An atrial fibrillation identification method, device, storage medium and computer device

Abstract

Embodiments of the present application disclose an atrial fibrillation detection method and device, a storage medium and a computer device. The method comprises: receiving an ECG signal sequence of a preset length and performing a preprocessing operation on the ECG signal; determining a QRS complex position based on the processed ECG signal; extracting a plurality of characteristic parameters according to the QRS complex position; and identifying and classifying the plurality of characteristic parameters based on a classifier. The present application simultaneously utilizes two features of an atrial fibrillation waveform, i.e., irregular changes in an RR interval and a baseline of a heartbeat interval being an irregular continuous high-frequency low-amplitude fibrillation wave, to detect and analyze an atrial fibrillation signal. Specifically, by determining a QRS complex position and extracting a plurality of characteristic parameters based on the position, the change rule of the RR interval and the related plurality of characteristic parameter values of the fibrillation wave are obtained. Further, a support vector machine (SVM) classifier is used to perform binary classification prediction on the above characteristic parameters, thereby achieving detection and identification of the atrial fibrillation signal and greatly improving the accuracy of the analysis result of the atrial fibrillation signal.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrocardiosignal data analysis, and in particular to a method and device for identifying atrial fibrillation, a storage medium and a computer device. BACKGROUND

[0002] Atrial fibrillation is a common type of arrhythmia in heart-related disease symptoms, which is referred to as atrial fibrillation. Atrial fibrillation is one of the important factors leading to many serious diseases, so it is crucial to accurately identify whether a patient has atrial fibrillation, the type of atrial fibrillation and the frequency of occurrence of atrial fibrillation. The most typical features of the waveform of atrial fibrillation mainly include: 1. P-wave disappearance; 2. QRS wave interval time is not the same, specifically, the interval time between two R waves is not the same, i.e. RR interval irregularly changes; 3. The heart beat baseline is irregular continuous high frequency low amplitude tremor wave and the like. Based on the characteristics of the waveform of atrial fibrillation, most of the current detection methods for atrial fibrillation are to analyze the disappearance of P-wave and the irregular change of RR interval. However, on the one hand, there is a tremor wave between two R waves, which is similar to P-wave, so it is difficult to analyze the atrial fibrillation signal by using the feature of P-wave disappearance; on the other hand, since the ventricular fibrillation signal, atrial flutter signal and other arrhythmia signals have similarities with the atrial fibrillation signal, the irregular change of RR interval also occurs in other arrhythmia types.

[0003] In summary, the analysis of the atrial fibrillation signal based on the two features of P-wave disappearance and irregular change of RR interval is not accurate. SUMMARY

[0004] Therefore, it is necessary to propose a method and device for identifying atrial fibrillation, a storage medium and a computer device to accurately detect and identify the atrial fibrillation signal in view of the above problems.

[0005] In a first aspect, the present application provides a method for identifying atrial fibrillation, which comprises:

[0006] receiving an ECG signal of a preset length and performing a preprocessing operation on the ECG signal;

[0007] determining the position of a QRS complex based on the preprocessed ECG signal;

[0008] extracting and fusing multiple feature parameters according to the position of the QRS complex;

[0009] identifying and classifying the multiple feature parameters based on a classifier to identify whether the ECG signal is an atrial fibrillation signal.

[0010] Optionally, the determination of the position of the QRS complex based on the preprocessed ECG signal comprises:

[0011] acquire a set of R-wave positions in the ECG signal based on a preset detection method;

[0012] search in preset intervals before and after the R-wave positions in the ECG signal based on the R-wave positions, and determine a Q-wave starting point and a J-point position corresponding to each R-wave.

[0013] Optionally, the multiple feature parameters include: an RR interval feature parameter, an f-wave feature parameter, a fusion feature parameter, and a reduced dimension feature parameter.

[0014] Optionally, the RR interval feature parameter includes: RR interval values of all adjacent R-waves, adjacent RR interval difference values, a mean value and a variance of the RR interval difference values.

[0015] The extraction method of the RR interval feature parameter includes: calculating the RR interval feature parameter based on adjacent sets of R-wave positions.

[0016] Optionally, the f-wave feature parameter includes: a mean value and a variance of the f-wave period, and a mean value and a variance of the f-wave amplitude.

[0017] The extraction method of the f-wave feature parameter includes: performing window end point integration in an interval from the J-point position to a Q-wave starting point of a next QRS wave.

[0018] Based on the window end point integration result, an over-zero point in an integrated signal in the interval is acquired, and a maximum value position between the over-zero points is determined; the maximum value position includes an f-wave maximum value position and an f-wave minimum value position.

[0019] The f-wave feature parameter is calculated based on the f-wave maximum value position and the f-wave minimum value position.

[0020] Optionally, a normalized feature parameter is obtained by using a normalization combination method based on the RR interval feature parameter and the f-wave feature parameter, and the fusion feature parameter is obtained by fusing the normalized feature parameter.

[0021] The fusion feature parameter is reduced in dimension to obtain the reduced dimension feature parameter.

[0022] Optionally, the classifier is a support vector machine classifier.

[0023] In a second aspect, the present application provides an atrial fibrillation recognition device, the device comprising:

[0024] a preprocessing module configured to receive an ECG signal of a preset length and perform a preprocessing operation on the ECG signal;

[0025] a processing module configured to determine a QRS complex position based on the preprocessed ECG signal;

[0026] an extraction module configured to extract and fuse multiple feature parameters according to the QRS complex position;

[0027] a recognition module configured to recognize and classify the multiple feature parameters based on a classifier to identify whether the ECG signal is an atrial fibrillation signal.

[0028] In a third aspect, the present application provides a computer readable storage medium storing a computer program, wherein the computer program, when executed by a processor, causes the processor to perform any of the above methods.

[0029] In a fourth aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the computer program, when executed by the processor, causes the processor to perform any of the above methods.

[0030] The embodiments of the present application have the following beneficial effects:

[0031] The present application provides an atrial fibrillation recognition method and device, a storage medium and a computer device, the method comprising: receiving an ECG signal sequence of a preset length and performing a preprocessing operation on the ECG signal; determining a QRS complex position based on the processed ECG signal; extracting multiple feature parameters according to the QRS complex position; and recognizing and classifying the multiple feature parameters based on a classifier. The present application simultaneously utilizes two features of an atrial fibrillation waveform, i.e., irregular changes in RR intervals and irregular continuous high-frequency low-amplitude tremor waves as the baseline of heartbeats, to detect and analyze an atrial fibrillation signal. Specifically, by determining a QRS complex position and extracting multiple feature parameters based on the position, the change rule of RR intervals and the values of related multiple feature parameters of tremor waves are obtained, and a support vector machine (SVM) classifier is further utilized to perform binary classification prediction on the above feature parameters, thereby realizing detection and recognition of an atrial fibrillation signal and greatly improving the accuracy of analysis results of an atrial fibrillation signal. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0033] In the drawings:

[0034] Figure 1 a flowchart of an atrial fibrillation recognition method provided in the embodiments of the present application;

[0035] Figure 2A method flowchart of step 103 provided in an embodiment of the present application;

[0036] Figure 3 Another method flowchart of step 103 provided in an embodiment of the present application;

[0037] Figure 4 A structure schematic diagram of an atrial fibrillation recognition device provided in an embodiment of the present application;

[0038] Figure 5 An internal structure diagram of a computer device in an embodiment is shown. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0040] It should be noted that the ECG signal refers to an electrocardiogram signal, which is a technology of recording the electrical activity change graph of each cardiac cycle of the heart from the body surface by an electrocardiograph, i.e., the signal is a sequence of electrical impulses that triggers the heart to beat. The QRS complex represents the potential change situation generated by the depolarization of the left and right ventricles. The QRS complex is composed of a single Q wave, an R wave, and an S wave, or a combination of these waveforms. Among them, the first QRS wave component above the reference horizontal line is called the R wave, the wave before the R wave downward is called the Q wave, and the S wave is the first downward wave after the R wave, i.e., even if the three waves do not exist at the same time, it is also called the QRS complex. The intersection point of the end of the QRS complex and the beginning of the ST segment is called the J point, which is the turning point of the sudden conversion of the QRS terminal to the ST segment, marking the end of ventricular depolarization and the beginning of ventricular repolarization. The f wave is an atrial fibrillation wave, and the frequency is generally 350-600H Z The RR interval between two R waves is not regular, which can reflect the degree of heart rate fluctuation of the current ECG signal, so the change of the RR interval in a period of time is counted as one of the main features of atrial fibrillation. The baseline part between the two R waves of the atrial fibrillation signal has a fibrillation wave, which generally has a high frequency, usually 350-600H Z , and its amplitude is lower than that of the R wave. The signal between the R waves of the normal sinus rhythm signal has the characteristics of small period and low amplitude, so the period and amplitude characteristics of the fibrillation wave are counted as a feature of atrial fibrillation, which can effectively distinguish the normal sinus rhythm signal.

[0041] In the embodiment of the present application, in order to effectively distinguish the atrial fibrillation signal from other arrhythmia signals such as ventricular fibrillation signal, atrial flutter signal and the like which have similarities with the atrial fibrillation signal, the embodiment of the present application extracts the main features that are most different from other similar signals, that is, the RR interval difference, the fibrillation wave period and amplitude and the statistical features, and then fuses these features to effectively improve the accuracy of atrial fibrillation recognition.

[0042] Please refer to Figure 1 The flow chart of the atrial fibrillation recognition method provided in the embodiment of the present application is as follows:

[0043] Step 101, receiving an ECG signal of a preset length and performing a preprocessing operation on the ECG signal.

[0044] It can be understood that the preset length can be 10 seconds. That is, the system receives the ECG signal sequence of which the input time length is more than 10 seconds and performs a preprocessing operation on the ECG signal. The purpose of the preprocessing operation is to filter out high-frequency noise interference, and therefore a 20Hz low-pass filter can be used to filter out high-frequency noise to reduce the interference of high-frequency noise on the detection of key points of the ECG signal. It should be noted that the actual size and type of the low-pass filter proposed here are only for illustration, and in actual implementation, it can be any low-pass filter that meets the requirements, which is not limited here. Z

[0045] Step 102, determining the position of the QRS complex based on the preprocessed ECG signal.

[0046] In the embodiment of the present application, the set of R wave positions in the ECG signal is obtained based on a preset detection method, and then the positions of the Q wave starting point and the J point corresponding to each R wave are determined based on the search in the preset interval before and after the R wave position in the ECG signal. The preset detection method is the Pan_Tompkins R wave detection method, and the size of the preset interval is 100ms. That is, the set of R wave positions in the ECG signal is obtained using the Pan_Tompkins R wave detection method, and the search is performed in a certain interval before and after the R wave position in the ECG signal, and the size of the search interval is set to 100ms, and the positions of the Q wave starting point and the J point corresponding to each R wave are determined in the interval.

[0047] Step 103, extracting and fusing multiple feature parameters according to the position of the QRS complex.

[0048] The multiple feature parameters include: RR interval feature parameter, f wave feature parameter, fused feature parameter and reduced dimension feature parameter. In the embodiment of the present application, based on the set of R waves and the positions of the Q wave starting point and the J point corresponding to the R waves obtained in step 102, please refer to Figure 2 ​For the method flowchart of step 103 provided in the embodiments of the present application, step 103 specifically comprises the following steps:

[0049] S1031. Calculate RR interval feature parameters. The RR interval feature parameters include: RR interval values of all adjacent R waves, adjacent RR interval difference values, mean and variance of RR interval difference values. The above-mentioned RR interval feature parameters are calculated based on the R wave set obtained in the foregoing.

[0050] S1032. Calculate f wave feature parameters. The f wave feature parameters include: mean and variance of f wave period, mean and variance of f wave amplitude. It should be noted that step S1032 and step S1031 can be performed simultaneously, that is, the f wave feature parameters are calculated at the same time when the RR interval feature parameters are calculated.

[0051] In step S1032, the following steps are included:

[0052] i. Perform window end point integration in the interval from the J point position of the current QRS wave to the Q wave starting point of the next QRS wave.

[0053] ii. Based on the window end point integration result, search for the zero-crossing points in the above-mentioned integrated signal in the interval, determine the maximum value position between the zero-crossing points, and the maximum value position is the f wave maximum value position and the f wave minimum value position.

[0054] iii. Based on the maximum value positions obtained in the foregoing, calculate the mean and variance of the f wave period and the mean and variance of the f wave amplitude, and then obtain the f wave feature parameters related to the period and the amplitude.

[0055] S1033. Based on the RR interval feature parameters and the f wave feature parameters, obtain normalized feature parameters by using a normalized combination method, and obtain fusion feature parameters by fusing the normalized feature parameters.

[0056] Specifically, the RR interval difference feature in the RR interval feature parameters and the f wave period feature in the f wave feature parameters are normalized and combined, and the RR interval difference variance feature in the RR interval feature parameters and the f wave period variance feature in the f wave feature parameters are normalized and combined; the two groups of features are concatenated and fused to obtain a fusion feature vector. It can be understood that the data after normalization and combination is the normalized feature parameter, and the fusion feature vector is the fusion feature parameter.

[0057] S1034. Reduce the dimension of the fusion feature parameter to obtain a reduced dimension feature parameter.

[0058] Specifically, the principal component analysis (PCA) is used to reduce the dimension of the fusion feature vector, and then a reduced dimension feature vector is obtained. The reduced dimension feature vector is the reduced dimension feature parameter.

[0059] Based on the above description of step 103, the embodiment of the present application provides an actual implementation of step 103, please refer to Figure 3 Another method flow chart of step 103 provided in the embodiment of the present application.

[0060] Specifically,

[0061] i. Obtain the R wave set in the ECG signal obtained by the Pan_Tompkins R wave detection method in step 102. Wherein, the R wave set also includes the Q wave starting point and J point position corresponding to the R wave.

[0062] ii. As shown in the left branch of Figure 3 , first calculate the RR interval value of all adjacent R waves in the R wave set, and second calculate the RR interval difference value and the mean and variance of the RR interval difference value to obtain the RR interval characteristic parameter.

[0063] iii. At the same time, as shown in the right branch of Figure 3 , the window end integration is performed in the interval from the J point position of the current QRS wave to the Q wave starting point of the next QRS wave, and then the zero-crossing point in the integrated signal in the interval is searched to determine the maximum position between the zero-crossing points, and the maximum position is the position of the maximum value and minimum value of the f wave, and then the mean and variance of the f wave period and the mean and variance of the amplitude are calculated to obtain the characteristic parameters of the period and amplitude of the f wave.

[0064] iv. Normalize and combine the RR interval difference value feature and the f wave period feature, normalize and combine the RR interval difference variance feature and the f wave period variance feature, and fuse the two groups of features to obtain a fusion feature vector.

[0065] v. Use principal component analysis (PCA) to reduce the dimension of the feature vector to obtain a reduced dimension feature vector.

[0066] Step 104, based on the classifier, the multi-feature parameters are recognized and classified to identify whether the ECG signal is an atrial fibrillation signal.

[0067] In the embodiment of the present application, the support vector machine (SVM) classifier is used to perform binary classification prediction on the reduced dimension feature vector obtained in step 103, and then the atrial fibrillation signal is identified.

[0068] In this embodiment, atrial fibrillation (AF) signals are detected and analyzed by analyzing two characteristics of the RR interval waveform: irregular changes in the RR interval and an irregular, continuous high-frequency, low-amplitude fibrillation wave at the baseline between heartbeats. Specifically, the location of the QRS complex wave is determined, and relevant RR interval characteristic parameters, f-wave characteristic parameters, fusion characteristic parameters, and dimensionality reduction characteristic parameters are acquired and calculated. Furthermore, a support vector machine (SVM) classifier is used to perform binary classification prediction on the above characteristic parameters, thereby achieving the detection and identification of AF signals and greatly improving the accuracy of the AF signal analysis results.

[0069] Based on the above method steps 101-104, this application provides an example of a specific embodiment. It is understood that this is merely an illustrative example and represents only one implementation of the above method. Any other implementations that conform to the above method but differ from the example provided also fall within the protection scope of this application. This example embodiment includes:

[0070] I. Input an ECG signal sequence X{x1,x2,…,x... that meets the specified time length. n}, where n is the number of sampling points, and the sequence length exceeds 10 seconds to ensure that a sufficiently effective R-wave can be detected. A 10th-order 20H sequence is used. Z The Butterworth low-pass filter removes high-frequency noise from the ECG signal, reducing the interference of high-frequency noise on the detection of key points in the signal.

[0071] II. Use the Pan_Tompkins R-wave detection method to detect the R-wave position and obtain the set of all R-wave positions in the ECG signal: Y{r1,r2,…,r m}, where m is the number of detected R waves. A search interval of 100ms is set based on the detected R wave positions. The search proceeds forward from the R wave position to determine the Q wave initiation position of the current QRS wave, and backwards from the R wave position to determine the J point position of the current QRS wave. This process is repeated to determine the Q wave initiation and J point positions corresponding to all R waves, resulting in the set of three key point positions: Z{(q1,r1,j1),(q2,r2,j2),…,(q m ,r m ,j m )}.

[0072] III. Based on the location set of the three key points obtained in step II, calculate the RR interval values ​​of all adjacent R waves starting from the second R wave, obtaining the RR interval value set RR{rr1,rr2,…,rr}. (m-1) Next, the mean and variance of the adjacent RR interval differences are calculated to obtain the RR interval characteristic parameters, including the set of RR interval differences: RR_Diff{rrd1,rrd2,…,rrd}(m-2)}, where rrd i =rr (i+1) -rrd i And the mean RR_Diff_Mean and variance RR_Diff_Var of the RR interval differences. Also, at the J point position of the current QRS wave. i The Q wave starting point q of the next QRS wave (i-1) Integrate within the interval up to the window endpoint, with the integration window length set to 100ms. The integral value is the cumulative sum of the differences between the signal value within the integration window and the value at the current endpoint of the integration window. After obtaining the integral, integrate within the interval (j... i ,q (i-1) The algorithm searches for all zero-crossing points of the integral signal within a given range, determining the extreme values ​​between these points. These extreme values ​​correspond to the maximum and minimum values ​​of the f-wave. The mean F_Mean and variance F_Var of the f-wave's period, and the mean F_Amp_Mean and variance F_Amp_Var of its amplitude, are then calculated to obtain the period and amplitude characteristic parameters of the f-wave. The RR interval difference characteristic and the f-wave period characteristic are normalized and combined, as are the RR interval difference variance characteristic and the f-wave period variance characteristic. These two sets of characteristic vectors are then concatenated to obtain a fused characteristic vector. Principal component analysis (PCA) is used to reduce the dimensionality of the characteristic vector, resulting in a dimensionality-reduced characteristic vector.

[0073] IV. The reduced feature vector of the ECG signal is fed into a support vector machine (SVM) classifier for binary classification prediction, thereby identifying whether the input ECG signal is an atrial fibrillation signal.

[0074] Please see Figure 4 This is a schematic diagram of an atrial fibrillation identification device provided in an embodiment of this application. The device includes:

[0075] The preprocessing module 401 is used to receive an ECG signal of a preset length and perform preprocessing operations on the ECG signal.

[0076] Processing module 402 is used to determine the position of the QRS composite wave based on the preprocessed ECG signal;

[0077] Extraction module 403 is used to extract and fuse multiple feature parameters based on the QRS composite wave position;

[0078] The identification module 404 is used to identify and classify multiple feature parameters based on a classifier in order to identify whether the ECG signal is an atrial fibrillation signal.

[0079] In this embodiment, the relevant content of the preprocessing module 401, processing module 402, extraction module 403, and recognition module 404 can be found in the following references. Figure 1 andFigure 2 The contents of the illustrated embodiments will not be repeated here.

[0080] In this application embodiment, a computer-readable storage medium is provided, storing a computer program, which, when executed by a processor, causes the processor to perform the method of any one of the above method embodiments.

[0081] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the method of any one of the above method embodiments.

[0082] Figure 5 An internal structural diagram of a computer device in one embodiment is shown. This computer device can specifically be a terminal, a server, or a gateway. Figure 5 As shown, the computer device includes a processor, memory, and network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program causes the processor to perform the steps in the above-described method embodiments. The internal memory may also store a computer program, which, when executed by the processor, causes the processor to perform the steps in the above-described method embodiments. Those skilled in the art will understand that... Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0083] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer readable storage medium, and when the program is executed, the processes of the above-mentioned embodiment methods can be included. Any reference to memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0084] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0085] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for identifying atrial fibrillation, characterized in that, The method includes: Receive an ECG signal of a preset length and perform preprocessing operations on the ECG signal; The location of the QRS composite wave is determined based on the preprocessed ECG signal; Multiple feature parameters are extracted and fused based on the QRS composite wave position; The ECG signal is classified based on the multi-feature parameters to identify whether it is an atrial fibrillation signal. The multiple feature parameters include: RR interval feature parameters, f-wave feature parameters, fusion feature parameters, and dimensionality reduction feature parameters; The characteristic parameters of the f-wave include: the mean and variance of the f-wave period, and the mean and variance of the f-wave amplitude; The method for extracting the f-wave characteristic parameters includes: performing window endpoint integration within the interval from the J-point position to the Q-wave starting point of the next QRS composite wave; Based on the window endpoint integration result, the zero-crossing points in the integral signal within the interval are obtained, and the extreme value positions between the zero-crossing points are determined; the extreme value positions include the position of the maximum value of the f-wave and the position of the minimum value of the f-wave. The characteristic parameters of the f-wave are calculated based on the location of the maximum value of the f-wave and the location of the minimum value of the f-wave. The determination of the QRS composite wave position based on the processed ECG signal includes: The set of R-wave positions in the ECG signal is obtained based on a preset detection method; Based on the R-wave position, search within the preset interval before and after the ECG signal to determine the Q-wave start point and J-point position corresponding to each R-wave; Based on the RR interval characteristic parameters and the f-wave characteristic parameters, normalized characteristic parameters are obtained using a normalized combination method, and the normalized characteristic parameters are then fused to obtain the fused characteristic parameters.

2. The method according to claim 1, characterized in that, The RR interval characteristic parameters include: the RR interval values ​​of all adjacent R waves, the difference between adjacent RR intervals, and the mean and variance of the RR interval difference; The method for extracting the RR interval characteristic parameters includes: calculating the RR interval characteristic parameters based on the set of adjacent R-wave positions.

3. The method according to claim 1, characterized in that, The dimensionality of the fused feature parameters is reduced to obtain the dimensionality-reduced feature parameters.

4. The method according to claim 1, characterized in that, The classifier is a support vector machine classifier.

5. An atrial fibrillation detection device, characterized in that, The device includes: The preprocessing module is used to receive an ECG signal of a preset length and perform preprocessing operations on the ECG signal; The processing module is used to determine the position of the QRS composite wave based on the preprocessed ECG signal; The extraction module is used to extract and fuse multiple feature parameters based on the QRS composite wave position; The identification module is used to identify and classify the multiple feature parameters based on a classifier in order to identify whether the ECG signal is an atrial fibrillation signal. The multiple feature parameters include: RR interval feature parameters, f-wave feature parameters, fusion feature parameters, and dimensionality reduction feature parameters; The characteristic parameters of the f-wave include: the mean and variance of the f-wave period, and the mean and variance of the f-wave amplitude; The method for extracting the f-wave characteristic parameters includes: performing window endpoint integration within the interval from the J-point position to the Q-wave starting point of the next QRS composite wave; Based on the window endpoint integration result, the zero-crossing points in the integral signal within the interval are obtained, and the extreme value positions between the zero-crossing points are determined; the extreme value positions include the position of the maximum value of the f-wave and the position of the minimum value of the f-wave. The characteristic parameters of the f-wave are calculated based on the location of the maximum value of the f-wave and the location of the minimum value of the f-wave. The determination of the QRS composite wave position based on the processed ECG signal includes: The set of R-wave positions in the ECG signal is obtained based on a preset detection method; Based on the R-wave position, search within the preset interval before and after the ECG signal to determine the Q-wave start point and J-point position corresponding to each R-wave; Based on the RR interval characteristic parameters and the f-wave characteristic parameters, normalized characteristic parameters are obtained using a normalized combination method, and the normalized characteristic parameters are then fused to obtain the fused characteristic parameters.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the steps of the method as described in any one of claims 1 to 4.

7. A computer device, comprising a memory and a processor, characterized in that, The memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 4.

Images