CN113705324A - Multi-class motor imagery electroencephalogram signal feature extraction and classification method based on decision tree and CSP-SVM - Google Patents
Multi-class motor imagery electroencephalogram signal feature extraction and classification method based on decision tree and CSP-SVM Download PDFInfo
- Publication number
- CN113705324A CN113705324A CN202110726186.9A CN202110726186A CN113705324A CN 113705324 A CN113705324 A CN 113705324A CN 202110726186 A CN202110726186 A CN 202110726186A CN 113705324 A CN113705324 A CN 113705324A
- Authority
- CN
- China
- Prior art keywords
- svm
- motor imagery
- csp
- decision tree
- imagery electroencephalogram
- 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
- 238000003066 decision tree Methods 0.000 title claims abstract description 38
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000000605 extraction Methods 0.000 title claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 16
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 239000013598 vector Substances 0.000 claims description 9
- 238000002507 cathodic stripping potentiometry Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000002790 cross-validation Methods 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 238000013507 mapping Methods 0.000 claims description 2
- 239000011159 matrix material Substances 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000012795 verification Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 210000004556 brain Anatomy 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000033764 rhythmic process Effects 0.000 description 2
- 206010008190 Cerebrovascular accident Diseases 0.000 description 1
- 208000012661 Dyskinesia Diseases 0.000 description 1
- 206010033799 Paralysis Diseases 0.000 description 1
- 208000006011 Stroke Diseases 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007177 brain activity Effects 0.000 description 1
- 230000002490 cerebral effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 201000010901 lateral sclerosis Diseases 0.000 description 1
- 208000005264 motor neuron disease Diseases 0.000 description 1
- 201000000585 muscular atrophy Diseases 0.000 description 1
- 230000002232 neuromuscular Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000003863 physical function Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 210000000278 spinal cord Anatomy 0.000 description 1
- 208000020431 spinal cord injury Diseases 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
- G06F2218/08—Feature extraction
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention requests to protect a multi-class motor imagery electroencephalogram signal feature extraction and classification method based on a decision tree and a CSP-SVM, and the method comprises the following steps: s1, preprocessing the N motor imagery electroencephalogram signals; s2, processing the preprocessed motor imagery electroencephalogram data by utilizing the one-to-many CSP, thereby constructing N spatial filters and acquiring electroencephalogram signal characteristics; s3, classifying the features by using an SVM; s4, selecting the combination of the optimal spatial filter and the SVM to construct decision tree branches according to the classification result, so that one class of motor imagery electroencephalogram signals can be distinguished, and the N classification problem is changed into an N-1 classification problem; and S5, repeatedly constructing branches of the decision tree by using the one-to-many CSP and the SVM until a decision tree capable of distinguishing all classes is constructed. The result shows that the method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signals has a better classification result by verifying the BCI compatibility III Data Sets IIIa in the public Data set and comparing the verification with related documents.
Description
Technical Field
The invention belongs to the field of signal processing and pattern recognition, and particularly relates to a multi-class motor imagery electroencephalogram signal feature extraction and classification method based on a decision tree and a CSP-SVM.
Background
With the rapid development of fields such as brain science, artificial intelligence, computer technology and the like, brain-computer interface technology has brought about extensive research and attention. The brain-computer interface technology avoids the traditional neuromuscular channel, and directly realizes the communication control between the human brain and a computer or other external auxiliary equipment. In the medical field, the brain-computer interface can help a patient losing physical functions (a patient suffering from paralysis caused by diseases such as spinal cord lateral sclerosis, spinal cord injury, cerebral apoplexy, muscular atrophy and the like) to solve the problem of dyskinesia, so that the patient has certain ability of self-care life again. In addition, the brain-computer interface is widely applied to the fields of traffic driving, man-machine interaction, engineering and the like.
The brain-computer interface technology comprises 3 parts of signal acquisition, signal processing and external control system. Firstly, recording the brain activity of a user through the acquisition of electroencephalogram signals; then, the acquired electroencephalogram signals are converted into corresponding control instructions through signal processing and output to an external control system; and finally, the external control system completes related activities according to the received instruction. The signal processing is the most important component in the brain-computer interface technology, and comprises preprocessing, feature extraction and classification. The accuracy of signal processing directly affects the reliability and safety of the whole brain-computer interface. At present, methods for extracting electroencephalogram signal features include Common Spatial Pattern (CSP), deep learning, power spectrum estimation, and the like. The CSP algorithm obtains good effect in the feature extraction of two types of motor imagery electroencephalogram signals. In 2018, Zhang Wenqing et al combined with S transformation and a common space mode to perform feature extraction and classify by using SVM, and the average accuracy rate obtained on a Data set BCI composition IV Data sets 2b is 92.8%. In 2020, Meng et al optimized the co-spatial mode by selecting channels and bands, and the experimental results showed that the method gave an average accuracy of 90.25% on the public Data set BCI composition III Data set IVa. However, when the CSP is applied to feature extraction of multi-class motor imagery electroencephalogram signals, as the CSP needs to construct a plurality of spatial filters, the extracted features are complex in spatial distribution and low in discrimination, and further the identification accuracy is low.
Disclosure of Invention
In order to solve the problems, the invention provides a multi-class motor imagery electroencephalogram signal feature extraction and classification method based on a decision tree and a CSP-SVM. In the method, the multi-classification problem is converted into a plurality of two-classification problems by utilizing the decision tree idea, so that the advantages of the CSP and the SVM on the two-classification problems are fully exerted, and the motor imagery electroencephalogram signal can be more accurately identified.
In order to achieve the purpose, the invention adopts the technical scheme that: the multi-class motor imagery electroencephalogram signal feature extraction and classification method based on the decision tree and the CSP-SVM comprises the following steps:
s1, preprocessing the N motor imagery electroencephalogram signals;
s2, processing the preprocessed motor imagery electroencephalogram signal by using the one-to-many CSP, thereby constructing N spatial filters and acquiring the final characteristics of the electroencephalogram signal;
s3, classifying the characteristics in the step S2 by using SVM;
s4, selecting the combination of the optimal space filter and the SVM to construct the branch of the decision tree according to the classification result;
and S5, repeatedly constructing branches of the decision tree by using the one-to-many CSP and the SVM until a decision tree capable of distinguishing all classes is constructed.
The invention has the following advantages and beneficial effects:
aiming at the problem that when CSP extracts multi-class motor imagery features, the extracted features are complex in spatial distribution and low in discrimination so that the recognition accuracy is low, a multi-class motor imagery electroencephalogram feature extraction and classification method based on a decision tree and a CSP-SVM is provided, an optimal spatial filter constructed by one-to-many CSPs is selected by using classification results of the SVM, and the multi-classification problem is converted into a plurality of two-classification problems by using a decision tree idea so as to improve the accuracy of multi-class motor imagery electroencephalogram classification.
Tests on the public Data set BCI coordinates Data sets IIIa show that an optimal spatial filter is selected, better feature vectors can be obtained, and robustness of classification of a hyperplane constructed by the SVM can be improved by converting multi-classification into a plurality of two-classification. In order to more clearly compare the classification effects of different methods on multi-class motor imagery electroencephalogram signals, the algorithm model provided by the invention is compared with other CSP and SVM-based methods, and the algorithm provided by the invention has stronger robustness and higher accuracy.
Drawings
FIG. 1 is a branch creation process of multi-class motor imagery electroencephalogram signal decision trees based on decision trees and CSP-SVM;
fig. 2 is a classification decision tree for N classes of motor imagery electroencephalogram signals.
Detailed Description
The technical solutions in the embodiments of the present invention will be described in detail and clearly with reference to the accompanying drawings. The described embodiments are only some of the embodiments of the present invention.
As shown in fig. 1, the scheme of the present invention comprises the following steps:
and S1, preprocessing the N motor imagery electroencephalogram signals. EEGLAB band-pass filtering processing is adopted for the original motor imagery electroencephalogram signals, and a band-pass frequency band of 8-30 Hz is set, so that mu rhythms (8-13 Hz) and beta rhythms (13-30 Hz) related to motor imagery are reserved.
S2, the N spatial filters are constructed using "one-to-many" CSP for processing. One of the motion imagination types is regarded as one type, and the other N-1 motion imagination types are regarded as the other type, and CSP processing is carried out so as to construct a spatial filter. For N classes of motor imagery electroencephalogram signals, each class of motor imagery and other N-1 classes of motor imagery need to be regarded as another class, so N spatial filters are constructed. Mapping the original motor imagery electroencephalogram signal through a spatial filter to obtain a new signal ZpIn order to increase the variance of the characteristic values, a new signal Z is appliedp(p ═ 1,2, … 2m) variance logarithm and normalization processing as final feature fp： Zp(p＝1,2,…2m)
Where var () represents variance, and m represents m rows before and after the projection matrix when constructing the spatial filter.
And S3, classifying the features by using SVM. The features obtained in S2 are classified by SVM and parameters (radial basis kernel parameters and error penalty factors) influencing the support vector machine are selected by a grid search method and a cross validation method. The final decision function expression of the SVM becomes:
where f (x) is a decision function, sign () denotes a sign function, λiIs the Lagrangian vector, yiFor training set sample labels, x is the independent variable of the decision function, i.e. the feature vector of the test set sample, xiIs the feature vector of the training set sample, b is the offset vector, K (x, x)i) As a function of the radial basis kernel, K (x, x)i) The expression of (a) is:
K(x,xi)＝exp(-γ||x-xi||2) (3)
wherein γ is the function bandwidth of the radial basis kernel function and γ > 0.
S4, selecting the combination of the optimal space filter and SVM to construct decision tree branch, the 'one-to-many' CSP processes N classification problem to construct N space filters, each of which takes one class as one class and the rest N-1 classes as another class, so each space filter can only distinguish one class of signals. Each class of spatial filter performs feature extraction on the electroencephalogram signals, then, one SVM is used as a classifier to classify the features extracted by each spatial filter, N classification results are obtained, and then the combination of the filter (the optimal filter) corresponding to the best classification result and the SVM is selected to construct decision tree branches.
S5, repeatedly constructing branches of the decision tree by using the one-to-many CSP and the SVM. The construction of each decision tree branch enables the N classification problem to be changed into the N-1 classification problem, the distinguishable categories are removed from the data set after each decision tree branch is constructed, then the decision tree branches are repeatedly constructed by utilizing the 'one-to-many' CSP and the SVM, namely the steps S2, S3 and S4 are repeated, but the 'one-to-many' CSP becomes the original CSP method when only two categories are left, because the 'one-to-many' CSP takes one category as one category and the other categories as the other categories, because only two categories are left at the moment, the other categories are one category in total, and at the moment, the final complete binary tree is obtained by the CSP and SVM decision tree branch to realize the classification of all the categories. FIG. 2 is a complete binary tree established by classifying N types of motor imagery tasks by the multi-type motor imagery electroencephalogram signal feature extraction and classification method based on the decision tree and the CSP-SVM.
The above examples are to be construed as merely illustrative and not limitative of the remainder of the disclosure. After reading the description of the invention, the skilled person can make various changes or modifications to the invention, and these equivalent changes and modifications also fall into the scope of the invention defined by the claims.
Claims (7)
1. The multi-class motor imagery electroencephalogram signal feature extraction and classification method based on the decision tree and the CSP-SVM is characterized by comprising the following steps:
s1, preprocessing the N motor imagery electroencephalogram signals;
s2, processing the preprocessed motor imagery electroencephalogram signal by using the one-to-many CSP, thereby constructing N spatial filters and acquiring the final characteristics of the electroencephalogram signal;
s3, classifying the characteristics in the step S2 by using SVM;
s4, selecting the combination of the optimal space filter and the SVM to construct the branch of the decision tree according to the classification result;
and S5, repeatedly constructing branches of the decision tree by using the one-to-many CSP and the SVM until a decision tree capable of distinguishing all classes is constructed.
2. The method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signal based on the decision tree and the CSP-SVM as claimed in claim 1, wherein: the preprocessing of the N classes of motor imagery electroencephalogram signals specifically comprises the following steps: and carrying out band-pass filtering processing on the original motor imagery electroencephalogram signals by EEGLAB, and setting a band-pass frequency band of 8-30 Hz.
3. The method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signal based on the decision tree and the CSP-SVM as claimed in claim 1, wherein: the step of constructing the N spatial filters is as follows: one type of motor imagery is regarded as one type, the other N-1 types of motor imagery are regarded as the other type, CSP processing is carried out to construct a spatial filter, and therefore, for N types of motor imagery electroencephalogram signals, each type of motor imagery and the other N-1 types of motor imagery need to be subjected to CSP processing to construct N spatial filters.
4. The method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signal based on the decision tree and the CSP-SVM as claimed in claim 1 or 3, wherein: mapping the original motor imagery electroencephalogram signal through a spatial filter to obtain a new signal ZpNew signal Zp(p ═ 1,2, … 2m) variance logarithm and normalization processing as final feature fp：
Where var () represents variance, and m represents m rows before and after the projection matrix when constructing the spatial filter.
5. The method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signal based on the decision tree and the CSP-SVM as claimed in claim 1, wherein: the classifying the features in step S2 by using SVM is specifically to classify the final features by using SVM and select parameters (radial basis kernel parameters and error penalty factors) affecting the support vector machine by using a grid search method and a cross validation method.
6. The method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signal based on the decision tree and the CSP-SVM as claimed in claim 5, wherein: the final decision function expression of the SVM is as follows:
where sign () represents a sign function, λiIs the Lagrangian vector, yiFor training set sample labels, x is the independent variable of the decision function, i.e. the test set sample feature vector, xiFor training set sample feature vectors, b is a bias vector, K (x, x)i) As a function of the radial basis kernel, K (x, x)i) The expression of (a) is:
K(x,xi)＝exp(-γ||x-xi||2)
where γ is the function bandwidth of the radial basis kernel function.
7. The method for extracting and classifying the characteristics of the multi-class motor imagery electroencephalogram signal based on the decision tree and the CSP-SVM as claimed in claim 1, wherein: the process of repeatedly building decision trees removes the categories that can be distinguished from all categories, and then repeatedly builds decision tree branches using "one-to-many" CSPs and SVMs until all categories can be distinguished.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110726186.9A CN113705324A (en) | 2021-06-29 | 2021-06-29 | Multi-class motor imagery electroencephalogram signal feature extraction and classification method based on decision tree and CSP-SVM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110726186.9A CN113705324A (en) | 2021-06-29 | 2021-06-29 | Multi-class motor imagery electroencephalogram signal feature extraction and classification method based on decision tree and CSP-SVM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113705324A true CN113705324A (en) | 2021-11-26 |
Family
ID=78648271
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110726186.9A Pending CN113705324A (en) | 2021-06-29 | 2021-06-29 | Multi-class motor imagery electroencephalogram signal feature extraction and classification method based on decision tree and CSP-SVM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113705324A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751563A (en) * | 2010-01-21 | 2010-06-23 | 上海大学 | EEG signal identification method based on LabWindows/CVI and Matlab hybrid programming |
| CN102184015A (en) * | 2011-05-16 | 2011-09-14 | 上海电机学院 | Method for quickly sorting EEG (Electroencephalo-graph) signal based on threshold analysis |
| CN102629156A (en) * | 2012-03-06 | 2012-08-08 | 上海大学 | Method for achieving motor imagery brain computer interface based on Matlab and digital signal processor (DSP) |
| CN104091172A (en) * | 2014-07-04 | 2014-10-08 | 北京工业大学 | Characteristic extraction method of motor imagery electroencephalogram signals |
| CN108960299A (en) * | 2018-06-15 | 2018-12-07 | 东华大学 | A kind of recognition methods of multiclass Mental imagery EEG signals |
| US20190113973A1 (en) * | 2012-09-14 | 2019-04-18 | Interaxon Inc | Systems and methods for collecting, analyzing, and sharing bio-signal and non-bio-signal data |
| WO2019140047A1 (en) * | 2018-01-10 | 2019-07-18 | The Trustees Of Princeton University | System and method for smart, secure, energy-efficient iot sensors |
| CN111338473A (en) * | 2020-02-18 | 2020-06-26 | 珠海爬山虎科技有限公司 | Brain electrical signal-based rehabilitation training method, computer device and computer-readable storage medium |
| CN113011239A (en) * | 2020-12-02 | 2021-06-22 | 杭州电子科技大学 | Optimal narrow-band feature fusion-based motor imagery classification method |
-
2021
- 2021-06-29 CN CN202110726186.9A patent/CN113705324A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751563A (en) * | 2010-01-21 | 2010-06-23 | 上海大学 | EEG signal identification method based on LabWindows/CVI and Matlab hybrid programming |
| CN102184015A (en) * | 2011-05-16 | 2011-09-14 | 上海电机学院 | Method for quickly sorting EEG (Electroencephalo-graph) signal based on threshold analysis |
| CN102629156A (en) * | 2012-03-06 | 2012-08-08 | 上海大学 | Method for achieving motor imagery brain computer interface based on Matlab and digital signal processor (DSP) |
| US20190113973A1 (en) * | 2012-09-14 | 2019-04-18 | Interaxon Inc | Systems and methods for collecting, analyzing, and sharing bio-signal and non-bio-signal data |
| CN104091172A (en) * | 2014-07-04 | 2014-10-08 | 北京工业大学 | Characteristic extraction method of motor imagery electroencephalogram signals |
| WO2019140047A1 (en) * | 2018-01-10 | 2019-07-18 | The Trustees Of Princeton University | System and method for smart, secure, energy-efficient iot sensors |
| CN108960299A (en) * | 2018-06-15 | 2018-12-07 | 东华大学 | A kind of recognition methods of multiclass Mental imagery EEG signals |
| CN111338473A (en) * | 2020-02-18 | 2020-06-26 | 珠海爬山虎科技有限公司 | Brain electrical signal-based rehabilitation training method, computer device and computer-readable storage medium |
| CN113011239A (en) * | 2020-12-02 | 2021-06-22 | 杭州电子科技大学 | Optimal narrow-band feature fusion-based motor imagery classification method |
Non-Patent Citations (3)
| Title |
|---|
| 卢官明等: "基于ＳＶＭ 的新生儿疼痛表情识别", 《南京邮电大学学报（ 自然科学版）》, vol. 28, no. 6, 31 December 2008 (2008-12-31), pages 6 - 11 * |
| 李丽君等: "基于四类运动想象任务的脑电信号识别", 《中国组织工程研究与临床康复》, vol. 15, no. 48, 30 November 2011 (2011-11-30), pages 9003 - 9006 * |
| 李丽君等: "基于四类运动想象任务的脑电信号识别", 《中国组织工程研究与临床康复》, vol. 15, no. 48, pages 9003 - 9006 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Miao et al. | Spatial-frequency feature learning and classification of motor imagery EEG based on deep convolution neural network | |
| CN105740887A (en) | Electroencephalogram feature classification method based on PSO-SVM (Particle Swarm Optimization-Support Vector Machine) | |
| CN105956624A (en) | Motor imagery electroencephalogram classification method based on space-time-frequency optimization feature sparse representation | |
| CN107961120B (en) | Intelligent wheelchair control system based on electroencephalogram control | |
| CN109299647B (en) | Vehicle control-oriented multitask motor imagery electroencephalogram feature extraction and mode recognition method | |
| CN113486752B (en) | Emotion recognition method and system based on electrocardiosignal | |
| Yang et al. | A novel method for identifying electrocardiograms using an independent component analysis and principal component analysis network | |
| Liu et al. | Improved GMM with parameter initialization for unsupervised adaptation of brain–computer interface | |
| CN109657646B (en) | Method and device for representing and extracting features of physiological time series and storage medium | |
| CN105550705A (en) | Electroencephalogram (EEG) signal recognition method based on improved self-training learning | |
| CN110074779A (en) | A kind of EEG signal identification method and device | |
| CN111310656A (en) | Single motor imagery electroencephalogram signal identification method based on multi-linear principal component analysis | |
| CN115221969A (en) | Motor imagery electroencephalogram signal identification method based on EMD data enhancement and parallel SCN | |
| CN115238796A (en) | Motor imagery electroencephalogram signal classification method based on parallel DAMSCN-LSTM | |
| Fabien et al. | Studying the use of fuzzy inference systems for motor imagery classification | |
| Tang et al. | Motor Imagery EEG Decoding Based on New Spatial-Frequency Feature and Hybrid Feature Selection Method. | |
| CN113662558B (en) | Intelligent classification method for distinguishing brain electrical blink artifacts and forehead electrode epilepsy-like discharge | |
| Tang et al. | Research on extraction and classification of EEG features for multi-class motor imagery | |
| Zhang et al. | A new convolutional neural network for motor imagery classification | |
| CN109144277B (en) | Method for constructing intelligent vehicle controlled by brain based on machine learning | |
| Ebrahimpour et al. | EEG-based motor imagery classification using wavelet coefficients and ensemble classifiers | |
| CN113705324A (en) | Multi-class motor imagery electroencephalogram signal feature extraction and classification method based on decision tree and CSP-SVM | |
| CN116919422A (en) | Multi-feature emotion electroencephalogram recognition model establishment method and device based on graph convolution | |
| Zhang et al. | An ECoG-based binary classification of BCI using optimized extreme learning machine | |
| CN114781461B (en) | Target detection method and system based on auditory brain-computer interface |
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 |