CN108926348B - Method and device for extracting atrial fibrillation signals - Google Patents

Abstract

The invention discloses a kind of extracting methods of atrial fibrillation signal, comprising: obtains ECG mixed signal, and pre-processes to ECG mixed signal；Ventricular cardiac signal is filtered out from pretreated ECG mixed signal, is obtained non-ventricular signal, is mixed with atrial fibrillation signal and noise signal in non-ventricular signal；Based on non-ventricular signal in autocorrelation domain can dtex sign, generate non-ventricular signal autocorrelation matrix；Noise signal is removed using autocorrelation matrix and least square method, to obtain the atrial fibrillation signal with higher degree and accuracy, convenient for distinguishing the type of atrial fibrillation signal, and the inducement of atrial fibrillation occurs.Correspondingly, extraction element, equipment and the computer readable storage medium of a kind of atrial fibrillation signal disclosed by the invention, similarly has above-mentioned technique effect.

Description

Method and device for extracting atrial fibrillation signals

technical field

The present invention relates to the technical field of signal processing, and more particularly, to a method, apparatus, device and computer-readable storage medium for extracting atrial fibrillation signals.

Background technique

Atrial fibrillation (abbreviated as atrial fibrillation) is the most common sustained arrhythmia, with an incidence of 0.4 to 1.0%. The incidence of atrial fibrillation increases with age, reaching as high as 8% in people over 80 years of age. The prevalence of atrial fibrillation is also closely related to coronary heart disease, hypertension and heart failure. With the increasing aging of modern society, the number of patients with atrial fibrillation will gradually increase. Among heart disease patients, the mortality rate of atrial fibrillation patients is twice that of other patients. Therefore, the diagnosis and treatment of atrial fibrillation has always been the focus of clinical and international electrophysiological research.

Body surface electrocardiogram (electrocardiograph, ECG) is an important method for the diagnosis of heart disease, and it is also an effective method for the diagnosis of atrial fibrillation. A mixture of atrial waves, ventricular waves, and other signals together make up the surface twelve-lead ECG. For a normal ECG signal, with normal sinus rhythm, electrical activity in the atria and ventricles produces P waves and QRST complexes, respectively, on the body surface. In the case of atrial fibrillation, the main manifestation of ECG is that the P wave disappears, and atrial flocculation excitation waves of different sizes and shapes appear, that is, the atrial fibrillation signal (f wave). Extracting the atrial fibrillation signal (f wave) from the ECG of the patient with atrial fibrillation disease can facilitate the doctor to timely determine whether the patient has atrial fibrillation, and can assist the doctor in diagnosing the type of atrial fibrillation that the patient suffers from.

Currently, the existing extraction methods are generally template cancellation methods (such as ABS (average beat subtraction) and STC (spatiotemporal cancellation)) and blind signal separation methods (such as independent component analysis, ICA). Among them, the template cancellation method is more sensitive to the morphological changes of the QRST complex, and the extracted atrial fibrillation signal is obviously distorted; as for the blind signal separation method, because it uses the high-order statistics of the signal, the separation effect is poor. The extracted atrial fibrillation signal is still mixed with other noise signals.

Therefore, how to make the extracted atrial fibrillation signal have high purity and accuracy is a problem to be solved by those skilled in the art.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a method, device, device and computer-readable storage medium for extracting atrial fibrillation signals, so as to extract atrial fibrillation signals with higher purity and accuracy.

To achieve the above purpose, the embodiments of the present invention provide the following technical solutions:

A method for extracting atrial fibrillation signals, comprising:

Obtaining the ECG mixed signal, and preprocessing the ECG mixed signal;

The ventricular signal is filtered from the preprocessed ECG mixed signal to obtain a non-ventricular signal, wherein the non-ventricular signal is mixed with atrial fibrillation signal and noise signal;

generating an autocorrelation matrix of the non-ventricular signal based on the separable features of the non-ventricular signal in the autocorrelation domain;

The atrial fibrillation signal is obtained by removing the noise signal using the autocorrelation matrix and the least squares method.

Wherein, the preprocessing of the ECG mixed signal includes:

The Gaussian white noise in the ECG mixed signal is removed by wavelet denoising, and the noise whose amplitude and modulus value is less than a preset threshold in the ECG mixed signal is removed by threshold denoising.

Wherein, generating the autocorrelation matrix of the non-ventricular signal based on the separable features of the non-ventricular signal in the autocorrelation domain includes:

Based on the separable features of the non-ventricular signal in the autocorrelation domain, an autocorrelation matrix in the target time period τ _j is generated

in, τ _j represents the time period when the value of the noise signal in the autocorrelation domain is zero, X(t) represents the non-ventricular signal at the time point t, and X(t-τ _j ) ^T represents the time period of the time point t-τ _j . The transposed matrix of the non-ventricular signal, H represents the mixing matrix of the atrial fibrillation signal f(t), W represents the mixing matrix of the noise signal j(t), E{·} represents the expectation operation, t represents point in time.

Wherein, using the autocorrelation matrix and the least squares method to remove the noise signal to obtain the atrial fibrillation signal includes:

Construct And solve the filter Q, use the filter Q to filter to obtain the first noise component

the first noise component with a component of the non-ventricular signal X ^* (t) Fitting using the least squares method to obtain the second noise component and use calculating the atrial fibrillation signal f _i (t);

in, Q ^T represents the transposed matrix of the filter Q, and f(t) represents the atrial fibrillation signal.

Wherein, after removing the noise signal by using the autocorrelation matrix and the least squares method to obtain the atrial fibrillation signal, the method further includes:

The atrial fibrillation signal is visualized.

A device for extracting atrial fibrillation signals, comprising:

an acquisition module for acquiring the ECG mixed signal and preprocessing the ECG mixed signal;

a filtering module, configured to filter the ventricular signal from the preprocessed ECG mixed signal to obtain a non-ventricular signal, wherein the non-ventricular signal is mixed with atrial fibrillation signal and noise signal;

a generating module, configured to generate an autocorrelation matrix of the non-ventricular signal based on the separable features of the non-ventricular signal in the autocorrelation domain;

A removal module, configured to remove the noise signal by using the autocorrelation matrix and the least squares method to obtain the atrial fibrillation signal.

Wherein, the acquisition module is specifically used for:

The Gaussian white noise in the ECG mixed signal is removed by wavelet denoising, and the noise whose amplitude and modulus value is less than a preset threshold in the ECG mixed signal is removed by threshold denoising.

Among them, it also includes:

The display module is used to visually display the atrial fibrillation signal.

A device for extracting atrial fibrillation signals, comprising:

memory for storing computer programs;

The processor is configured to implement the steps of the method for extracting atrial fibrillation signals according to any one of the above when executing the computer program.

A computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, implements the steps of the method for extracting atrial fibrillation signals described in any one of the above.

It can be seen from the above solutions that the method for extracting atrial fibrillation signals provided by the embodiments of the present invention includes: acquiring an ECG mixed signal, and preprocessing the ECG mixed signal; filtering out the ventricle from the preprocessed ECG mixed signal signal, and obtain a non-ventricular signal, the non-ventricular signal is mixed with atrial fibrillation signal and noise signal; based on the separable feature of the non-ventricular signal in the autocorrelation domain, the autocorrelation matrix of the non-ventricular signal is generated; The autocorrelation matrix and the least squares method are used to remove the noise signal to obtain the atrial fibrillation signal.

It can be seen that, in order to extract high-purity atrial fibrillation signals, the method firstly preprocesses the obtained ECG mixed signals, and filters out ventricular signals from the preprocessed ECG mixed signals to obtain non-ventricular signals. The separable features of the non-ventricular signal in the autocorrelation domain and the least squares method remove the noise signal in the non-ventricular signal, so as to obtain atrial fibrillation with high accuracy. Among them, the method further filters out the noise signal on the basis of the separated non-ventricular signal, so that the atrial fibrillation signal with high purity and accuracy can be extracted, which is convenient for distinguishing the type of atrial fibrillation signal and the occurrence of atrial fibrillation. tremor triggers.

Correspondingly, the apparatus, device, and computer-readable storage medium for extracting atrial fibrillation signals provided by the embodiments of the present invention also have the above technical effects.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings without creative efforts.

1 is a flowchart of a method for extracting atrial fibrillation signals disclosed in an embodiment of the present invention;

2 is a flowchart of another method for extracting atrial fibrillation signals disclosed in an embodiment of the present invention;

3 is a schematic diagram of a device for extracting atrial fibrillation signals disclosed in an embodiment of the present invention;

4 is a schematic diagram of a device for extracting atrial fibrillation signals disclosed in an embodiment of the present invention;

5 is a waveform diagram of an ECG mixed signal disclosed in an embodiment of the present invention in the time domain;

FIG. 6 is a waveform diagram of the atrial fibrillation signal extracted from FIG. 5 in the time domain.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Embodiments of the present invention disclose a method, device, device and computer-readable storage medium for extracting atrial fibrillation signals, so as to extract atrial fibrillation signals with high purity and accuracy.

Referring to FIG. 1, a method for extracting atrial fibrillation signals provided by an embodiment of the present invention includes:

S101, obtaining an ECG mixed signal, and preprocessing the ECG mixed signal;

The ECG mixed signal is an electrocardiogram signal composed of atrial signal, ventricular signal and other noise signals. If a patient develops atrial fibrillation, there is atrial fibrillation signal in the atrial signal of the patient. Therefore, in this embodiment, the obtained ECG mixed signal is an ECG signal mixed with atrial fibrillation signals. In order to provide good preconditions for the subsequent extraction steps, the obtained ECG mixed signal may be preprocessed, that is, to The obtained ECG mixed signal is filtered and denoised.

S102, filtering the ventricular signal from the preprocessed ECG mixed signal to obtain a non-ventricular signal, and the non-ventricular signal is mixed with atrial fibrillation signal and noise signal;

It should be noted that the ventricular signal conforms to a super-Gaussian distribution, while the atrial signal conforms to a sub-Gaussian distribution. Among them, the kurtosis value of the atrial signal is close to zero, so it can be used as a quasi-Gaussian signal. Therefore, the ventricular signal conforming to the Gaussian distribution can be accurately filtered out, and the remaining signals are the atrial signal and other noise signals, wherein the noise signals include thermal noise signals.

S103, based on the separable features of the non-ventricular signal in the autocorrelation domain, generate an autocorrelation matrix of the non-ventricular signal;

S104, remove the noise signal by using the autocorrelation matrix and the least square method to obtain the atrial fibrillation signal.

It should be noted that the non-ventricular signal is composed of atrial signal and other noise signals, and atrial fibrillation signal exists in the atrial signal. That is, the non-ventricular signal is mixed with atrial fibrillation signal and noise signal.

Among them, the separable feature in the autocorrelation domain means that the signals in a certain signal set have their own unique feature delay sets. At this time, the signals in the signal set can be said to be separable in the autocorrelation domain, that is, through the processing in the autocorrelation domain. The method can realize the separation of each signal. On the other hand, the non-ventricular signal can be separated according to its separable features in the autocorrelation domain, and combined with the least squares method to remove the noise signal, so as to obtain the atrial fibrillation signal with high purity and accuracy.

In actual processing, the autocorrelation domain features that meet the needs of non-ventricular signal processing can be extracted according to the characteristics of different non-ventricular signals, or the autocorrelation domain features of corresponding signals can be designed in advance according to the needs of non-ventricular signal processing. At the same time, the use of autocorrelation domain separable features for signal separation has many advantages, such as the simplicity of implementation, the interpretability of decomposition forms and decomposition results, and the small storage space occupied.

It can be seen that this embodiment provides a method for extracting atrial fibrillation signals. In order to extract a high-purity atrial fibrillation signal, the method first preprocesses the obtained ECG mixed signal, and mixes the obtained ECG signals from the preprocessed ECG signals. The ventricular signal is filtered from the signal to obtain the non-ventricular signal, and then the noise signal in the non-ventricular signal is removed based on the separable feature of the non-ventricular signal in the autocorrelation domain and the least squares method, so as to obtain a high-accuracy atrial fibrillation. Among them, the method further filters out the noise signal on the basis of the separated non-ventricular signal, so that the atrial fibrillation signal with high purity and accuracy can be extracted, which is convenient for distinguishing the type of atrial fibrillation signal and the occurrence of atrial fibrillation. tremor triggers.

The embodiment of the present invention discloses another method for extracting atrial fibrillation signals. Compared with the previous embodiment, this embodiment further describes and optimizes the technical solution.

Referring to FIG. 2, another method for extracting atrial fibrillation signals provided by an embodiment of the present invention includes:

S201, obtaining the ECG mixed signal, and preprocessing the ECG mixed signal;

S202, filtering the ventricular signal from the preprocessed ECG mixed signal to obtain a non-ventricular signal, and the non-ventricular signal is mixed with atrial fibrillation signal and noise signal;

S203, based on the separable features of the non-ventricular signal in the autocorrelation domain, generate an autocorrelation matrix of the non-ventricular signal;

S204, remove the noise signal by using the autocorrelation matrix and the least squares method to obtain the atrial fibrillation signal;

S205. Visually display the atrial fibrillation signal.

In this embodiment, after obtaining the atrial fibrillation signal with high purity and accuracy according to the above steps, in order to facilitate people to identify the type of the atrial fibrillation signal, the obtained atrial fibrillation signal may be displayed visually. That is, the waveform of the acquired atrial fibrillation signal in the time domain is displayed in the visualization window, and if necessary, it can also be played dynamically, so that people can identify the type of atrial fibrillation signal based on this waveform, and explore the cause of atrial fibrillation. incentives.

Of course, the obtained atrial fibrillation signals with high purity and accuracy can also be stored centrally, so as to provide experimental data that can be based on for medical research.

It can be seen that this embodiment provides another method for extracting atrial fibrillation signals. In order to extract a high-purity atrial fibrillation signal, the method first preprocesses the obtained ECG mixed signal, and extracts the ECG signal from the preprocessed ECG signal. The ventricular signal is filtered out of the mixed signal to obtain the non-ventricular signal, and then the noise signal in the non-ventricular signal is removed based on the separable feature of the non-ventricular signal in the autocorrelation domain and the least squares method, so as to obtain a high-accuracy atrial fibrillation. Among them, the method further filters out the noise signal on the basis of the separated non-ventricular signal, so that the atrial fibrillation signal with high purity and accuracy can be extracted; at the same time, the visual display of the atrial fibrillation signal can facilitate People identify the types of atrial fibrillation signals, and the triggers for developing atrial fibrillation.

Based on any of the foregoing embodiments, it should be noted that the preprocessing of the ECG mixed signal includes:

The Gaussian white noise in the ECG mixed signal is removed by wavelet denoising, and the noise whose amplitude and modulus value is less than a preset threshold in the ECG mixed signal is removed by threshold denoising.

Of course, adaptive filtering denoising, Wiener filtering or other denoising methods can also be used to remove the noise signal in the ECG mixed signal.

Based on any of the foregoing embodiments, it should be noted that the generating the autocorrelation matrix of the non-ventricular signal based on the separable features of the non-ventricular signal in the autocorrelation domain includes:

Based on the separable features of the non-ventricular signal in the autocorrelation domain, an autocorrelation matrix in the target time period τ _j is generated

in, τ _j represents the time period when the value of the noise signal in the autocorrelation domain is zero, X(t) represents the non-ventricular signal at the time point t, and X(t-τ _j ) ^T represents the time period of the time point t-τ _j . The transposed matrix of the non-ventricular signal, H represents the mixing matrix of the atrial fibrillation signal f(t), W represents the mixing matrix of the noise signal j(t), E{·} represents the expectation operation, t represents point in time.

Based on any of the foregoing embodiments, it should be noted that the step of removing the noise signal by using the autocorrelation matrix and the least squares method to obtain the atrial fibrillation signal includes:

Construct And solve the filter Q, use the filter Q to filter to obtain the first noise component

the first noise component with a component of the non-ventricular signal X ^* (t) Fitting using the least squares method to obtain the second noise component and use calculating the atrial fibrillation signal f _i (t);

in, Q ^T represents the transposed matrix of the filter Q, and f(t) represents the atrial fibrillation signal.

Based on any of the foregoing embodiments, it should be noted that the filtering out the ventricular signal from the preprocessed ECG mixed signal includes: filtering out the ventricular signal from the preprocessed ECG mixed signal by using an independent component analysis method. Among them, the independent component analysis method can be an improved algorithm for the independent component analysis, for example, a fast independent component analysis method.

An apparatus for extracting an atrial fibrillation signal provided by an embodiment of the present invention is introduced below. The apparatus for extracting an atrial fibrillation signal described below and a method for extracting an atrial fibrillation signal described above can be referred to each other.

Referring to FIG. 3 , an apparatus for extracting atrial fibrillation signals provided by an embodiment of the present invention includes:

an acquisition module 301, configured to acquire an ECG mixed signal, and preprocess the ECG mixed signal;

A filtering module 302, configured to filter out the ventricular signal from the preprocessed ECG mixed signal to obtain a non-ventricular signal, wherein the non-ventricular signal is mixed with atrial fibrillation signal and noise signal;

a generating module 303, configured to generate an autocorrelation matrix of the non-ventricular signal based on the separable features of the non-ventricular signal in the autocorrelation domain;

A removing module 304 is configured to remove the noise signal by using the autocorrelation matrix and the least squares method to obtain the atrial fibrillation signal.

Wherein, the acquisition module is specifically used for:

The Gaussian white noise in the ECG mixed signal is removed by wavelet denoising, and the noise whose amplitude and modulus value is less than a preset threshold in the ECG mixed signal is removed by threshold denoising.

Among them, it also includes:

The display module is used to visually display the atrial fibrillation signal.

Wherein, the acquisition module is specifically used for:

The Gaussian white noise in the ECG mixed signal is removed by wavelet denoising, and the noise whose amplitude and modulus value is less than a preset threshold in the ECG mixed signal is removed by threshold denoising.

Wherein, the generation module is specifically used for:

Based on the separable features of the non-ventricular signal in the autocorrelation domain, an autocorrelation matrix in the target time period τ _j is generated

in, τ _j represents the time period when the value of the noise signal in the autocorrelation domain is zero, X(t) represents the non-ventricular signal at the time point t, and X(t-τ _j ) ^T represents the time period of the time point t-τ _j . The transposed matrix of the non-ventricular signal, H represents the mixing matrix of the atrial fibrillation signal f(t), W represents the mixing matrix of the noise signal j(t), E{·} represents the expectation operation, t represents point in time.

Wherein, the removal module includes:

A first noise component calculation module for constructing And solve the filter Q, use the filter Q to filter to obtain the first noise component

The second noise component calculation module is used to calculate the first noise component with a component of the non-ventricular signal X ^* (t) Fitting using the least squares method to obtain the second noise component and use calculating the atrial fibrillation signal f _i (t);

in, Q ^T represents the transposed matrix of the filter Q, and f(t) represents the atrial fibrillation signal.

The following describes a device for extracting atrial fibrillation signals provided by embodiments of the present invention. The device for extracting atrial fibrillation signals described below and the method and apparatus for extracting atrial fibrillation signals described above may be referred to each other.

Referring to FIG. 4 , a device for extracting atrial fibrillation signals provided by an embodiment of the present invention includes:

memory 401 for storing computer programs;

The processor 402 is configured to implement the steps of the method for extracting atrial fibrillation signals according to any of the foregoing embodiments when executing the computer program.

Preferably, the device may also include:

The acquisition device is used to collect the ECG mixed signal, amplify the ECG mixed signal with appropriate power, and then convert it into a digital signal;

The display device is used for displaying and playing the extracted atrial fibrillation signal.

Wherein, the memory is also used for:

Store ECG mixed signals, and data such as intermediate parameters and atrial fibrillation signals in the extraction process.

Wherein, the processor is also used for:

Coordinate and manage acquisition devices, storage devices and display devices.

Specifically, the memory is a DRR memory, the processor is a digital signal processing chip (DSP) or a programmable logic array (FPGA); the acquisition device has a sampler (such as a stethoscope), an analog-to-digital converter and an analog signal Amplifier; the display device has a speaker, a display, a transmission module and other devices, and the transmission module can be: a WIFI module, a Bluetooth module or a zigbee communication module.

The following describes a computer-readable storage medium provided by an embodiment of the present invention. The computer-readable storage medium described below and the method, apparatus, and device for extracting atrial fibrillation signals described above can be referred to each other.

A computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, implements the steps of the method for extracting atrial fibrillation signals according to any of the foregoing embodiments.

Based on any of the foregoing embodiments, it should be noted that the method for extracting atrial fibrillation signals provided in this specification can be implemented according to the following steps.

S501, the ECG mixed signal is preprocessed to obtain a mixed signal X(t);

S502, ICA is used to separate the ventricular signal and the non-ventricular signal in the ECG mixed signal, and the obtained non-ventricular signal is denoted as X ^* (t), and the calculation model is: X ^* (t)=Hf(t)+Wj( t);

The non-ventricular signal at this time is a mixed signal of atrial fibrillation signal and other interfering signals.

S503, select a suitable τ _j ∈ T _nj , so that R _J (τ _j )=0;

Since the interfering signal j(t) mixed in the non-ventricular signal is close to the Gaussian signal, the value of its autocorrelation domain tends to zero except when τ=t _max , so a suitable τ _j ∈T _nj ( Typically τ>0 and τ≠t _max ) such that R _J (τ _j )=0.

S504. Use Calculate the autocorrelation matrix of the non-ventricular signal at τ = τ _j

S505, use solve Obtain filter Q.

S506, use Filter processing: .

S507, will with one of the components of X ^* (t) Perform a least squares fit to get

S508, use Calculate the atrial fibrillation signal f _i (t).

Wherein, if the waveform diagram of the ECG mixed signal in the time domain is shown in FIG. 5 , using the atrial fibrillation signal extraction method provided in this specification, please refer to FIG. 6 for the waveform diagram of the extracted atrial fibrillation signal in the time domain. It can be seen from Figure 5 that the ventricular signal and the atrial fibrillation signal are mixed together in the time domain. As can be seen from Figure 6, the waveform of the atrial fibrillation signal is clearly visible, and the mixed noise signal has been significantly eliminated.

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

The above description of the disclosed embodiments enables any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. A method for extracting atrial fibrillation signals comprises the following steps:

acquiring an ECG mixed signal and preprocessing the ECG mixed signal;

filtering ventricular signals from the preprocessed ECG mixed signals to obtain non-ventricular signals, wherein atrial fibrillation signals and noise signals are mixed in the non-ventricular signals;

it is characterized by comprising:

generating an autocorrelation matrix of the non-ventricular signal based on separable features of the non-ventricular signal in an autocorrelation domain;

and removing the noise signal by utilizing the autocorrelation matrix and a least square method to obtain the atrial fibrillation signal.

2. The method for extracting atrial fibrillation signals according to claim 1, wherein the preprocessing the ECG mixed signals comprises:

removing Gaussian white noise in the ECG mixed signal through wavelet denoising, and removing noise of which the amplitude modulus value is smaller than a preset threshold value in the ECG mixed signal through threshold denoising.

3. The method of claim 1, wherein generating the autocorrelation matrix of the non-ventricular signal based on separable features of the non-ventricular signal in an autocorrelation domain comprises:

generating a target time period τ based on separable characteristics of the non-ventricular signal in an autocorrelation domain_jInner autocorrelation matrix

Wherein,τ_jrepresenting a time period in which the value of the noise signal in the autocorrelation domain is zero, X (t) representing the non-ventricular signal at time t, X (t- τ)_j)^TRepresents t-tau_jA transposed matrix of the non-ventricular signals at a time point, H represents a mixing matrix of the atrial fibrillation signals f (t), W represents a mixing matrix of the noise signals j (t), E {. cndot.) represents an expectation operation, and t represents a time point.

4. The method for extracting atrial fibrillation signals according to claim 3, wherein the removing the noise signals by using the autocorrelation matrix and a least square method to obtain the atrial fibrillation signals comprises:

construction ofSolving a filter Q, and filtering by using the filter Q to obtain a first noise component

Combining the first noise componentAnd the non-ventricular signal X^*One component of (t)Fitting by using the least square method to obtain a second noise componentAnd useCalculating the atrial fibrillation signal f_i(t)；

Wherein,Q^Ta transposed matrix representing the filter Q, f (t) representing the atrial fibrillation signal.

5. The method for extracting atrial fibrillation signals according to any one of claims 1 to 4, wherein after the step of removing the noise signals by using the autocorrelation matrix and the least square method to obtain the atrial fibrillation signals, the method further comprises the following steps:

and visually displaying the atrial fibrillation signals.

6. An extraction apparatus of atrial fibrillation signals, comprising:

the acquisition module is used for acquiring an ECG mixed signal and preprocessing the ECG mixed signal;

the filtering module is used for filtering ventricular signals from the preprocessed ECG mixed signals to obtain non-ventricular signals, and atrial fibrillation signals and noise signals are mixed in the non-ventricular signals;

it is characterized by comprising:

a generating module for generating an autocorrelation matrix of the non-ventricular signal based on separable features of the non-ventricular signal in an autocorrelation domain;

and the removing module is used for removing the noise signals by utilizing the autocorrelation matrix and a least square method to obtain the atrial fibrillation signals.

7. The apparatus according to claim 6, wherein the obtaining module is specifically configured to:

removing Gaussian white noise in the ECG mixed signal through wavelet denoising, and removing noise of which the amplitude modulus value is smaller than a preset threshold value in the ECG mixed signal through threshold denoising.

8. The apparatus for extracting atrial fibrillation signals according to claim 6 or 7, further comprising:

and the display module is used for visually displaying the atrial fibrillation signals.

9. An extraction apparatus of atrial fibrillation signals, comprising:

a memory for storing a computer program;

a processor for implementing the steps of the method for extracting atrial fibrillation signals according to any one of claims 1 to 5 when said computer program is executed.

10. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a computer program which, when being executed by a processor, carries out the steps of the method for extracting atrial fibrillation signals according to any one of claims 1 to 5.