CN111543984B - Method for removing ocular artifacts of electroencephalogram signals based on SSDA - Google Patents

Method for removing ocular artifacts of electroencephalogram signals based on SSDA Download PDF

Info

Publication number
CN111543984B
CN111543984B CN202010285650.0A CN202010285650A CN111543984B CN 111543984 B CN111543984 B CN 111543984B CN 202010285650 A CN202010285650 A CN 202010285650A CN 111543984 B CN111543984 B CN 111543984B
Authority
CN
China
Prior art keywords
eeg
ssda
representing
model
signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010285650.0A
Other languages
Chinese (zh)
Other versions
CN111543984A (en
Inventor
胡章芳
刘鹏飞
罗元
张毅
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN202010285650.0A priority Critical patent/CN111543984B/en
Publication of CN111543984A publication Critical patent/CN111543984A/en
Application granted granted Critical
Publication of CN111543984B publication Critical patent/CN111543984B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing

Abstract

The invention discloses an SSDA-based electroencephalogram signal ocular artifact removal method, which comprises the following steps: s1, carrying out normalization processing on the pure electroencephalogram signals, and inputting the signals into an SSDA model; s2, continuously adjusting model parameters without training SSDA according to the minimum error between the reconstructed EEG signal and the pure EEG signal; s3, normalizing the EEG signals containing the electro-oculogram (EOG) artifact, inputting the trained SSDA model, and performing anti-normalization processing on the output data to obtain the EEG without the EOG. The method can avoid using the ocular artifacts as reference signals in the process of removing the ocular artifacts, and ensure the effectiveness and the real-time property when removing the ocular artifacts.

Description

Method for removing ocular artifacts of electroencephalogram signals based on SSDA (steady state data acquisition)
Technical Field
The invention belongs to the field of brain electrical signal processing in a brain-computer interface, and particularly relates to an SSDA (steady state digital-analog converter) -based method for removing ocular artifacts of brain electrical signals.
Background
Currently, a non-invasive research mode is generally adopted in brain-computer interface research, but the non-invasive research mode has certain disadvantages, and due to the fact that the electrodes are directly pasted on the surface of a cerebral cortex, in the process of signal acquisition, a subject can blink, and electro-oculogram (EOG) artifacts can be brought. The electrical signal generated by brain activity is weak, and the existence of artifacts can cover up the real electroencephalogram signal, so that the quality of the electroencephalogram signal is sharply reduced. Research shows that certain overlap exists between the frequency spectrums of the ocular artifacts and the electroencephalogram signals in certain frequency ranges, and the ocular artifacts and the electroencephalogram signals cannot be removed directly through a filter. Therefore, how to effectively remove the ocular artifacts from the electroencephalogram signals is a key step of electroencephalogram signal analysis. At present, blind source separation methods such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA) are often applied to ocular artifact removal, but the methods need to analyze a large number of electroencephalogram channels, and the signal separation process is time-consuming. In addition, blind source separation is easy to cause the loss of effective components of the electroencephalogram signals and increase the similarity among all channels. With the development of a neural network, an FLN-RBF algorithm and an FLNN-ANFIS are proposed, the two algorithms utilize good approximation performance of the neural network, and a good ocular artifact removal effect is achieved, but the two algorithms need additional ocular electrodes to acquire ocular signals to serve as reference signals of the network, and the subsequent integrated application is not facilitated. The invention considers the strong learning ability and signal reconstruction ability of self-coding, and uses the pure brain electrical signal to train the model, and can learn the characteristics of the pure brain electrical signal, so the pure brain electrical signal can be reconstructed from the brain electrical signal containing the ocular artifacts without using extra ocular electrical signal as a reference signal. In order to reduce the complexity of self-coding, the invention further improves the model and provides a stack-type sparse denoising self-coding SSDA model.
Disclosure of Invention
The present invention is directed to solving the above problems of the prior art. An eye electrical artifact removing method based on an SSDA electroencephalogram signal is provided. The technical scheme of the invention is as follows:
an ocular artifact removal method for brain electrical signals based on SSDA comprises the following steps:
and S1, in an off-line stage, taking the pure electroencephalogram signal as a training set, carrying out normalization processing, inputting a stacked sparse denoising self-coding SSDA model for training, carrying out pre-training on the stacked sparse denoising self-coding SSDA model, wherein the SSDA is composed of two sparse denoising self-coding (SDAs) in an end-to-end connection mode, the output of the first SDA is the input of the second SDA, and the output of the second SDA is subjected to inverse normalization processing, so that the reconstructed electroencephalogram signal is obtained.
S2, obtaining an error between the reconstructed EEG signal and the pure EEG signal to minimize the error, continuously training SSDA, and finely adjusting the model parameters according to a gradient descent method;
and S3, in an online stage, acquiring the electroencephalogram signals containing the ocular artifacts, normalizing the electroencephalogram signals, inputting the normalized EEG signals into the SSDA model trained in the step S2, and performing inverse normalization processing on output data to obtain the electroencephalogram signals without the ocular artifacts.
Further, the calculation formula of the normalization of the pure electroencephalogram signal in the step S1 is as follows:
Figure GDA0003663202780000021
in the middle, EEGstd(i) Represents the value after normalization, i represents the number of electroencephalogram sampling points, j ∈ 1,2orgRepresenting the original value before normalization.
Further, in step S2, the SSDA is trained according to the error between the reconstructed EEG signal and the clean electroencephalogram signal, which is as follows:
the pre-training process comprises the following steps:
(1) randomly zeroing the normalized EEG signal according to a proportion to destroy the normalized EEG signal to obtain
Figure GDA0003663202780000022
Represents one sample, n represents the number of samples;
(2) carrying out random initialization on w and b, and obtaining the mapping h of the first hidden layer of the model according to the formulas (3) to (5)(1)；
h＝f(wx+b) (3)
Figure GDA0003663202780000023
Figure GDA0003663202780000031
In the formulas (3) to (5), h represents the value of the hidden layer, x represents the electroencephalogram sequence, and JDAE(w, b) represents a loss function of sparse noise reduction self-coding,
Figure GDA0003663202780000032
f (-) and g (-) represent mapping functions for encoding and decoding, respectively, and are usually non-linear, w represents a weight matrix between the input layer and the hidden layer, wTRepresenting the weight matrix between the hidden layer and the output layer, b and b' representing the bias vectors of the hidden layer and the output layer, respectively, n representing the number of samples of the input,
Figure GDA0003663202780000033
representing contaminated input data, hiRepresenting hidden layer vectors, w and b represent weight and offset vectors respectively;
(3) for hidden layer h(1)Performing thinning treatment to obtain h according to the formulas (6) and (7)(1)And determines model parameters w1,b1Finishing the training of the first SDA;
Figure GDA0003663202780000034
Figure GDA0003663202780000035
in the formula, beta is the weight of sparse penalty factor, s2The number of hidden layer neurons after the sparse layer, KL is relative entropy, rho is sparse parameter,
Figure GDA0003663202780000036
Average activation of training set, lambda is regularization parameter weight,
Figure GDA0003663202780000037
As weight, s, between the input layer and the hidden layerlThe number of hidden layer neurons after regularization. EEG (electroencephalogram)nstd(i) For de-normalised EEG signals, EEGorg(i) And EEGout(i) Respectively representing signals input and output by the model, and n represents the number of signal sampling points;
(4) will hide the layer h(1)Is used as the input of the second SDA, and the value of (w) trained by the last SDA is used1,b1Replace the random parameter, w1、b1Respectively representing the weight and offset value between the input layer and the output layer of the first SDA, and repeating steps (2) and (3) to determine the output and parameter { w } of the second SDA2,b2}，w2、b2Representing the weights and offset values between the second SDA input layer and the hidden layer, respectively. So far, the two SDAs in the model are trained, namely the pre-training process of the SSDA is completed, and then the SSDA is finely adjusted to enable the parameter value to be optimal on the whole network.
Further, the fine tuning process specifically includes:
(1) for Δ w1And Δ b1Initialization is performed, Δ w1、Δb1Delta values representing the weight and bias between the first SDA input layer and the hidden layer, respectively. Let Δ w1＝0，Δb1＝0；
(2) Calculated by back propagation algorithm
Figure GDA0003663202780000041
And
Figure GDA0003663202780000042
and
Figure GDA0003663202780000043
are respectively indicated
Figure GDA0003663202780000044
And
Figure GDA0003663202780000045
contrast dispersion values representing weight values and bias values, respectively;
(3) order to
Figure GDA0003663202780000046
Δwl、ΔblDelta values representing the weights and biases of the ith SDA input layer and hidden layer, respectively, where l ∈ {1,2 };
(4) order to
Figure GDA0003663202780000047
(5) The parameters are updated in such a way that,
Figure GDA0003663202780000048
where α represents the learning rate.
At this point, the fine tuning process of the SSDA model is completed, that is, the training process of the entire model is completed, and the parameters at this time are optimized on the entire model.
Further, in step S3, the EEG signal containing the ocular electrical EOG artifact is normalized, the trained SSDA model is input, and the output data is subjected to the inverse normalization processing, so as to obtain the EEG with EOG removed, specifically as follows:
carrying out normalization processing on the EEG containing the EOG artifact, wherein the normalization calculation is shown as a formula (8):
Figure GDA0003663202780000049
in the formula, EEGstd(i) Representing values after normalization, EEGorgRepresenting the original value before normalization;
inputting the normalized signal into the trained SSDA model, and then performing inverse normalization processing on the output value of the model to obtain the EEG signal without EOG, wherein the process of inverse normalization calculation is shown as formula (9):
Figure GDA0003663202780000051
in the formula, EEGnstd(i) Representing de-normalized EEG signals, i.e. EEG signals after removal of ocular artefacts, EEGinAnd EEGout(i) Respectively representing signals input and output by the model, i represents the number of electroencephalogram signal sampling points, and j belongs to 1, 2.
The invention has the following advantages and beneficial effects:
the invention provides an SSDA-based electroencephalogram signal ocular artifact removal method, which can learn the characteristics of a pure electroencephalogram signal through strong learning capability and signal reconstruction capability of a self-coding network on the premise of not using an ocular artifact signal as a reference signal, further reconstruct a pure electroencephalogram signal from an electroencephalogram signal containing the ocular artifact, achieve the purpose of removing the ocular artifact and remove the ocular artifact of the electroencephalogram signal of any channel. The method comprises the following specific steps: firstly, taking a pure brain electrical signal as a training set, carrying out normalization processing, inputting a stack-type sparse denoising self-coding (SSDA) model and carrying out pre-training, wherein the SSDA model consists of two sparse denoising self-coding Systems (SDAs) in a head-to-tail connection mode, the output of the first SDA is the input of the second SDA, and the output of the second SDA is subjected to inverse normalization processing, so that the reconstructed brain electrical signal is obtained. And secondly, acquiring an error between the reconstructed electroencephalogram signal and the pure electroencephalogram signal to minimize the error, continuously training the SSDA, and finely adjusting the model parameters according to a gradient descent method to finish the training of the SSDA model. And finally, normalizing the electroencephalogram signals containing the ocular artifacts, inputting the normalized electroencephalogram signals into the trained SSDA model, and performing inverse normalization processing on output data to obtain the electroencephalogram signals without the ocular artifacts. Compared with other methods, the method can not only reduce the time for removing the ocular artifacts in the electroencephalogram signal, but also improve the signal-to-noise ratio of the electroencephalogram signal.
Drawings
FIG. 1 is a block diagram of an SSDA model provided in a preferred embodiment of the present invention;
FIG. 2 is a flow chart of a method for removing ocular artifacts based on SSDA electroencephalogram signals provided by the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described in detail and clearly with reference to the accompanying drawings. The described embodiments are only some of the embodiments of the present invention.
The technical scheme for solving the technical problems is as follows:
an ocular artifact removal method for brain electrical signals based on SSDA comprises the following steps:
and S1, taking the pure electroencephalogram signal as a training set, carrying out normalization processing, and inputting the signal into the SSDA model.
Figure GDA0003663202780000061
In the middle, EEGstd(i) Represents the value after normalization, i represents the number of electroencephalogram sampling points, j ∈ 1,2orgRepresenting the original value before normalization.
And S2, continuously adjusting model parameters without training SSDA according to the minimum error between the reconstructed EEG signal and the pure EEG signal.
The pre-training process comprises the following steps:
(1) randomly zeroing and destroying the normalized EEG signal according to a certain proportion to obtain
Figure GDA0003663202780000062
Represents one sample, n represents the number of samples;
(2) performing random initialization on w and b, and obtaining the mapping h of the first hidden layer of the model according to the formulas (19) - (21)(1)。
h＝f(wx+b) (19)
Figure GDA0003663202780000063
Figure GDA0003663202780000071
In the formulae (19) to (21), h represents the value of the hidden layer, x represents the electroencephalogram sequence, and JDAE(w, b) represents a loss function of sparse noise reduction self-coding,
Figure GDA0003663202780000072
f (-) and g (-) represent mapping functions for encoding and decoding, respectively, and are usually non-linear, w represents a weight matrix between the input layer and the hidden layer, wTRepresenting a weight matrix between the hidden layer and the output layer, b and b' representing offset vectors of the hidden layer and the output layer, respectively, n representing the number of samples of the input,
Figure GDA0003663202780000073
representing contaminated input data, hiRepresenting hidden layer vectors, w and b represent weight and offset vectors respectively;
(3) for hidden layer h(1)Performing thinning treatment to obtain h according to the formulas (22) and (23)(1)And determines model parameters w1,b1To complete the training of the first SDA.
Figure GDA0003663202780000074
Figure GDA0003663202780000075
In the formula, beta is the weight of sparse penalty factor, s2The number of hidden layer neurons after the sparse layer, KL is relative entropy, rho is sparse parameter,
Figure GDA0003663202780000076
Average activation of training set, lambda is regularization parameter weight,
Figure GDA0003663202780000077
As weight, s, between the input layer and the hidden layerlThe number of hidden layer neurons after regularization. EEG (electroencephalogram)nstd(i) For inverse normalized EEG signals, EEGorg(i) And EEGout(i) Respectively representing the input and output signals of the model, and the number of n-generation samples;
(4) will hide the layer h(1)Is used as the input of the second SDA, and the value of (w) trained by the last SDA is used1,b1Replacing the random parameter, repeating steps (2) and (3) to determine the output of the second SDA and the parameter { w }2,b2}. At this point, the two SDAs in the model are trained, and the pre-training process of the SSDA is completed. The SSDA is then fine-tuned to optimize the parameter values throughout the network.
And (3) fine adjustment process:
(1) for Δ w1And Δ b1Initialization is performed, Δ w1、Δb1Delta values representing the weight and bias between the first SDA input layer and the hidden layer, respectively. Let Δ w1＝0，Δb1＝0；
(2) Calculated by back propagation algorithm
Figure GDA0003663202780000081
And
Figure GDA0003663202780000082
and
Figure GDA0003663202780000083
are respectively indicated
Figure GDA0003663202780000084
And
Figure GDA0003663202780000085
contrast dispersion values representing weight values and bias values, respectively;
(3) order to
Figure GDA0003663202780000086
Δwl、ΔblDelta values representing the weights and biases of the ith SDA input layer and hidden layer, respectively, where l ∈ {1,2 };
(4) order to
Figure GDA0003663202780000087
(5) The parameters are updated in such a way that,
Figure GDA0003663202780000088
where α represents the learning rate.
At this point, the fine tuning process of the SSDA model is completed, that is, the training process of the entire model is completed, and the parameters at this time are optimized on the entire model.
S3, the EEG signal containing the electro-oculogram (EOG) artifact is normalized, the trained SSDA model is input, and the output data is processed by inverse normalization, so that the EEG without the EOG is obtained.
Figure GDA0003663202780000089
In the formula, EEGstd(i) Representing values after normalization, EEGorgRepresenting the original value before normalization.
And inputting the normalized signal into the SSDA model after training, and then performing inverse normalization processing on the output value of the model to obtain the EEG signal after EOG removal. The process of the denormalization calculation is shown in equation (25):
Figure GDA0003663202780000091
in the formula, EEGnstd(i) Representing de-normalized EEG signals, i.e. EEG signals after removal of ocular artefacts, EEGorg(i) And EEGout(i) Respectively representing signals input and output by the model, i represents the number of electroencephalogram signal sampling points, and j belongs to 1, 2.
The above examples are to be construed as merely illustrative and not limitative of the remainder of the disclosure. After reading the description of the invention, the skilled person can make various changes or modifications to the invention, and these equivalent changes and modifications also fall into the scope of the invention defined by the claims.

Claims (4)

1. An SSDA-based electroencephalogram signal ocular artifact removal method is characterized by comprising the following steps:
s1, in an off-line stage, taking a pure electroencephalogram signal as a training set, carrying out normalization processing, inputting a stack-type sparse denoising self-coding SSDA model and carrying out pre-training, wherein the SSDA is composed of two sparse denoising self-coding SDAs (self-coding data acquisition systems) in an end-to-end connection mode, the output of the first SDA is the input of the second SDA, and the output of the second SDA is subjected to inverse normalization processing to obtain a reconstructed electroencephalogram signal;
s2, obtaining an error between the reconstructed EEG electroencephalogram signal and the pure electroencephalogram signal to minimize the error, continuously training SSDA, and finely adjusting the model parameters according to a gradient descent method;
s3, in an online stage, acquiring an electroencephalogram signal containing ocular artifacts, normalizing the electroencephalogram signal, inputting the normalized electroencephalogram signal into the SSDA model trained in the step S2, and performing inverse normalization processing on output data to obtain an electroencephalogram signal without the ocular artifacts;
in step S2, the SSDA is trained according to the error between the reconstructed EEG signal and the clean EEG signal, which is as follows:
the pre-training process comprises the following steps:
(1) randomly zeroing the normalized EEG signal according to a proportion to destroy the normalized EEG signal to obtain
Figure FDA0003663202770000011
Figure FDA0003663202770000012
Represents one sample, n represents the number of samples;
(2) putting w and b inLine random initialization, and obtaining the mapping h of the first hidden layer of the model according to the formulas (3) to (5)(1)；
h＝f(wx+b) (3)
Figure FDA0003663202770000013
Figure FDA0003663202770000014
In the formulas (3) to (5), h represents the value of the hidden layer, x represents the electroencephalogram sequence, and JDAE(w, b) represents a loss function of sparse noise reduction self-coding,
Figure FDA0003663202770000015
f (-) and g (-) represent mapping functions for encoding and decoding, respectively, and are usually non-linear, w represents a weight matrix between the input layer and the hidden layer, wTRepresenting the weight matrix between the hidden layer and the output layer, b and b' representing the bias vectors of the hidden layer and the output layer, respectively, n representing the number of samples of the input,
Figure FDA0003663202770000021
representing contaminated input data, hiRepresenting hidden layer vectors, w and b represent weight and offset vectors respectively;
(3) for hidden layer h(1)Performing thinning treatment to obtain h according to the formulas (6) and (7)(1)And determines model parameters w1,b1Finishing the training of the first SDA;
Figure FDA0003663202770000022
Figure FDA0003663202770000023
in the formula, beta is the weight of sparse penalty factor, s2The number of hidden layer neurons after sparseness, KL is relative entropy, rho is a sparse parameter,
Figure FDA0003663202770000024
Average activation of training set, lambda is regularization parameter weight,
Figure FDA0003663202770000025
As weight, s, between the input layer and the hidden layerlEEG for the number of hidden layer neurons after regularizationnstd(i) For inverse normalized EEG signals, EEGorg(i) And EEGout(i) Respectively representing signals input and output by the model, and n represents the number of signal samples;
(4) will hide the layer h(1)Is used as the input of the second SDA, and the value of (w) trained by the last SDA is used1,b1Replace the random parameter, w1、b1Respectively representing the weight and offset value between the input layer and the output layer of the first SDA, and repeating steps (2) and (3) to determine the output and parameter { w } of the second SDA2,b2}，w2、b2And respectively representing the weight and the offset value between the second SDA input layer and the hidden layer, so far, finishing the pre-training process of the SSDA after the two SDAs in the model are trained, and then finely adjusting the SSDA to ensure that the parameter value is optimal on the whole network.
2. The SSDA-based method for removing ocular artifacts from electroencephalogram signals according to claim 1, wherein the pure electroencephalogram signal normalization in step S1 is calculated by the following formula:
Figure FDA0003663202770000031
in the formula, EEGstd(i) Representing the value after normalization, i representing the number of electroencephalogram sampling points, j ∈1,2,...,i，EEGorgRepresenting the original value before normalization.
3. The method for removing ocular artifacts of brain electrical signals based on SSDA according to claim 1, wherein the fine tuning process specifically comprises:
(1) for Δ w1And Δ b1Initialization is performed, Δ w1、Δb1Incremental values representing the weight and offset between the first SDA input layer and the hidden layer, respectively, let Δ w1＝0，Δb1＝0；
(2) Calculated by back propagation algorithm
Figure FDA0003663202770000032
And
Figure FDA0003663202770000033
Figure FDA0003663202770000034
and
Figure FDA0003663202770000035
are respectively indicated
Figure FDA0003663202770000036
And
Figure FDA0003663202770000037
Figure FDA0003663202770000038
contrast dispersion values representing weight values and bias values, respectively; j. the design is a squareSSDARepresenting a stack type sparse denoising self-coding SSDA model;
(3) order to
Figure FDA0003663202770000039
Δwl、ΔblRepresenting the ith SDA input layer and hidden layer weights and offsets, respectivelyIs determined, where l ∈ {1,2 };
(4) order to
Figure FDA00036632027700000310
(5) The parameters are updated in such a way that,
Figure FDA00036632027700000311
wherein α represents a learning rate;
at this point, the fine tuning process of the SSDA model is completed, that is, the training process of the entire model is completed, and the parameters at this time are optimized on the entire model.
4. The method for removing ocular artifacts of electroencephalogram signals based on SSDA according to claim 1, wherein the EEG signals containing ocular EOG artifacts in step S3 are normalized, the trained SSDA model is input, and the output data is subjected to the inverse normalization processing, so as to obtain the EEG with EOG removed, specifically as follows:
carrying out normalization processing on the EEG containing the EOG artifact, wherein the normalization calculation is shown as a formula (8):
Figure FDA0003663202770000041
in the formula, EEGstd(i) Representing values after normalization, EEGorgRepresenting the original value before normalization;
inputting the normalized signal into the trained SSDA model, and then performing inverse normalization processing on the output value of the model to obtain the EEG signal without EOG, wherein the process of inverse normalization calculation is shown as formula (9):
Figure FDA0003663202770000042
in the formula, EEGnstd(i) Representing de-normalized EEG signals, i.e. after removal of ocular artefactsEEG signal of (1), EEGinAnd EEGout(i) Respectively representing signals input and output by the model, i represents the number of electroencephalogram signal sampling points, and j belongs to 1, 2.
CN202010285650.0A 2020-04-13 2020-04-13 Method for removing ocular artifacts of electroencephalogram signals based on SSDA Active CN111543984B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010285650.0A CN111543984B (en) 2020-04-13 2020-04-13 Method for removing ocular artifacts of electroencephalogram signals based on SSDA

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010285650.0A CN111543984B (en) 2020-04-13 2020-04-13 Method for removing ocular artifacts of electroencephalogram signals based on SSDA

Publications (2)

Publication Number Publication Date
CN111543984A CN111543984A (en) 2020-08-18
CN111543984B true CN111543984B (en) 2022-07-01

Family

ID=71998148

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010285650.0A Active CN111543984B (en) 2020-04-13 2020-04-13 Method for removing ocular artifacts of electroencephalogram signals based on SSDA

Country Status (1)

Country Link
CN (1) CN111543984B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112245756A (en) * 2020-10-16 2021-01-22 郑州大学 Attention training method based on single-channel electroencephalogram
CN112472107B (en) * 2020-12-09 2022-04-19 中国科学技术大学 Electroencephalogram artifact removing method and device
CN114081503A (en) * 2021-11-18 2022-02-25 江苏科技大学 Method for removing ocular artifacts in electroencephalogram signals

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101474070A (en) * 2009-01-21 2009-07-08 电子科技大学 Method for removing ocular artifacts in brain-electrical signal
CN104473635A (en) * 2014-12-16 2015-04-01 重庆邮电大学 Left-right hand motor imagery electroencephalogram characteristic extraction method mixing wavelet and common spatial pattern
WO2016154298A1 (en) * 2015-03-23 2016-09-29 Temple University-Of The Commonwealth System Of Higher Education System and method for automatic interpretation of eeg signals using a deep learning statistical model
CN106529476A (en) * 2016-11-11 2017-03-22 重庆邮电大学 Deep stack network-based electroencephalogram signal feature extraction and classification method
US10003483B1 (en) * 2017-05-03 2018-06-19 The United States Of America, As Represented By The Secretary Of The Navy Biologically inspired methods and systems for automatically determining the modulation types of radio signals using stacked de-noising autoencoders
CN108734391A (en) * 2018-05-08 2018-11-02 重庆大学 Electric-gas integrated energy system probability energy flow computational methods based on storehouse noise reduction autocoder
US10147442B1 (en) * 2015-09-29 2018-12-04 Amazon Technologies, Inc. Robust neural network acoustic model with side task prediction of reference signals
CN109559281A (en) * 2017-09-26 2019-04-02 三星电子株式会社 Image denoising neural network framework and its training method
CN209063895U (en) * 2018-12-03 2019-07-05 安阳师范学院 Tired driver combined of multi-sensor information drives early warning and pro-active intervention system
CN109977810A (en) * 2019-03-12 2019-07-05 北京工业大学 Brain electricity classification method based on HELM and combination PTSNE and LDA Fusion Features
CN110353672A (en) * 2019-07-15 2019-10-22 西安邮电大学 Eye artefact removal system and minimizing technology in a kind of EEG signals
CN110490816A (en) * 2019-07-15 2019-11-22 哈尔滨工程大学 A kind of underwater Heterogeneous Information data noise reduction

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090062680A1 (en) * 2007-09-04 2009-03-05 Brain Train Artifact detection and correction system for electroencephalograph neurofeedback training methodology
CN109620218A (en) * 2019-01-29 2019-04-16 杭州妞诺科技有限公司 Brain wave intelligence screening method and system

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101474070A (en) * 2009-01-21 2009-07-08 电子科技大学 Method for removing ocular artifacts in brain-electrical signal
CN104473635A (en) * 2014-12-16 2015-04-01 重庆邮电大学 Left-right hand motor imagery electroencephalogram characteristic extraction method mixing wavelet and common spatial pattern
WO2016154298A1 (en) * 2015-03-23 2016-09-29 Temple University-Of The Commonwealth System Of Higher Education System and method for automatic interpretation of eeg signals using a deep learning statistical model
US10147442B1 (en) * 2015-09-29 2018-12-04 Amazon Technologies, Inc. Robust neural network acoustic model with side task prediction of reference signals
CN106529476A (en) * 2016-11-11 2017-03-22 重庆邮电大学 Deep stack network-based electroencephalogram signal feature extraction and classification method
US10003483B1 (en) * 2017-05-03 2018-06-19 The United States Of America, As Represented By The Secretary Of The Navy Biologically inspired methods and systems for automatically determining the modulation types of radio signals using stacked de-noising autoencoders
CN109559281A (en) * 2017-09-26 2019-04-02 三星电子株式会社 Image denoising neural network framework and its training method
CN108734391A (en) * 2018-05-08 2018-11-02 重庆大学 Electric-gas integrated energy system probability energy flow computational methods based on storehouse noise reduction autocoder
CN209063895U (en) * 2018-12-03 2019-07-05 安阳师范学院 Tired driver combined of multi-sensor information drives early warning and pro-active intervention system
CN109977810A (en) * 2019-03-12 2019-07-05 北京工业大学 Brain electricity classification method based on HELM and combination PTSNE and LDA Fusion Features
CN110353672A (en) * 2019-07-15 2019-10-22 西安邮电大学 Eye artefact removal system and minimizing technology in a kind of EEG signals
CN110490816A (en) * 2019-07-15 2019-11-22 哈尔滨工程大学 A kind of underwater Heterogeneous Information data noise reduction

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Motor imagery EEG feature extraction method based on multi-feature fusion;Liu Pengfei;《Journal of Computer Applications》;20200210;全文 *
Removing EOG artifacts from EEG signal using noise-assisted multivariate empirical mode decomposition;Zahan, S;《2016 2nd International Conference on Electrical, Computer & Telecommunication Engineering (ICECTE). Proceedings》;20170413;全文 *
多特征融合的运动想象脑电特征提取方法;刘鹏飞;《计算机应用》;20190919;第616-620页 *
脑电信号中眼电伪迹自动识别与去除方法研究;李佳庆;《计算机工程与应用》;20170824;第148-152页 *

Also Published As

Publication number Publication date
CN111543984A (en) 2020-08-18

Similar Documents

Publication Publication Date Title
Sun et al. A novel end-to-end 1D-ResCNN model to remove artifact from EEG signals
CN109784242B (en) Electroencephalogram signal denoising method based on one-dimensional residual convolution neural network
CN111543984B (en) Method for removing ocular artifacts of electroencephalogram signals based on SSDA
Peng et al. Removal of ocular artifacts in EEG—An improved approach combining DWT and ANC for portable applications
CN110353672B (en) System and method for removing eye artifacts in electroencephalogram signals
Azzerboni et al. Neural-ICA and wavelet transform for artifacts removal in surface EMG
Li et al. Automatic removal of ocular artifact from EEG with DWT and ICA Method
CN111657936B (en) Signal denoising method based on wavelet transformation and total variation regularization
CN114648048B (en) Electrocardiosignal noise reduction method based on variational self-coding and PixelCNN model
CN114224360A (en) EEG signal processing method and equipment based on improved EMD-ICA and storage medium
CN106805969B (en) Electroencephalogram relaxation degree identification method and device based on Kalman filtering and wavelet transformation
Li et al. Signal denoising based on the Schrödinger operator's eigenspectrum and a curvature constraint
CN113723171A (en) Electroencephalogram signal denoising method based on residual error generation countermeasure network
CN117064405A (en) Single-channel electroencephalogram signal artifact removal method, equipment and medium
CN112861625A (en) Method for determining stacking denoising autoencoder model
CN117235576A (en) Method for classifying motor imagery electroencephalogram intentions based on Riemann space
CN114757236B (en) Electroencephalogram signal denoising optimization method and system based on TQWT and SVMD
Akhtar et al. Focal artifact removal from ongoing EEG–a hybrid approach based on spatially-constrained ICA and wavelet de-noising
Bhattacharyya et al. Ocular artifacts elimination from multivariate EEG signal using frequency-spatial filtering
CN115462803A (en) BG-Attention-based electroencephalogram signal denoising method, device and storage medium
CN114947883A (en) Time-frequency domain information fusion deep learning electroencephalogram noise reduction method
CN111493864B (en) EEG signal mixed noise processing method, equipment and storage medium
Zhao et al. GTSception: a deep learning eeg emotion recognition model based on fusion of global, time domain and frequency domain feature extraction
Molla et al. EEG signal enhancement using multivariate wavelet transform application to single-trial classification of event-related potentials
Islam et al. Wavelet-based Artifact Removal Algorithm for EEG Data by Optimizing Mother Wavelet and Threshold Parameters

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant