Heart health status detection method, device and electronic equipment

Abstract

The present disclosure relates to a method, device and electronic device for detecting heart health status. The method includes: determining multiple first feature sets and multiple second feature sets in a short-term feature set based on a heart sound signal to be measured, wherein each first feature set is determined based on the signal amplitude of the heart sound signal in a corresponding window at each sampling point, and each second feature set is determined based on the duration of the first heart sound interval, the systolic interval, the second heart sound interval and the diastolic interval of each cardiac cycle in the heart sound signal in the corresponding window; determining a time domain feature set, a frequency domain feature set and a Poincaré map-related parameter set in a long-term feature set based on the heart sound signal to be measured, wherein the time domain feature set is determined based on the duration of each cardiac cycle, the frequency domain feature set is determined based on the signal power, and the Poincaré map-related parameter set is determined based on the Poincaré map generated based on the duration of each cardiac cycle; and inputting the short-term feature set and the long-term feature set into a set heart health status detection model to obtain a heart health status result.

Description

Technical Field

[0001] The present disclosure relates to heart state detection technology, and more specifically, to a heart health state detection method, device, and electronic device. Background Art

[0002] Cardiovascular disease is the leading cause of death worldwide. Cardiac auscultation using a stethoscope is one of the most important and cost-effective tools for pre-screening for cardiovascular disease. Two basic heart sound components can be heard during cardiac auscultation: the first heart sound (S1) and the second heart sound (S2). S1 is produced when the mitral and tricuspid valves close, while S2 is produced when the aortic and pulmonary valves close. In addition to these heart sounds, murmurs can also be heard during auscultation. Clinicians can identify abnormal heart sound characteristics such as murmurs, changes in heart sound intensity, and extra heart sounds through auscultation, enabling preliminary screening for various types of cardiovascular disease.

[0003] In recent years, computer-assisted auscultation systems based on phonocardiogram (PCG) signals have attracted increasing attention. However, although such computer-assisted auscultation systems have improved diagnostic efficiency, the accuracy of diagnostic results is relatively poor. Summary of the Invention

[0004] An object of the present invention is to provide a new technical solution for a method for detecting a heart health status.

[0005] According to a first aspect of the present invention, a method for detecting a heart health state is provided, comprising:

[0006] Determine a short-term feature set based on the heart sound signal to be measured, wherein the short-term feature set includes a plurality of first feature sets and a plurality of second feature sets, each first feature set is determined based on the signal amplitude of the heart sound signal in a corresponding window at each sampling point, and each second feature set is determined based on the duration of a first heart sound interval, a systolic interval, a second heart sound interval, and a diastolic interval in each cardiac cycle in the heart sound signal in the corresponding window, and the heart sound signal in each window includes signals corresponding to a preset number of cardiac cycles;

[0007] Determining a long-term feature set based on the heart sound signal to be measured, wherein the long-term feature set includes a time domain feature set, a frequency domain feature set, and a Poincare map-related parameter set, wherein the time domain feature set is determined based on the duration of each cardiac cycle, the frequency domain feature set includes signal power in a low-frequency range, signal power in a high-frequency range, and a ratio of signal power in the low-frequency range to signal power in the high-frequency range, and the Poincare map-related parameter set is determined based on a Poincare map generated based on the duration of each cardiac cycle;

[0008] The first feature set, the second feature set, the time domain feature set, the frequency domain feature set and the Poincare map related parameter set are input into a set heart health status detection model to obtain a heart health status result.

[0009] Optionally, determining a short-term feature set based on the heart sound signal to be measured includes:

[0010] Based on the heart sound signal to be measured, the mean, median, standard deviation, skewness and kurtosis of the signal amplitude of the heart sound signal at each sampling point in the corresponding window are determined to form a plurality of first feature sets.

[0011] Optionally, determining a short-term feature set based on the heart sound signal to be measured includes:

[0012] Based on the heart sound signal to be measured, determine the average duration of the first heart sound interval, the average duration of the systolic interval, the average duration of the second heart sound interval, the average duration of the diastolic interval, the average duration of the cardiac cycle, the ratio of the average duration of the systolic interval to the average duration of the diastolic interval, the ratio of the average duration of the systolic interval to the average duration of the cardiac cycle, and the average duration of the diastolic interval to the average duration of the cardiac cycle in the heart sound signal within the corresponding window to form multiple second feature sets.

[0013] Optionally, determining the long-term feature set based on the heart sound signal to be measured includes:

[0014] Based on the heart sound signal to be measured, the mean of the cardiac cycle duration, the standard deviation of the cardiac cycle duration, the root mean square of the difference between the durations of two adjacent cardiac cycles, and the ratio of the number of cardiac cycles with a difference between the durations of two adjacent cardiac cycles greater than a preset threshold to the number of all cardiac cycles are determined to form a time domain feature set.

[0015] Optionally, determining the long-term feature set based on the heart sound signal to be measured includes:

[0016] Performing a fast Fourier transform on the heart sound signal to be measured to obtain a power spectral density function;

[0017] Obtain preset frequency resolution, preset low frequency range, and preset high frequency range;

[0018] The signal power of the low frequency range is determined according to the power spectral density function, the preset frequency resolution and the preset low frequency range; the signal power of the high frequency range is determined according to the power spectral density function, the preset frequency resolution and the preset high frequency range; and the ratio of the signal power of the low frequency range to the signal power of the high frequency range is determined according to the signal power of the low frequency range and the signal power of the high frequency range to form a frequency domain feature set.

[0019] Optionally, determining the long-term feature set based on the heart sound signal to be measured includes:

[0020] generating a Poincare map based on the duration of each cardiac cycle in the heart sound signal to be measured;

[0021] The length of the minor axis of the Poincaré map, the length of the major axis of the Poincaré map, and the ratio of the length of the minor axis of the Poincaré map to the length of the major axis of the Poincaré map are determined as parameters related to the Poincaré map.

[0022] Optionally, before determining the short-term feature set based on the heart sound signal to be measured, the method further includes:

[0023] determining the duration of each cardiac cycle based on the heart sound signal to be measured;

[0024] Based on the duration of each cardiac cycle, screening out cardiac cycles with a duration greater than a preset duration;

[0025] The signals corresponding to the cardiac cycles whose duration is greater than the preset duration are eliminated from the heart sound signals to be measured, so as to obtain the processed heart sound signals to be measured.

[0026] According to a second aspect of the present invention, there is provided a device for detecting a heart health state, comprising:

[0027] a short-term feature set determination module, configured to determine a short-term feature set based on the heart sound signal to be measured, wherein the short-term feature set includes a plurality of first feature sets and a plurality of second feature sets, each first feature set is determined based on the signal amplitude of the heart sound signal in a corresponding window at each sampling point, and each second feature set is determined based on the duration of a first heart sound interval, a systolic interval, a second heart sound interval, and a diastolic interval in each cardiac cycle in the heart sound signal in the corresponding window, and the heart sound signal in each window includes signals corresponding to a preset number of cardiac cycles;

[0028] a long-term feature set determination module, configured to determine a long-term feature set based on the heart sound signal to be measured, wherein the long-term feature set includes a time-domain feature set, a frequency-domain feature set, and a Poincare map-related parameter set, wherein the time-domain feature set is determined based on the duration of each cardiac cycle, the frequency-domain feature set includes the signal power in the low-frequency range, the signal power in the high-frequency range, and the ratio of the signal power in the low-frequency range to the signal power in the high-frequency range, and the Poincare map-related parameter set is determined based on the Poincare map generated based on the duration of each cardiac cycle;

[0029] The detection module is used to input the first feature set, the second feature set, the time domain feature set, the frequency domain feature set and the Poincare map related parameter set into the set heart health status detection model to obtain a heart health status result.

[0030] According to a third aspect of the present invention, a heart health status detection device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the computer program is used to control the processor to operate to execute the method according to any one of the first aspects.

[0031] According to a fourth aspect of the present invention, an electronic device is provided, comprising a heart health status detection device as described in the second aspect or the third aspect and a sensor for collecting heart sound signals, wherein the sensor for collecting heart sound signals is connected to the heart health status detection device.

[0032] The present disclosure provides a method for detecting heart health status. Based on the heart sound signal to be measured, a short-time feature set and a long-time feature set are determined respectively. Then, based on the short-time feature set and the long-time feature set, a set heart health status detection model is used to obtain a heart health status result. The short-time feature set is determined based on the heart sound signal in each window. The heart sound signal in each window is a short-time heart sound signal obtained by continuous segmentation from the heart sound signal to be measured. The short-time features extracted based on the short-time heart sound signal constitute a short-time feature set. The short-time features can sensitively capture transient events and provide an effective data basis for the detection of subsequent heart health status results. In addition, the short-time The time feature set includes feature sets of multiple dimensions, namely the first feature set and the second feature set, which provide a more comprehensive data basis for the subsequent detection of heart health status results. The long-term feature set is determined based on the entire heart sound signal to be tested. The features in the long-term feature set can quantify the overall trend, and based on the overall trend, the accuracy of identifying chronic heart diseases can be improved. The long-term feature set also includes feature sets of multiple dimensions, namely the time domain feature set, the frequency domain feature set and the Poincare map related parameter set, which provide a more comprehensive data basis for the subsequent detection of heart health status results. Using the set heart health status detection model, the detection efficiency and accuracy can be improved.

[0033] Features and advantages of the embodiments of the present specification will become apparent from the following detailed description of exemplary embodiments of the present specification with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the specification and, together with the description, serve to explain the principles of the embodiments of the specification.

[0035] Figure 1 FIG. 4 is a flow chart of a method for detecting heart health status according to an embodiment of the present invention.

[0036] Figure 2 FIG. 1 is a schematic diagram of a heart sound signal in a window of a heart sound signal to be measured according to an embodiment of the present invention.

[0037] Figure 3 FIG. 4 is a structural diagram of a device for detecting heart health status according to an embodiment of the present invention.

[0038] Figure 4 FIG. 4 is a structural diagram of a device for detecting heart health status according to an embodiment of the present invention.

[0039] Figure 5 FIG. 1 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] Various exemplary embodiments of the present specification will now be described in detail with reference to the accompanying drawings.

[0041] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the embodiments of this specification, its application, or uses.

[0042] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0043] In order to solve the above technical problems, the present disclosure provides a method for detecting the health status of the heart. Based on the heart sound signal to be measured, a short-time feature set and a long-time feature set are determined respectively. Then, based on the short-time feature set and the long-time feature set, a set heart health status detection model is used to obtain the heart health status result. The short-time feature set is determined based on the heart sound signal in each window. The heart sound signal in each window is a short-time heart sound signal obtained by continuous segmentation from the heart sound signal to be measured. The short-time features extracted based on the short-time heart sound signal constitute a short-time feature set. The short-time features can sensitively capture transient events and provide an effective data basis for the detection of subsequent heart health status results. In addition, the short-term feature set includes feature sets of multiple dimensions, namely the first feature set and the second feature set, which provide a more comprehensive data basis for the subsequent detection of heart health status results. The long-term feature set is determined based on the entire heart sound signal to be tested. The features in the long-term feature set can quantify the overall trend. Based on the overall trend, the accuracy of identifying chronic heart diseases can be improved. The long-term feature set also includes feature sets of multiple dimensions, namely the time domain feature set, the frequency domain feature set and the Poincare map related parameter set, which provide a more comprehensive data basis for the subsequent detection of heart health status results. Using the set heart health status detection model, the detection efficiency and accuracy can be improved.

[0044] In one embodiment of the present invention, a method for detecting heart health status is provided. Figure 1 As shown, the heart health status detection method of this embodiment includes the following steps S110 to S130.

[0045] Step S110: Determine a short-time feature set based on the heart sound signal to be measured, wherein the short-time feature set includes multiple first feature sets and multiple second feature sets, each first feature set is determined based on the signal amplitude of the heart sound signal in the corresponding window at each sampling point, and each second feature set is determined based on the duration of the first heart sound interval, the systolic period, the second heart sound interval and the diastolic period in each cardiac cycle in the heart sound signal in the corresponding window, and the heart sound signal in each window includes signals corresponding to a preset number of cardiac cycles.

[0046] The short-term feature set is determined based on the heart sound signals within each window, which are continuously segmented from the heart sound signal to be measured. The short-term features extracted from these short-term heart sound signals form a short-term feature set. These short-term features can sensitively capture transient events such as the S3 heart sound and premature beats, providing an effective data foundation for subsequent cardiac health status assessment.

[0047] The short-term feature set includes feature sets of multiple dimensions, namely the first feature set and the second feature set, which provide a more comprehensive data basis for the subsequent detection of heart health status results.

[0048] In some embodiments, before step S110 , the method further includes: continuously dividing the heart sound signal to be measured into heart sound signals within a plurality of windows using a preset sliding window.

[0049] The length of the preset sliding window is not fixed, and the number of cardiac cycles included in the heart sound signal in each window obtained by segmentation based on the preset sliding window is the same.

[0050] In some embodiments, step S110 specifically includes: based on the heart sound signal to be measured, determining the mean, median, standard deviation, skewness and kurtosis of the signal amplitude of the heart sound signal in the corresponding window at each sampling point to form multiple first feature sets.

[0051] Taking the heart sound signal within a window as an example, the mean value of the signal amplitude at each sampling point is determined based on the following calculation formula:

[0052]

[0053] Among them, μ is the mean value of the amplitude of the heart sound signal at each sampling point in the window, X i is the signal amplitude of the heart sound signal in the window at each sampling point, and N is the number of heart sound signals in the window at each sampling point.

[0054] Taking the heart sound signal within a window as an example, the median of the signal amplitude at each sampling point is determined based on the following calculation formula:

[0055]

[0056] Where m is the median value of the heart sound signal amplitude at each sampling point within the window, is the signal amplitude of the (N+1) / 2th sampling point after the signal amplitudes of the heart sound signal at each sampling point in the window are arranged in ascending or descending order. is the signal amplitude of the heart sound signal at the N / 2th sampling point after the signal amplitudes of the sampling points in the window are arranged in ascending or descending order, It is the signal amplitude of the (N / 2)+1th sampling point after the signal amplitudes of the heart sound signals at each sampling point in the window are arranged in ascending or descending order, and N is the number of heart sound signals at each sampling point in the window.

[0057] Taking the heart sound signal within a window as an example, the standard deviation of the signal amplitude at each sampling point is determined based on the following calculation formula:

[0058]

[0059] Among them, σ is the standard deviation of the signal amplitude of the heart sound signal at each sampling point in the window, X i is the amplitude of the heart sound signal at each sampling point within the window, μ is the mean amplitude of the heart sound signal at each sampling point within the window, and N is the number of heart sound signals at each sampling point within the window. μ can be calculated by referring to the above formula.

[0060] Taking the heart sound signal within a window as an example, the skewness of the signal amplitude at each sampling point is determined based on the following calculation formula:

[0061]

[0062] Among them, skew is the skewness of the signal amplitude of the heart sound signal in the window at each sampling point, X i is the amplitude of the heart sound signal at each sampling point within the window, μ is the mean amplitude of the heart sound signal at each sampling point within the window, and N is the number of heart sound signals at each sampling point within the window. μ can be calculated by referring to the above formula.

[0063] Taking the heart sound signal within a window as an example, the kurtosis of the signal amplitude at each sampling point is determined based on the following calculation formula:

[0064]

[0065] Among them, kurt is the peak value of the heart sound signal at each sampling point in the window, X i is the amplitude of the heart sound signal at each sampling point within the window, μ is the mean amplitude of the heart sound signal at each sampling point within the window, and N is the number of heart sound signals at each sampling point within the window. μ can be calculated by referring to the above formula.

[0066] In some embodiments, step S110 specifically includes: based on the heart sound signal to be measured, determining the average duration of the first heart sound interval, the average duration of the systolic interval, the average duration of the second heart sound interval, the average duration of the diastolic interval, the average duration of the cardiac cycle, the ratio of the average duration of the systolic interval to the average duration of the diastolic interval, the ratio of the average duration of the systolic interval to the average duration of the cardiac cycle, the average duration of the diastolic interval and the average duration of the cardiac cycle in the heart sound signal within the corresponding window to form multiple second feature sets.

[0067] Based on the heart sound signal in each window, the durations of the first heart sound interval, the systolic interval, the second heart sound interval, and the diastolic interval in each cardiac cycle are determined based on a hidden semi-Markov model.

[0068] according to Figure 2 As shown, the heart sound signal in each window includes signals corresponding to three cardiac cycles. Each cardiac cycle includes the first heart sound interval, the systolic period, the second heart sound interval and the diastolic period. Figure 2 As shown, the signal corresponding to the first heart sound interval is the heart sound signal in the blue area, the signal corresponding to the systolic period is the heart sound signal in the orange area, the signal corresponding to the second heart sound interval is the heart sound signal in the green area, and the signal corresponding to the diastolic period is the heart sound signal in the purple area. Figure 2 As shown, the red circle is the peak point of the first heart sound, and the yellow circle is the peak point of the second heart sound.

[0069] by Figure 2 Taking the heart sound signal within a window shown as an example, the duration of the first heart sound interval, the duration of the systolic interval, the duration of the second heart sound interval, and the duration of the diastolic interval in each cardiac cycle are determined based on a hidden semi-Markov model. For each cardiac cycle, the duration of the first heart sound interval, the duration of the systolic interval, the duration of the second heart sound interval, and the duration of the diastolic interval are added together to obtain the duration of the corresponding cardiac cycle. Then, based on the duration of each first heart sound interval, each systolic interval, each second heart sound interval, each diastolic interval, and each cardiac cycle duration, the feature values ​​in the second feature set are determined.

[0070] Step S120: Determine a long-term feature set based on the heart sound signal to be measured, wherein the long-term feature set includes a time domain feature set, a frequency domain feature set, and a Poincare map related parameter set. The time domain feature set is determined based on the duration of each cardiac cycle, the frequency domain feature set includes the signal power in the low-frequency range, the signal power in the high-frequency range, and the ratio of the signal power in the low-frequency range to the signal power in the high-frequency range, and the Poincare map related parameter set is determined based on the Poincare map generated based on the duration of each cardiac cycle.

[0071] The long-term feature set is determined based on the entire heart sound signal to be measured. The features in the long-term feature set can quantify the overall trend, for example, heart rate variability. Based on the overall trend, the accuracy of identifying chronic heart diseases can be improved.

[0072] The long-term feature set includes feature sets of multiple dimensions, namely the time domain feature set, frequency domain feature set and Poincare map related parameter set, providing a more comprehensive data basis for the subsequent detection of heart health status results.

[0073] In some embodiments, step S120 specifically includes: based on the heart sound signal to be measured, determining the mean of the cardiac cycle duration, the standard deviation of the cardiac cycle duration, the root mean square of the difference between the durations of two adjacent cardiac cycles, and the ratio of the number of cardiac cycles whose difference between the durations of two adjacent cardiac cycles is greater than a preset threshold to the number of all cardiac cycles to form a time domain feature set.

[0074] The mean value of the duration of the cardiac cycle in the heart sound signal to be measured is determined based on the following calculation formula:

[0075]

[0076] Among them, mRR is the mean duration of the cardiac cycle in the heart sound signal to be measured, RR i is the duration of the i-th cardiac cycle in the heart sound signal to be measured, and M is the number of cardiac cycles in the heart sound signal to be measured.

[0077] The standard deviation of the duration of the cardiac cycle in the heart sound signal to be measured is determined based on the following calculation formula:

[0078]

[0079] Among them, SDNN is the standard deviation of the duration of the heartbeat cycle in the heart sound signal to be measured, mRR is the mean duration of the heartbeat cycle in the heart sound signal to be measured, and RR i is the duration of the i-th cardiac cycle in the heart sound signal to be measured, and M is the number of cardiac cycles in the heart sound signal to be measured. The calculation method of the mean mRR of the duration of each cardiac cycle in the heart sound signal to be measured can refer to the above calculation formula.

[0080] The root mean square of the difference between the durations of two adjacent cardiac cycles is determined based on the following formula:

[0081]

[0082] Among them, RMSSD is the root mean square of the difference between the durations of two adjacent cardiac cycles in the heart sound signal to be measured, RR i is the duration of the i-th cardiac cycle in the heart sound signal to be measured, RR i+1is the duration of the i+1th cardiac cycle in the heart sound signal to be measured, and M is the number of cardiac cycles in the heart sound signal to be measured.

[0083] The ratio of the number of cardiac cycles whose duration difference between two adjacent cardiac cycles is greater than a preset threshold to the number of all cardiac cycles is determined based on the following calculation formula:

[0084]

[0085] Where pNN is the ratio of the number of cardiac cycles whose duration difference between two adjacent cardiac cycles is greater than the preset threshold to the number of all cardiac cycles, RR i is the duration of the i-th cardiac cycle in the heart sound signal to be measured, RR i+1 is the duration of the i+1th cardiac cycle in the heart sound signal to be measured, M is the number of cardiac cycles in the heart sound signal to be measured, and α is the preset threshold. The preset threshold can be set according to needs.

[0086] In some embodiments, step S120 specifically includes: performing fast Fourier transform on the heart sound signal to be measured to obtain a power spectral density function; obtaining a preset frequency resolution, a preset low-frequency range, and a preset high-frequency range; determining the signal power of the low-frequency range according to the power spectral density function, the preset frequency resolution, and the preset low-frequency range, determining the signal power of the high-frequency range according to the power spectral density function, the preset frequency resolution, and the preset high-frequency range, and determining the ratio of the signal power of the low-frequency range to the signal power of the high-frequency range according to the signal power of the low-frequency range and the signal power of the high-frequency range to form a frequency domain feature set.

[0087] The signal power in the low-frequency range is determined based on the following calculation formula:

[0088]

[0089] Wherein, LF is the signal power in the low-frequency range, P(f) is the power spectrum density function obtained by fast Fourier transforming the heart sound signal to be measured, Δf is the preset frequency resolution, and [f1, f2] is the preset low-frequency range.

[0090] The signal power in the high frequency range is determined based on the following calculation formula:

[0091]

[0092] Wherein, HF is the signal power in the high frequency range, P(f) is the power spectrum density function obtained by fast Fourier transforming the heart sound signal to be measured, Δf is the preset frequency resolution, and [f3, f4] is the preset high frequency range.

[0093] Based on the following calculation formula, the ratio of the signal power in the low frequency range to the signal power in the high frequency range is determined.

[0094]

[0095] Here, β is the ratio of the signal power in the low-frequency range to the signal power in the high-frequency range.

[0096] The preset low frequency range [f1, f2] and the preset high frequency range [f3, f4] can be set according to needs. For example, the preset low frequency range is [0.04Hz, 0.15Hz] and the preset high frequency range is [0.15Hz, 0.4Hz].

[0097] In some embodiments, step S120 specifically includes: generating a Poincaré map based on the duration of each cardiac cycle in the heart sound signal to be measured; determining the short axis length of the Poincaré map, the long axis length of the Poincaré map, and the ratio of the short axis length of the Poincaré map to the long axis length of the Poincaré map as parameters related to the Poincaré map.

[0098] For the heart sound signal to be measured, the duration of each cardiac cycle is RR1, RR2, ..., RR M Based on the duration of each cardiac cycle, a two-dimensional scatter plot is drawn, where the horizontal coordinate of each point in the plot is the duration of the current cardiac cycle, and the vertical coordinate is the duration of the next cardiac cycle. This two-dimensional scatter plot is a Poincare map.

[0099] Determine the minor axis length of the Poincaré map based on the following calculation formula:

[0100]

[0101] Among them, sd1 is the short axis length of the Poincare map, VAR() is the variance function, RR i is the duration of the i-th cardiac cycle in the heart sound signal to be measured, RR i+1 The input value of the VAR() function is the duration of each cardiac cycle, i.e., RR1, RR2, ..., RR M .

[0102] Determine the length of the major axis of the Poincaré map based on the following calculation formula:

[0103]

[0104] Among them, sd2 is the length of the major axis of the Poincaré map, VAR() is the variance function, RR i is the duration of the i-th cardiac cycle in the heart sound signal to be measured, RR i+1 The input value of the VAR() function is the duration of each cardiac cycle, i.e., RR1, RR2, ..., RR M .

[0105] Based on the following calculation formula, determine the ratio of the minor axis length of the Poincaré map to the major axis length of the Poincaré map,

[0106]

[0107] Here, γ is the ratio of the length of the minor axis of the Poincaré map to the length of the major axis of the Poincaré map.

[0108] In step S130, the first feature set, the second feature set, the time domain feature set, the frequency domain feature set, and the Poincare map-related parameter set are input into the set heart health status detection model to obtain a heart health status result. The heart health status result is one of abnormality and no abnormality.

[0109] In some embodiments, the heart health status detection model is determined based on training with a training sample set and validation with a test sample set. Each sample in the training sample set and the test sample set includes a short-term feature set and a long-term feature set determined based on different heart sound signals, and a heart health status result corresponding to each heart sound signal.

[0110] The heart health status detection model to be trained is a machine learning network model, such as a support vector machine. First, the heart health status detection model to be trained is trained using a training sample set, so that the trained heart health status detection model learns the correlation between the short-term feature set, the long-term feature set, and the corresponding heart health status results in each training sample. Then, the trained heart health status detection model is verified using a test sample set. If the output of the heart health status detection model meets preset requirements, training of the heart health status detection model is stopped, and the trained heart health status detection model is used as the set heart health status detection model.

[0111] In some embodiments, before determining the short-term feature set based on the heart sound signal to be measured, the method also includes: determining the duration of each cardiac cycle based on the heart sound signal to be measured; screening out abnormal cardiac cycles based on the duration of each cardiac cycle, wherein the duration of the abnormal cardiac cycle is greater than a preset duration; eliminating the signal corresponding to the abnormal cardiac cycle in the heart sound signal to be measured to obtain a processed heart sound signal to be measured.

[0112] Based on each cardiac cycle, the duration of the first heart sound interval, the duration of the systolic interval, the duration of the second heart sound interval, and the duration of the diastolic interval are added together to obtain the duration of the corresponding cardiac cycle.

[0113] The preset duration is determined based on a normal range of cardiac cycle duration values.

[0114] In this embodiment, abnormal cardiac cycles in the heart sound signal to be measured are deleted to eliminate the interference of abnormal cardiac cycles on the entire heart sound signal to be measured, so that the short-term feature set and long-term feature set determined based on the processed heart sound signal to be measured are more accurate.

[0115] In some embodiments, before determining the short-term feature set based on the heart sound signal to be measured, the method further includes normalizing and filtering the heart sound signal to be measured to obtain a processed heart sound signal to be measured. Specifically, a Butterworth filter can be used to filter out noise and artifacts that affect signal quality, thereby improving the overall signal quality. Butterworth filters include, but are not limited to, high-pass filters, low-pass filters, and band-pass filters.

[0116] One embodiment of the present invention provides a device for detecting heart health status. Figure 3 As shown, the heart health status detection device includes a short-term feature set determination module 310 , a long-term feature set determination module 320 and a detection module 330 .

[0117] The short-time feature set determination module 310 is used to determine the short-time feature set based on the heart sound signal to be measured, wherein the short-time feature set includes multiple first feature sets and multiple second feature sets, each first feature set is determined based on the signal amplitude of the heart sound signal in the corresponding window at each sampling point, and each second feature set is determined based on the duration of the first heart sound interval, the systolic period, the second heart sound interval and the diastolic period included in each cardiac cycle in the heart sound signal in the corresponding window, and the heart sound signal in each window includes signals corresponding to a preset number of cardiac cycles.

[0118] The long-term feature set determination module 320 is used to determine the long-term feature set based on the heart sound signal to be measured, wherein the long-term feature set includes a time domain feature set, a frequency domain feature set and a Poincare map related parameter set, wherein the time domain feature set is determined based on the duration of each cardiac cycle, the frequency domain feature set includes the signal power in the low-frequency range, the signal power in the high-frequency range, and the ratio of the signal power in the low-frequency range to the signal power in the high-frequency range, and the Poincare map related parameter set is determined based on the Poincare map generated based on the duration of each cardiac cycle.

[0119] The detection module 330 is used to input the first feature set, the second feature set, the time domain feature set, the frequency domain feature set and the Poincare map related parameter set into the set heart health status detection model to obtain a heart health status result.

[0120] In some embodiments, the short-time feature set determination module 310 is used to determine the mean, median, standard deviation, skewness and kurtosis of the signal amplitude of the heart sound signal in the corresponding window at each sampling point based on the heart sound signal to be measured, so as to form multiple first feature sets.

[0121] In some embodiments, the short-time feature set determination module 310 is used to determine, based on the heart sound signal to be measured, the average duration of the first heart sound interval, the average duration of the systolic interval, the average duration of the second heart sound interval, the average duration of the diastolic interval, the average duration of the cardiac cycle, the ratio of the average duration of the systolic interval to the average duration of the diastolic interval, the ratio of the average duration of the systolic interval to the average duration of the cardiac cycle, the average duration of the diastolic interval and the average duration of the cardiac cycle in the heart sound signal within the corresponding window to form multiple second feature sets.

[0122] In some embodiments, the long-term feature set determination module 320 is used to determine the mean of the cardiac cycle duration, the standard deviation of the cardiac cycle duration, the root mean square of the difference between the durations of two adjacent cardiac cycles, and the ratio of the number of cardiac cycles whose difference between the durations of two adjacent cardiac cycles is greater than a preset threshold to the number of all cardiac cycles based on the heart sound signal to form a time domain feature set.

[0123] In some embodiments, the long-term feature set determination module 320 is used to obtain a power spectral density function by fast Fourier transforming the heart sound signal to be measured; obtain a preset frequency resolution, a preset low-frequency range, and a preset high-frequency range; determine the signal power of the low-frequency range according to the power spectral density function, the preset frequency resolution, and the preset low-frequency range, determine the signal power of the high-frequency range according to the power spectral density function, the preset frequency resolution, and the preset high-frequency range, and determine the ratio of the signal power of the low-frequency range to the signal power of the high-frequency range according to the signal power of the low-frequency range and the signal power of the high-frequency range, so as to form a frequency domain feature set.

[0124] In some embodiments, the long-term feature set determination module 320 is used to generate a Poincaré map based on the duration of each cardiac cycle in the heart sound signal to be measured; determine the short axis length of the Poincaré map, the long axis length of the Poincaré map, and the ratio of the short axis length of the Poincaré map to the long axis length of the Poincaré map as parameters related to the Poincaré map.

[0125] In some embodiments, the device further includes an abnormal cardiac cycle rejection module. The abnormal cardiac cycle rejection module is configured to determine the duration of each cardiac cycle based on the heart sound signal to be measured; based on the duration of each cardiac cycle, screen out cardiac cycles with a duration greater than a preset duration; and reject signals corresponding to cardiac cycles with a duration greater than the preset duration from the heart sound signal to be measured, thereby obtaining a processed heart sound signal to be measured.

[0126] One embodiment of the present invention provides a device for detecting heart health status. Figure 4As shown, the heart health status detection device includes a memory 420 and a processor 410. The memory 420 stores a computer program, which is used to control the processor 410 to operate so as to execute the heart health status detection method provided according to any of the above embodiments.

[0127] The processor 410 is configured to execute computer instructions, which may be written using an instruction set of an architecture such as x86, Arm, RISC, MIPS, or SSE. The memory 420 may include, for example, ROM (read-only memory), RAM (random access memory), or a non-volatile memory such as a hard disk, although these are not limited herein.

[0128] An embodiment of the present invention provides an electronic device. Figure 5 As shown, the electronic device includes the heart health status detection device provided by any of the above embodiments and a sensor for collecting heart sound signals.

[0129] according to Figure 5 As shown, the sensor for collecting heart sound signals is connected to the heart health status detection device.

[0130] The electronic device can be any of the following: a watch, a bracelet, or a ring.

[0131] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0132] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] The embodiments of this specification may be systems, methods, and / or computer program products. The computer program product may include a computer-readable storage medium carrying computer instructions for causing a processor to implement various aspects of the embodiments of this specification.

[0134] A computer-readable storage medium can be a tangible device that can retain and store computer instructions for use by a computer instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punched card or raised-in-groove structure on which computer instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber-optic cable), or an electrical signal transmitted through an electrical wire.

[0135] The computer instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer instructions from the network and forwards the computer instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0136] The flowcharts and block diagrams in the accompanying drawings show the possible implementation architectures, functions and operations of the systems, methods and computer program products according to multiple embodiments of this specification. In this regard, each box in the flowchart or block diagram can represent a module, program segment or part of a computer instruction, and the module, program segment or part of a computer instruction contains one or more executable computer instructions for implementing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or can be implemented using a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, implementation by software, and implementation by a combination of software and hardware are all equivalent.

[0137] The embodiments of the present specification have been described above. The above description is illustrative and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to existing technologies, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A heart health status detection device, characterized in that: include: a short-term feature set determination module, configured to determine a short-term feature set based on the heart sound signal to be measured, wherein the short-term feature set includes a plurality of first feature sets and a plurality of second feature sets, each first feature set is determined based on the signal amplitude of the heart sound signal in a corresponding window at each sampling point, and each second feature set is determined based on the duration of a first heart sound interval, a systolic interval, a second heart sound interval, and a diastolic interval in each cardiac cycle in the heart sound signal in the corresponding window, and the heart sound signal in each window includes signals corresponding to a preset number of cardiac cycles; a long-term feature set determination module, configured to determine a long-term feature set based on the heart sound signal to be measured, wherein the long-term feature set includes a time-domain feature set, a frequency-domain feature set, and a Poincare map-related parameter set, wherein the time-domain feature set is determined based on the duration of each cardiac cycle, the frequency-domain feature set includes the signal power in the low-frequency range, the signal power in the high-frequency range, and the ratio of the signal power in the low-frequency range to the signal power in the high-frequency range, and the Poincare map-related parameter set is determined based on the Poincare map generated based on the duration of each cardiac cycle; The detection module is used to input the first feature set, the second feature set, the time domain feature set, the frequency domain feature set and the Poincare map related parameter set into the set heart health status detection model to obtain a heart health status result.

2. The device according to claim 1, characterized in that The short-time feature set determination module is used to determine the mean, median, standard deviation, skewness and kurtosis of the signal amplitude of the heart sound signal in the corresponding window at each sampling point based on the heart sound signal to be measured, so as to form multiple first feature sets.

3. The device according to claim 1, characterized in that The short-time feature set determination module is used to determine the average duration of the first heart sound interval, the average duration of the systolic interval, the average duration of the second heart sound interval, the average duration of the diastolic interval, the average duration of the cardiac cycle, the ratio of the average duration of the systolic interval to the average duration of the diastolic interval, the ratio of the average duration of the systolic interval to the average duration of the cardiac cycle, and the ratio of the average duration of the diastolic interval to the average duration of the cardiac cycle in the heart sound signal within the corresponding window based on the heart sound signal to be measured, so as to form multiple second feature sets.

4. The device according to claim 1, characterized in that The long-term feature set determination module is used to determine the mean of the cardiac cycle duration, the standard deviation of the cardiac cycle duration, the root mean square of the difference between the durations of two adjacent cardiac cycles, and the ratio of the number of cardiac cycles with a difference between the durations of two adjacent cardiac cycles greater than a preset threshold to the number of all cardiac cycles based on the heart sound signal to be measured, so as to form a time domain feature set.

5. The device according to claim 1, characterized in that The long-term feature set determination module is used to obtain a power spectral density function by fast Fourier transforming the heart sound signal to be measured; obtain a preset frequency resolution, a preset low-frequency range, and a preset high-frequency range; determine the signal power of the low-frequency range according to the power spectral density function, the preset frequency resolution, and the preset low-frequency range, determine the signal power of the high-frequency range according to the power spectral density function, the preset frequency resolution, and the preset high-frequency range, and determine the ratio of the signal power of the low-frequency range to the signal power of the high-frequency range according to the signal power of the low-frequency range and the signal power of the high-frequency range, so as to form a frequency domain feature set.

6. The device according to claim 1, characterized in that The long-term feature set determination module is used to generate a Poincaré map based on the duration of each cardiac cycle in the heart sound signal to be measured; determine the short axis length of the Poincaré map, the long axis length of the Poincaré map, and the ratio of the short axis length of the Poincaré map to the long axis length of the Poincaré map as parameters related to the Poincaré map.

7. The device according to any one of claims 1 to 6, characterized in that: The device also includes an abnormal cardiac cycle elimination module, wherein: The abnormal cardiac cycle elimination module is used to determine the duration of each cardiac cycle based on the heart sound signal to be measured; based on the duration of each cardiac cycle, screen out cardiac cycles with a duration greater than a preset duration; eliminate the signals corresponding to the cardiac cycles with a duration greater than the preset duration in the heart sound signal to be measured, and obtain the processed heart sound signal to be measured.

8. A heart health status detection device, characterized in that: The invention comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is used to control the processor to operate to perform a heart health status detection method, wherein: The heart health status detection method includes: determining a short-term feature set based on a heart sound signal to be detected, wherein the short-term feature set includes a plurality of first feature sets and a plurality of second feature sets, each first feature set is determined based on the signal amplitude of the heart sound signal in a corresponding window at each sampling point, and each second feature set is determined based on the duration of a first heart sound interval, a systolic period, a second heart sound interval, and a diastolic period included in each cardiac cycle in the heart sound signal in the corresponding window, and the heart sound signal in each window includes signals corresponding to a preset number of cardiac cycles; Determining a long-term feature set based on the heart sound signal to be measured, wherein the long-term feature set includes a time domain feature set, a frequency domain feature set, and a Poincare map-related parameter set, wherein the time domain feature set is determined based on the duration of each cardiac cycle, the frequency domain feature set includes signal power in a low-frequency range, signal power in a high-frequency range, and a ratio of signal power in the low-frequency range to signal power in the high-frequency range, and the Poincare map-related parameter set is determined based on a Poincare map generated based on the duration of each cardiac cycle; The first feature set, the second feature set, the time domain feature set, the frequency domain feature set and the Poincare map related parameter set are input into a set heart health status detection model to obtain a heart health status result.

9. An electronic device, characterized in that: include: The heart health status detection device and the sensor for collecting heart sound signals according to claim 7 or 8, wherein: The sensor for collecting heart sound signals is connected to the heart health status detection device.

Images