EEG signal spectrum leakage suppression method and system
Through the combination of wavelet transformation and deep learning models, spectrum leakage in EEG signals is identified and suppressed, and the problem of inaccurate signal analysis in the prior art is solved, achieving more efficient signal processing and more accurate diagnostic support.
Patent Information
- Application Number
- CN202510067916.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-16
- Publication Date
- 2025-05-13
AI Technical Summary
The prior art is difficult to effectively suppress spectrum leakage when processing EEG signals, resulting in inaccurate signal analysis.
Wavelet transform is used to decompose the EEG signal into components on different scales, and a deep learning model is constructed, using the wavelet coefficient as input feature to identify and suppress the components of spectrum leakage, and finally the improved components are restored to the time domain through inverse wavelet transform.
It improves the accuracy of spectrum leakage inhibition, enhances the generalization ability of the model, and promotes the development of neuroscience research and clinical diagnosis.
Smart Images
Figure CN119988949A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of signal processing, and in particular to a method and system for suppressing EEG signal spectrum leakage. Background Art
[0002] As an important carrier reflecting brain neural activity, the spectral characteristics of electroencephalogram (EEG) signals are of great significance for neuroscience research and clinical diagnosis.
[0003] However, in practical applications, EEG signals are often interfered by spectrum leakage, resulting in inaccurate signal analysis. Spectral leakage refers to the uneven distribution of signal energy in the frequency domain, with some energy leaking to other frequency bands, thus affecting the judgment of the true spectrum characteristics of the signal.
[0004] In the prior art, the methods for suppressing EEG signal spectrum leakage mainly include windowing function, filtering, etc. However, these methods often fail to achieve ideal results when processing complex and changeable EEG signals. Although windowing function can reduce spectrum leakage, it will cause distortion of signal time domain characteristics; filtering method may filter out useful signal components, resulting in information loss.
[0005] Therefore, a more effective and accurate EEG signal spectrum leakage suppression method is needed to overcome the shortcomings of the prior art. Summary of the invention
[0006] The main purpose of the present invention is to solve the technical problem of inaccurate EEG signal spectrum leakage suppression in the prior art. A method for suppressing EEG signal spectrum leakage comprises the following steps: Acquiring EEG signals and preprocessing the EEG signals; The preprocessed EEG signal is subjected to wavelet transform to decompose it into components at different scales. The wavelet transform formula is: ; Where W(a,b) represents the wavelet transform coefficient, a represents the scale parameter, b represents the displacement parameter, and ψ(t) represents the wavelet basis function; Build a deep learning model and use the components after wavelet decomposition as input features; Apply the trained deep learning model to new EEG signal data, identify the components of spectral leakage, and output the suppressed signal; The suppressed signal is subjected to inverse wavelet transform, and the improved components in the frequency domain are restored to the time domain to obtain the processed EEG signal; the formula for inverse wavelet transform is: Where Cψ represents the admissibility condition constant of the wavelet basis function.
[0007] The step of acquiring an EEG signal and preprocessing the EEG signal comprises: The collected EEG signal is preprocessed, including denoising and filtering steps. The preprocessed signal is expressed as: Among them, x(t) represents the preprocessed signal, s(t) represents the original EEG signal, and n(t) represents the noise.
[0008] The pre-processed EEG signal is subjected to wavelet transformation to decompose it into components at different scales, including: The signal x(t) is decomposed once, that is, the signal x(t) is convolved with a low-pass filter and a high-pass filter to obtain low-frequency and high-frequency components c1(t) and d1(t).
[0009] The low-frequency component c1(t) is decomposed again by wavelet to obtain c2(t) and d2(t).
[0010] Continue decomposition: Continue to decompose the low-frequency part until the required number of decomposition layers is reached.
[0011] The steps of constructing the deep learning model are: Constructing a convolutional neural network deep learning model, extracting the corresponding wavelet coefficients for each decomposed frequency band, and using the wavelet coefficients as input features of the convolutional neural network deep learning model; Using a large amount of EEG signal data for training, and using appropriate Adam and SGD optimization algorithms to update model parameters, the training process of the deep learning model can be expressed as: Among them, θ∗ represents the optimal model parameters, L represents the loss function, yi represents the true label, f(xi;θ) represents the model prediction result, and N represents the number of samples.
[0012] The inverse wavelet transform is performed on the suppressed signal to restore the improved components in the frequency domain to the time domain to obtain the processed EEG signal, including: The low-frequency and high-frequency components obtained by wavelet transform of each layer are combined, and the low-frequency and high-frequency components of each layer are suppressed, and the decomposed components of the signal are formed after combination; For each layer of wavelet decomposition, the low-frequency components are upsampled, which means doubling the sampling rate of the signal, that is, inserting zero values into each sample; the high-frequency components are upsampled; the shape of the signal after upsampling will be restored to the same sampling frequency as the original signal.
[0013] Each layer of the wavelet transform has a pair of low-pass and high-pass filters; the inverse wavelet transform performs filtering operations on the upsampled signal through low-pass and high-pass filters; The reconstructed signals of each layer are added layer by layer to finally obtain a completely reconstructed signal.
[0014] A second aspect of the present invention provides an EEG signal spectrum leakage suppression system, comprising: An acquisition module, acquiring EEG signals and preprocessing the EEG signals; The wavelet transform module performs wavelet transform on the preprocessed EEG signal and decomposes it into components at different scales; Model building module, builds a deep learning model and uses the components after wavelet decomposition as input features; The spectrum leakage identification module applies the trained deep learning model to new EEG signal data, identifies the components of spectrum leakage, and outputs the suppressed signal; The signal suppression module performs inverse wavelet transform on the suppressed signal, restores the improved components in the frequency domain to the time domain, and obtains the processed EEG signal.
[0015] A third aspect of the present invention provides an electronic device, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via lines; the at least one processor calls the instructions in the memory so that the electronic device executes the above-mentioned EEG signal spectrum leakage suppression method.
[0016] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned EEG signal spectrum leakage suppression method.
[0017] Compared with the prior art, the present invention has the following beneficial effects: (1) Improved accuracy of spectrum leakage suppression: Combining wavelet transform with deep learning methods can more accurately identify and suppress spectrum leakage in EEG signals, thereby improving the accuracy of signal analysis.
[0018] (2) Enhanced generalization ability of the model: The deep learning model can learn complex nonlinear relationships and adapt to the spectral leakage characteristics of EEG signals in different individuals and different acquisition environments, thereby improving the generalization ability of the model.
[0019] (3) Promoted the development of neuroscience research and clinical diagnosis: The present invention provides more reliable EEG signal data for neuroscience research, which helps to reveal the mysteries of brain activity. At the same time, it also provides a more accurate basis for clinical diagnosis, helping doctors to detect diseases earlier and formulate treatment plans. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 A flow chart of a method for suppressing EEG signal spectrum leakage provided by an embodiment of the present invention; DETAILED DESCRIPTION The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0021] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the method for suppressing EEG signal spectrum leakage in the embodiment of the present invention includes: A method for suppressing EEG signal spectrum leakage comprises the following steps: 1. Acquire EEG signals and pre-process the EEG signals; The collected EEG signal is preprocessed, including denoising and filtering steps. The preprocessed signal is expressed as: Among them, x(t) represents the preprocessed signal, s(t) represents the original EEG signal, and n(t) represents the noise.
[0022] 2. Perform wavelet transform on the preprocessed EEG signal to decompose it into components at different scales. The wavelet transform formula is: ; Where W(a,b) represents the wavelet transform coefficient, a represents the scale parameter, b represents the displacement parameter, and ψ(t) represents the wavelet basis function; Perform a decomposition of the signal x(t), that is, convolve the signal x(t) with a low-pass filter and a high-pass filter to obtain low-frequency and high-frequency components c1(t) and d1(t). Perform a wavelet decomposition of the low-frequency component c1(t) again to obtain c2(t) and d2(t). Continue decomposition: Continue to decompose the low-frequency part until the required number of decomposition layers is reached.
[0023] 3. Build a deep learning model and use the components after wavelet decomposition as input features; Construct a convolutional neural network deep learning model. For each decomposed frequency band, extract its corresponding wavelet coefficients, and use the wavelet coefficients as the input features of the convolutional neural network deep learning model. Use a large amount of EEG signal data for training, and use appropriate Adam and SGD optimization algorithms to update model parameters. The training process of the deep learning model can be expressed as: Among them, θ∗ represents the optimal model parameters, L represents the loss function, yi represents the true label, f(xi;θ) represents the model prediction result, and N represents the number of samples.
[0024] 4. Apply the trained deep learning model to new EEG signal data, identify the components of spectrum leakage, and output the suppressed signal; 5. Perform inverse wavelet transform on the suppressed signal to restore the improved components in the frequency domain to the time domain to obtain the processed EEG signal; the formula for inverse wavelet transform is: Where Cψ represents the admissibility condition constant of the wavelet basis function.
[0025] The low-frequency and high-frequency components obtained by the wavelet transform of each layer are combined, and the low-frequency components and high-frequency components are suppressed for each layer, and the decomposed components of the signal are formed after the combination; for the wavelet decomposition of each layer, the low-frequency components are upsampled, and upsampling doubles the sampling rate of the signal, that is, each sample is inserted with a zero value; the high-frequency components are upsampled; the shape of the upsampled signal will be restored to the same sampling frequency as the original signal. The wavelet transform of each layer has a pair of low-pass and high-pass filters; the inverse wavelet transform filters the upsampled signal through low-pass and high-pass filters; the reconstructed signals of each layer are added layer by layer, and finally a completely reconstructed signal is obtained.
[0026] The following is a specific embodiment of the present invention: 1. Signal preprocessing: De-noise and filter the collected EEG signals to reduce the impact of noise on subsequent analysis. Common denoising methods can be used, such as wavelet denoising, Kalman filtering, etc.
[0027] 2. Wavelet transform: Select appropriate wavelet basis functions (such as Daubechies wavelet, Morlet wavelet, etc.) to perform wavelet transform on the preprocessed EEG signal and decompose it into components at different scales. Different wavelet basis functions and decomposition levels can be selected according to actual needs.
[0028] 3. Deep learning model training: Build deep learning models such as deep convolutional neural networks (CNN) or recurrent neural networks (RNN), and use a large amount of EEG signal data for training. During the training process, strategies such as cross-validation and early stopping can be used to prevent overfitting, and appropriate optimization algorithms (such as Adam, SGD, etc.) can be used to update model parameters.
[0029] 4. Spectral leakage suppression: Apply the trained deep learning model to the new EEG signal data, identify the components of spectrum leakage, and output the suppressed signal. The signal can be restored in the time domain by performing an inverse wavelet transform on the model output.
[0030] 5. Post-processing and verification: Perform post-processing steps such as smoothing and denoising on the suppressed signal to improve the signal quality. At the same time, verify the suppression effect by comparing it with the actual EEG signal data. Common evaluation indicators (such as signal-to-noise ratio, mean square error, etc.) can be used to evaluate the suppression effect.
[0031] The above describes the EEG signal spectrum leakage suppression method in the embodiment of the present invention. The following describes the EEG signal spectrum leakage suppression device in the embodiment of the present invention: An acquisition module, acquiring EEG signals and preprocessing the EEG signals; The wavelet transform module performs wavelet transform on the preprocessed EEG signal and decomposes it into components at different scales; Model building module, builds a deep learning model and uses the components after wavelet decomposition as input features; The spectrum leakage identification module applies the trained deep learning model to new EEG signal data, identifies the components of spectrum leakage, and outputs the suppressed signal; The signal suppression module performs inverse wavelet transform on the suppressed signal, restores the improved components in the frequency domain to the time domain, and obtains the processed EEG signal.
[0032] An embodiment of the present invention also provides an electronic device, which may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) (for example, one or more processors) and memories, and one or more storage media for storing applications or data (for example, one or more mass storage devices). Among them, the memory and the storage medium may be short-term storage or permanent storage. The program stored in the storage medium may include one or more modules, each module may include a series of instruction operations in the electronic device. Furthermore, the processor may be configured to communicate with the storage medium and execute a series of instruction operations in the storage medium on the electronic device.
[0033] The electronic device may also include one or more power supplies, one or more wired or wireless network interfaces, one or more input and output interfaces, and / or one or more operating systems, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that the electronic device structure in this embodiment does not constitute a limitation on the electronic device, and may include more or fewer components, or combine certain components, or arrange components differently.
[0034] An embodiment of the present invention provides a structure of an electronic device, which may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) (for example, one or more processors) and memories, and one or more storage media for storing applications or data (for example, one or more mass storage devices). Among them, the memory and the storage medium can be short-term storage or permanent storage. The program stored in the storage medium may include one or more modules, each module may include a series of instruction operations in the electronic device. Furthermore, the processor can be configured to communicate with the storage medium and execute a series of instruction operations in the storage medium on the electronic device.
[0035] The electronic device may also include one or more power supplies, one or more wired or wireless network interfaces, one or more input and output interfaces, and / or one or more operating systems, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that the electronic device structure does not constitute a limitation on the electronic device, and may include more or less components than the aforementioned, or combine certain components, or arrange the components differently.
[0036] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer executes the steps of the aforementioned method.
[0037] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.
[0038] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.
[0039] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for suppressing spectrum leakage of an EEG signal, characterized in that: The following steps are involved: Acquiring EEG signals and preprocessing the EEG signals; The preprocessed EEG signal is subjected to wavelet transform to decompose it into components at different scales. The wavelet transform formula is: ； Where W(a,b) represents the wavelet transform coefficient, a represents the scale parameter, b represents the displacement parameter, and ψ(t) represents the wavelet basis function; Build a deep learning model and use the components after wavelet decomposition as input features; Apply the trained deep learning model to new EEG signal data, identify the components of spectral leakage, and output the suppressed signal; The suppressed signal is subjected to inverse wavelet transform, and the improved components in the frequency domain are restored to the time domain to obtain the processed EEG signal; the formula for inverse wavelet transform is: Where Cψ represents the admissibility condition constant of the wavelet basis function.
2. The method for suppressing EEG signal spectrum leakage according to claim 1, characterized in that: The step of acquiring an EEG signal and preprocessing the EEG signal comprises: The collected EEG signal is preprocessed, including denoising and filtering steps. The preprocessed signal is expressed as: Among them, x(t) represents the preprocessed signal, s(t) represents the original EEG signal, and n(t) represents the noise.
3. The method for suppressing EEG signal spectrum leakage according to claim 1, characterized in that: The pre-processed EEG signal is subjected to wavelet transformation to decompose it into components at different scales, including: The signal x(t) is decomposed once, that is, the signal x(t) is convolved with a low-pass filter and a high-pass filter to obtain low-frequency and high-frequency components c1(t) and d1(t). The low-frequency component c1(t) is decomposed again by wavelet to obtain c2(t) and d2(t). Continue decomposition: Continue to decompose the low-frequency part until the required number of decomposition layers is reached.
4. The method for suppressing EEG signal spectrum leakage according to claim 1, characterized in that: The steps of constructing the deep learning model are: Constructing a convolutional neural network deep learning model, extracting the corresponding wavelet coefficients for each decomposed frequency band, and using the wavelet coefficients as input features of the convolutional neural network deep learning model; Using a large amount of EEG signal data for training, and using appropriate Adam and SGD optimization algorithms to update model parameters, the training process of the deep learning model can be expressed as: Among them, θ∗ represents the optimal model parameters, L represents the loss function, yi represents the true label, f(xi;θ) represents the model prediction result, and N represents the number of samples.
5. The method for suppressing EEG signal spectrum leakage according to claim 1, characterized in that: The inverse wavelet transform is performed on the suppressed signal to restore the improved components in the frequency domain to the time domain to obtain the processed EEG signal, including: The low-frequency and high-frequency components obtained by wavelet transform of each layer are combined, and the low-frequency and high-frequency components of each layer are suppressed, and the decomposed components of the signal are formed after combination; For each layer of wavelet decomposition, the low-frequency components are upsampled, which means doubling the sampling rate of the signal, that is, inserting zero values into each sample; the high-frequency components are upsampled; the shape of the signal after upsampling will be restored to the same sampling frequency as the original signal. Each layer of wavelet transform has a pair of low-pass and high-pass filters; the inverse wavelet transform performs filtering operations on the upsampled signal through low-pass and high-pass filters; The reconstructed signals of each layer are added layer by layer to finally obtain a completely reconstructed signal.
6. An EEG signal spectrum leakage suppression system, characterized in that: The system comprises: An acquisition module, acquiring EEG signals and preprocessing the EEG signals; The wavelet transform module performs wavelet transform on the preprocessed EEG signal and decomposes it into components at different scales; Model building module, builds a deep learning model and uses the components after wavelet decomposition as input features; The spectrum leakage identification module applies the trained deep learning model to new EEG signal data, identifies the components of spectrum leakage, and outputs the suppressed signal; The signal suppression module performs inverse wavelet transform on the suppressed signal, restores the improved components in the frequency domain to the time domain, and obtains the processed EEG signal.
7. An electronic device, comprising a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory to enable the electronic device to execute each step of the EEG signal spectrum leakage suppression method as described in any one of claims 1-5.
8. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the method for suppressing EEG signal spectrum leakage as described in any one of claims 1 to 5 are implemented.