CN104679249B - A kind of Chinese brain-computer interface implementation method based on DIVA models - Google Patents
A kind of Chinese brain-computer interface implementation method based on DIVA models Download PDFInfo
- Publication number
- CN104679249B CN104679249B CN201510100475.2A CN201510100475A CN104679249B CN 104679249 B CN104679249 B CN 104679249B CN 201510100475 A CN201510100475 A CN 201510100475A CN 104679249 B CN104679249 B CN 104679249B
- Authority
- CN
- China
- Prior art keywords
- diva
- models
- eeg signals
- chinese
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000000354 decomposition reaction Methods 0.000 claims description 20
- 210000004556 brain Anatomy 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 17
- 238000004458 analytical method Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 8
- 238000012880 independent component analysis Methods 0.000 claims description 7
- 238000012706 support-vector machine Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 4
- 230000033764 rhythmic process Effects 0.000 claims description 4
- 230000002123 temporal effect Effects 0.000 claims description 4
- 230000001953 sensory effect Effects 0.000 claims description 3
- 210000000337 motor cortex Anatomy 0.000 claims description 2
- 230000001537 neural effect Effects 0.000 abstract description 6
- 238000000605 extraction Methods 0.000 abstract description 4
- 210000005036 nerve Anatomy 0.000 abstract description 4
- 238000012549 training Methods 0.000 description 7
- 238000005457 optimization Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 210000000056 organ Anatomy 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 230000001755 vocal effect Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004070 electrodeposition Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000007830 nerve conduction Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000008288 physiological mechanism Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000026676 system process Effects 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Abstract
The invention discloses a kind of Chinese brain-computer interface implementation method based on DIVA models, nerve signal is identified during by sound pronunciation to getting, so as to control to drive the motion of VODER.The present invention based on DIVA models, the collection of the EEG signals of Chinese speech pronunciation has been carried out using noninvasive mode and denoising, feature extraction and classification and identification has been carried out to signal, by recognition result be input into DIVA models in, by DIVA models by with brain-computer interface(BCI)Combination, construct meet Chinese speech sounding rule, with real physiologic meaning Chinese speech generation with obtain neural computation model so that for further construct " thought reader " with Chinese's Thinking Characteristics establish theory and practice basis.
Description
Technical field
The invention discloses a kind of Chinese brain-computer interface implementation method based on DIVA models, it is related to EEG Processing skill
Art field.
Background technology
By the thought process " reading " in brain out, this be always one of mankind dream.Boston University's voice reality
Test DIVA (the Directions Into Velocities of that Frank Gang Se professors (Guenther.F.H.) of room is proposed
Articulators) adaptive neural network model helps people to realize this dream.The main purpose of DIVA models be for
Inquire into the physiological mechanism when mankind produce voice.Sound is sent by the model of nerve conduction mechanism control vocal organs, and
Target error is adjusted by sensory system.Feedback control system (feedforward control are contained in DIVA models
System), the action of voice signal control vocal organs is spread out of by voice mapping ensemblen (speech sound map) and produces language
Sound;Another kind is feedback control system (feedback control system), and pronunciation is adjusted by auditory sensation area and somesthetic area
When with target voice produced by error.DIVA models are related to speech production and voice to manage in can emulating and describing relevant brain
Solve the correlation function in region.It can be said that it is a kind of in order to generate word, syllable or phoneme, for controlling simulation sound channel to move
Adaptive neural network model.The network model learns the motion of control simulation sound channel in a computer, corresponding to produce
Voice.After study is completed, model can produce any combination of voice.DIVA network models are many voices for studying for a long period of time
Influence, expected coarticulation and the coarticulation of carrying down of speed are carried when generation phenomenon includes the equivalent, context change of motion, speaks
A unified explanation is supplied.
Professor Gang Se goes out a kind of instrument of neural analysis system (voice brain machine interface system) based on DIVA modellings, allows
User only need to simply think of the language of oneself desired expression, and its content can be just directly changed into language by speech synthesis system
Sound.It is combined by with brain-computer interface technology, user can directly control sound to export, and its reaction speed is than famous scientist
The fast many of typewriting processing mode that Huo Jin is presently using.This neural analysis system is made up of two parts:BCI and voice
Synthesis system DIVA.In BCI, the producing method of EEG signals is a kind of wireless nerve electrode, for being chronically implanted the big of patient
Cortex, and the nerve signal for detecting then be used to drive continuous " motion " of VODER, for patient provides real-time language
Sound is exported.
At present, domestic researchers want to design the Chinese neural analysis system that " can read " Chinese thinking process
System, then learn English and Chinese in the difference of the processing mechanism of phonation deutocerebrum and find the brain for being suitable for Chinese speech pronunciation
Electric signal sorting technique is exactly two Important Problems in Chinese neural analysis system design.
The content of the invention
The technical problems to be solved by the invention are:For the defect of prior art, there is provided a kind of based on DIVA models
Chinese brain-computer interface implementation method, nerve signal is identified during by sound pronunciation to getting, so as to control to drive language
The motion of sound synthesizer.
The present invention uses following technical scheme to solve above-mentioned technical problem:
A kind of Chinese brain-computer interface implementation method based on DIVA models, it is characterised in that specific steps include:
The EEG signals of step one, acquisition people when Bopomofo pronunciation is carried out, the EEG signals to collecting enter line number
Word is filtered, and removes nonsensical EEG signals composition;
Step 2, predefining according to DIVA models, the selection brain area electrodes institute related to Chinese speech pronunciation task
The EEG signals data for collecting;
Step 3, EEG signals are carried out with independent component analysis, choose the EEG signals composition related with task of pronouncing, gone
Except the noise signal unrelated with pronunciation behavior;
Step 4,6 layers of WAVELET PACKET DECOMPOSITION are carried out to the signal after denoising using Db2 wavelet functions, obtain EEG signals
Characteristic vector;
Step 5, support vector machine classifier is trained using the characteristic vector obtained by WAVELET PACKET DECOMPOSITION;
Step 6, the EEG signals pronounced after user are entered using trained support vector machine classifier
Row identification;
Step 7, the recognition result produced using step 6 adjust the user interface provided in DIVA models to channel parameters
It is configured, so as to drive the DIVA models to carry out pronunciation task.
As present invention further optimization scheme, in the step one, 128 passage Neuralynx damage-free type brains are used
Electrographic recording analysis system carries out the acquisition of EEG signals.
As present invention further optimization scheme, in the step 3, EEGLAB tool boxes runICA methods pair are used
EEG signals carry out independent component analysis, and the noise signal of removal includes the dynamic signal of electro-ocular signal, flesh.
Used as present invention further optimization scheme, in the digital filtering in the step one, rejection frequency is in 0.3-
Signal beyond 30Hz scopes.
As present invention further optimization scheme, in the step 2, pronounce with Chinese speech in the DIVA models
Related brain area domain includes top bottom and the superior temporal gyrus in left side, uses Neuralynx damage-free type eeg recording analysis systems
Data acquisition is carried out, the data to the 42nd channel acquisition are processed.
As present invention further optimization scheme, in the step 4, in the case where no sensory stimuli is input into, will
The electric signal of the 9-13Hz of one kind produced in motor cortex is defined as the rhythm and pace of moving things, using 6 layers of decomposition of wavelet packet, by signal in orthogonal ground
Decompose 32 separate bands so that the sinusoidal signal of the rhythm and pace of moving things is incorporated into corresponding second band, obtain the decomposition of the second layer
Coefficient and energy are used as characteristic vector.
The present invention uses above technical scheme compared with prior art, with following technique effect:
The present invention has carried out the collection of EEG signals of Chinese speech pronunciation using noninvasive mode simultaneously based on DIVA models
Denoising, feature extraction and classification and identification are carried out to signal, recognition result is input into DIVA models, it is logical by DIVA models
Cross and brain-computer interface (BCI) combination, construct and meet Chinese speech sounding rule, the Chinese language with real physiologic meaning
Sound generates the neural computation model with acquisition, so that further to construct " the thought reader " with Chinese's Thinking Characteristics
Establish theory and practice basis.
Brief description of the drawings
Fig. 1 is the 128 passage Neuralynx eeg recordings analysis scanning cap electrode location drawing.
Fig. 2 is DIVA models to WAVELET PACKET DECOMPOSITION schematic diagram.
Fig. 3 is supporting vector machine model recognition result figure.
Fig. 4 is that Chinese speech brain machine interface system processes EEG signals and controls the flow chart of DIVA models pronunciation.
Specific embodiment
Embodiments of the present invention are described below in detail, the example of the implementation method is shown in the drawings, wherein ad initio
Same or similar element or element with same or like function are represented to same or similar label eventually.Below by ginseng
The implementation method for examining Description of Drawings is exemplary, is only used for explaining the present invention, and is not construed as limiting the claims.
Technical scheme is described in further detail below in conjunction with the accompanying drawings:
The present invention is extracted using the definition of DIVA models to voice EEG signals, and language must be carried out using WAVELET PACKET DECOMPOSITION
The feature extraction of sound EEG signals, is trained using characteristic vector to SVM classifier, and uses the SVMs after training
Model is identified to voice EEG signals, so that it is determined that pronunciation phonemes, finally control DIVA models to carry out voice output, including
Following steps:
Step 1, brain when being pronounced to Chinese person using 128 passage Neuralynx eeg recordings analysis systems of damage-free type
Electric signal is acquired, and carries out the digital filtering treatment of 0.3-30Hz, removes nonsensical EEG signals;
Step 2, in by the amended DIVA models of studies in China person, pair brain area domain related to Chinese speech pronunciation
Activation be made that clearly definition, these regions include left side top bottom and superior temporal gyrus, in the pillow in left side return,
In Neuralynx eeg recording analysis systems, these regions are 42,37, No. 28 electrode positions, so brain machine interface system is chosen
The EEG signals that these electrodes are collected are processed.
Step 3, EEG signals are carried out with independent component analysis, choose the EEG signals composition related with task of pronouncing, gone
Except the eye electricity unrelated with pronunciation behavior, the noise signal such as flesh is dynamic.Assuming that source signal is:
S (t)=[s1(t),s2(t),...sN(t)]T (1)
Wherein, T represents transposition computing, s1(t)...sNT () is signal phasor, t is the sampling time, and N is sweared for unknown source signal
The quantity of amount, then the EEG signals for collecting are:
Q (t)=[q1(t),q2(t),...qm(t)]T (2)
Wherein, m is number of electrodes, q1(t),q2(t),...qmT () is the EEG signals vector that each electrode is collected, and
EEG signals are made up of N number of unknown source signal, and each source signal is statistical iteration, and has space information weight
Weight, can be write as:
Q (t)=AS (t) (3)
In formula, A is M × N-dimensional full rank hybrid matrix, and each row of A are that each unknown source signal has fixed space to believe
Breath weight.By solving N × M sequency spectrum hybrid separation matrix Ws, it is acted on observation signal U gained and estimate signal X'=WU,
Most approached under the meaning of statistical iteration with position source signal X because eye electricity flesh is dynamic etc., interference signal is generally sub-Gaussian signals, institute
Super-Gaussian and sub-Gaussian signal is distinguished to use cost function regulation W to cause each component of X' independent as much as possible, so as to
By useful component in eeg data and eye electricity with flesh move etc. interference signal separate.
Step 4, for any wavelet series:
Wavelet packet component can be decomposed into:
Wherein, Djx(t1) it is signal x (t1) wavelet series,J wavelet packet subspace of n yardsticks is represented,It is chi
Degree function, X1It is shift factor, k is the number of plies of WAVELET PACKET DECOMPOSITION, and m is k layers and decomposes resulting decomposed signal, and j is resolution ratio
Level.For signal X (t), if j=0, k=0 and m=0, primary signal X (t) itself under level of resolution j is represented,
Being designated as X1. can carry out multi-level Wavelet Transform bag decomposition, if decomposing 1 time, i.e. k=1, m=0,1, in the 1st layer of WAVELET PACKET DECOMPOSITION
To decomposed signal X2 and X3.If decomposing 2 times, i.e. k=2, m=0,1,2,3, decomposed in the 2nd layer of WAVELET PACKET DECOMPOSITION
Signal X4, X5 and X6, X7, the like, we carry out 6 layers of decomposition of wavelet packet to EEG signals in step 4, by signal in orthogonal
Decompose 32 separate bands, extract the characteristic vector in second frequency band.
Step 5, hypothesis Xi, i=1,2 ... K1It is K1Individual training feature vector, class label is Y ∈ -1 ,+1, then supported
Vector machine is exactly to solve following optimization problem:
Wherein, W1It is weight vector, (xk,yk) it is sample point, C>0 is wrong punishment parameter, and it can be by cross validation mistake
Journey is selected.ξkIt is slack variable, b is deviation.Because it is found that two class data boundaries with | | W1||2It is inverse be directly proportional, so mesh
The Section 1 of scalar functions is that, for maximizing classification boundaries, the Section 2 in object function is regularization term, and representing allows not
Training mistake in the case of can dividing.
Step 6, using recognition result adjustment DIVA models in provide user interface channel parameters are configured so that
Driving DIVA models carries out pronunciation task.
As shown in figure 1, in DIVA models, the definition to Chinese sound pronunciation brain associative brain regions, these regions include
Top bottom and the superior temporal gyrus in left side are included, is returned in the pillow in left side.
In 128 passage Neuralynx damage-free type eeg recording analysis systems, 42,37 are chosen according to DIVA models,
The EEG signals of 28 passages carry out data processing.
As shown in Fig. 2 carrying out the schematic diagram of WAVELET PACKET DECOMPOSITION to signal, the schematic diagram carries out 4 layers of decomposition, X8- to signal
X15 is the 4th layer of decomposed signal.
As shown in figure 3, in the present invention to one group of EEG signals, preceding 20 pronunciations are as training sample, afterwards 20 pronunciations
As test sample, the result figure of classification and the identification of vector machine classifier is supported.
As shown in figure 4, the specific workflow of Chinese speech brain machine interface system includes data acquisition in the present invention,
Pretreatment, feature extraction, pattern-recognition, application interface, DIVA models pronounce these parts.
Below, a specific embodiment of the invention is given:
Step 1, use 128 passage Neuralynx eeg recordings analysis systems of damage-free type to 1 healthy mother tongue for
The male of Chinese carries out eeg signal acquisition, 40 Chinese phonetic alphabet a and 40 Chinese phonetic alphabet j are carried out altogether amounts to the pronunciation of 80 times appointing
Business, each phonation continues 600ms, and carries out 0.3-50Hz to the EEG signals for collecting using eeg recording analysis system
Digital filtering.
Step 2, the definition according to DIVA models to Chinese sound pronunciation associative brain regions, are read using EEGLAB tool boxes
The EEG signals for collecting, choose the EEG signals of 42,37,28 passages, and delete remaining EEG signals.
Step 3, using the RunICA methods in EEGLAB tool boxes EEG signals are carried out with independent component analysis, extracted and hair
The related EEG signals composition of sound task.
Step 4, in MATLAB environment, using Db2 wavelet functions to by the brain telecommunications after independent component analysis denoising
The decomposition of 6 floor number is carried out, signal in orthogonal is decomposed 32 separate bands, i.e. 0-7.8125Hz, 7.8125-
15.6250Hz ... .. is just incorporated in corresponding second band 9 to 13Hz sinusoidal signal into, obtains the decomposition of the second layer
Coefficient and energy are used as characteristic vector.
Step 5, the SVM classifier in LibSVM tool boxes is carried out using the characteristic vector obtained by WAVELET PACKET DECOMPOSITION
Training and recognize, preceding 20 characteristic vectors as training sample, 20 as test sample, using SVM classifier by training set
[0,1] interval is normalized to test set, whole sample space is divided into two classes, EEG signals when Chinese phonetic alphabet a pronounces are special
Levy as the 0th class, EEG signals feature when Chinese phonetic alphabet J pronounces is used as the 1st class.
Step 6, channel parameters user interface in DIVA models is controlled by classification results, so as to drive DIVA moulds
Type is pronounced.
Embodiments of the present invention are explained in detail above in conjunction with accompanying drawing, but the present invention is not limited to above-mentioned implementation
Mode, in the ken that those of ordinary skill in the art possess, can also be on the premise of present inventive concept not be departed from
Make a variety of changes.The above, is only presently preferred embodiments of the present invention, and any formal limit is not made to the present invention
System, although the present invention is disclosed above with preferred embodiment, but is not limited to the present invention, any to be familiar with this professional skill
Art personnel, without departing from the scope of the present invention, when using the technology contents of the disclosure above make it is a little change or
The Equivalent embodiments of equivalent variations are modified to, as long as being without departing from technical solution of the present invention content, according to technology reality of the invention
Matter, within the spirit and principles in the present invention, any simple modification, equivalent and the improvement made to above example
Deng still falling within the protection domain of technical solution of the present invention.
Claims (5)
1. a kind of Chinese brain-computer interface implementation method based on DIVA models, it is characterised in that specific steps include:
The EEG signals of step one, acquisition people when Bopomofo pronunciation is carried out, the EEG signals to collecting carry out digital filter
Ripple, removes nonsensical EEG signals composition;
Step 2, predefining according to DIVA models, the selection brain area electrodes related to Chinese speech pronunciation task are gathered
The EEG signals data for arriving;Wherein:The brain area domain related to Chinese speech pronunciation task includes the top in left side in DIVA models
Hypophyll and superior temporal gyrus, carry out data acquisition, to the 42nd channel acquisition using Neuralynx damage-free type eeg recording analysis systems
Data processed;
Step 3, EEG signals are carried out with independent component analysis, choose the EEG signals composition related with pronunciation task, remove and
The unrelated noise signal of pronunciation behavior;
Step 4,6 layers of WAVELET PACKET DECOMPOSITION are carried out to the signal after denoising using Db2 wavelet functions, obtain the feature of EEG signals
Vector;
Step 5, support vector machine classifier is trained using the characteristic vector obtained by WAVELET PACKET DECOMPOSITION;
Step 6, the EEG signals pronounced after user are known using trained support vector machine classifier
Not;
Step 7, the recognition result produced using step 6 are adjusted the user interface provided in DIVA models and channel parameters are entered
Row is set, so as to drive the DIVA models to carry out pronunciation task.
2. a kind of Chinese brain-computer interface implementation method based on DIVA models as claimed in claim 1, it is characterised in that:It is described
In step one, the acquisition of EEG signals is carried out using 128 passage Neuralynx damage-free type eeg recording analysis systems.
3. a kind of Chinese brain-computer interface implementation method based on DIVA models as claimed in claim 1, it is characterised in that:Institute
State in step 3, EEG signals are carried out with independent component analysis, the noise letter of removal using EEGLAB tool boxes runICA methods
Number include the dynamic signal of electro-ocular signal, flesh.
4. a kind of Chinese brain-computer interface implementation method based on DIVA models as claimed in claim 1, it is characterised in that:Institute
In stating the digital filtering in step one, signal of the rejection frequency beyond 0.3-30Hz scopes.
5. a kind of Chinese brain-computer interface implementation method based on DIVA models as claimed in claim 1, it is characterised in that:It is described
In step 4, in the case where no sensory stimuli is input into, the electric signal definition of the 9-13Hz of one kind that will be produced in motor cortex
It is the rhythm and pace of moving things, using 6 layers of decomposition of wavelet packet, signal in orthogonal is decomposed 32 separate bands so that the sinusoidal signal of the rhythm and pace of moving things is drawn
It is grouped into corresponding second band, obtains the decomposition coefficient and energy of the second layer as characteristic vector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510100475.2A CN104679249B (en) | 2015-03-06 | 2015-03-06 | A kind of Chinese brain-computer interface implementation method based on DIVA models |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510100475.2A CN104679249B (en) | 2015-03-06 | 2015-03-06 | A kind of Chinese brain-computer interface implementation method based on DIVA models |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104679249A CN104679249A (en) | 2015-06-03 |
CN104679249B true CN104679249B (en) | 2017-07-07 |
Family
ID=53314442
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510100475.2A Expired - Fee Related CN104679249B (en) | 2015-03-06 | 2015-03-06 | A kind of Chinese brain-computer interface implementation method based on DIVA models |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104679249B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105054928A (en) * | 2015-07-17 | 2015-11-18 | 张洪振 | Emotion display equipment based on BCI (brain-computer interface) device electroencephalogram acquisition and analysis |
CN110007769A (en) * | 2019-04-16 | 2019-07-12 | 山东建筑大学 | A kind of AC system and method based on asynchronous brain-computer interface |
CN113031766B (en) * | 2021-03-15 | 2022-09-23 | 哈尔滨工业大学 | Method for decoding Chinese pronunciation through electroencephalogram |
CN113143293B (en) * | 2021-04-12 | 2023-04-07 | 天津大学 | Continuous speech envelope nerve entrainment extraction method based on electroencephalogram source imaging |
CN113180695B (en) * | 2021-04-20 | 2024-04-05 | 西安交通大学 | Brain-computer interface signal classification method, system, equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102708288A (en) * | 2012-04-28 | 2012-10-03 | 东北大学 | Brain-computer interface based doctor-patient interaction method |
CN102999154A (en) * | 2011-09-09 | 2013-03-27 | 中国科学院声学研究所 | Electromyography (EMG)-based auxiliary sound producing method and device |
CN103310272A (en) * | 2013-07-02 | 2013-09-18 | 南京邮电大学 | Articulation method of Directions Into of Articulators (DIVA) neural network model improved on basis of track action knowledge base |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10136862B2 (en) * | 2012-05-30 | 2018-11-27 | The Board Of Trustees Of The Leland Stanford Junior University | Method of sonifying brain electrical activity |
-
2015
- 2015-03-06 CN CN201510100475.2A patent/CN104679249B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102999154A (en) * | 2011-09-09 | 2013-03-27 | 中国科学院声学研究所 | Electromyography (EMG)-based auxiliary sound producing method and device |
CN102708288A (en) * | 2012-04-28 | 2012-10-03 | 东北大学 | Brain-computer interface based doctor-patient interaction method |
CN103310272A (en) * | 2013-07-02 | 2013-09-18 | 南京邮电大学 | Articulation method of Directions Into of Articulators (DIVA) neural network model improved on basis of track action knowledge base |
Non-Patent Citations (1)
Title |
---|
基于汉语音位发音想象的脑机接口研究;杨晓芳,江铭虎;《中文信息学报》;20140930;第28卷(第5期);第13-24页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104679249A (en) | 2015-06-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104679249B (en) | A kind of Chinese brain-computer interface implementation method based on DIVA models | |
Guenther | Neural control of speech | |
Krishna et al. | Speech recognition with no speech or with noisy speech | |
CN103366618A (en) | Scene device for Chinese learning training based on artificial intelligence and virtual reality | |
CN110575141A (en) | Epilepsy detection method based on generation countermeasure network | |
CN113031766A (en) | Method for decoding Chinese pronunciation through electroencephalogram | |
Besedová et al. | Classification of brain activities during language and music perception | |
Wang et al. | Speech neuromuscular decoding based on spectrogram images using conformal predictors with Bi-LSTM | |
Wu et al. | Information-dense actions as contexts | |
CN107219927A (en) | A kind of mode identification method and eeg signal processing method and the intelligent home control system based on brain wave | |
Wang et al. | Simulation experiment of bci based on imagined speech eeg decoding | |
CN103310272B (en) | Based on the DIVA neural network model manner of articulation that sound channel action knowledge base is improved | |
Krishna et al. | Continuous Silent Speech Recognition using EEG | |
Wingfield et al. | On the similarities of representations in artificial and brain neural networks for speech recognition | |
Gómez-Vilda et al. | Time-frequency representations in speech perception | |
Chen et al. | Content classification with electroglottograph | |
CN113625870A (en) | Brain-computer interface control system and method based on language imagination | |
Huynh et al. | An investigation of ensemble methods to classify electroencephalogram signaling modes | |
Srisuwan et al. | Three steps of neuron network classification for EMG-based Thai tones speech recognition | |
Al-Anbary et al. | A Survey of Eeg Signals Prepossessing and Classification for Imagined Speech Application | |
Seleznyev et al. | Recognition of mentally pronounced Russian phonemes using convolutional Neural Networks and electroencephalography data | |
Kröger et al. | Phonetotopy within a neurocomputational model of speech production and speech acquisition | |
Shaobai et al. | Research on the mechanism for phonating stressed English syllables based on DIVA model | |
Naik et al. | Biomedical Signal Processing: A Modern Approach | |
Renner et al. | Forces of Nature: Dynamism and Agency in German Romanticism |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20200511 Address after: 710000 25 / F, block D, Tsinghua Science Park, Keji 2nd Road, Zhangba Street office, hi tech Zone, Xi'an City, Shaanxi Province Patentee after: Xi'an Huinao Intelligent Technology Co.,Ltd. Address before: 210003 Gulou District, Jiangsu, Nanjing new model road, No. 66 Patentee before: NANJING UNIVERSITY OF POSTS AND TELECOMMUNICATIONS |
|
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170707 |