CN114492536A - Muscle impedance signal separation method - Google Patents
Muscle impedance signal separation method Download PDFInfo
- Publication number
- CN114492536A CN114492536A CN202210136712.0A CN202210136712A CN114492536A CN 114492536 A CN114492536 A CN 114492536A CN 202210136712 A CN202210136712 A CN 202210136712A CN 114492536 A CN114492536 A CN 114492536A
- Authority
- CN
- China
- Prior art keywords
- signal
- matrix
- separation
- algorithm
- signals
- 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.)
- Pending
Links
- 238000000926 separation method Methods 0.000 title claims abstract description 61
- 210000003205 muscle Anatomy 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 claims abstract description 34
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000001514 detection method Methods 0.000 claims abstract description 5
- 230000005284 excitation Effects 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 66
- 230000006870 function Effects 0.000 claims description 27
- 230000002087 whitening effect Effects 0.000 claims description 20
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 9
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 210000001519 tissue Anatomy 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- OELFLUMRDSZNSF-OFLPRAFFSA-N (2R)-2-[[oxo-(4-propan-2-ylcyclohexyl)methyl]amino]-3-phenylpropanoic acid Chemical compound C1CC(C(C)C)CCC1C(=O)N[C@@H](C(O)=O)CC1=CC=CC=C1 OELFLUMRDSZNSF-OFLPRAFFSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000004070 electrodeposition Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 208000018360 neuromuscular disease Diseases 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 206010033675 panniculitis Diseases 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 210000004003 subcutaneous fat Anatomy 0.000 description 1
- 210000004304 subcutaneous tissue Anatomy 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Abstract
The invention provides a muscle impedance signal separation method, which is characterized in that an excitation electric signal harmless to a human body acts on a part to be detected, a signal detection module obtains a muscle impedance signal Z, and a real part and an imaginary part of the Z are respectively taken as a resistance signal and a reactance signal; on the premise that the direction of the electrode array is unchanged, different mixed signals are obtained by changing the distance between the current electrode and the voltage electrode, the size of the electrode or the size of the electrode; transmitting the mixed signal to an upper computer for blind source separation processing, and separating out muscle layer source resistivity and reactance rate; in the technical scheme, in the data processing process, the data preprocessing process is simplified, so that the rapid and stable signal separation can be realized.
Description
Technical Field
The invention relates to the technical field of muscle impedance signal separation, in particular to a muscle impedance signal separation method.
Background
An EIM (electrical Impedance mapping) based on surface electrodes is a bioelectrical Impedance technology for evaluating muscle states, and changes of muscle Impedance signals reflect changes of electrical characteristics of muscle states, and is widely applied to diagnosis and evaluation of neuromuscular diseases. However, there are skin, fat, muscle, bone and other tissue layers under the surface electrode, and these subcutaneous tissue layers greatly interfere with the detection of muscle layer signals, and reduce the sensitivity of the muscle impedance signals to reflect the muscle condition, so it is necessary to perform real-time signal separation on the muscle impedance, restore the impedance rate characteristic of the muscle layer, and improve the accuracy of the EIM technology for detecting the muscle state change. The variational modal decomposition is often used for separating physiological signals, but the separation effect is extremely dependent on the preset modal number, and both the variational modal decomposition and the kurtosis function are very sensitive to noise and are easily influenced by the environment. With the further development of wavelet transform technology, empirical wavelet transform is widely used in signal separation, and has good time-frequency analysis capability, but at the same time, the performance of the method depends on the accuracy of wavelet base selection, the signal processing process is complex, and the problems of high data redundancy and the like exist.
Disclosure of Invention
In view of the above, an object of the present invention is to provide a muscle impedance signal separation method, which simplifies the data preprocessing process and can realize fast and stable signal separation.
In order to achieve the purpose, the invention adopts the following technical scheme: a muscle impedance signal separation method comprises the steps that an excitation electric signal harmless to a human body acts on a part to be detected, a signal detection module obtains a muscle impedance signal Z, and a real part and an imaginary part of the Z are taken as resistance signals and reactance signals respectively; on the premise that the direction of the electrode array is unchanged, different mixed signals are obtained by changing the distance between the current electrode and the voltage electrode, the size of the electrode or the size of the electrode; transmitting the mixed signal to an upper computer for blind source separation processing, and separating out muscle layer source resistivity and reactance rate;
the estimated signals y are linked to a demixing matrix W to form a loss function ρ (y, W) whose mathematical expectation j (W) ═ E { ρ (y, W) } denotes the respective signal y separatediThe mutual independence measure is taken as a blind source separation criterion, and the expression of the risk function R (W) is as follows:
J(W)=E{ρ(y,W)}=-log|det(W)|-∑i=1E{log(qi(yi)}1-1
wherein W represents a demixing matrix, qi(yi) Is the product of the edge probability density functions of the components of y, and indicates separation when equation 1-1 attains a minimumAre independent of each other;
optimizing the separation matrix by using a gradient method, and performing differential operation on the separation matrix by using a risk function (formula 1-1) first, wherein the random gradient of the risk function is represented as:
in the formula W-TIs a transposed inverse matrix of the unmixing matrix W (T is a matrix transpose symbol, -number represents the inverse of the matrix), x represents the mixed signal;
the blind separation algorithm based on the natural gradient is an improved algorithm on the random gradient, and a parameter space is expanded to a Riemann space from an Euclidean space; the natural gradient method eliminates inverse matrix calculation in random gradient and reduces the complexity of the algorithm; the fastest descent direction in riemann space is represented as:
in the formula of U-1(W) is expressed as the inverse of the riemann matrix, and this expression represents the fastest descent direction in the riemann space, i.e. the natural gradient; for simple calculation, use WTW to approximate U-1(W), then the natural gradient is obtained as:
under the natural gradient criterion of mutual information, the unmixing matrix is updated as follows:
the matrix W (k) is the kth iteration of the unmixing matrix; μ is the step size; i is an identity matrix; phi (·) is phi1(.),...,φn(.)]TIs a fractional function; phi is ak(.) is oneNon-linear function, with element yk(ii) related; y iskIn the case of sub-Gaussian signals, y is usually taken3,ykWhen the signal is a superss signal, the signal is taken as tanhy; in the method, y is selected3As a non-linear activation function;
signal whitening is a signal preprocessing mode, and the whitening processing of data can remove the correlation among observed signals and simplify the extraction process of independent components; in the method, whitening is carried out to remove the second-order correlation of signals and simplify the complexity of an algorithm; the method correlates the pre-whitening of the signal with the unmixing matrix to obtain a pre-whitening of the mixed signal:
finally combining equations 1-5 with equations 1-6 yields an iterative algorithm having the formula:
on the basis of a natural gradient target function, combining a blind source algorithm and a whitening algorithm together to obtain an equivalent self-adaptive separation algorithm, so that the algorithm is not influenced by a mixing matrix and an initial value of a separation matrix, and an algorithm formula can continuously carry out iterative correction to obtain a final separation matrix W;
all the mixed signals pass through the finally obtained separation matrix W, so that each independent component can be obtained, and the impedance rate of the muscle can be obtained.
Compared with the prior art, the invention has the following beneficial effects:
the method separates the impedance rate of the muscle layer by performing signal separation on the mixed muscle impedance signal of the unknown source signal. The influence of other tissues on the muscle impedance signal is reduced, and the sensitivity of detecting the muscle change is improved. Compared with a learning type signal separation method constructed by a model, the method has wider application range, does not need to extract relevant characteristics from the source signal, is simpler and more convenient, and can separate the mixed signal of unknown source signals. Compared with the Fastic algorithm, the method avoids the complex processes of preprocessing such as data whitening, high-order decorrelation and the like, and the processes are included in the iteration step. The algorithm has faster convergence speed and better stability, and cannot be influenced by the outside and a mixed matrix. Compared with the empirical wavelet transform, the method is simpler to realize, the algorithm steps are relatively fewer, the iteration step length is adjustable, and the difficulty of muscle impedance signal separation is reduced. Therefore, the method can meet the requirement of muscle impedance signal separation, and the algorithm shows good separation effect and stability.
Drawings
FIG. 1 is a schematic block diagram of blind source separation in a preferred embodiment of the present invention;
FIG. 2 is an estimated resistivity and a source resistivity of a fat layer in a preferred embodiment of the invention;
FIG. 3 is an estimated resistivity of the muscle layer and source resistivity in a preferred embodiment of the invention;
FIG. 4 is a graph of estimated and source reactance rates for a fat layer in a preferred embodiment of the present invention;
FIG. 5 is an estimated reactance rate and a source reactance rate for a muscle layer in a preferred embodiment of the present invention;
FIG. 6 is a block diagram of a detailed flow of a blind source separation algorithm in a preferred embodiment of the present invention;
FIG. 7 illustrates the principle of muscle impedance signal separation in a preferred embodiment of the present invention.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application; as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The method aims to improve the contribution rate of the muscle layer in the muscle impedance signal to the signal and reduce the interference of subcutaneous fat and the like to the muscle impedance signal, and provides a high-efficiency and rapid muscle impedance signal separation method. The system injects harmless current signals into a human body, then collects corresponding response signals to obtain muscle impedance signals, and transmits the muscle impedance signals to an upper computer for signal separation. The signal separation process is to regard the obtained mixed muscle impedance signals as the product of the source signals and the mixing matrix, find the demixing matrix by adopting an iterative contraction method, and multiply all the mixed signals with the final demixing matrix, thereby separating the muscle impedance rate of the muscle tissue. FIG. 1 is a schematic diagram of the blind source separation method. And in the iteration part, the preprocessing step of the mixed signal is associated with the iteration of the unmixing matrix, so that the data preprocessing process is omitted, and the unmixing step is simplified. The iterative algorithm adopts a serial updating method, and the weight coefficient of the iterative algorithm is continuously updated by taking the natural gradient of mutual information as the optimization criterion of the unmixing matrix, so that the self-adaptive adjustment function of the iterative algorithm is realized, the convergence speed is accelerated, and finally, the fast and stable signal separation is realized. FIGS. 2-5 show the electrical conductivity and reactance separated by this method for the fat and muscle layers, which are highly coincident with the source conductivity and reactance.
A blind source separation method provided by a muscle impedance signal separation method mainly comprises the steps of obtaining a muscle impedance mixed signal, a blind source separation criterion, a separation matrix updating module, a signal whitening module and a final iterative algorithm. And applying a stable high-frequency excitation electric signal harmless to a human body to the part to be detected, obtaining a muscle impedance signal Z by a signal detection module, and taking a real part and an imaginary part of the Z as resistance and reactance signals respectively. In order to obtain the effective impedance Z including each tissue, different mixed signals can be obtained by changing the distance between the current electrode and the voltage electrode, the size of the electrode, or both, on the premise that the direction of the electrode array is kept unchanged. And transmitting the mixed signal to an upper computer for blind source separation processing, and separating out the source resistivity and the reactance rate of the muscle layer.
And (2) relating the estimated signals y and the unmixing matrix W to construct a loss function rho (y, W), wherein the mathematical expectation J (W) of the loss function E { rho (y, W) } represents a measure of the mutual independence between the separated signals, and the measure is used as a blind source separation criterion, and the expression of a risk function R (W) is as follows:
J(W)=E{ρ(y,W)}=-log|det(W)|-∑i=1E{log(qi(yi))} 1-1
wherein W represents a demixing matrix, qi(yi) Is the product of the edge probability density functions of the components of y, and when equation 1-1 obtains a minimum value, it indicates that the separated source signals are independent of each other.
The separation matrix is optimized by using a gradient method, and the risk function (formula 1-1) firstly performs differential operation on the separation matrix, so that the random gradient of the risk function can be represented as:
in the formula W-TIs the transposed inverse of the unmixing matrix W (T is the matrix transpose, -number represents the inverse of the matrix), x represents the mixed signal.
The blind separation algorithm based on natural gradient is an improved algorithm on random gradient, and can be understood as extending the parameter space from Euclidean space to Riemannian space. The natural gradient method eliminates inverse matrix calculation in random gradient and reduces algorithm complexity. The fastest descent direction in riemann space can be expressed as:
in the formula of U-1(W) is expressed as the inverse of the Riemannian matrix, and this expression represents the fastest descent direction in Riemannian space, i.e. the natural gradient. To operate simplyW may be usedTW to approximate U-1(W), then the natural gradient is obtained as:
under the natural gradient criterion of mutual information, the unmixing matrix can be updated as:
W(k+1)=W(k)-μΔJ=W(k)+μ[I-φ(yk)yk T]W(k) 1-5
the matrix W (k) is the kth iteration of the unmixing matrix; μ is the step size; i is an identity matrix; phi ()) (phi ═ phi1(.),...,φn(.)]TIs a fractional function; phi is ak(.) is a non-linear function, and element ykIt is related. y iskIn the case of sub-Gaussian signals, y is usually taken3,ykWhen the signal is a superss signal, it is taken as tanhy. In the method, y is selected3As a non-linear activation function.
Signal whitening is a signal preprocessing mode, and whitening processing on data can remove correlation among observed signals and simplify the extraction process of independent components. In the method, whitening is performed to remove second-order correlation of the signal and simplify the complexity of the algorithm. The method correlates the pre-whitening of the signal with the unmixing matrix to obtain a pre-whitening of the mixed signal:
finally combining equations 1-5 with equations 1-6 yields an iterative algorithm having the formula:
on the basis of a natural gradient objective function, a blind source algorithm and a whitening algorithm are combined together to obtain an equivalent self-adaptive separation algorithm, so that the algorithm is not influenced by a mixing matrix and an initial value of a separation matrix, an algorithm formula can be continuously subjected to iterative correction, and a final separation matrix W is obtained. The block diagram for the description of the whole algorithm is shown in fig. 6.
All the mixed signals pass through the finally obtained separation matrix W, so that each independent component can be obtained, and the impedance rate of the muscle can be obtained. The process of acquiring the unmixed source signal from the mixed signal is as shown in fig. 7.
The product use process or mode: firstly, a muscle impedance mixed signal is obtained through operations such as electrode position and distance transformation, the signal is transmitted to an upper computer, data preprocessing is carried out through an iterative algorithm, and an initial matrix of the iterative algorithm is obtained. The system is then adaptively updated by a blind source algorithm to obtain a new separation matrix. And obtaining a final separation matrix W after all the signals are received, and multiplying all the mixed signals by W to obtain independent source signals.
Claims (1)
1. A muscle impedance signal separation method is characterized in that an excitation electric signal harmless to a human body acts on a part to be detected, a signal detection module obtains a muscle impedance signal Z, and a real part and an imaginary part of the Z are taken as a resistance signal and a reactance signal respectively; on the premise that the direction of the electrode array is unchanged, different mixed signals are obtained by changing the distance between the current electrode and the voltage electrode, the size of the electrode or the size of the electrode; transmitting the mixed signal to an upper computer for blind source separation processing to separate out the source resistivity and the reactance rate of a muscle layer;
the estimated signals y are linked to a demixing matrix W to form a loss function ρ (y, W) whose mathematical expectation j (W) ═ E { ρ (y, W) } denotes the respective signal y separatediThe mutual independence measure is taken as a blind source separation criterion, and the expression of the risk function R (W) is as follows:
J(W)=E{ρ(y,W)}=-log|det(W)|-∑i=1E{log(qi(yi))} 1-1
wherein W represents a demixing matrix, qi(yi) Is the product of the edge probability density functions of the components of y, and when the formula 1-1 obtains the minimum value, the separated source signals are mutually independent;
optimizing the separation matrix by using a gradient method, and performing differential operation on the separation matrix by using a risk function formula (1-1) to obtain a random gradient of the risk function:
in the formula W-TIs a transposed inverse matrix of the unmixing matrix W (T is a matrix transpose symbol, -number represents the inverse of the matrix), x represents the mixed signal;
the blind separation algorithm based on the natural gradient is an improved algorithm on the random gradient, and a parameter space is expanded to a Riemann space from an Euclidean space; the natural gradient method eliminates inverse matrix calculation in random gradient and reduces the complexity of the algorithm; the fastest descent direction in riemann space is represented as:
in the formula of U-1(W) is expressed as the inverse of the riemann matrix, and this expression represents the fastest descent direction in the riemann space, i.e. the natural gradient; for simple calculation, use WTW to approximate U-1W, then obtain
Under the natural gradient criterion of mutual information, the unmixing matrix is updated as follows:
W(k+1)=W(k)-μΔJ=W(k)+μ[I-φ(yk)yk T]W(k) 1-5
the matrix W (k) is the kth iteration of the unmixing matrix; μ is the step size; i is an identity matrix; phi (·) is phi1(.),...,φn(.)]TIs a fractional function; phi is ak(.) is a non-linear function, and element yk(ii) related; y iskIn the case of sub-Gaussian signals, y is usually taken3;ykWhen the signal is a superss signal, the signal is taken as tanhy; in the method, y is selected3As a non-linear activation function;
signal whitening is a signal preprocessing mode, and the whitening processing of data can remove the correlation among all observed signals and simplify the extraction process of independent components; in the method, whitening is carried out to remove the second-order correlation of signals and simplify the complexity of an algorithm; the method correlates the pre-whitening of the signal with the unmixing matrix to obtain a pre-whitening of the mixed signal:
finally combining equations 1-5 with equations 1-6 yields an iterative algorithm having the formula:
on the basis of a natural gradient target function, combining a blind source algorithm and a whitening algorithm together to obtain an equivalent self-adaptive separation algorithm, so that the algorithm is not influenced by a mixing matrix and an initial value of a separation matrix, and an algorithm formula can continuously carry out iterative correction to obtain a final separation matrix W;
all the mixed signals pass through the finally obtained separation matrix W, so that each independent component can be obtained, and the impedance rate of the muscle can be obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210136712.0A CN114492536A (en) | 2022-02-15 | 2022-02-15 | Muscle impedance signal separation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210136712.0A CN114492536A (en) | 2022-02-15 | 2022-02-15 | Muscle impedance signal separation method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114492536A true CN114492536A (en) | 2022-05-13 |
Family
ID=81480618
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210136712.0A Pending CN114492536A (en) | 2022-02-15 | 2022-02-15 | Muscle impedance signal separation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114492536A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101163443A (en) * | 2005-02-15 | 2008-04-16 | 明尼苏达大学董事会 | Pathology assessment with impedance measurements using convergent bioelectric lead fields |
CN102631195A (en) * | 2012-04-18 | 2012-08-15 | 太原科技大学 | Single-channel blind source separation method of surface electromyogram signals of human body |
CN103295193A (en) * | 2013-05-10 | 2013-09-11 | 天津理工大学 | Cross-power spectrum based blind source separation method |
AU2020103605A4 (en) * | 2020-11-23 | 2021-02-04 | Sichuan University of Science and Engineering | An Anti-collision Method of the RFID System in the Under-determined State |
-
2022
- 2022-02-15 CN CN202210136712.0A patent/CN114492536A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101163443A (en) * | 2005-02-15 | 2008-04-16 | 明尼苏达大学董事会 | Pathology assessment with impedance measurements using convergent bioelectric lead fields |
CN102631195A (en) * | 2012-04-18 | 2012-08-15 | 太原科技大学 | Single-channel blind source separation method of surface electromyogram signals of human body |
CN103295193A (en) * | 2013-05-10 | 2013-09-11 | 天津理工大学 | Cross-power spectrum based blind source separation method |
AU2020103605A4 (en) * | 2020-11-23 | 2021-02-04 | Sichuan University of Science and Engineering | An Anti-collision Method of the RFID System in the Under-determined State |
Non-Patent Citations (1)
Title |
---|
高跃明: ""基于肌阻抗图的肌肉疲劳检测电极配置优化研究"", 《电气技术》, 31 July 2020 (2020-07-31) * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220346655A1 (en) | Continuous non-invasive monitoring of a pregnant human subject | |
Yin et al. | ECG monitoring system integrated with IR-UWB radar based on CNN | |
CN107811626A (en) | A kind of arrhythmia classification method based on one-dimensional convolutional neural networks and S-transformation | |
AU2016252353B2 (en) | Continuous non-invasive monitoring of a pregnant human subject | |
CN106108889A (en) | Electrocardiogram classification method based on degree of depth learning algorithm | |
CN111887858B (en) | Ballistocardiogram signal heart rate estimation method based on cross-modal mapping | |
CN107980151B (en) | Access control system based on electrocardio authentication and authentication method thereof | |
JP2018512243A5 (en) | ||
CN113111831A (en) | Gesture recognition technology based on multi-mode information fusion | |
CN115795346A (en) | Classification and identification method of human electroencephalogram signals | |
CN111000551A (en) | Heart disease risk diagnosis method based on deep convolutional neural network model | |
JP2007522536A (en) | Dynamic blind signal separation | |
CN107550491A (en) | A kind of multi-class Mental imagery classifying identification method | |
CN109222964A (en) | atrial fibrillation detection device and storage medium | |
Caytak et al. | Bioimpedance spectroscopy processing and applications | |
CN109077721A (en) | atrial fibrillation detection device and storage medium | |
CN113967022B (en) | Individual self-adaption-based motor imagery electroencephalogram characteristic characterization method | |
Phinyomark et al. | Applications of variance fractal dimension: A survey | |
CN113128384B (en) | Brain-computer interface software key technical method of cerebral apoplexy rehabilitation system based on deep learning | |
CN112401905B (en) | Natural action electroencephalogram recognition method based on source localization and brain network | |
CN114492536A (en) | Muscle impedance signal separation method | |
CN108836316B (en) | Electrocardiosignal R wave extraction method based on BP neural network | |
CN108388345B (en) | Brain electrode optimization method based on wavelet multi-resolution complex network and application thereof | |
CN105662391B (en) | A kind of feature extraction and classifying system and method for stomach magnetic signal | |
Liutkus et al. | Source separation for target enhancement of food intake acoustics from noisy recordings |
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 |