Atrial fibrillation detection system, atrial fibrillation detection method and processor

Abstract

The invention discloses an atrial fibrillation detection system, an atrial fibrillation detection method and a processor. In the scheme, a signal acquisition module acquires a pulse wave signal of a target object within a preset duration and generates a corresponding pulse oscillogram; the first data processing module performs time domain feature analysis on the pulse oscillogram to obtain a first time domain feature value and a second time domain feature value, and determines a multi-dimensional feature joint value based on the first time domain feature value and the second time domain feature value; a second data processing module performs data distribution feature analysis on the pulse oscillogram to obtain a target feature value when determining that the target object has an arrhythmia condition based on the multi-dimensional feature joint value; and the recognition module recognizes the arrhythmia type of the target object according to the target feature value to obtain a recognition result. By optimizing the pulse wave analysis algorithm of the atrial fibrillation, accurate recognition of the atrial fibrillation can be completed through single pulse wave signal measurement, and the auxiliary value of the recognition result on medication and treatment of medical workers is effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of medical devices, and particularly to an atrial fibrillation detection system, an atrial fibrillation detection method, and a processor. Background Art

[0002] Atrial fibrillation (AF) is also known as atrial flutter. It is caused by various pathological reasons resulting in irregular atrial excitation, with an irregular rhythm and the atria losing effective contraction function, belonging to a type of arrhythmia.

[0003] Traditional atrial fibrillation detection methods usually rely on the pulse wave signal of the brachial artery. By collecting multiple groups of pulse wave signals during the pressurization and decompression processes of the cuff, and calculating the average pulse wave interval of multiple groups of pulse wave signals to evaluate the AF risk. However, the above methods only rely on a single - dimension index (such as the average pulse wave interval) to evaluate the AF risk, and it is difficult to effectively distinguish AF from similar pulse waveforms generated by arrhythmias such as ventricular / atrial premature contractions, resulting in a relatively high misdiagnosis rate of AF, thus unable to effectively assist doctors in medication and treatment, and to a certain extent, affecting the rapid recovery of individuals.

[0004] In addition, in order to ensure the accuracy of the recognition results, traditional detection methods usually need to repeatedly measure multiple groups of pulse wave signals, resulting in a cumbersome and time - consuming detection process, seriously affecting the user experience. Summary of the Invention

[0005] Based on the above problems, the present application provides an atrial fibrillation detection system, an atrial fibrillation detection method, and a processor, aiming to achieve accurate identification of atrial fibrillation with a single measurement of the pulse wave signal, and effectively improve the auxiliary value of the recognition results for medical workers in medication and treatment.

[0006] The embodiments of the present application disclose the following technical solutions:

[0007] In the first aspect of the present application, an atrial fibrillation detection system is provided. The atrial fibrillation detection system includes a signal acquisition module, a first data processing module, a second data processing module, and an identification module;

[0008] The signal acquisition module is configured to collect the pulse wave signal of the target object within a preset time period and generate a corresponding pulse waveform diagram;

[0009] The first data processing module is configured to perform time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain feature value and a second time-domain feature value, and determine a multi-dimensional feature joint value based on the first time-domain feature value and the second time-domain feature value; the first time-domain feature value is used to measure the degree of abnormality of the heart rate of the target object, and the second time-domain feature value is used to measure the degree of abnormality of the heart beat amplitude of the target object;

[0010] The second data processing module is configured to perform data distribution feature analysis on the pulse waveform diagram to obtain a target feature value when it is determined based on the multi-dimensional feature joint value that the target object has an arrhythmia condition; the target feature value is used to evaluate whether the type of arrhythmia of the target object is atrial fibrillation;

[0011] The recognition module is configured to recognize the type of arrhythmia of the target object according to the target feature value to obtain a recognition result.

[0012] In an alternative implementation, the first data processing module is configured to:

[0013] Extract the peak extreme points of each pulse wave from the pulse waveform diagram;

[0014] Based on each of the peak extreme points, determine a plurality of pulse intervals corresponding to the pulse waveform diagram; the pulse interval is used to represent the time interval corresponding to two consecutive pulse beats of the target object;

[0015] Determine the first time-domain feature value based on the plurality of pulse intervals; there is a positive correlation between the first time-domain feature value and the probability that the target object has an arrhythmia condition.

[0016] In an alternative implementation, the first data processing module is further configured to:

[0017] Determine the pulse peak value of each pulse wave based on the peak extreme points of each pulse wave, and construct a pulse peak value sequence based on the pulse peak values of each pulse wave;

[0018] According to the first value and the maximum value in the pulse peak value sequence, perform linear interpolation on the first half of the waveform in the pulse waveform diagram to obtain a first interpolation sequence; the starting point of the first half of the waveform is the peak extreme point corresponding to the first value, and the ending point of the first half of the waveform is the peak extreme point corresponding to the maximum value; the first interpolation sequence includes the interpolation values corresponding to each of the peak extreme points in the first half of the waveform;

[0019] Perform the linear interpolation on the latter half of the waveform in the pulse waveform diagram according to the maximum value and the tail value in the pulse peak sequence, to obtain a second interpolation sequence; the starting point of the latter half of the waveform is the peak extreme point corresponding to the maximum value, and the ending point of the former half of the waveform is the peak extreme point corresponding to the tail value; the second interpolation sequence includes the interpolation values corresponding to the respective peak extreme points in the latter half of the waveform.

[0020] Calculate the difference between the interpolation value corresponding to each peak extreme point in the former half of the waveform and the pulse peak corresponding to this peak extreme point, to obtain a first difference vector.

[0021] Calculate the difference between the interpolation value corresponding to each peak extreme point in the latter half of the waveform and the pulse peak corresponding to this peak extreme point, to obtain a second difference vector.

[0022] Merge the first difference vector and the second difference vector, and determine the second time domain eigenvalue based on the merged difference vector; there is a positive correlation between the second time domain eigenvalue and the probability that the target object has the arrhythmia condition.

[0023] In an alternative implementation, the second data processing module is configured to:

[0024] Judge whether the multi-dimensional feature combined value is greater than a first preset threshold.

[0025] If the multi-dimensional feature combined value is greater than the first preset threshold, determine that the target object has the arrhythmia condition.

[0026] If the multi-dimensional feature combined value is less than or equal to the first preset threshold, determine that the target object does not have the arrhythmia condition.

[0027] In an alternative implementation, the second data processing module is further configured to:

[0028] Obtain a plurality of pulse intervals corresponding to the pulse waveform diagram.

[0029] Calculate the average value of the plurality of pulse intervals, and determine the histogram interval boundaries according to the average value.

[0030] Divide the plurality of pulse intervals into a plurality of intervals according to the histogram interval boundaries.

[0031] Count the number of data in each interval, and construct a statistical frequency vector based on the number of data.

[0032] Determine the target eigenvalue based on the statistical frequency vector and the total amount of data of the pulse intervals.

[0033] In an alternative implementation, the recognition module is further configured to:

[0034] Determine whether the target feature value is greater than a second preset threshold;

[0035] If the target feature value is greater than the second preset threshold, determine the recognition result as the first recognition result; the first recognition result is used to characterize that the arrhythmia type of the target object is atrial fibrillation;

[0036] If the target feature value is less than or equal to the second preset threshold, determine the recognition result as the second recognition result; the second recognition result is used to characterize that the arrhythmia type of the target object is non - atrial fibrillation.

[0037] In an alternative implementation, the signal acquisition module is configured to acquire the pulse wave signal by the oscillometric method or the photoplethysmography method.

[0038] In a second aspect of the present application, there is provided a method for detecting atrial fibrillation, which is characterized in that it is applied to the atrial fibrillation detection system, and the method includes:

[0039] Acquire the pulse wave signal of a target object within a preset time period and generate a corresponding pulse waveform diagram;

[0040] Perform time - domain feature analysis on the pulse waveform diagram to obtain a first time - domain feature value and a second time - domain feature value, and determine a multi - dimensional feature joint value based on the first time - domain feature value and the second time - domain feature value; the first time - domain feature value is used to measure the abnormal degree of the heart rate of the target object, and the second time - domain feature value is used to measure the abnormal degree of the heart beat amplitude of the target object;

[0041] When it is determined based on the multi - dimensional feature joint value that the target object has an arrhythmia condition, perform data distribution feature analysis on the pulse waveform diagram to obtain a target feature value; the target feature value is used to evaluate whether the arrhythmia type of the target object is atrial fibrillation;

[0042] Identify the arrhythmia type of the target object according to the target feature value to obtain a recognition result.

[0043] In a third aspect of the present application, there is provided a computer - readable storage medium, in which a computer program is stored, and when the computer program is run by a processor, the above - mentioned method for detecting atrial fibrillation is implemented.

[0044] In a fourth aspect of the present application, there is provided a processor for running a computer program, and when the computer program runs, it executes the above - mentioned method for detecting atrial fibrillation.

[0045] Compared with the prior art, the present application has the following beneficial effects:

[0046] In the technical solution of the present application, first, a signal acquisition module collects the pulse wave signal of the target object within a preset time period and generates a corresponding pulse waveform diagram; subsequently, a first data processing module performs time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain feature value and a second time-domain feature value. Since the first time-domain feature value can be used to measure the abnormality degree of the target object's heart rate, and the second time-domain feature value can be used to measure the abnormality degree of the target object's heart beat amplitude, the multi-dimensional feature joint value obtained based on the above two feature values can comprehensively evaluate whether the target object has arrhythmia in terms of the two-dimensional indexes of heart rate and heart beat amplitude, effectively avoiding the misjudgment problem caused by relying solely on a single-dimensional index for atrial fibrillation assessment;

[0047] Then, on the premise of determining that the target object has arrhythmia, a second data processing module further performs data distribution feature analysis on the pulse waveform diagram to obtain a target feature value for distinguishing atrial fibrillation from other similar arrhythmias. Through this feature value, it can effectively reflect whether the arrhythmia type of the target object is atrial fibrillation; finally, an identification module accurately identifies the arrhythmia type of the target object according to the target feature value to obtain an identification result, realizing the accurate identification of atrial fibrillation through a single pulse wave signal measurement, thereby improving the atrial fibrillation detection efficiency of the system, and further providing a fast and reliable diagnosis basis for medical workers, reducing the risk of over-treatment or delayed treatment caused by missed diagnosis or misdiagnosis, especially suitable for primary medical institutions, home self-testing and dynamic health monitoring scenarios, and improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0049] Figure 1 It is a schematic structural diagram of an atrial fibrillation detection system provided by an embodiment of the present application;

[0050] Figure 2 It is a schematic diagram of a pulse waveform diagram provided by an embodiment of the present application;

[0051] Figure 3 It is a flowchart of an atrial fibrillation detection method provided by an embodiment of the present application;

[0052] Figure 4Flow chart of a process for determining a first time-domain eigenvalue provided by an embodiment of the present application;

[0053] Figure 5 Flow chart of a process for determining a second time-domain eigenvalue provided by an embodiment of the present application;

[0054] Figure 6 Flow chart of a process for determining a histogram distribution eigenvalue provided by an embodiment of the present application. Detailed implementation manners

[0055] As described above, traditional atrial fibrillation detection methods usually rely on the pulse wave signal of the arm artery. Multiple groups of pulse wave signals are collected during the pressurization and decompression processes of the cuff, and the average pulse wave interval of multiple groups of pulse wave signals is calculated to evaluate the atrial fibrillation risk. However, the above method only relies on a single-dimensional index (such as the average pulse wave interval) to evaluate the atrial fibrillation risk, and it is difficult to effectively distinguish atrial fibrillation from similar pulse waveforms generated by arrhythmias such as ventricular / atrial premature contractions, resulting in a relatively high misdiagnosis rate of atrial fibrillation, thus unable to effectively assist doctors in medication and treatment, and to a certain extent, affecting the rapid recovery of individuals. In addition, in order to ensure the accuracy of the recognition result, traditional detection methods usually need to repeatedly measure multiple groups of pulse wave signals, resulting in a cumbersome and time-consuming detection process, seriously affecting the user experience.

[0056] After research, the inventors proposed an atrial fibrillation detection system, an atrial fibrillation detection method, and a processor. In this solution, first, a signal acquisition module collects the pulse wave signal of the target object within a preset time period and generates a corresponding pulse waveform diagram; subsequently, a first data processing module performs time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain eigenvalue and a second time-domain eigenvalue. Since the first time-domain eigenvalue can be used to measure the abnormal degree of the heart rate of the target object, and the second time-domain eigenvalue can be used to measure the abnormal degree of the heart rate amplitude of the target object, the multi-dimensional feature combined value obtained based on the above two eigenvalues can comprehensively evaluate whether the target object has arrhythmia in the two-dimensional indexes of heart rate and heart rate amplitude, effectively avoiding the misjudgment problem caused by relying only on a single-dimensional index for atrial fibrillation evaluation;

[0057] Then, on the premise of determining that the target object has an arrhythmia condition, the second data processing module further analyzes the data distribution characteristics of the pulse waveform diagram to obtain target characteristic values for distinguishing atrial fibrillation from other similar arrhythmias. Whether the arrhythmia type of the target object is atrial fibrillation can be effectively reflected by this characteristic value. Finally, the recognition module accurately identifies the arrhythmia type of the target object according to the target characteristic value to obtain a recognition result, realizing the accurate recognition of atrial fibrillation through a single pulse wave signal measurement, thereby improving the atrial fibrillation detection efficiency of the system. Furthermore, it provides a rapid and reliable diagnostic basis for medical workers, reducing the risks of over-treatment or delayed treatment caused by missed diagnosis or misdiagnosis. It is especially suitable for primary medical institutions, home self-testing, and dynamic health monitoring scenarios, improving the user experience.

[0058] Keyword Definition:

[0059] Pulse Wave: It refers to the pressure wave generated after injecting blood into the aorta during each cardiac contraction. This pressure wave propagates along the arterial wall. The pulse wave not only reflects the pumping function of the heart but is also closely related to the elasticity of blood vessels and blood pressure levels. Its waveform, amplitude, and propagation speed are often used to evaluate cardiovascular health status.

[0060] Pulse-to-Pulse Interval: Abbreviated as PPI, it refers to the time interval between two consecutive pulse waves, usually in milliseconds (ms). It should be noted that the unit of PPI calculated in this application is sampling points and has not been converted to time.

[0061] Pulse Peak: Refers to Pulse peak, abbreviated as PP. Usually, it refers to the measured value at the highest point of each pulse waveform when detecting or recording the pulse wave. The pulse peak reflects the maximum intensity of the pressure wave generated by each cardiac contraction (ventricular ejection) on the arterial system.

[0062] To enable those skilled in the art to better understand the solution of this application, the following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0063] System Embodiment

[0064] An embodiment of this application provides an atrial fibrillation detection system, where Figure 1 is a schematic structural diagram of an atrial fibrillation detection system provided by an embodiment of this application, as Figure 1As shown, the system includes: a signal acquisition module 11, a first data processing module 12, a second data processing module 13, and an identification module 14. From Figure 1 the connection relationships between several modules can be seen.

[0065] Among them, the signal acquisition module 11 is used to acquire the pulse wave signal of the target object within a preset duration and generate a corresponding pulse waveform diagram.

[0066] In the embodiment of the present application, the signal acquisition module can acquire the pulse wave signal of the target object within a preset duration by the oscillometric method or the photoplethysmography method. Among them, when acquiring the pulse wave signal by the oscillometric method, the signal acquisition module may include a cuff, a pipeline, a tracheal joint, a gas pump, a pressure sensor, and the cuff includes an airbag and a sleeve body that wraps the airbag; the signal acquisition module applies a controllable pressure to the detection part of the target object through the inflation and expansion of the cuff, and acquires the pulse wave signal of the target object in combination with the pressure sensor during the pressurization or decompression process of the airbag. It should be noted that the pulse wave signal of the brachial artery of the target object is acquired by the oscillometric method. Since the brachial artery is closer to the heart, the oscillometric method can be used in scenarios with higher accuracy requirements for measurement and the identification scenario of persistent atrial fibrillation.

[0067] Optionally, when acquiring the pulse wave signal by the photoplethysmography method, the signal acquisition module may include a photoelectric sensor. The signal acquisition module can acquire the pulse wave signal of the wrist or finger of the target object through the photoelectric sensor. Since the wrist and finger are far from the heart, the blood vessels are branched and thin, and the photoplethysmography method is easily affected by factors such as ambient light and individual differences, and measurement errors are likely to occur. Therefore, the photoplethysmography method can be used in scenarios with lower accuracy requirements for measurement, such as ring and watch products. In addition, the photoplethysmography method can be used to acquire the pulse fluctuation situation within a period of time, so the photoplethysmography method has a higher accuracy rate for the identification of paroxysmal atrial fibrillation.

[0068] In the embodiment of the present application, after the signal acquisition module 11 acquires the pulse wave signal of the target object within a preset duration, the pulse wave signal can be subjected to segment segmentation processing, and the processed pulse wave signal can be denoised by low-pass filtering, filtering out the data fluctuations caused by the sensor accuracy, and generating Figure 2 the shown pulse waveform diagram, where Figure 2 the x-axis in represents time, and the y-axis represents the amplitude of the pulse wave (for example, the actual pressure reading in the cuff or the change amount of the light absorption amount used to indirectly reflect the blood flow situation).

[0069] The first data processing module 12 is configured to perform time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain feature value and a second time-domain feature value, and determine a multi-dimensional feature joint value based on the first time-domain feature value and the second time-domain feature value; the first time-domain feature value is used to measure the degree of abnormality of the heart rate of the target object, and the second time-domain feature value is used to measure the degree of abnormality of the heart beat amplitude of the target object.

[0070] To solve the problem of high misdiagnosis rate of atrial fibrillation caused by the atrial fibrillation risk assessment relying only on a single-dimensional index in traditional atrial fibrillation detection methods, in the embodiments of the present application, the atrial fibrillation detection system performs time-domain feature analysis on the pulse waveform diagram through the first data processing module 12 to obtain a first time-domain feature value for measuring the degree of abnormality of the heart rate of the target object, and a second time-domain feature value for measuring the degree of abnormality of the heart beat amplitude of the target object, and obtains a multi-dimensional feature joint value based on the above two feature values, realizing a comprehensive assessment of whether the target object has arrhythmia in the two-dimensional indexes of heart rate and heart beat amplitude, thereby effectively avoiding the recognition error introduced by a single-dimensional index.

[0071] In the embodiments of the present application, the first data processing module can obtain the multi-dimensional feature joint value by calculating the product of the first time-domain feature value and the second time-domain feature value. It should be noted that the multi-dimensional feature joint value of the non-arrhythmia condition is usually low, so that the arrhythmia condition and the non-arrhythmia condition can be accurately identified through the multi-dimensional feature joint value, avoiding the problem of misjudgment.

[0072] Specifically, the first data processing module 12 is configured to extract the peak extreme points of each pulse wave from the pulse waveform diagram; then, based on the peak extreme points, determine a plurality of pulse intervals corresponding to the pulse waveform diagram, where the pulse interval is used to represent the time interval corresponding to two consecutive pulse beats of the target object; finally, determine the first time-domain feature value based on the plurality of pulse intervals; there is a positive correlation between the first time-domain feature value and the probability of the target object having an arrhythmia condition.

[0073] In the embodiments of the present application, the first data processing module can extract Figure 2 the peak extreme points of each pulse wave shown in (i.e., Figure 2 the position where the lower triangle in is located), and calculate the time interval between two consecutive pulse beats (i.e., two consecutive peaks in the pulse waveform diagram) based on the peak extreme points to obtain the pulse interval; then the first data processing module determines the first time-domain feature value based on the calculated plurality of pulse intervals. Among them, the first data processing module can calculate the first time-domain feature value σ through the following formula (1).

[0074]

[0075] where N is the total number of extreme points; x i is the i-th pulse interval; is the average value of all pulse intervals.

[0076] Optionally, in the embodiments of the present application, the greater the first time-domain eigenvalue, the greater the probability that the target object has an arrhythmia condition (such as atrial fibrillation), and the smaller the first time-domain eigenvalue, the smaller the probability that the target object has an arrhythmia condition.

[0077] It should be noted that in the prior art, abnormal waveform segments need to be removed when calculating the pulse interval, that is, the interval calculated by the prior art is the interval between normal sinus beats; while in the present application, abnormal waveform segments are not removed when calculating the pulse interval, and the calculated pulse interval is the interval between all heart beats, so as to effectively capture the arrhythmia phenomenon of the user in the heart rate frequency dimension.

[0078] The first data processing module 12 is further configured to determine the pulse peak value of each pulse wave based on the peak extreme points of each pulse wave, and construct a pulse peak value sequence based on the pulse peak values of each pulse wave; then, according to the first value and the maximum value in the pulse peak value sequence, perform linear interpolation on the first half of the waveform in the pulse waveform diagram to obtain a first interpolation sequence; the starting point of the first half of the waveform is the peak extreme point corresponding to the first value, and the ending point of the first half of the waveform is the peak extreme point corresponding to the maximum value; the first interpolation sequence includes the interpolation values corresponding to each peak extreme point in the first half of the waveform; and according to the maximum value and the last value in the pulse peak value sequence, perform linear interpolation on the second half of the waveform in the pulse waveform diagram to obtain a second interpolation sequence; the starting point of the second half of the waveform is the peak extreme point corresponding to the maximum value, and the ending point of the first half of the waveform is the peak extreme point corresponding to the last value; the second interpolation sequence includes the interpolation values corresponding to each peak extreme point in the second half of the waveform; then calculate the difference between the interpolation value corresponding to each peak extreme point in the first half of the waveform and the pulse peak value corresponding to the peak extreme point to obtain a first difference vector; and calculate the difference between the interpolation value corresponding to each peak extreme point in the second half of the waveform and the pulse peak value corresponding to the peak extreme point to obtain a second difference vector; finally, merge the first difference vector and the second difference vector, and determine the second time-domain eigenvalue based on the merged difference vector; there is a positive correlation between the second time-domain eigenvalue and the probability that the target object has an arrhythmia condition.

[0079] In the embodiments of the present application, the pulse peak value (i.e., the amplitude of the pulse wave) of the pulse wave is Figure 2 the measured value corresponding to the peak extreme point shown in, and the first data processing module can construct a pulse peak value sequence K = [p1, p2, p3,..., p K ,..., p n, where the pulse peak p1 is the measured value corresponding to the peak extreme point a, the pulse peak p2 is the measured value corresponding to the peak extreme point b, the pulse peak p3 is the measured value corresponding to the peak extreme point c, ..., and the pulse peak p N is the measured value corresponding to the peak extreme point n. Thus, the first data processing module can use the pulse peak p1 as the first value, and the pulse peak p n as the last value, and use the largest pulse peak p K in the pulse peak sequence K as the maximum value, which is the measured value corresponding to the peak extreme point k.

[0080] Furthermore, the first data processing module can perform linear interpolation on the first half of the pulse waveform shown in Figure 2 through the following formula (2):

[0081]

[0082] where intl i is the interpolation value corresponding to the i-th peak extreme point in the first half of the waveform. For example, intl1 is the interpolation value corresponding to the peak extreme point corresponding to the first value (i.e., the first peak extreme point); intl2 is the interpolation value corresponding to the second peak extreme point; intl K is the interpolation value corresponding to the peak extreme point corresponding to the maximum value (the K-th peak extreme point); p1 is the first value, M is the pulse peak of the K-th peak extreme point (i.e., the maximum value); pos_max is the position corresponding to the maximum value (i.e., K).

[0083] The first data processing module can also perform linear interpolation on the second half of the pulse waveform shown in Figure 2 through the following formula (3):

[0084]

[0085] where int2 i is the interpolation value corresponding to the i-th peak extreme point in the second half of the waveform. For example, int2 K is the interpolation value corresponding to the peak extreme point corresponding to the maximum value; int2 K+1 is the interpolation value corresponding to the (K + 1)-th peak extreme point; int2 N is the interpolation value corresponding to the peak extreme point corresponding to the last value; p n is the last value, M is the pulse peak of the K-th peak extreme point (i.e., the maximum value); pos_max is the position corresponding to the maximum value (i.e., K).

[0086] The first interpolation sequence obtained through the above formula (2) and formula (3) is [intl1, intl2, ..., intl K , and the second interpolation sequence is [int2K , int2 K+1 ,..., int2 N ; Then the first data processing module can calculate the difference between the interpolation value corresponding to each peak extreme point in the first half of the waveform and the pulse peak value corresponding to the peak extreme point, to obtain the first difference vector: [p1 - intl1, p2 - intl2,..., p K - intl K . And calculate the difference between the interpolation value corresponding to each peak extreme point in the second half of the waveform and the pulse peak value corresponding to the peak extreme point, to obtain the second difference vector: [p K - int2 K , p K+1 - int2 K+1 ,..., p N - int2 N . Then, the first data processing module can merge the first difference vector and the second difference vector, and calculate the standard deviation of the merged difference vector to obtain the second time-domain eigenvalue; there is a positive correlation between the second time-domain eigenvalue and the probability that the target object has an arrhythmia condition, that is, the larger the second time-domain eigenvalue, the greater the probability that the target object has an arrhythmia condition, and the smaller the second time-domain eigenvalue, the smaller the probability that the target object has an arrhythmia condition.

[0087] It should be noted that the second time-domain eigenvalue obtained based on the pulse peak values of each pulse wave can measure the abnormal degree of the heartbeat amplitude of the target object, so as to be able to evaluate whether the target object has an arrhythmia phenomenon in terms of this dimensional index of the heartbeat amplitude.

[0088] The second data processing module 13 is used to perform data distribution feature analysis on the pulse waveform diagram when determining that the target object has an arrhythmia condition based on the multi-dimensional feature joint value, to obtain the target eigenvalue; wherein, the target eigenvalue is used to evaluate whether the type of arrhythmia of the target object is atrial fibrillation.

[0089] Because atrial extrasystoles and other arrhythmia symptoms have similar pulse wave trend characteristics to atrial fibrillation, traditional atrial fibrillation detection methods often find it difficult to effectively distinguish the pulse waveforms of atrial fibrillation from those of atrial extrasystoles. Based on this, the inventors discovered that the PPI histogram of atrial extrasystoles exhibits a bimodal distribution, while the PPI histogram of atrial fibrillation exhibits a broad unimodal distribution. Therefore, the inventors proposed using the histogram distribution feature value HistFea (i.e., the target feature value) to distinguish atrial fibrillation from atrial extrasystoles. Specifically, when the target subject is determined to have an arrhythmia based on the multidimensional feature joint value, the second data processing module can perform data distribution feature analysis on the pulse waveform graph and calculate the histogram distribution feature value that can distinguish atrial fibrillation from atrial extrasystoles. Thus, the histogram distribution feature value can effectively reflect whether the target subject's arrhythmia type is atrial fibrillation or atrial extrasystoles.

[0090] The second data processing module 13 is used to determine whether the multidimensional feature joint value is greater than the first preset threshold; if the multidimensional feature joint value is greater than the first preset threshold, the second data processing module can determine that the target object has an arrhythmia condition; if the multidimensional feature joint value is less than or equal to the first preset threshold, the second data processing module can determine that the target object does not have an arrhythmia condition.

[0091] In the embodiment of the present application, the first preset threshold is a value preset to reflect the dividing line between normal heart rhythm and arrhythmia. For example, when the multi-dimensional feature joint value is greater than the first preset threshold 1×10 5 When the multi-dimensional feature joint value is less than or equal to the first preset threshold value 1×10 5 When the target object has no arrhythmia, the second data processing module can determine that the target object has no arrhythmia.

[0092] It should be noted that by judging whether the joint value of the multi-dimensional features is greater than the first preset threshold, it is determined whether the target object has arrhythmia. This can comprehensively evaluate whether the target object has arrhythmia based on the two dimensional indicators of heart rate and heart amplitude, effectively avoiding the misjudgment problem caused by relying solely on a single dimensional indicator for evaluation, thereby improving the accuracy of arrhythmia recognition.

[0093] The second data processing module 13 is also used to obtain multiple pulse intervals corresponding to the pulse waveform; then calculate the average value of the multiple pulse intervals and determine the histogram interval boundary based on the average value; the second data processing module can divide the multiple pulse intervals into multiple intervals based on the histogram interval boundary; then count the number of data in each interval and construct a statistical frequency vector based on the data number; finally, determine the target feature value based on the statistical frequency vector and the total amount of data of the pulse interval.

[0094] In an embodiment of the present application, the second data processing module can calculate the mean of multiple pulse intervals corresponding to the obtained pulse waveform diagram and round down to obtain the average mean(PPI) of the multiple pulse intervals; then determine the histogram interval boundaries based on a preset interval value and the average value. For example, with an interval of 15 (i.e., the preset interval value), the determined histogram interval boundaries are mean(PPI) - 30 and mean(PPI) + 30. Then the second data processing module can divide the multiple pulse intervals into multiple intervals according to the histogram interval boundaries and count the number of data Hist in each interval, and construct a statistical frequency vector {Hist1, Hist2, Hist3,...} based on the number of data. Finally, the second data processing module can calculate the target feature value HistFea through the following formula (4).

[0095]

[0096] Where, |PPI| is the total amount of data of the pulse interval.

[0097] The recognition module 14 is used to recognize the arrhythmia type of the target object according to the target feature value to obtain a recognition result.

[0098] In an embodiment of the present application, the recognition module 14 can identify whether the arrhythmia type of the target object is atrial fibrillation or non - atrial fibrillation (such as premature atrial contraction) by determining whether the target feature value is greater than a second preset threshold, thus realizing the accurate identification of atrial fibrillation through a single - pulse wave signal measurement, improving the atrial fibrillation detection efficiency of the system, and further providing a rapid and reliable diagnostic basis for medical workers, reducing the risk of over - treatment or delayed treatment caused by missed diagnosis or misdiagnosis. It is especially suitable for primary medical institutions, home self - testing and dynamic health monitoring scenarios, improving the user experience.

[0099] Specifically, the recognition module can be used to determine whether the target feature value is greater than a second preset threshold; if the target feature value is greater than the second preset threshold, the recognition module can determine that the recognition result is the first recognition result; where the first recognition result is used to represent that the arrhythmia type of the target object is atrial fibrillation; if the target feature value is less than or equal to the second preset threshold, the recognition module can determine that the recognition result is the second recognition result; where the second recognition result is used to represent that the arrhythmia type of the target object is non - atrial fibrillation.

[0100] In an alternative embodiment, the atrial fibrillation detection system may further include a user interaction module, where the user interaction module includes an atrial fibrillation display unit, a blood pressure display unit, and a key feedback unit. The atrial fibrillation display unit is used to display the atrial fibrillation recognition result to the user, and the blood pressure display unit is used to display the blood pressure measurement result to the user.

[0101] Through the atrial fibrillation detection system provided by this application, the time-domain feature analysis method is used to comprehensively evaluate whether the target object has arrhythmia in terms of two-dimensional indicators of heart rate and heart rate amplitude, effectively avoiding the misjudgment problem caused by relying solely on a single-dimensional indicator for atrial fibrillation assessment; the data distribution feature analysis method is used to effectively distinguish the pulse waveform diagrams of atrial fibrillation from other similar arrhythmias; the accurate identification of atrial fibrillation can be completed through a single pulse wave signal measurement, thereby improving the atrial fibrillation detection efficiency of the system, and further providing a fast and reliable diagnosis basis for medical workers, reducing the risk of over-treatment or delayed treatment caused by missed diagnosis or misdiagnosis. It is especially suitable for primary medical institutions, home self-testing, and dynamic health monitoring scenarios, improving the user experience.

[0102] Method Embodiment

[0103] An embodiment of the method for detecting atrial fibrillation provided by this application is applied to the atrial fibrillation detection system in the above system embodiment. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0104] See Figure 3 , which is a flowchart of a method for detecting atrial fibrillation provided by an embodiment of this application. This method is applied to an atrial fibrillation detection system. As shown in Figure 3, the method includes the following steps:

[0105] Step S301, collect the pulse wave signal of the target object within a preset time period and generate a corresponding pulse waveform diagram.

[0106] In the embodiment of this application, the atrial fibrillation detection system can collect the pulse wave signal of the target object within a preset time period by the oscillometric method or the photoplethysmography method.

[0107] Step S302, perform time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain feature value and a second time-domain feature value, and determine a multi-dimensional feature joint value based on the first time-domain feature value and the second time-domain feature value.

[0108] In step S302, the first time-domain feature value is used to measure the abnormal degree of the heart rate of the target object, and the second time-domain feature value is used to measure the abnormal degree of the heart rate amplitude of the target object.

[0109] In the embodiment of this application, the atrial fibrillation detection system can determine the first time-domain feature value through the process shown in Figure 4 , and this process includes the following steps:

[0110] Step S401: extracting the peak extreme points of each pulse wave from the pulse waveform diagram.

[0111] Step S402: determining a plurality of pulse intervals corresponding to the pulse waveform based on each peak extreme point.

[0112] In step S402, the pulse interval is used to represent the time interval corresponding to two consecutive pulse beats of the target object.

[0113] Step S403: determining a first time-domain characteristic value based on a plurality of pulse intervals.

[0114] In step S403 , the first time-domain feature value is positively correlated with the probability of the target subject having an arrhythmia condition.

[0115] It should be noted that the existing technology needs to eliminate abnormal waveform segments when calculating the pulse interval, that is, the existing technology calculates the interval between normal sinus beats; while the present application does not eliminate abnormal waveform segments when calculating the pulse interval, and the calculated pulse interval is the interval between all heartbeats, thereby effectively capturing the user's arrhythmia in the heart rate dimension.

[0116] Furthermore, the atrial fibrillation detection system can be Figure 5 The process shown in is used to determine the second time domain eigenvalue, which includes the following steps:

[0117] Step S501 : determining the pulse peak value of each pulse wave based on the peak extreme value point of each pulse wave, and constructing a pulse peak value sequence based on the pulse peak value of each pulse wave.

[0118] Step S502 : performing linear interpolation on the first half of the pulse waveform in the pulse waveform diagram according to the first value and the maximum value in the pulse peak sequence to obtain a first interpolation sequence.

[0119] In step S502, the starting point of the first half waveform is the peak extreme point corresponding to the first value, and the ending point of the first half waveform is the peak extreme point corresponding to the maximum value; the first interpolation sequence includes the interpolation values corresponding to each peak extreme point in the first half waveform.

[0120] Step S503 , performing linear interpolation on the second half of the pulse waveform in the pulse waveform diagram according to the maximum value and tail value in the pulse peak sequence to obtain a second interpolation sequence.

[0121] In step S503, the starting point of the second half waveform is the peak extreme point corresponding to the maximum value, and the ending point of the first half waveform is the peak extreme point corresponding to the tail value; the second interpolation sequence includes the interpolation values corresponding to each peak extreme point in the second half waveform.

[0122] Step S504: Calculate the difference between the interpolation value corresponding to each peak extreme point in the first half of the waveform and the pulse peak value corresponding to this peak extreme point, to obtain the first difference vector;

[0123] Step S505: Calculate the difference between the interpolation value corresponding to each peak extreme point in the second half of the waveform and the pulse peak value corresponding to this peak extreme point, to obtain the second difference vector;

[0124] Step S506: Combine the first difference vector and the second difference vector, and determine the second time-domain eigenvalue based on the combined difference vector.

[0125] In step S506, there is a positive correlation between the second time-domain eigenvalue and the probability that the target object has an arrhythmia condition.

[0126] It should be noted that the second time-domain eigenvalue obtained based on the pulse peak values of each pulse wave can measure the abnormal degree of the heartbeat amplitude of the target object, so as to be able to evaluate whether the target object has an arrhythmia phenomenon in terms of this dimensional index of the heartbeat amplitude.

[0127] In the embodiment of the present application, the first data processing module can obtain the multi-dimensional feature combined value by calculating the product of the first time-domain eigenvalue and the second time-domain eigenvalue. It should be noted that the multi-dimensional feature combined value of the non-arrhythmia condition is relatively low, so that the arrhythmia condition and the non-arrhythmia condition can be accurately identified through the multi-dimensional feature combined value, avoiding the problem of misjudgment.

[0128] Step S303: When it is determined that the target object has an arrhythmia based on the multi-dimensional feature combined value, perform data distribution feature analysis on the pulse waveform diagram to obtain the target eigenvalue.

[0129] In step S303, the target eigenvalue is used to evaluate whether the type of arrhythmia of the target object is atrial fibrillation.

[0130] In the embodiment of the present application, the atrial fibrillation detection system can determine whether the target object has an arrhythmia condition by judging whether the multi-dimensional feature combined value is greater than the first preset threshold; if the multi-dimensional feature combined value is greater than the first preset threshold, it is determined that the target object has an arrhythmia condition; if the multi-dimensional feature combined value is less than or equal to the first preset threshold, it is determined that the target object does not have an arrhythmia condition.

[0131] It should be noted that by judging whether the multi-dimensional feature combined value is greater than the first preset threshold to determine whether the target object has an arrhythmia condition, it is possible to comprehensively evaluate whether the target object has an arrhythmia phenomenon in terms of these two dimensional indexes of heart rate and heartbeat amplitude, effectively avoiding the misjudgment problem caused by relying only on a single dimensional index for evaluation, thereby improving the recognition accuracy of arrhythmia.

[0132] Further, the atrial fibrillation detection system can determine the histogram distribution eigenvalue (i.e., the target eigenvalue) through the process shown in Figure 6 The process includes the following steps:

[0133] Step S601: Obtain multiple pulse intervals corresponding to the pulse waveform diagram.

[0134] Step S602: Calculate the average value of the multiple pulse intervals, and determine the histogram interval boundaries based on the average value.

[0135] Step S603: Divide the multiple pulse intervals into multiple intervals according to the histogram interval boundaries.

[0136] Step S604: Count the number of data in each interval, and construct a statistical frequency vector based on the number of data.

[0137] Step S605: Determine the target eigenvalue based on the statistical frequency vector and the total amount of pulse interval data.

[0138] It should be noted that by analyzing the data distribution characteristics of the pulse waveform diagram, the histogram distribution eigenvalue (i.e., the target eigenvalue) that can distinguish atrial fibrillation from premature atrial contractions is calculated, and the histogram distribution eigenvalue can effectively reflect whether the arrhythmia type of the target object is atrial fibrillation or premature atrial contractions.

[0139] Step S304: Identify the arrhythmia type of the target object according to the target eigenvalue to obtain the identification result.

[0140] In the embodiment of the present application, the atrial fibrillation detection system can determine the arrhythmia type of the target object by judging whether the target eigenvalue is greater than the second preset threshold; if the target eigenvalue is greater than the second preset threshold, it is determined that the identification result is the first identification result; the first identification result is used to represent that the arrhythmia type of the target object is atrial fibrillation; if the target eigenvalue is less than or equal to the second preset threshold, it is determined that the identification result is the second identification result; the second identification result is used to represent that the arrhythmia type of the target object is non-atrial fibrillation.

[0141] Through the atrial fibrillation detection method provided by this application, the time-domain feature analysis method is used to comprehensively evaluate whether there is arrhythmia in the target object in terms of two-dimensional indicators of heart rate and heart amplitude, effectively avoiding the misjudgment problem caused by relying solely on a single-dimensional indicator for atrial fibrillation evaluation; the data distribution feature analysis method is used to effectively distinguish the pulse waveform diagrams of atrial fibrillation from other similar cardiac arrhythmias; it realizes the accurate identification of atrial fibrillation through a single pulse wave signal measurement, thereby improving the atrial fibrillation detection efficiency of the system, and further providing medical workers with fast and reliable diagnostic basis, reducing the risk of over-treatment or delayed treatment caused by missed diagnosis or misdiagnosis. It is especially suitable for primary medical institutions, home self-testing, and dynamic health monitoring scenarios, improving the user experience.

[0142] Embodiment of storage medium

[0143] The embodiment of this application provides a computer-readable storage medium, on which a program is stored. When the program is executed by a processor, it implements some or all of the steps in the atrial fibrillation detection method described in the foregoing method embodiment of this application. The storage medium can be various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc.

[0144] Embodiment of processor

[0145] The embodiment of this application provides a processor for running a program. When the program is running, it executes some or all of the steps in the atrial fibrillation detection method described in the foregoing method embodiment.

[0146] It should be noted that each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key point of each embodiment is to illustrate the differences from other embodiments. In particular, for the method embodiment, since it is basically similar to the system embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiment. The method embodiments described above are only illustrative. The units described as separate components may or may not be physically separated, and the components indicated as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0147] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An atrial fibrillation detection system, characterized in that, The atrial fibrillation detection system includes a signal acquisition module, a first data processing module, a second data processing module, and an identification module; The signal acquisition module is configured to acquire the pulse wave signal of the target object within a preset time period and generate a corresponding pulse waveform diagram; The first data processing module is configured to perform time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain feature value and a second time-domain feature value, and determine a multi-dimensional feature combined value based on the first time-domain feature value and the second time-domain feature value; The first time-domain feature value is used to measure the abnormality degree of the heart rate of the target object, and the second time-domain feature value is used to measure the abnormality degree of the heart rate amplitude of the target object; The second data processing module is configured to perform data distribution feature analysis on the pulse waveform diagram to obtain a target feature value when it is determined based on the multi-dimensional feature combined value that the target object has an arrhythmia condition; The target feature value is used to evaluate whether the arrhythmia type of the target object is atrial fibrillation; The identification module is configured to identify the arrhythmia type of the target object according to the target feature value to obtain an identification result.

2. The system according to claim 1, wherein The first data processing module is used for: Extracting the peak extreme points of each pulse wave from the pulse waveform diagram; Based on each of the peak extreme points, determining a plurality of pulse intervals corresponding to the pulse waveform diagram; the pulse interval is used to represent the time interval corresponding to two consecutive pulse beats of the target object; Determining the first time-domain feature value based on the plurality of pulse intervals; There is a positive correlation between the first time-domain feature value and the probability that the target object has an arrhythmia condition.

3. The system according to claim 2, wherein The first data processing module is further used for: Determining the pulse peak value of each pulse wave based on each of the peak extreme points of the pulse wave, and constructing a pulse peak value sequence based on the pulse peak values of each pulse wave; According to the first value and the maximum value in the pulse peak value sequence, performing linear interpolation on the first half of the waveform in the pulse waveform diagram to obtain a first interpolation sequence; the starting point of the first half of the waveform is the peak extreme point corresponding to the first value, and the ending point of the first half of the waveform is the peak extreme point corresponding to the maximum value; the first interpolation sequence includes the interpolation values corresponding to each of the peak extreme points in the first half of the waveform; According to the maximum value and the last value in the pulse peak value sequence, performing the linear interpolation on the second half of the waveform in the pulse waveform diagram to obtain a second interpolation sequence; the starting point of the second half of the waveform is the peak extreme point corresponding to the maximum value, and the ending point of the first half of the waveform is the peak extreme point corresponding to the last value; the second interpolation sequence includes the interpolation values corresponding to each of the peak extreme points in the second half of the waveform; Calculating the difference between the interpolation value corresponding to each of the peak extreme points in the first half of the waveform and the pulse peak value corresponding to the peak extreme point to obtain a first difference vector; Calculating the difference between the interpolation value corresponding to each of the peak extreme points in the second half of the waveform and the pulse peak value corresponding to the peak extreme point to obtain a second difference vector; Merge the first difference vector and the second difference vector, and determine the second time-domain eigenvalue based on the merged difference vector; There is a positive correlation between the second time-domain eigenvalue and the probability that the target object has the arrhythmia condition.

4. The system according to claim 1, characterized in that, The second data processing module is configured to: Judge whether the multi-dimensional feature joint value is greater than a first preset threshold; If the multi-dimensional feature joint value is greater than the first preset threshold, determine that the target object has the arrhythmia condition; If the multi-dimensional feature joint value is less than or equal to the first preset threshold, determine that the target object does not have the arrhythmia condition.

5. The system according to claim 1, wherein The second data processing module is further configured to: Obtain a plurality of pulse intervals corresponding to the pulse waveform diagram; Calculate the average value of the plurality of pulse intervals, and determine the histogram interval boundary according to the average value; Divide the plurality of pulse intervals into a plurality of intervals according to the histogram interval boundary; Count the number of data in each interval, and construct a statistical frequency vector based on the number of data; Determine the target eigenvalue based on the statistical frequency vector and the total amount of data of the pulse interval.

6. The system according to claim 1, wherein The recognition module is further configured to: Judge whether the target eigenvalue is greater than a second preset threshold; If the target eigenvalue is greater than the second preset threshold, determine that the recognition result is the first recognition result; The first recognition result is used to characterize that the arrhythmia type of the target object is atrial fibrillation; If the target eigenvalue is less than or equal to the second preset threshold, determine that the recognition result is the second recognition result; The second recognition result is used to characterize that the arrhythmia type of the target object is non-atrial fibrillation.

7. The system according to claim 1, characterized in that, The signal acquisition module is configured to acquire the pulse wave signal by an oscillometric method or a photoplethysmography method.

8. A method for detecting atrial fibrillation, characterized in that, Applied to the atrial fibrillation detection system according to any one of claims 1-7 above, the method includes: Acquire the pulse wave signal of the target object within a preset time period, and generate a corresponding pulse waveform diagram; Perform time-domain feature analysis on the pulse waveform diagram to obtain a first time-domain eigenvalue and a second time-domain eigenvalue, and determine a multi-dimensional feature joint value based on the first time-domain eigenvalue and the second time-domain eigenvalue; the first time-domain eigenvalue is used to measure the abnormal degree of the heart rate of the target object, and the second time-domain eigenvalue is used to measure the abnormal degree of the heart rate amplitude of the target object; When it is determined that the target object has an arrhythmia condition based on the multi-dimensional feature joint value, perform data distribution feature analysis on the pulse waveform diagram to obtain a target eigenvalue; the target eigenvalue is used to evaluate whether the arrhythmia type of the target object is atrial fibrillation; Identify the arrhythmia type of the target object according to the target eigenvalue to obtain a recognition result.

9. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, and when the computer program is run by a processor, the atrial fibrillation detection method described in claim 8 is implemented.

10. A processor, characterized in that, For running a computer program, when the computer program runs, it executes the atrial fibrillation detection method described in claim 8.

Images