CN108852348A - The collection point sort method and system of scalp brain electricity - Google Patents
The collection point sort method and system of scalp brain electricity Download PDFInfo
- Publication number
- CN108852348A CN108852348A CN201810455255.5A CN201810455255A CN108852348A CN 108852348 A CN108852348 A CN 108852348A CN 201810455255 A CN201810455255 A CN 201810455255A CN 108852348 A CN108852348 A CN 108852348A
- Authority
- CN
- China
- Prior art keywords
- data
- eeg signals
- characteristic
- serial number
- dimensional
- 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
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]
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Psychology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Psychiatry (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses the collection point sort methods and system of a kind of scalp brain electricity.The collection point sort method includes：Data are read in；Data prediction：Original EEG signals are subjected to artefact removal and noise suppressed processing, export EEG signals after the processing of three-dimensional matrice；Feature extraction：Extract time domain and frequency domain character, the brain electrical characteristic data of feature serial number and test number (TN) serial number and channel position composition three-dimensional matrice；Characteristic signature：Brain electrical characteristic data is subjected to feature pool, feature clustering, data reforming, obtains the brain electric signature data of the two-dimensional matrix recorded by test number (TN) serial number and channel position；Relevance ranking：Correlation between brain electric signature data and EEG signals classification is calculated and sorted, the channel position of one-dimensional vector is exported；As a result it exports.This method not will lead to the calculating process of random diverging, improves the performance of EEG-BCI system, reduces data scale, versatile, be conducive to for other methods being included in system.
Description
Technical field
The invention belongs to eeg signal acquisition technical fields, and in particular to a kind of collection point sort method of scalp brain electricity
And system.
Background technique
Brain-computer interface (brain-computer interface, BCI) is the directly interactive new paragon of brain machine, can be with
Brain activity signal relevant to psychological activity is collected, is then suitable instruction by these signal resolutions, and these are instructed
It is sent to computer.The collected EEG signals of the scalp of people be called scalp brain electricity (electroencephalography,
EEG), the scalp brain electricity of damage-free type has the advantages that low-risk, low cost, flexibility height etc. are suitable for practical application, in BCI
It is widely applied.Usually, EEG data is the multichannel letter collected from the electrode for being placed on the multiple sites of scalp
Number, it reflects the active characteristics of the multiple cortexes of brain.
The core of BCI system (EEG-BCI) based on EEG is the parsing or classification of EEG signal, it is usually by four necessity
Part forms：Signal Pretreatment, feature extraction, channel selecting and classification.Wherein, the basis of the channel selecting of core is EEG
Collection point sequence.The purpose of EEG collection point sequence is the EEG feature removed in uncorrelated and redundant channel, and then is contracted
Short system time reduces equipment cost, improves EEG resolution speed and efficiency.
There are several bottlenecks in practical application by existing EEG-BCI.Firstly, existing EEG-BCI has ignored EEG parsing
Computation complexity extends the application process of EEG-BCI, and more seriously, too long preparation and test period may be unfavorable for
The mental concentration of subject, and then reduce the signal-to-noise ratio of EEG；Secondly, seldom considering result calculating process in some iteration frames
Convergence, lead to calculating process that is random, even dissipating；Finally, existing EEG-BCI designs excessively fine, poor universality,
Narrowed scalability, is unfavorable for for other methods being included in system.
Summary of the invention
The purpose of the present invention is to provide the collection point sort methods and system of a kind of scalp brain electricity, to overcome existing
The shortcomings that EEG-BCI application time is too long, signal-to-noise ratio is low, calculating process is random and poor universality.
To realize said one or multiple purposes, in one embodiment of the invention, the present invention provides a kind of heads
The collection point sort method of skin brain electricity, this method include：Data read in step：By EEG signals classification and by test number (TN) sequence
Number, channel position and time point serial number record original EEG signals, in the form of the one-dimensional vector of structuring and three-dimensional matrice read
Enter in data prediction step；Data prediction step：The original brain for the three-dimensional matrice form that step inputs will be read in from data
Electric signal carries out artefact removal and noise suppressed processing, exports EEG signals after the processing of three-dimensional matrice form；Feature extraction step
Suddenly：Using time domain parameter algorithm, wavelet transformation, Fourier transformation, from EEG signals after the processing that data prediction step inputs
Middle extraction temporal signatures, frequency domain character or time-frequency domain binding characteristic will be remembered in EEG signals after processing with the time point serial number
The voltage value signal conversion of record is characterized the characteristic value of serial number record, the test after the feature serial number and processing in EEG signals
Number serial number and channel position form the brain electrical characteristic data of three-dimensional matrice form together；Characteristic signature step：It will be mentioned from feature
It takes the brain electrical characteristic data inputted in step to carry out feature pool, feature clustering processing, obtains test number (TN) serial number and channel sequence
The signature of number corresponding signal characteristic, and handled through data reforming, obtain the two dimension recorded by test number (TN) serial number and channel position
The brain electric signature data of matrix form；Relevance ranking step：By the brain of the two-dimensional matrix form inputted from characteristic signature step
Correlation is calculated and is sorted between electric signature data and EEG signals classification, and the one-dimensional vector form after output sequence is led to
Road serial number；And result exports step：Export the channel position of the one-dimensional vector form inputted from relevance ranking step.
In a preferred embodiment of the invention, the present invention provides a kind of collection point sequence sides of scalp brain electricity
Method, it is using flip-flop removal step, bandpass filtering step, down-sampling step that original brain is electric in data prediction step
Signal carries out artefact removal and noise suppressed processing.
In a preferred embodiment of the invention, the present invention provides a kind of collection point sequence sides of scalp brain electricity
Method carries out the bandpass filtering step using bandpass filter.
In a preferred embodiment of the invention, the present invention provides a kind of collection point sequence sides of scalp brain electricity
Method carries out the down-sampling step using region average, intermediate value.
In a preferred embodiment of the invention, the present invention provides a kind of collection point sequence sides of scalp brain electricity
Method, in characteristic signature step, the feature clustering processing is selected from one of the processing of k mean cluster, spectral clustering processing or more
Kind.
In a preferred embodiment of the invention, the present invention provides a kind of collection point sequence sides of scalp brain electricity
Method is carried out correlation using classification, recurrence and is calculated, is ranked up using bubbling method in relevance ranking step.
In another embodiment of the present invention, the present invention provides a kind of sequences of the collection point of scalp brain electricity is
System, including：Data read in module, data preprocessing module, characteristic extracting module, characteristic signature module, relevance ranking module
And result output module；The data read in module：For by the original EEG signals collected and EEG signals classification
It is read in data preprocessing module in the form of the three-dimensional matrice of structuring and one-dimensional vector；The data preprocessing module：With
It is handled in the original EEG signals for reading in the three-dimensional matrice form that module inputs from data are carried out artefact removal and noise suppressed,
EEG signals after the processing of acquisition three-dimensional matrice form；The characteristic extracting module：Using time domain parameter algorithm, after extraction process
Temporal signatures, frequency domain character or time-frequency domain binding characteristic in EEG signals will be remembered in original EEG signals with time point serial number
The voltage value signal conversion of record is characterized the characteristic value of serial number record, and together with original test number (TN) serial number and channel position
Form the brain electrical characteristic data of three-dimensional matrice form；The characteristic signature module：For brain electrical characteristic data to be carried out feature pool
Change, feature clustering processing, obtain the signature that test number (TN) serial number and channel position correspond to signal characteristic, and through data reforming at
Reason obtains the brain electric signature data of the two-dimensional matrix form recorded by test number (TN) serial number and channel position；The correlation row
Sequence module：It is directed to the row sparse regression optimization problem of feature weight matrix by solving, obtains the weight in each channel, and by institute
There is channel to be ranked up according to the sequence of weight from big to small；And the result output module：For exporting one after sorting
The channel position of dimensional vector form.
In a preferred embodiment of the invention, the present invention provides a kind of sequences of the collection point of scalp brain electricity is
It unites, it is using flip-flop removal step, bandpass filtering step, down-sampling step that original brain is electric in the data preprocessing module
Signal carries out artefact removal and noise suppressed processing.
Wherein, it is read in step in above-mentioned data, the original EEG signals and brain electricity is acquired using brain wave acquisition equipment
Signal classification.Electric brain wave acquisition equipment is selected from Neuroscan, newly opens up, one of BioSemi.
Wherein, the EEG signals classification is one-dimensional categorization vector matched with EEG signals, refers to each brain telecommunications
Number any class belonged to, is the true tag manually marked, is not signal to be processed, does not need to handle.
Wherein, above-mentioned relevance ranking step explanation：One two-dimensional matrix includes N number of sample, M feature, N number of sample point
Belong to C classification, the task of general modfel identification is just to discriminate between N number of sample and belongs to that classification (" classification " task), usually
For, not all M feature is all helpful to correctly classifying, and relevance ranking step is then to arrange M feature
Sequence, in the top larger to classification help, ranking behind, it is small to help classification.
Compared with prior art, the present invention has the advantages that：
(1) present invention introduces theoretical convergent feature selection approach, and channel is ranked up and is selected, not will lead to it is random,
The calculating process of diverging improves EEG classification accuracy, further improves the performance of EEG-BCI system.
(2) the brain electrical characteristic data inputted from characteristic extraction step is subjected to feature pool, feature clustering in the present invention
Processing can carry out compressive features by channel, and then reduce data scale.
(3) the present invention provides the EEG classfying frames that one is suitable for most of common EEG processing and mode identification method
Frame, it is versatile, be conducive to for other methods being included in system, improve the property of EEG-BCI system by Integration ofTechnology for after
It can lay a good foundation.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application.Attached
In figure：
Fig. 1 is the flow chart of the collection point sort method of middle scalp brain electricity according to embodiments of the present invention.
Fig. 2 is the interior of data prediction step in the collection point sort method of middle scalp brain electricity according to embodiments of the present invention
Portion's process flow diagram.
Fig. 3 is the inside of characteristic extraction step in the collection point sort method of middle scalp brain electricity according to embodiments of the present invention
Process flow diagram.
Fig. 4 is the inside of characteristic signature step in the collection point sort method of middle scalp brain electricity according to embodiments of the present invention
Process flow diagram.
Fig. 5 is the structural block diagram of the collection point ordering system of middle scalp brain electricity according to embodiments of the present invention.
Specific embodiment
Further progress detailed description is made to technical solution of the present invention below, it is to be understood that protection scope of the present invention
It is not limited by the specific implementation.
Unless otherwise explicitly stated, otherwise in entire disclosure and claims, term " includes " or its change
Changing such as "comprising" or " including " etc. will be understood to comprise stated element or component, and not exclude other members
Part or other component parts.
Fig. 1 is the flow chart of the collection point sort method of middle scalp brain electricity according to embodiments of the present invention.As shown in Figure 1,
A kind of collection point sort method of scalp brain electricity, including：By EEG signals classification and by test number (TN) serial number, channel position and
The original EEG signals of time point serial number record read in data prediction step in the form of the one-dimensional vector of structuring and three-dimensional matrice
In rapid, i.e., data read in step 101；Then, the original EEG signals for the three-dimensional matrice form that step inputs will be read in from data
Artefact removal and noise suppressed processing are carried out, EEG signals, i.e. data prediction step after the processing of three-dimensional matrice form are exported
102；Next, using time domain parameter algorithm, wavelet transformation, Fourier transformation, after the processing that data prediction step inputs
Temporal signatures, frequency domain character or time-frequency domain binding characteristic are extracted in EEG signals, by EEG signals after processing with the time
The voltage value signal conversion of point serial number record is characterized the characteristic value of serial number record, EEG signals after the feature serial number and processing
In test number (TN) serial number and channel position form the brain electrical characteristic data 103 of three-dimensional matrice form together；It again, will be from feature
The brain electrical characteristic data inputted in extraction step carries out feature pool, feature clustering processing, obtains test number (TN) serial number and channel
Serial number corresponds to the signature of signal characteristic, and handles through data reforming, obtains recorded by test number (TN) serial number and channel position two
Tie up the brain electric signature data of matrix form, i.e. characteristic signature step 104；Further, the two dimension that will be inputted from characteristic signature step
Correlation is calculated and is sorted between the brain electric signature data of matrix form and EEG signals classification, one-dimensional after output sequence
The channel position of vector form, i.e. relevance ranking step 105；Finally, output inputted from relevance ranking step it is one-dimensional to
The channel position of amount form, i.e. result export step 106.
Wherein, as shown in Fig. 2, in above-mentioned data prediction step 102 input be with test number (TN) serial number, channel position
With the original EEG signals of the three-dimensional matrice form of time point serial number record, output is three-dimensional matrice form treated brain
Electric signal.By taking the signal of one-dimensional vector form in the channel wherein once tested as an example, concrete operations include three parts：1)
The signal of vector form, is subtracted their mean value, can remove the flip-flop in signal by flip-flop removal, is obtained
The signal that value is 0；2) bandpass filtering, using the bandpass filter of certain window, by translation, gradually by signal with task without
The low frequency and radio-frequency component of pass filter out；3) signal of vector form is taken a number every several points, cast out by down-sampling
The number of taking-up is lined up the form of vector again, then is placed back into original test number (TN) and channel by other numbers
Obtain treated EEG signals.
As shown in figure 3, what is inputted in features described above extraction step 103 is with test number (TN) serial number, channel position and time
Three-dimensional matrice form treated the EEG signals of point serial number record, output be three-dimensional matrice form brain electrical feature number
According to.By taking the signal of one-dimensional vector form in the channel wherein once tested as an example, concrete operations include three parts：1) it seeks
Differential signal calculates 0 rank of the signal to time point, 1 rank and 2 rank differential of vector form, signal one-dimensional vector shape can be obtained
Amplitude Characteristics (0 rank differential signal), high-frequency characteristic (1 rank differential signal) and the frequency variation characteristic (2 rank differential signal) of formula；2)
Variance yields is sought, 0 rank of one-dimensional vector form, the variance yields of 1 rank and 2 rank differential signals are calculated separately；3) logarithm Gauss is sought
Approximation calculates the logarithm of 0 rank, 1 rank and 2 rank differential signal variance yields, seeks random point that logarithm can obey original signal
Cloth is approximately Gaussian Profile, is conducive to subsequent channel sequence.Then, obtained logarithm Gaussian approximation value is lined up length is 3
Vector form, then be placed back into original test number (TN) and channel, can be obtained the brain electrical feature number of three-dimensional matrice form
According to.
As shown in figure 4, what is inputted in features described above signature step 105 is with test number (TN) serial number, channel position and feature
Serial number record three-dimensional matrice form brain electrical characteristic data, output be two-dimensional matrix form brain electric signature data.With it
In in a channel once testing for the signal of one-dimensional vector form, concrete operations include three parts：1) feature pool, will
All test number (TN)s, all channels all features put together according to certain sequence；2) feature clustering carries out all features
Cluster, and use feature generic as its signature；3) data reforming, by the signature of all features according to test number (TN), logical
Road serial number is rearranged as a two-dimensional matrix, can be obtained the brain electric signature data of two-dimensional matrix form.
Above-mentioned relevance ranking step input is in the form of the two-dimensional matrix that test number (TN) serial number and channel position record
Brain electric signature data and one-dimensional vector form EEG signals classification, output be one-dimensional vector form sequence after channel
Serial number.If respectively indicating brain electric signature data and EEG signals classification with matrix X and vector y, C is channel number, and yi is vector y
I-th of element, Ul：For the l row of matrix U, | | | | 2 indicate 2 norms of vectors or matrix, and yi=z indicates i-th test
Belong to z class.The concrete operations of relevance ranking include six parts：1) y is extended to classification matrix Y, if yi=z, then Y
For i-th row in addition to z-th of numerical value is 1, other are 0；2) it enables A=[X I], I is unit matrix, and initialization D is unit matrix；3)
First of diagonal element for calculating U=D-1AT (AD-1AT) -1Y, and updating D is 0.5 | | Ul:||2-1；4) return step 2), directly
Maximum number of iterations is changed very little or reached to U；5) to the preceding C row of U, the norm of each row vector is calculated, as corresponding logical
The weight in road；6) all weights are ranked up according to sequence from big to small, and obtain the channel position after corresponding sequence,
And then carry out result output.
Fig. 5 is the structural block diagram of the collection point ordering system of middle scalp brain electricity according to embodiments of the present invention.Such as Fig. 5 institute
Show, the collection point ordering system of scalp brain electricity includes：Data read in module, data preprocessing module, characteristic extracting module, spy
Levy signature blocks, relevance ranking module and result output module；Wherein, data read in the original that module is used to collect
Beginning EEG signals and EEG signals classification read in data preprocessing module in the form of the three-dimensional matrice of structuring and one-dimensional vector
In；Data preprocessing module is used to the original EEG signals for reading in the three-dimensional matrice form that module inputs from data carrying out artefact
Removal and noise suppressed processing, obtain EEG signals after the processing of three-dimensional matrice form；Characteristic extracting module utilizes time domain parameter
Algorithm, temporal signatures, frequency domain character or time-frequency domain binding characteristic after extraction process in EEG signals, will be in original EEG signals
By time point serial number record voltage value signal conversion characterized by serial number record characteristic value, and with original test number (TN) serial number
Form the brain electrical characteristic data of three-dimensional matrice form together with channel position；Characteristic signature module be used for by brain electrical characteristic data into
Row feature pool, feature clustering processing, obtain test number (TN) serial number and channel position corresponds to the signature of signal characteristic, and through data
Reformation processing obtains the brain electric signature data of the two-dimensional matrix form recorded by test number (TN) serial number and channel position；Correlation
Sorting module：It is directed to the row sparse regression optimization problem of feature weight matrix by solving, obtains the weight in each channel, and will
All channels are ranked up according to the sequence of weight from big to small；As a result output module is used to export the one-dimensional vector shape after sequence
The channel position of formula.
Wherein, step, bandpass filtering step, down-sampling step are removed using flip-flop in above-mentioned data preprocessing module
Original EEG signals are subjected to artefact removal and noise suppressed processing.
The aforementioned description to specific exemplary embodiment of the invention is in order to illustrate and illustration purpose.These descriptions
It is not wishing to limit the invention to disclosed precise forms, and it will be apparent that according to the above instruction, can much be changed
And variation.The purpose of selecting and describing the exemplary embodiment is that explaining specific principle of the invention and its actually answering
With so that those skilled in the art can be realized and utilize a variety of different exemplary implementation schemes of the invention and
Various chooses and changes.The scope of the present invention is intended to be limited by claims and its equivalents.
Claims (8)
1. a kind of collection point sort method of scalp brain electricity, which is characterized in that include the following steps：
Data read in step：By EEG signals classification and the original recorded by test number (TN) serial number, channel position and time point serial number
Beginning EEG signals are read in data prediction step in the form of the one-dimensional vector of structuring and three-dimensional matrice；
Data prediction step：The original EEG signals that the three-dimensional matrice form that step inputs is read in from data are subjected to artefact
Except handling with noise suppressed, EEG signals after the processing of three-dimensional matrice form are exported；
Characteristic extraction step：Using time domain parameter algorithm, wavelet transformation, Fourier transformation, inputted from data prediction step
Temporal signatures, frequency domain character or time-frequency domain binding characteristic are extracted after processing in EEG signals, by EEG signals after processing with institute
The voltage value signal conversion for stating time point serial number record is characterized the characteristic value of serial number record, the feature serial number and processing hindbrain
Test number (TN) serial number and channel position in electric signal form the brain electrical characteristic data of three-dimensional matrice form together；
Characteristic signature step：Feature pool will be carried out, from feature clustering from the brain electrical characteristic data that inputs in characteristic extraction step
Reason, obtains test number (TN) serial number and channel position corresponds to the signature of signal characteristic, and handles through data reforming, obtains secondary by testing
The brain electric signature data of number sequence number and the two-dimensional matrix form of channel position record；
Relevance ranking step：By the brain electric signature data and EEG signals of the two-dimensional matrix form inputted from characteristic signature step
Correlation is calculated and is sorted between classification, the channel position of the one-dimensional vector form after output sequence；And
As a result step is exported：Export the channel position of the one-dimensional vector form inputted from relevance ranking step.
2. collection point sort method according to claim 1, which is characterized in that in data prediction step, use
Flip-flop removes step, bandpass filtering step, down-sampling step and original EEG signals is carried out artefact removal and noise suppressed
Processing.
3. collection point sort method according to claim 2, which is characterized in that carry out the band using bandpass filter
Pass filtering step.
4. collection point sort method according to claim 2, which is characterized in that carry out institute using region average, intermediate value
State down-sampling step.
5. collection point sort method according to claim 1, which is characterized in that in characteristic signature step, the spy
It levies clustering processing and is selected from one of the processing of k mean cluster, spectral clustering processing or a variety of.
6. collection point sort method according to claim 1, which is characterized in that in relevance ranking step, use
Classification, recurrence carry out correlation and are calculated, and are ranked up using bubbling method.
7. a kind of collection point ordering system of scalp brain electricity, which is characterized in that including：Data read in module, data prediction
Module, characteristic extracting module, characteristic signature module, relevance ranking module and result output module；
The data read in module：For by the original EEG signals collected and EEG signals classification with the three-dimensional of structuring
The form of matrix and one-dimensional vector is read in data preprocessing module；
The data preprocessing module：For by from data read in module input three-dimensional matrice form original EEG signals into
The removal of row artefact and noise suppressed processing, obtain EEG signals after the processing of three-dimensional matrice form；
The characteristic extracting module：Temporal signatures, frequency domain character using time domain parameter algorithm, after extraction process in EEG signals
Or time-frequency domain binding characteristic, by serial number is remembered characterized by the voltage value signal conversion of time point serial number record in original EEG signals
The characteristic value of record, and form together with original test number (TN) serial number and channel position the brain electrical feature number of three-dimensional matrice form
According to；
The characteristic signature module：For brain electrical characteristic data to be carried out feature pool, feature clustering processing, test number (TN) is obtained
Serial number and channel position correspond to the signature of signal characteristic, and handle through data reforming, obtain by test number (TN) serial number and channel sequence
Number record two-dimensional matrix form brain electric signature data；
The relevance ranking module：It is directed to the row sparse regression optimization problem of feature weight matrix by solving, obtains each
The weight in channel, and all channels are ranked up according to the sequence of weight from big to small；And
The result output module：For exporting the channel position of the one-dimensional vector form after sorting.
8. collection point ordering system according to claim 7, which is characterized in that used in the data preprocessing module
Flip-flop removes step, bandpass filtering step, down-sampling step and original EEG signals is carried out artefact removal and noise suppressed
Processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810455255.5A CN108852348A (en) | 2018-05-14 | 2018-05-14 | The collection point sort method and system of scalp brain electricity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810455255.5A CN108852348A (en) | 2018-05-14 | 2018-05-14 | The collection point sort method and system of scalp brain electricity |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108852348A true CN108852348A (en) | 2018-11-23 |
Family
ID=64334337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810455255.5A Pending CN108852348A (en) | 2018-05-14 | 2018-05-14 | The collection point sort method and system of scalp brain electricity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108852348A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111067515A (en) * | 2019-12-11 | 2020-04-28 | 中国人民解放军军事科学院军事医学研究院 | Intelligent airbag helmet system based on closed-loop control technology |
| CN112036229A (en) * | 2020-06-24 | 2020-12-04 | 宿州小马电子商务有限公司 | Intelligent bassinet electroencephalogram signal channel configuration method with demand sensing function |
| CN113397562A (en) * | 2021-07-20 | 2021-09-17 | 电子科技大学 | Sleep spindle wave detection method based on deep learning |
| CN114533084A (en) * | 2022-02-09 | 2022-05-27 | 北京师范大学 | Electroencephalogram feature extraction method and device, electronic equipment and storage medium |
| CN115105095A (en) * | 2022-08-29 | 2022-09-27 | 成都体育学院 | Electroencephalogram signal-based movement intention identification method, system and equipment |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103077205A (en) * | 2012-12-27 | 2013-05-01 | 浙江大学 | Method for carrying out semantic voice search by sound stimulation induced ERP (event related potential) |
| CN103340623A (en) * | 2013-06-27 | 2013-10-09 | 南方医科大学 | Method for extracting evoked potentials under high stimulation ratio |
| CN103340624A (en) * | 2013-07-22 | 2013-10-09 | 上海交通大学 | Method for extracting motor imagery electroencephalogram characteristics on condition of few channels |
| CN103885445A (en) * | 2014-03-20 | 2014-06-25 | 浙江大学 | Brain-controlling animal robot system and brain-controlling method of animal robot |
| CN104545901A (en) * | 2015-01-29 | 2015-04-29 | 中国科学院电子学研究所 | Electroencephalogram detecting system |
| CN105249961A (en) * | 2015-11-02 | 2016-01-20 | 东南大学 | Real-time driving fatigue detection system and detection method based on Bluetooth electroencephalogram headset |
| CN105395194A (en) * | 2015-12-14 | 2016-03-16 | 中国人民解放军信息工程大学 | Electroencephalogram (EEG) channel selection method assisted by functional magnetic resonance imaging |
| US20170003332A1 (en) * | 2012-05-09 | 2017-01-05 | Cardioinsight Technologies, Inc. | Channel integrity detection |
| CN106491125A (en) * | 2016-11-04 | 2017-03-15 | 广州视源电子科技股份有限公司 | Electroencephalogram state identification method and device |
| CN106491083A (en) * | 2016-10-11 | 2017-03-15 | 天津大学 | Head-wearing type intelligent wearing number of electrodes optimization and application for brain status monitoring |
| CN107184207A (en) * | 2017-05-12 | 2017-09-22 | 河海大学常州校区 | The CHANNEL OPTIMIZATION method of epilepsy EEG automatic detections based on sparse common space pattern |
-
2018
- 2018-05-14 CN CN201810455255.5A patent/CN108852348A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170003332A1 (en) * | 2012-05-09 | 2017-01-05 | Cardioinsight Technologies, Inc. | Channel integrity detection |
| CN103077205A (en) * | 2012-12-27 | 2013-05-01 | 浙江大学 | Method for carrying out semantic voice search by sound stimulation induced ERP (event related potential) |
| CN103340623A (en) * | 2013-06-27 | 2013-10-09 | 南方医科大学 | Method for extracting evoked potentials under high stimulation ratio |
| CN103340624A (en) * | 2013-07-22 | 2013-10-09 | 上海交通大学 | Method for extracting motor imagery electroencephalogram characteristics on condition of few channels |
| CN103885445A (en) * | 2014-03-20 | 2014-06-25 | 浙江大学 | Brain-controlling animal robot system and brain-controlling method of animal robot |
| CN104545901A (en) * | 2015-01-29 | 2015-04-29 | 中国科学院电子学研究所 | Electroencephalogram detecting system |
| CN105249961A (en) * | 2015-11-02 | 2016-01-20 | 东南大学 | Real-time driving fatigue detection system and detection method based on Bluetooth electroencephalogram headset |
| CN105395194A (en) * | 2015-12-14 | 2016-03-16 | 中国人民解放军信息工程大学 | Electroencephalogram (EEG) channel selection method assisted by functional magnetic resonance imaging |
| CN106491083A (en) * | 2016-10-11 | 2017-03-15 | 天津大学 | Head-wearing type intelligent wearing number of electrodes optimization and application for brain status monitoring |
| CN106491125A (en) * | 2016-11-04 | 2017-03-15 | 广州视源电子科技股份有限公司 | Electroencephalogram state identification method and device |
| CN107184207A (en) * | 2017-05-12 | 2017-09-22 | 河海大学常州校区 | The CHANNEL OPTIMIZATION method of epilepsy EEG automatic detections based on sparse common space pattern |
Non-Patent Citations (1)
| Title |
|---|
| HAN, JIUQI,等: "A Fast, Open EEG Classification Framework Based on Feature Compression and Channel Ranking", 《FRONTIERS IN NEUROSCIENCE》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111067515A (en) * | 2019-12-11 | 2020-04-28 | 中国人民解放军军事科学院军事医学研究院 | Intelligent airbag helmet system based on closed-loop control technology |
| CN111067515B (en) * | 2019-12-11 | 2022-03-29 | 中国人民解放军军事科学院军事医学研究院 | Intelligent airbag helmet system based on closed-loop control technology |
| CN112036229A (en) * | 2020-06-24 | 2020-12-04 | 宿州小马电子商务有限公司 | Intelligent bassinet electroencephalogram signal channel configuration method with demand sensing function |
| CN112036229B (en) * | 2020-06-24 | 2024-04-19 | 宿州小马电子商务有限公司 | Intelligent bassinet electroencephalogram signal channel configuration method with demand sensing function |
| CN113397562A (en) * | 2021-07-20 | 2021-09-17 | 电子科技大学 | Sleep spindle wave detection method based on deep learning |
| CN114533084A (en) * | 2022-02-09 | 2022-05-27 | 北京师范大学 | Electroencephalogram feature extraction method and device, electronic equipment and storage medium |
| CN115105095A (en) * | 2022-08-29 | 2022-09-27 | 成都体育学院 | Electroencephalogram signal-based movement intention identification method, system and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108852348A (en) | The collection point sort method and system of scalp brain electricity | |
| CN110693493B (en) | Epilepsia electroencephalogram prediction feature extraction method based on convolution and recurrent neural network combined time multiscale | |
| CN104771163B (en) | EEG feature extraction method based on CSP and R CSP algorithms | |
| CN110353673B (en) | Electroencephalogram channel selection method based on standard mutual information | |
| CN103971124B (en) | Multi-class motor imagery brain electrical signal classification method based on phase synchronization | |
| CN110657985B (en) | Gearbox fault diagnosis method and system based on singular value spectrum manifold analysis | |
| CN109657642A (en) | A kind of Mental imagery Method of EEG signals classification and system based on Riemann's distance | |
| CN109497990B (en) | Electrocardiosignal identity recognition method and system based on canonical correlation analysis | |
| CN109598222B (en) | EEMD data enhancement-based wavelet neural network motor imagery electroencephalogram classification method | |
| CN109480834A (en) | A kind of Method of EEG signals classification based on quick multiple dimension empirical mode decomposition | |
| CN113536882B (en) | Multi-class motor imagery electroencephalogram signal feature extraction and classification method | |
| CN111310656A (en) | Single motor imagery electroencephalogram signal identification method based on multi-linear principal component analysis | |
| CN112353397A (en) | Electrocardiogram signal identity recognition method | |
| CN117520891A (en) | Motor imagery electroencephalogram signal classification method and system | |
| CN108573207A (en) | EMD and CSP merges most optimum wavelengths space filtering brain electrical feature extracting method | |
| CN107550484B (en) | Magnetocardiogram signal quality evaluation method and system | |
| CN111191509B (en) | Electroencephalogram signal feature extraction and classification method based on SCSP-LDA | |
| CN113128384B (en) | Brain-computer interface software key technical method of cerebral apoplexy rehabilitation system based on deep learning | |
| CN112336369B (en) | Coronary heart disease risk index evaluation system of multichannel heart sound signals | |
| CN111310680B (en) | Radiation source individual identification method based on deep learning | |
| Yuwono et al. | Feature selection with Lasso for classification of ischemic strokes based on EEG signals | |
| CN116340825A (en) | Method for classifying cross-tested RSVP (respiratory tract protocol) electroencephalogram signals based on transfer learning | |
| CN116439663A (en) | Sleep stage system based on self-supervision learning and multi-view learning | |
| Saeed et al. | A Slantlet based statistical features extraction for classification of normal, arrhythmia, and congestive heart failure in electrocardiogram | |
| CN114048784A (en) | Electroencephalogram signal fatigue feature extraction method based on common spatial mode fuzzy wavelet packet |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181123 |
|
| RJ01 | Rejection of invention patent application after publication |