of the following QRS wave.
[0061] S223, calculating the median RR interval of the patient's exercise electrocardiogram using Formula 2.
[0062]
[0063] Among them, M RR It is the middle value of the RR interval of the patient's exercise electrocardiogram, T1 is the peak time of the previous QRS wave, and T2 is the peak time of the following QRS wave.
[0064] Specifically, the QRS complex wave in the electrocardiogram represents the change state of the cardiac potential of ventricular depolarization, and is composed of a downward Q wave, a wide and high-amplitude R wave, and a downward S wave. The QRS complex wave has the most obvious waveform in the electrocardiogram, with a wide waveform and a large slope, so it is the easiest to identify. It is often used as the basis for electrocardiogram signal research and is widely used.
[0065] Optionally, in addition to calculating the two exercise data, the average RR interval of the patient's exercise ECG and the median RR interval of the patient's exercise ECG, the standard deviation of the entire sinus RR interval, the standard deviation of the mean sinus RR interval every 5 minutes, the average of the standard deviation of the mean sinus RR interval every 5 minutes, the ratio of the standard deviation of the RR interval to the average RR interval, the standard deviation of the difference between adjacent NN intervals throughout the entire process, the root mean square value of the difference between adjacent NN intervals throughout the entire process, and the difference between the longest RR interval and the shortest RR interval can also be calculated.
[0066] These motion data are all detection data of the patient's motion electrocardiogram in the time domain. Since these motion data have a reference interval, when the calculated motion data is not in the reference interval, it can be considered that the patient's heart rate variability is large.
[0067] In this embodiment, the patient's heart rate variability is determined by calculating the patient's motion data in the time domain and using the calculated patient's motion data as one of the criteria for evaluating the patient's heart rate variability.
[0068] In one embodiment of the present application, the S220 further includes:
[0069] S224, setting the RR difference threshold.
[0070] S225, counting the number of adjacent RR intervals in the patient's exercise electrocardiogram that are greater than or equal to the RR difference threshold.
[0071] S226, calculating the ratio of the number of adjacent RR intervals greater than or equal to the RR difference threshold in the patient's exercise electrocardiogram to the total number of heartbeats in the patient's exercise electrocardiogram using Formula 3.
[0072]
[0073] Wherein, W is the ratio of the number of adjacent RR intervals greater than or equal to the RR difference threshold in the patient's exercise ECG to all heartbeats in the patient's exercise ECG, g is the number of adjacent RR intervals greater than or equal to the RR difference threshold in the patient's exercise ECG, and G is all heartbeats in the patient's exercise ECG.
[0074] S227, returning to randomly select a patient exercise electrocardiogram and the patient exercise data corresponding to the patient exercise electrocardiogram from the electrocardiogram database, until all patient exercise electrocardiograms in the electrocardiogram database are selected.
[0075] In this embodiment, the motion data of the patient's exercise electrocardiogram in the time domain are calculated, and these motion data are used together with the motion data calculated in the above embodiments as one of the bases for judging the patient's heart rate variability.
[0076] In one embodiment of the present application, the S220 further includes:
[0077] S227, randomly select a patient exercise electrocardiogram from the electrocardiogram data.
[0078] S228, performing Fourier transform on the selected patient exercise electrocardiogram to convert the selected patient exercise electrocardiogram into a patient exercise frequency domain image.
[0079] Specifically, the patient's exercise electrocardiogram can be Fourier transformed by formula 7:
[0080] (F(\omega)=\int_{-\infty}^{-\infty}f(t)e^{-j\omegat}dt) Formula 7.
[0081] S229, acquiring a low-frequency range power of the patient motion frequency domain image and a high-frequency range power of the patient motion frequency domain image based on the patient motion frequency domain image.
[0082] S229-A, calculate the high- and low-frequency power ratio of the patient's motion frequency domain image using Formula 4.
[0083]
[0084] Among them, Q is the high-to-low frequency power ratio of the patient's motion frequency domain image, LF is the low-frequency range power of the motion frequency domain image, and HF is the high-frequency range power of the patient's motion frequency domain image.
[0085] S229-B, returns to randomly select a patient exercise ECG from the ECG data until all patient exercise ECGs in the ECG database are selected, and obtains the high- and low-frequency power ratios of each patient exercise ECG.
[0086] In this embodiment, since the sine wave is the only waveform that exists in the frequency domain, that is, the sine wave is a description of the frequency domain, and Fourier transform is a method that can analyze sine wave signals, Fourier transform is used to analyze the motion data of the patient's exercise electrocardiogram in the frequency domain.
[0087] By performing Fourier transform on the patient's exercise ECG, the patient's exercise ECG is converted from the time domain to the frequency domain, and the patient's motion data is calculated in the frequency domain. Similarly, after calculating the patient's motion data in the frequency domain, the motion data in the frequency domain will also be used as one of the bases for judging the patient's heart rate variability.
[0088] The reason why it is necessary to calculate the patient's motion data in the time domain and frequency domain respectively is to make the analysis of the patient's motion electrocardiogram more comprehensive, thereby ensuring a more accurate assessment of the patient's heart rate variability.
[0089] In one embodiment of the present application, between S229 and S229-A, further comprising:
[0090] K100, the threshold of the patient's motion frequency domain image is calculated by formula 5.
[0091]
[0092] Wherein, λ is the threshold of the patient motion frequency domain image, σ is the standard deviation of the noise, and N is the signal length of the patient motion frequency domain image.
[0093] K200, performs soft threshold processing on the patient's motion frequency domain image to remove noise from the patient's motion frequency domain image.
[0094] In this embodiment, the patient motion frequency domain image is denoised by wavelet filtering, thereby reducing the noise in the waveform of the patient motion frequency domain image to ensure that the motion data in the frequency domain direction is more reliable.
[0095] The wavelet filtering method is to decompose the signal into multiple wavelet coefficients of different scales, and then remove the noise by thresholding the wavelet coefficients. Since the patient's motion electrocardiogram has been converted into the patient's motion frequency domain image by Fourier transform in the above embodiment, it is only necessary to perform thresholding on the wavelet coefficients of each scale of the patient's motion frequency domain image and reconstruct the processed wavelet coefficients to complete the denoising process.
[0096] In one embodiment of the present application, the S230 includes:
[0097] S231 calculates the patient's heart rate variability using formula 6.
[0098] Res=A×RR+B×MRR+C×Q Formula 6.
[0099] Where Res is the patient's heart rate variability, RR is the average RR interval of the patient's exercise electrocardiogram, and M RRis the median RR interval of the patient's exercise electrocardiogram, Q is the high-to-low frequency power ratio of the patient's exercise frequency domain image, A is the weight coefficient of the average RR interval of the patient's exercise electrocardiogram, B is the weight coefficient of the median RR interval of the patient's exercise electrocardiogram, and C is the weight coefficient of the high-to-low frequency power ratio of the patient's exercise frequency domain image.
[0100] In this embodiment, different weight coefficients are set for the motion data in the time domain and the motion data in the frequency domain respectively to balance the influence of the motion data in the time domain and the motion data in the frequency domain on the final heart rate variability of the patient.
[0101] Since the motion data is not limited to the motion data provided in the aforementioned embodiment, when actually calculating the patient's heart rate variability, other motion data can also be added to the calculation formula 6 of the patient's heart rate variability, and a corresponding weight coefficient can be assigned to each motion data, thereby ensuring that the patient's heart rate variability finally calculated has confidence for different motion data.
[0102] In one embodiment of the present application, the S400 includes:
[0103] S410, setting a qualified threshold and a qualified number.
[0104] S420, inputting the exercise electrocardiograms of all patients in the test set into the trained heart rate analysis model in sequence.
[0105] S430, obtaining the patient's heart rate variability of each patient's exercise electrocardiogram output by the trained heart rate analysis model.
[0106] S440, determining whether the heart rate variability of each patient's exercise electrocardiogram is qualified, and adding 1 to the qualified number whenever there is a patient's exercise electrocardiogram with a qualified heart rate variability.
[0107] S450, determining whether the qualified number is greater than or equal to a qualified threshold.
[0108] S460: If the qualified number is greater than or equal to the qualified threshold, the trained heart rate analysis model is qualified.
[0109] S470, if the number of qualified results is less than the qualified threshold, the division ratio of the training set and the test set is returned to be set, and all patient exercise electrocardiograms in the electrocardiogram database are divided into the training set and the test set according to the division ratio, until the number of qualified results is greater than or equal to the qualified threshold.
[0110] In this embodiment, the test set is used to verify whether the trained heart rate analysis model has the ability to output the patient's heart rate analysis report based on the input patient's exercise electrocardiogram. When the trained heart rate analysis model has the ability to output the patient's heart rate analysis report based on the input patient's exercise electrocardiogram, the training of the heart rate analysis model can be considered completed.
[0111] If the trained heart rate analysis model still does not have the ability to output a patient heart rate analysis report based on the input patient exercise electrocardiogram, then it is necessary to re-divide the training set and test set to retrain the heart rate analysis model. When re-dividing the training set and test set, the division ratio of the training set should be increased to increase the number of training samples contained in the training set.
[0112] like Figure 2 As shown, in one embodiment of the present application, a system for measuring heart rate variability during exercise is provided. The system for measuring heart rate variability during exercise includes an acquisition terminal 100 , an acquisition module 200 , and a processor 300 .
[0113] The acquisition terminal 100 is worn at the patient's acquisition location. The acquisition module 200 is disposed at the acquisition terminal 100, and the patient's exercise electrocardiogram is acquired through the acquisition module 200 in the acquisition terminal 100. The acquisition module 200 is in communication connection with the processor 300, and the processor 300 performs the exercise heart rate variability measurement method as described in any of the aforementioned embodiments on the patient's exercise electrocardiogram.
[0114] In this embodiment, Figure 3 As shown in the figure, the horizontal coordinate of the collected patient exercise electrocardiogram is time, in ms, and the vertical coordinate is voltage, in mV. The waveform of the patient's exercise electrocardiogram contains P wave, QRS complex, T wave and U wave, so the patient's exercise data can be directly calculated based on the waveform of the patient's exercise electrocardiogram.
[0115] Among them, the initial P wave corresponds to the excitation process of the atrium. A complete cardiac cycle starts from the sinoatrial node. The onset of the P wave represents the process in which the sinoatrial node transmits the cardiac excitement to the atrium so that the entire atrium is stimulated.
[0116] The QRS complex represents the changing state of the heart's potential during ventricular depolarization. Since the various parts of the ventricles are stimulated in a certain order, the QRS complex wave is a comprehensive waveform that represents the excitation process of several parts of the ventricles. It is composed of a downward Q wave, a wide and high-amplitude R wave, and a downward S wave.
[0117] The average RR interval of the patient's exercise electrocardiogram calculated by formula 1 is exactly the time between two QRS complexes.
[0118] The technical features of the above-described embodiments may be arbitrarily combined, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0119] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for measuring heart rate variability during exercise, characterized in that: The exercise heart rate variability measurement method comprises: Collect multiple patient exercise electrocardiograms, and divide the multiple patient exercise electrocardiograms into a training set and a test set; Creating a heart rate analysis model, inputting a training set into the heart rate analysis model to train the heart rate analysis model, and obtaining a trained heart rate analysis model; the trained heart rate analysis model has the ability to output a patient heart rate analysis report based on an input patient exercise electrocardiogram; Input the test set into the trained heart rate analysis model to obtain the patient's heart rate analysis report output by the trained heart rate analysis model; record the patient's heart rate analysis report output by the trained heart rate analysis model as a predicted heart rate report; Determine whether the prediction training report is qualified. If the prediction training report is qualified, the heart rate analysis model training is considered completed; The step of inputting the training set into the heart rate analysis model to train the heart rate analysis model comprises: Obtain corresponding patient motion data based on each patient motion electrocardiogram in the training set; Combined with each patient's motion data, the patient's heart rate detection data is calculated in the time domain and frequency domain respectively; The patient's heart rate variability is output based on the patient's heart rate detection data.
2. The method for measuring heart rate variability during exercise according to claim 1, characterized in that: The method of collecting multiple patient exercise electrocardiograms and dividing the multiple patient exercise electrocardiograms into a training set and a test set includes: Create an ECG database; Collect multiple patient exercise electrocardiograms, and put all collected patient exercise electrocardiograms into an electrocardiogram database; The division ratio between the training set and the test set is set, and all patient exercise ECGs in the ECG database are divided into the training set and the test set according to the division ratio.
3. The method for measuring heart rate variability during exercise according to claim 2, characterized in that: The step of obtaining corresponding patient motion data based on each patient motion electrocardiogram in the training set includes: Randomly select a patient's exercise ECG from the training set; Based on the electrocardiogram waveform recorded by the selected patient exercise electrocardiogram, the patient exercise data is collected, and a coupling relationship between the patient exercise data and the patient exercise electrocardiogram is established; the patient exercise data includes the patient's target heart rate, the exercise termination heart rate, the percentage of the termination heart rate to the patient's target heart rate, and the heart rate in the recovery period for 1 minute; Return to randomly select a patient exercise ECG from the training set until all patient exercise ECGs in the ECG database have been selected, and obtain the patient exercise data corresponding to each patient exercise ECG.
4. The method for measuring heart rate variability during exercise according to claim 3, characterized in that: The step of combining each patient's motion data and calculating the patient's heart rate detection data in the time domain and the frequency domain respectively includes: Randomly select a patient exercise electrocardiogram and patient exercise data corresponding to the patient exercise electrocardiogram from the electrocardiogram data; The average RR interval of the patient's exercise electrocardiogram was calculated by formula 1; RR=T2-T1 Formula 1; Where RR is the average RR interval of the patient's exercise electrocardiogram, T1 is the peak time of the preceding QRS wave, and T2 is the peak time of the following QRS wave; The median RR interval of the patient's exercise electrocardiogram was calculated using formula 2; Among them, M RR It is the middle value of the RR interval of the patient's exercise electrocardiogram, T1 is the peak time of the previous QRS wave, and T2 is the peak time of the following QRS wave.
5. The method for measuring heart rate variability during exercise according to claim 4, characterized in that: The step of combining each patient's motion data and calculating the patient's heart rate detection data in the time domain and the frequency domain respectively further includes: Set the RR difference threshold; Count the number of adjacent RR intervals in the patient's exercise electrocardiogram that are greater than or equal to the RR difference threshold; The ratio of the number of adjacent RR intervals greater than or equal to the RR difference threshold in the patient's exercise electrocardiogram to all the heartbeats in the patient's exercise electrocardiogram is calculated by formula 3; Wherein, W is the ratio of the number of adjacent RR intervals greater than or equal to the RR difference threshold in the patient's exercise electrocardiogram to the total number of heartbeats in the patient's exercise electrocardiogram, g is the number of adjacent RR intervals greater than or equal to the RR difference threshold in the patient's exercise electrocardiogram, and G is the total number of heartbeats in the patient's exercise electrocardiogram; Return to randomly select a patient exercise electrocardiogram and the patient exercise data corresponding to the patient exercise electrocardiogram from the electrocardiogram database until all patient exercise electrocardiograms in the electrocardiogram database have been selected.
6. The method for measuring heart rate variability during exercise according to claim 5, characterized in that: The step of combining each patient's motion data and calculating the patient's heart rate detection data in the time domain and the frequency domain respectively further includes: A patient exercise ECG is randomly selected from the ECG data; Performing Fourier transformation on the selected patient's exercise electrocardiogram to convert the selected patient's exercise electrocardiogram into a patient's exercise frequency domain image; Acquire a low-frequency range power of the patient's motion frequency domain image and a high-frequency range power of the patient's motion frequency domain image based on the patient's motion frequency domain image; The high- and low-frequency power ratio of the patient's motion frequency domain image is calculated by formula 4; Among them, Q is the high-to-low frequency power ratio of the patient's motion frequency domain image, LF is the low-frequency range power of the motion frequency domain image, and HF is the high-frequency range power of the patient's motion frequency domain image; Return to randomly select a patient exercise ECG from the ECG data until all patient exercise ECGs in the ECG database are selected, and obtain the high- and low-frequency power ratios of each patient exercise ECG.
7. The method for measuring heart rate variability during exercise according to claim 6, characterized in that: In the step of obtaining the low-frequency range power and the high-frequency range power of the patient motion frequency domain image based on the patient motion frequency domain image, and calculating the high-to-low frequency power ratio of the patient motion frequency domain image by a formula, the method further includes: The threshold of the patient's motion frequency domain image is calculated by formula 5; Where λ is the threshold of the patient motion frequency domain image, σ is the standard deviation of the noise, and N is the signal length of the patient motion frequency domain image; The patient motion frequency domain image is subjected to soft threshold processing to remove noise from the patient motion frequency domain image.
8. The method for measuring exercise heart rate variability according to claim 7, characterized in that: The step of outputting the patient's heart rate variability based on the patient's heart rate detection data comprises: The patient's heart rate variability is calculated using Formula 6; Res=A×RR+B×MRR+C×Q Formula 6; Where Res is the patient's heart rate variability, RR is the average RR interval of the patient's exercise electrocardiogram, and M RR is the median RR interval of the patient's exercise electrocardiogram, Q is the high-to-low frequency power ratio of the patient's exercise frequency domain image, A is the weight coefficient of the average RR interval of the patient's exercise electrocardiogram, B is the weight coefficient of the median RR interval of the patient's exercise electrocardiogram, and C is the weight coefficient of the high-to-low frequency power ratio of the patient's exercise frequency domain image.
9. The method for measuring heart rate variability during exercise according to claim 8, characterized in that: The determining whether the prediction training report is qualified includes: Set the pass threshold and pass number; All patient exercise ECGs in the test set are sequentially input into the trained heart rate analysis model; Obtain the patient's heart rate variability for each patient's exercise electrocardiogram output by the trained heart rate analysis model; Determine whether the patient's heart rate variability of each patient's exercise electrocardiogram is qualified, and when there is a patient's heart rate variability of the patient's exercise electrocardiogram that is qualified, increase the qualified number by 1; Determine whether the qualified number is greater than or equal to the qualified threshold; If the qualified number is greater than or equal to the qualified threshold, the heart rate analysis model is qualified after training; If the number of qualified results is less than the qualified threshold, the division ratio of the set training set and the test set is returned, and all patient exercise ECGs in the ECG database are divided into the training set and the test set according to the division ratio until the number of qualified results is greater than or equal to the qualified threshold.
10. A system for measuring heart rate variability during exercise, characterized in that: The exercise heart rate variability measurement system comprises: A collection terminal, which is worn at the patient's collection site; A collection module, which is arranged at a collection end, and collects the patient's exercise electrocardiogram through the collection module in the collection end; A processor, wherein the acquisition module is communicatively connected to the processor, and the processor is used to perform the exercise heart rate variability measurement method according to any one of claims 1 to 9 on the patient's exercise electrocardiogram.