Patents
Literature
39 results about "Eeg classification" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Emotion recognition system based on depth learning and brain-computer interface as well as application
ActiveCN111616721AImplement feedbackAccurate acquisitionSensorsPsychotechnic devicesPortable EEGAcquisition apparatus
The invention discloses an emotion recognition system based on depth learning and a brain-computer interface as well as application. The emotion recognition system comprises portable EEG acquisition equipment, an emotion EEG classification module and an emotion classification display module, wherein the portable EEG acquisition equipment acquires emotion EEG signals from the brain of a subject; the emotion EEG classification module analyzes the emotion EEG signals, establishes a relation model between the EEG signals and emotions, and performs emotion classification on the input emotion EEG signals through the relation model; and the emotion classification module displays recognized emotion types and prompts the emotion state of the subject. The emotion recognition system based on depth learning and the brain-computer interface as well as application can realize accurate acquisition, effective identification and accurate classification of the emotion EEG signals, directly promote the emotion state and realize an emotion state monitoring function. The corresponding relation between the EEG signals and the emotion types is directly established, and feedback of emotions is realized.
Owner:TIANJIN UNIV +1
Electroencephalogram classification detection device based on lacuna characteristics
InactiveCN103190904AMeet the requirements for online classificationFast trainingDiagnostic recording/measuringSensorsFeature extractionAudio power amplifier
An electroencephalogram (EEG) classification detection device based on lacuna characteristics belongs to the technical field of electroencephalogram automatic detection. The EEG classification detection device comprises a multi-way EEG amplifier, a data collection card and a computer which are sequentially connected through a circuit. A signal preprocessing module, a signal segmentation module, a lacuna characteristic extraction module, a Bayes linear discriminant analysis classification module and a threshold judgment module are built in the computer. The multi-way EEG amplifier first amplifies EEG signals, then the data collection card collects the EEG signals and transmits the signals to the computer, finally the modules in the computer are utilized to conduct preprocessing and segmentation on the EEG signals and calculate the lacuna characteristics of the signals, a Bayes linear discriminant analysis classification device is utilized to classify the EEG lacuna characteristics, and the threshold judgment module is used for marking the classification and obtaining a result. The EEG classification detection device has the advantages of being simple in characteristic operation, high in practice and classification speed, high in classification accuracy and capable of achieving good classification detection effect.
Owner:SHANDONG UNIV
Motion imagination EEG classification processing method based on sparse representation classification algorithm
InactiveCN106446762AReduce computing pressureSimplify complexityCharacter and pattern recognitionBandpass filteringAlgorithm
The invention belongs to the field of EEG signal identification and artificial intelligence, and particularly relates to a motion imagination EEG classification processing method based on a sparse representation classification algorithm. The invention includes that a subject wears a scalp electrode cap with wireless or wired transmission on head in an environment with less external interference and imagines the movements based on the prompting, the EEG cap detects the EEG signals of the subject, and then the EEG signals are subjected to preliminary bandpass filtering, and the filtered EEG signals are imported to a host computer to be stored; and the host computer software organizes and marks the imported original EEG signals and makes a sample set and a test set for classification. According to the invention, only a large dictionary matrix is calculated and constructed once, the subsequent processing is calculated by means of sparse representation coefficient matrix, so that the data size of multi-channel motion imagination EEG information is greatly reduced, the data calculation pressure is reduced, and the operation speed is improved.
Owner:HARBIN ENG UNIV
EMD and gaussian kernel function SVM-based EEG emotion classification method
InactiveCN107273841ASolving linear inseparability problemsCharacter and pattern recognitionSignal classificationDecomposition
The invention discloses an EMD and gaussian kernel function SVM-based EEG emotion classification method. Aiming at the problem that the accuracy of EEG signal classification is not high, an Empirical Mode Decomposition (EMD) technology and SVMs are combined, first EMD is performed on an EEG signal to obtain a plurality of modal components, each modal component contains effective information of different frequencies, then frequency energy is used as a quantitative criterion of each modal component, i.e., each EEG signal can obtain different characteristic values, and the characteristic values are used as characteristic values of an EEG sequence to perform a next step of sample value classification. Experiments show that the EMD and gaussian kernel function SVM-based EEG classification method can improve the accuracy of EEG signal classification.
Owner:BEIJING UNIV OF TECH
Prior knowledge-based multi-modal ethnic identity sense quantitative measurement method
InactiveCN108256579AAccurate classificationImprove objectivityCharacter and pattern recognitionDiagnostic recording/measuringFeature extractionSurvey result
The invention relates to a prior knowledge-based multi-modal ethnic identity sense quantitative measurement method and belongs to the intelligent pattern recognition and ethnology research crossing filed. According to the measurement method, ethnic minority emotion pictures are adopted as basic materials; ethnic minority identity sense survey questionnaire questions are converted into the induction materials of electrophysiological signals; after the EEG information of subjects is acquired, the EEG information implying identity senses is classified through a fuzzy neural network learning-basedEEG emotion feature extraction and classification algorithm; questionnaire survey results and EEG classification results are fused, so that ethnic identity sense quantitative measurement is realized;and the objective feeling of the subjects for the ethnic identity sense and national identity sense can be automatically analyzed. With the method of the invention adopted, the subjective influence of questionnaire surveys can be decreased, the objectivity of identity sense classification can be enhanced, and ethnic identity sense quantitative classification is more accurate and more efficient.
Owner:MINZU UNIVERSITY OF CHINA
EEG classification method based on Fisher discrimination sparse extreme learning machine
InactiveCN108122004AImprove accuracyDiscriminatingCharacter and pattern recognitionLearning machineHuman–machine interface
The invention proposes an EEG classification method based on a Fisher discrimination sparse extreme learning machine. The method comprises the steps: firstly training a structural dictionary through aFisher criterion; secondly obtaining a more discriminative sparse coefficient according to the dictionary, and obtaining a more effective feature signal; finally carrying out the classification of anew feature signal through an extreme learning machine algorithm, thereby improving the accuracy of multi-motion imaginary task classification. The method is good in application prospect in the fieldof brain-machine interfaces.
Owner:HANGZHOU DIANZI UNIV
Brain function network feature extraction method based on dynamic directional transfer function
ActiveCN111528836AImprove adaptabilityImprove dynamic characteristicsDiagnostic recording/measuringSensorsNetwork connectionMachine learning
The invention discloses a brain function network feature extraction method based on a dynamic directional transfer function. The method mainly comprises the steps of firstly, performing preprocessingsuch as common average reference and lead optimization on an original motor imagery electroencephalogram signal; secondly, calculating a network connection edge of the preprocessed electroencephalogram signal by adopting a proposed DDTF algorithm, and respectively constructing brain function networks of different frequency bands; calculating network characteristic parameter outflow information andinformation flow gain according to the brain function network, and fusing two characteristic parameters in series to serve as characteristic vectors to be sent into a support vector machine for characteristic evaluation; and finally, determining an optimal parameter and an optimal frequency band according to a recognition rate closed loop to obtain a final classification result. The method is used for constructing the motor imagery brain function network, the network parameters obtained through calculation are used for MI-EEG feature extraction, the method not only can accurately describe change characteristics of MI-EEG in a frequency domain, but also accurately reflect a dynamic evolution process of BFN, and improvement of the MI-EEG classification accuracy is greatly facilitated.
Owner:BEIJING UNIV OF TECH
Symbol transfer entropy and brain network feature calculation method based on time-frequency energy
ActiveCN112932505AReflect the differenceComply with nonlinear dynamic characteristicsDiagnostic recording/measuringSensorsBiologyContinuous wavelet
The invention discloses a symbol transfer entropy and brain network feature calculation method based on time-frequency energy, which comprises the following steps: firstly, preprocessing collected motor imagery electroencephalogram signals (MI-EEG) based on common average reference; then, continuous wavelet transform is carried out on each lead MI-EEG, a time-frequency-energy matrix of each lead MI-EEG is obtained, time-energy sequences corresponding to each frequency in a frequency band closely related to motor imagery are spliced in sequence, and a one-dimensional time-frequency energy sequence of the lead is obtained; further, symbol transfer entropy between any two lead time-frequency energy sequences is calculated, a brain connectivity matrix is constructed, and matrix elements are optimized by using a Pearson feature selection algorithm; and finally, calculating the degree and the middle centrality of the brain function network, and forming a feature vector for MI-EEG classification. The result shows that the frequency domain feature and the nonlinear feature of the MI-EEG can be effectively extracted, and compared with a traditional feature extraction method based on the brain function network, the method has obvious advantages.
Owner:BEIJING UNIV OF TECH
Machine learning model robustness evaluation method based on noise data
ActiveCN110717602ARobustRich cognitionKernel methodsSustainable transportationOriginal dataData acquisition
The invention provides a machine learning model robustness evaluation method based on noise data. The method comprises the steps of original data set processing, noise data acquisition, model training, model prediction, accuracy reduction ratio calculation and model robustness evaluation. The original data set processing comprises the steps of collecting an original data set with a correct percentage label, and dividing an original training set and an original test set by adopting 10 times of 10-fold cross validation. The noise data acquisition comprises the following steps: on the basis of anoriginal training set, extracting t'= | D | * alpha data by adopting a stratified sampling method, and replacing a label of the data with an error label, and alpha is a noise data rate. Model training comprises the step of respectively inputting an original training set and a training set mixed with noise data to respectively construct an original model and a new model based on a common classification algorithm. Model prediction includes performing accuracy evaluation on an original model and a new model based on an original test set. Accuracy decline ratio calculation includes calculating arate of decline in accuracy of the new model relative to the original model. Model robustness evaluation comprises the steps of comparing the size of the rate of accuracy reduction in the transverse direction and the longitudinal direction, measuring the robustness of the model, and achieving the standard of judging the robustness of the model.
Owner:NANJING UNIV
Video awakening degree classification method and device and computer device
InactiveCN109190658AImprove classification accuracyCharacter and pattern recognitionEeg dataClassification methods
The invention discloses a video awakening degree classification method and device and a computer device, wherein the classification method comprises the following steps: step 1, collecting an EEG signal when a tester watches a video, and obtaining EEG data; step 2: standardizing the obtained EEG data according to columns; step 3, carrying out feature selection on the standardized EEG data to obtain EEG classification features; step 4, classifying the video according to the obtained EEG classification features. The embodiment of the invention classifies the video according to the EEG classification features by standardizing the collected data, and then extracts the EEG classification features from the EEG signals by using the SBMLR algorithm for the standardized processed data, thereby improving the accuracy of the awakening degree classification of the video.
Owner:CHINA ACADEMY OF ELECTRONICS & INFORMATION TECH OF CETC
EEG classification method adaptive to different sampling frequencies
PendingCN109009097ASimple and Effective CompletionImprove universalityDiagnostic recording/measuringSensorsSignal classificationSample length
The invention relates to an EEG (electroencephalogram) classification method adaptive to different sampling frequencies, and relates to a signal classification method. The EEG classification method adaptive to different sampling frequencies includes the steps: 1) construction of a CNN-E classification model based on a convolutional neural network; and 2) training and testing method for sample dataof different lengths. The model can be used to learn and classify EEG signals at different sampling frequencies, and can be adaptive to signals of different lengths. The model and a traditional classification method based on feature extraction can analyze the possible problems in the classification of EEG signals at different sampling frequencies. The network model CNN-E can be adaptive to various lengths of data by means of autonomous learning of the sample data, and by means of a simple and effective completion method. The experimental results show that the network model CNN-E can achieve well classification effect and preferable universality for classification of EEG signals at the same sampling frequency, classification of EEG signals at different sampling frequencies, and classification of EEG signals with different sample lengths.
Owner:XIAMEN UNIV
An epilepsy electroencephalogram classification method based on an iterative graph convolutional neural network
The application discloses a method for classifying epilepsy electroencephalogram (EEG) based on an iterative graph convolutional neural network, calculates the node similarity and distance similarity of epilepsy EEG data as an original graph structure, introduces a multi-head graph attention mechanism for node similarity measurement learning, iteratively optimizes the parameters of the graph structure and the graph convolutional neural network, finds the optimal graph structure and achieves the optimal epilepsy EEG classification effect. Experiments are conducted on TUEP bipolar and unipolar montage datasets and TUAB and MPI LEMON joint datasets to verify the effectiveness of the method. The method has better epilepsy EEG classification effect and obtains more accurate electroencephalogram structure.
Owner:CSSC HUMAN FACTORS ENG RES INST (QINGDAO) CO LTD
EEG classification method based on helm combined with ptsne and lda feature fusion
ActiveCN109977810BImprove classification accuracyAdding first-order normal form constraintsCharacter and pattern recognitionBiologyMachine learning
The invention discloses a motor imagery EEG classification method based on HELM combined with PTSNE manifold and LDA feature fusion, and improves the classification accuracy. In terms of feature extraction, on the one hand, using PCA combined with LDA to extract linear features can not only eliminate noise, but also consider the label information of training data; Complex nonlinear intrinsic manifold features. In terms of feature classification, the HELM algorithm with high classification accuracy is used to classify and identify motor imagery EEG signals.
Owner:BEIJING UNIV OF TECH
Emotion recognition system and application based on deep learning and brain-computer interface
ActiveCN111616721BImplement feedbackAccurate acquisitionSensorsPsychotechnic devicesPortable EEGMedicine
An emotion recognition system and application based on deep learning and brain-computer interface, comprising a sequentially connected portable EEG acquisition device, an emotional EEG classification module and an emotion classification display module, the portable EEG acquisition device collects data from the brain of a subject Emotional EEG signal, the emotional EEG classification module analyzes the emotional EEG signal, establishes a relationship model between the emotional EEG signal and emotion, and through the relationship model, the input emotional EEG signal is analyzed. Emotional classification, the emotional classification display module displays the recognized emotional types and prompts the emotional state of the subject. The emotion recognition system and application based on deep learning and brain-computer interface of the present invention can realize accurate acquisition, effective identification and correct classification of emotional EEG signals, intuitively prompt emotional states, and realize the function of emotional state monitoring. The invention directly establishes the corresponding relationship between the EEG signal and the emotion type, and realizes the feedback to the emotion.
Owner:TIANJIN UNIV +1
Target recognition methods based on electroencephalogram signals in natural reading environment
A target recognition method based on an electroencephalogram (EEG) signal in a natural reading environment is provided, including: presenting a fuzzy semantic target to a subject through a user interaction window on a display device; acquiring EEG signals of the subject via a wireless EEG acquisition device to obtain a target EEG signal corresponding to the fuzzy semantic target; determining a binary classification result corresponding to the fuzzy semantic target based on the target EEG signal through a trained EEG classification model; in response to the binary classification result indicating that the fuzzy semantic target is recognized, determining a semantic category to which the fuzzy semantic target belongs; and controlling, based on the semantic category, the display device to highlight annotations on the user interaction window; wherein highlighting annotations includes highlighting a text or an image related to the fuzzy semantic target and displaying associated information of the semantic category.
Owner:TIANJIN UNIV
Rsvp eeg classification method based on spatiotemporal progressive attention model
The application discloses a kind of based on fast sequence visual presentation electroencephalogram classification method of spatiotemporal progressive attention model, comprising: obtaining the electroencephalogram data to be classified;The electroencephalogram data to be classified is input to the trained spatiotemporal progressive attention model and is processed, obtains the category of electroencephalogram data to be classified;Wherein, the trained spatiotemporal progressive attention model with the preset data after acquisition and processing as training data set, to extract the spatial feature and time feature of electroencephalogram data as the purpose, after training to initial spatiotemporal progressive attention model, obtain.The electroencephalogram data is obtained by collecting subject in the mode of fast sequence visual presentation and watching image sequence, and the image sequence includes infrared small target image and non-target image.The present application can maximize the use of electrode and time segment useful information, while minimizing irrelevant information.
Owner:XIDIAN UNIV
Motor imagery EEG classification method based on sparse representation of space-time-frequency optimized features
ActiveCN105956624BImprove accuracyImprove robustnessCharacter and pattern recognitionSignal classificationClassification methods
The invention discloses a motor imagery EEG classification method based on the sparse representation of space-time-frequency optimized features, which mainly uses linear discriminant criteria to select the most favorable lead, time segment and frequency segment for classification, and extracts EEG features through a co-space pattern algorithm. Finally, classification is performed according to the feature sparse representation. The invention includes electroencephalogram signal preprocessing, lead selection, time-frequency block selection, feature extraction and feature classification. The results show that the method of the present invention can effectively select the most favorable lead, time segment and frequency segment for classification, and the sparse representation of the features extracted by the co-space pattern algorithm can achieve better classification effect. Compared with the existing algorithms, this method can automatically select the most beneficial space-time-frequency parameters for classification, and combine the features in the optimal time-frequency block, which is conducive to improving the accuracy of motor imagery EEG signal classification.
Owner:SOUTHEAST UNIV
Electroencephalogram classification detection device based on lacuna characteristics
InactiveCN103190904BMeet the requirements for online classificationFast trainingDiagnostic recording/measuringSensorsFeature extractionAudio power amplifier
An electroencephalogram (EEG) classification detection device based on lacuna characteristics belongs to the technical field of electroencephalogram automatic detection. The EEG classification detection device comprises a multi-way EEG amplifier, a data collection card and a computer which are sequentially connected through a circuit. A signal preprocessing module, a signal segmentation module, a lacuna characteristic extraction module, a Bayes linear discriminant analysis classification module and a threshold judgment module are built in the computer. The multi-way EEG amplifier first amplifies EEG signals, then the data collection card collects the EEG signals and transmits the signals to the computer, finally the modules in the computer are utilized to conduct preprocessing and segmentation on the EEG signals and calculate the lacuna characteristics of the signals, a Bayes linear discriminant analysis classification device is utilized to classify the EEG lacuna characteristics, and the threshold judgment module is used for marking the classification and obtaining a result. The EEG classification detection device has the advantages of being simple in characteristic operation, high in practice and classification speed, high in classification accuracy and capable of achieving good classification detection effect.
Owner:SHANDONG UNIV
EEG intent decoding method based on non-negative cp decomposition model
ActiveCN111012335BImprove classification accuracyEfficient decodingMedical data miningKernel methodsPattern recognitionDecoding methods
The present invention provides an EEG intent decoding method based on a non-negative CP decomposition model. The method extracts the time component features of the EEG of different subjects in the boundary avoidance task, optimizes the feature dimension by using 2-DPCA, and uses The support vector machine classifies the features, which can reflect the differences in the EEG of the subject in different states, and the accuracy of the EEG classification of a single subject is high; and this method uses the interaction between the various modes of the EEG, and uses the channel The components and frequency components obtain the time component characteristics of the EEG, and the characteristics of the obtained EEG time components are well separable. By optimizing its dimension, it can effectively detect the left and right hand movements in the boundary avoidance task. EEG intent to decode.
Owner:YANSHAN UNIV
Training methods and applications of EEG signal classification models for weak asymmetric visual stimuli
This invention belongs to the field of electroencephalogram (EEG) signal classification technology, and relates to a training method and application of an EEG signal classification model for weak asymmetric visual stimuli. The training method includes: acquiring a frequency domain positive frequency signal dataset based on the original signals and dividing it into a training set, a test set, and a validation set; constructing an EEG signal classification model and initializing a mode decomposition matrix with propagable gradients; training the model involves: dividing the training set into at least two mutually exclusive subsets, iteratively updating the mode decomposition matrix using backpropagation of the center band loss, bandwidth loss, reconstruction loss, and classification loss for each mutually exclusive subset, and iteratively updating the parameters of the model's convolutional neural network and fully connected layers using backpropagation of the classification loss. The validation set is used to verify the model's discriminative performance, and the model parameters with the best discriminative performance are saved. The test set is used to evaluate the model with the optimal discriminative performance parameters. This invention can significantly improve the discriminative performance in EEG classification tasks involving weak asymmetric visual stimuli.
Owner:TIANJIN UNIV
A kind of electroencephalogram signal classification method, device, equipment and medium
ActiveCN109766845BImprove accuracyCharacter and pattern recognitionVisual evoked potentialsSignal classification
The present application provides a method for classifying EEG signals, the method comprising: obtaining an EEG signal sequence to be identified; inputting the EEG signal sequence to be identified into an EEG signal classification model to obtain the EEG signal sequence to be identified EEG classification information, the EEG classification information includes motor imagery EEG signals and steady-state visual evoked potential EEG signals.
Owner:XUANWU HOSPITAL OF CAPITAL UNIV OF MEDICAL SCI +1
Voting strategy classification method of motor imagery EEG signal based on extremely fast learning machine
ActiveCN103413050BReduce randomnessImprove classification accuracySpecial data processing applicationsLearning machineFeature vector
The invention belongs to the field of mode recognition and a brain-machine interface and discloses a motor imagery electroencephalogram voting strategy sorting method based on extreme learning machines. The motor imagery electroencephalogram voting strategy sorting method comprises the following steps: dividing an original motor imagery electroencephalogram into S sections of sub-signals; carrying out dimensionality reduction on each section of sub-signal by a principal component analysis method; carrying out secondary dimensionality reduction on a feature vector subjected to the dimensionality reduction by a linear discrimination analysis method; carrying out the same processing on the S sections of sub-signals to finally obtain S (K-1)-dimensional feature vectors, and combining the S (K-1)-dimensional feature vectors to finally obtain an S*(K-1)-dimensional feature; and transmitting the S*(K-1)-dimensional feature into a plurality of ELM (Extreme Learning Machine) sorting devices so as to obtain a final sorting result by utilizing a voting sorting strategy. The invention provides a voting sorting strategy based on the ELMs; compared with a traditional multi-time ELM average accuracy scheme, the method provided by the invention has the advantages that the sorting accuracy is improved under the condition of not influencing the training sorting low consumption.
Owner:BEIJING UNIV OF TECH
Multi-sensory-mode BCI-VR control method and system and VR equipment
PendingCN113552941AHigh precisionImprove user experienceInput/output for user-computer interactionCharacter and pattern recognitionPhysical medicine and rehabilitationFeature extraction
The invention discloses a multi-sensory-mode BCI-VR control method and system and VR equipment, and the method comprises the steps: obtaining an EEG signal which is triggered by a VR scene and comprises a plurality of sensory signals corresponding to different sensory modes; performing feature extraction and classification recognition processing on the sensory signals to obtain EEG classification information corresponding to the sensory signals; and combining the EEG classification information to obtain an identification result of the EEG signal. According to the method, the traditional VR equipment is combined with the multi-sensory-mode BCI, and the SSVEP stimulation is fused to simulate the real environment, so that the adaptability and fatigue feeling of direct repeated stimulation on the brain of the user caused by the fact that the SSVEP single-visual-mode BCI depends on target object flicker stimulation are relieved, the visual fatigue is relieved, and the use experience of the user is improved.
Owner:上海厉鲨科技有限公司
Mass face library retrieval method based on man-machine collaboration
PendingCN112949369AQuick searchExact searchDigital data information retrievalCharacter and pattern recognitionCognitive capabilityImage retrieval
The invention discloses a massive face image library retrieval method based on man-machine collaboration. The low-level features extracted by computer vision are different from the high-level semantic features captured by human vision, so that the computer vision retrieval result cannot reach the expectation of people. Face images are small in difference in low-level contour features, and similar faces are difficult to distinguish only depending on computer vision. The method comprises the following steps: 1, establishing an image library; 2, training an EEG classification model by using an EEG signal generated by watching the portrait by a person; and 3, performing online iteration by using an EEG classification model, and retrieving a target image required by the testee from the image data. The face image retrieval method is different from a traditional face retrieval method, fast, accurate and robust massive face image retrieval is achieved by combining the powerful cognitive ability of the human brain with the fast computing ability and the massive storage ability of a computer, and the face image retrieval method has remarkable application value.
Owner:HANGZHOU DIANZI UNIV
SSVEP EEG classification method based on convolutional neural model augmented with EMD data
ActiveCN109299751BReduce fatigueImprove usabilityCharacter and pattern recognitionNeural architecturesEeg dataData transformation
The present invention relates to the algorithm of brain-computer interface, artificial intelligence and pattern recognition, more specifically, relates to the SSVEP EEG classification method based on the convolutional neural model enhanced by EMD data, which can be applied to the fields of medical equipment, human-computer interaction, robot control, etc. . This method first preprocesses the original EEG data. After preprocessing, the empirical mode decomposition method is used to decompose the original EEG data, and a large number of artificial EEG data that conform to the time-frequency domain characteristics of the original EEG signal are generated by mixing. The original EEG data is merged and used for parameter training of the neural network, so as to achieve the effect of effectively training network parameters with a small amount of EEG data. Finally, the complex Morlet wavelet transform is used to generate the EEG tensor, and the original time domain data is converted into a tensor dictionary integrating time domain, frequency domain and spatial information as the input of the neural network, and the convolutional neural network model is used to enhance the data. EEG training set for classification.
Owner:NANKAI UNIV
Feature weight fused mixed attention prototype network electroencephalogram classification method
The invention discloses a mixed attention prototype network electroencephalogram classification method fusing feature weights. The method comprises the steps that electroencephalogram data are collected; the method comprises the following steps: preprocessing collected EEG original data, extracting differential entropy features, and carrying out standardization and sample sampling on the extracted features to generate a support set and a query set; inputting the preprocessed features into a prototype network, extracting an embedded representation of a source domain / target domain, and calculating a feature weight and a channel-frequency band weight through a double attention layer; calculating prototype representation, and fusing the feature attention weight and the channel-frequency band attention weight to obtain a final weighted distance; a Grad-CAM visualization mechanism is introduced, and the distribution consistency of weight distribution is analyzed; and inputting the fused representation into a classifier, calculating classification loss according to a prediction score, and outputting a cross-period EEG electroencephalogram classification result, the cross-period EEG electroencephalogram classification method solves the problems of traditional black box and weak domain adaptation, and effectively improves the cross-period EEG classification precision, robustness and interpretability.
Owner:HANGZHOU DIANZI UNIV
Brain wave detection and classification method and device, program product and electronic equipment
PendingCN121891025AGuaranteed accuracyStrong complementarityBiological modelsSensorsEeg dataFeature extraction
The invention provides a brain wave detection and classification method and device, a program product and electronic equipment, and relates to the technical field of data processing. The brain wave detection and classification method comprises the following steps: acquiring to-be-detected electroencephalogram data, and determining convolution features of the to-be-detected electroencephalogram data under various downsampling rates; based on the pooling kernel corresponding to each down-sampling rate, pooling processing is carried out on the convolution features of the to-be-detected electroencephalogram data under each down-sampling rate, and pooling features of the to-be-detected electroencephalogram data under each down-sampling rate are obtained; wherein the pooling kernel corresponding to each down-sampling rate is negatively correlated with the down-sampling rate; and based on the pooling features of the to-be-detected electroencephalogram data under each down-sampling rate, brain wave waveform classification is carried out, and a classification detection result corresponding to the to-be-detected electroencephalogram data is obtained. The comprehensiveness and sufficiency of brain wave feature extraction can be improved to a certain extent, and the accuracy of brain wave classification result detection is improved.
Owner:HANGZHOU NETZHIYI INNOVATION TECH CO LTD +1
An EEG Signal Classification Method Based on Improved Deep Residual Group Convolutional Network
ActiveCN113065526BReflect EEG characteristicsImprove accuracyCharacter and pattern recognitionNeural architecturesSignal classificationTest set
The invention belongs to the field of pattern recognition and EEG signal processing, and relates to an EEG signal classification method based on an improved ResNeXt network; it includes four parts: EEG signal acquisition, preprocessing, feature extraction, and training ResNeXt classification network; training ResNeXt classification network is Refers to: dividing the training set and test set; building an improved ResNeXt EEG signal classification network; training the improved ResNeXt EEG signal classification network; building an improved ResNeXt EEG signal classification network refers to: improving on the basis of ResNeXt, and convolving the groups The middle layer of the convolutional layer of each block module increases the direct connection operation, speeds up the speed of model convergence, reduces the test error of the model, and improves the generalization ability; the invention speeds up the convergence speed of the classification model, compared with the convolutional neural network brain The electrical classification model, the improved ResNeXt classification model is easier to optimize, effectively improves the gradient explosion problem in the deep training model, and can greatly deepen the number of layers of the network while avoiding the degradation of the classification model.
Owner:JILIN UNIV
EEG classification model generation method, device and electronic equipment
Owner:SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
EEG classification transfer learning method and system based on Euclidean alignment and Procuses analysis
ActiveCN111832427AInput/output for user-computer interactionCharacter and pattern recognitionPattern recognitionCentering matrix
The invention discloses an EEG classification transfer learning method and system based on Euclidean alignment and Procuses analysis, and belongs to the field of brain-computer interfaces based on motor imagery. The method comprises the following steps: performing covariance alignment on previous user and new user feature matrixes; performing mean value alignment on the feature matrixes of the previous users and the new users; calculating the category center of the previous user according to the label of the previous user and the feature matrix of the previous user after the mean value is aligned, and calculating the category center of the new user according to a pseudo label of the new user and the feature matrix of the new user after the mean value is aligned; constructing a previous user and new user category center matrix, and calculating a rotation matrix for aligning the category center of the new user with the category center of the corresponding previous user through an orthogonal Principal analysis method; and multiplying the feature matrix and the rotation matrix of the new user after mean value alignment to obtain finally aligned new user data.
Owner:HUAZHONG UNIV OF SCI & TECH