116 results about "EEG feature" patented technology

A system for calibrating and/or verifying system performance of a remote portable EEG system having at least one EEG sensor. Embodiments of the invention can provide various reference signals to calibrate and quality control the remote performance of the data acquisition EEG system. In addition a calibration cable connects a reference signal source to the EEG sensors to enable remote calibration and quality control assessment. Further, a diagnostic biomarker is included to assess the state or function of a subject's brain and enables the classification, prognosis, diagnosis, monitoring of treatment, or response to therapy applied to the brain by measuring any one of a list of candidate features extracted from a given cognitive or sensory task, and measuring changes in the EEG feature and task combination over time, among multiple states, or compared to a normative database.

The invention discloses an extraction method for brain-computer interface system imagination action EEG signal features, in particular to an EEG feature extraction method based on a wavelet transform and a BP neural network. The invention takes the energy change caused by imagination action thinking to be a feature distinguishing the imagination movements of a left hand and a right hand, respectively calculates the point-to-point average power of the entire samplings of the EEG signal obtained from C3 and C4 channels by the left hand and the right hand through the imagination (thereinafter called as C3 and C4 of the left hand and the right hand) within 0 to 9s according to the average power formula. A time window is arranged, a discrete dyadic wavelet transform is made to the data of a section provided with the window, an approximation signal a6 on a sixth size is selected to be taken as a signal feature; a BP neural network is used as a classifier to classify. The method of the invention adopting the wavelet transform and the BP neural network to extract the potential of the imagination movement helps to improve the signal/noise ratio and the identification correction rate of the potential of the imagination action; in addition, the wavelet transform is a linear transform, has a quick calculation speed, and is suitable for on-line analysis.

The invention discloses a teaching method for improving attention. The teaching method includes the steps of: acquiring brain wave data of a student; acquiring brain wave data of a user acquired by anEEG acquisition device, and calculating an attention value on the basis of the brain wave data; switching to a training mode if the attention value is less than a first preset threshold; and acquiring an EEG feature during the user training, and outputting and displaying a corresponding animation effect on the basis of the EEG feature so as to adjust the user's attention. The invention also discloses a teaching device for improving attention and a computer readable storage medium. The invention can train the user's attention while monitoring the user's attention, thereby improving the user'sattention.

The invention discloses an emotion recognition method of multi-channel electroencephalogram (EEG) data and an electronic device. The method includes the steps: sending frequency domain features of channels of extracted EEG data into a sequence feature extractor so as to obtain hidden-layer feature vectors of the channels of the EEG data, capturing association between the EEG channels and emotion categories, and sending the association and the hidden-layer feature vectors to an automatic encoder so as to obtain probability value vectors of the emotion category corresponding to the EEG data. Long-term and short-term memory networks are utilized to learn the non-linear relationship between the EEG channels on the EEG channel sequence, the high-dimensional EEG features related to the emotion categories are extracted, the key information in the multi-channel EEG signals are selected automatically, and EEG feature field differences which are caused by instability of the EEG data and environmental change are reduced.

The invention provides a visual-auditory evoked emotion recognition method and a system based on an EEG signal. The method comprises the following steps: firstly, the EEG signal generated based on thevisual-auditory evoked emotion is collected and preprocessed through a band-pass filter. Then the multichannel EEG signals are decomposed by multivariate empirical mode decomposition, and the effective intrinsic mode functions are selected to extract the features of emotional EEG signals. Then, the sequence floating forward selection method is used as the search strategy to select and delete feature sets, and the filter and encapsulator are used as the evaluation criteria of the optimal feature subsets to filter the extracted emotional EEG features. Then the selected feature subset is input to support vector machine for classification, and the classification results are obtained. Finally, according to the classification results, the emotion recognition results are obtained, and the emotion recognition is realized. On the basis of exploring the law of emotional electroencephalogram, the invention carries out the research on the emotion recognition method of a plurality of induced modes, and effectively improves the recognition accuracy.

The invention discloses an attention evaluation method and an attention evaluation system. The attention evaluation method comprises the following steps: acquiring electroencephalogram (EEG), an electrocardiosignal and behavioral data; converting the EEG from a time domain to a frequency domain, and extracting an EEG feature vector; performing cepstrum processing on the electrocardiosignal, and extracting an electrocardiosignal feature vector; processing the behavior data, and extracting a behavior feature vector; and processing the EEG feature vector, electrocardiosignal feature vector and the behavior feature vector by adopting a PCVM (Production Centered Virtual Manufacturing) probability classification algorithm, thereby obtaining the evaluation result. With the adoption of the technical scheme, the EEG, electrocardiosignal and behavioral data of a testee can be comprehensively processed and analyzed, the comprehensive feature vectors are subjected to classified counting, the problem that the attention evaluation accuracy cannot be improved is solved, and accurate and stable attention level prediction results of the testee can be obtained.

The invention discloses a fine-grained visualization system and method for emotional electroencephalography (EEG), and solves the technical problem of how to display fine-grained information in the emotional EEG. The system is connected with a data acquisition module, a data preprocessing module, a feature extraction module and a network training control module in sequence; an expression atlas provides a target image; the network training control module and an affective computing generative adversarial network (AC-GAN) module complete the training of an AC-GAN; a network forward execution module controls to complete the generation of fine-grained expressions. The method comprises the steps of collecting emotional EEG data, preprocessing the EEG data, extracting EEG features, constructing the AC-GAN, preparing the expression atlas, training the AC-GAN, and obtaining a fine-grained facial expression generation result. The emotional EEG is directly visualized into facial expressions withthe fine-grained information which can be directly recognized, and the visualization system is used for interactive enhancement and experience optimization of rehabilitation equipment, emotional robots, VR devices and the like with a brain-computer interface.

An electronic apparatus for establishing prediction model based on electroencephalogram (EEG). The electronic apparatus is configured for: acquiring an EEG signal segment related to an epilepsy patient; dividing each EEG signal into EEG components according to a predetermined window size; retrieving datasets corresponding to EEG features from the EEG components of each EEG signal segment; acquiring statistical feature values of each dataset of each EEG signal segment; determining a gain ratio of each of the statistical feature values of each EEG signal segment based on the statistical feature values corresponding to each of the EEG features; selecting specific statistical feature values from the statistical feature values according to the gain ratio of each of the statistical feature values of each EEG signal segment; establishing a prediction model based on the specific statistical feature values of the epilepsy patient.

The invention discloses a multi-modal brain-computer interface (BCI) method and system based on synchronic compound limb imaginary movements. The method comprises the following steps: designing seven types of multi-limb participated synchronic compound limb imaginary movements related to hands, feet and a body; irritating a subject by adopting task models of four stages, collecting electroencephalogram data and carrying out preprocessing; carrying out feature extraction and mode recognition on the preprocessed electroencephalogram data by three different cospace mode algorithms, so as to acquire an electroencephalogram feature vector of a single task; inputting the electroencephalogram feature vector of the single task into a training classifier of a support vector machine, and then predicting a spatial feature from a test set. The invention establishes a novel multi-modal BIC pattern based on the synchronic compound limb imaginary movements, the dilemmas that the species of the existing imaginary movement patterns are limited and disaccord with actual movements are got rid of, big instruction set output of information controlled by an MI (Motor imagery)-BCI system is realized, furthermore, a novel pathway is explored and a novel method is provided for promoting the practical application of a brain-computer interface in rehabilitation engineering.

The invention relates to a multi-lead correlation analysis electroencephalo-graph (EEG) feature extraction method. In multi-class motor imagery task recognition, EEG signal features of brain areas activated by a specific motor imagery task are effectively extracted, and effectively extracting the EEG signal features of the brain areas activated by the specific motor imagery task is a key problem to improve a recognition rate. With the multi-lead correlation analysis EEG feature extraction method, firstly multi-lead motor imagery EEG signals are extracted, then a correlation coefficient between every two lead EEG signals is analyzed to obtain a correlation parameter matrix, next a row variance of each correlation parameter matrix, the ratio values of the sum of all the row variances, and natural logarithms of all the row variances are calculated, obtained results are used as characteristic vectors of the EEG signals, and finally the characteristic vectors are input into a classifier to complete classifying recognition of multi-class motor imagery tasks. With the multi-lead correlation analysis EEG feature extraction method, not only can the EEG signal features of the brain areas activated by the specific motor imagery task at the same time can be fully extracted, influences on characteristic parameters can be reduced to a large extent, wherein the influences are caused by EEG signal individual differences, and further the problem that insufficience problem of electrode choosing can be solved.

The invention discloses an electroencephalogram signal emotion recognition method based on VMD and WPD, and the method comprises the steps: carrying out the variational mode decomposition of obtainedelectroencephalogram signal data, and obtaining a variational mode component; wavelet packet decomposition and reconstruction are performed on the variational mode components, so that reconstructed signals under beta and gamma frequency bands in the electroencephalogram signal frequency band can be obtained; wavelet packet entropies, improved multi-scale sample entropies, fractal dimensions and first-order difference values of the reconstructed signals are calculated respectively, and feature vectors for electroencephalogram signal emotion recognition are formed; and sending the feature vectorinto a classifier to carry out classification and identification of the emotion state. Compared with the prior art, the method has the advantages that better time-frequency resolution is obtained, nonlinear features of EEG signals can be better captured, EEG frequency bands more related to emotion can be obtained, a good EEG feature basis is provided for emotion recognition, and then a better emotion recognition effect is obtained.

Based on combination between wavelet transform and a Fisher criterion, the invention provides an EEG feature extraction method specific to a P300 Speller Brain-machine interface, mainly comprising the steps of: configuring a wavelet transform matrix for an EEG data segment of a specific length according to a wavelet function and a wavelet decomposition level designated by a user; utilizing a wavelet transform matrix to map the EEG data segment to a wavelet domain, and utilizing the Fisher criterion to search for a projection axis which can distinguish different types of data to a maximum degree in the wavelet domain; utilizing the projection axis to extract a plurality of rows from the wavelet transform matrix to form a feature extraction matrix; mapping a corresponding EEG data segment to a feature vector by a feature extraction matrix of each channel; and splicing feature vectors on all channels to form a feature vector for each time of stimulation. The method stimulates a corresponding EEG data segment each time to calculate a feature vector, and can reduce times of stimulation repetition under the condition of meeting accuracy rate requirements so as to increases a character transfer rate.

The invention relates to a multi-task motor imagination EEG feature extraction and pattern recognition method for vehicle control. The multi-task imagination EEG signals were collected by EEG amplifier and transmitted to PC. The frequency domain and spatial feature combinations of imaginary EEG were extracted by Welch power spectrum and one-to-one common spatial pattern algorithm. A plurality of GMM classifiers are constructed according to the categories to which the training set data belongs. The original EEG signal is passed through the GMM classifier, and the probability density is comparedwith the set credible threshold. The samples below the credible threshold are classified twice by artificial neural network, and the final classification result is transmitted to the vehicle throughwireless serial port to realize the real-time movement of the vehicle. The invention utilizes welch power spectrum and CSP to extract frequency domain and spatial domain features related to motion imagination, utilizes GMM and artificial neural network two-level classifier, effectively improves real-time performance of vehicle control and safety of vehicle driving, and lays a foundation for practical application of brain-controlled vehicle.

The invention discloses an electroencephalogram-electromyographic signal fusion recognition method based on deep learning, and the method comprises the steps: carrying out the high-dimensional coding and sequence decoding of a preprocessed electroencephalogram signal through a recurrent neural network based on a deep self-attention mechanism, and obtaining an electroencephalogram classification feature matrix; performing intensity feature extraction on the preprocessed electromyographic signals to obtain electromyographic feature vectors; and performing feature fusion on the electroencephalogram classification feature matrix and the myoelectricity feature vector through an unsupervised sparse auto-encoder to generate a final instruction. According to the method, a multi-head self-attention mechanism and a coding-decoding model are combined and applied to the field of electroencephalogram feature extraction, so that the electroencephalogram classification and recognition precision is improved; the problem that the electromyographic signals have individual differences and position differences is solved by utilizing a self-adaptive method, and the action intensity is estimated by extracting the intensity characteristics of the electromyographic signals; and an unsupervised sparse auto-encoder is utilized to encode and decode different bioelectricity characteristics, and the fusion of electroencephalogram and myoelectricity characteristics is realized by an efficient and robust method.

The invention discloses a brain electrical signal EEG feature extraction method based on a convolution neural network, comprising the following steps of: S1, designing a brain electrical data collection experiment scheme of color recognition; and setting a cycle with three test pictures and three transitional pictures in full black, wherein the time of the test pictures is t1, the time of the transition pictures is t2, the test pictures in each cycle are separately red, green, blue primary colors, and the red, green, blue primary colors appear in random order, so that the time of one cycle is3t1 + 3t2, each subject is tested N cycles, and the total time is N*(3t1 + 3t2). The purpose of setting the transitional pictures is to eliminate the visual residue when switching the test pictures. The invention avoids the problem of extracting the orders in the brain electrical signal features by using an AR model, and the features extracted by the convolution neural network obtain the satisfactory results in color recognition.

The invention discloses a cross-subject EEG fatigue state classification method based on generative adversarial domain self-adaptation. The method comprises the following steps: firstly, acquiring andpre-treating data, and removing artifacts; secondly, performing EEG feature extraction through PSD, and acquiring a two-dimensional sample matrix from a three-dimensional EEG time sequence; distinguishing a source domain data set and a target domain data set to obtain a training set and a test set which are not overlapped; training a classification model GDANN by using part of label-free target domain data and random data conforming to Gaussian distribution; and finally, evaluating the accuracy of the classification result by using a confusion matrix. The generative adversarial network and the domain invariant thought are further combined, the problem that EEG signal data sets are rare and difficult to obtain is solved, the problem that source domain data and target domain data are not matched is balanced, negative migration is avoided to a certain extent, a high-precision cross-subject fatigue detection classifier is trained, and the method is expected to have a wide application prospect in actual brain-computer interaction.

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.

The invention discloses a motor imagery electroencephalogram feature extraction method based on a matrix variable Gaussian model. In the prior art, a motor imagery electroencephalogram feature extraction method is insufficient in precision. The method comprises the following steps: 1, performing electroencephalogram test to establish a sample set; 2, performing filter bank common space modal operation on the training sample set x; 3, constructing an inter-class weight matrix and an intra-class weight matrix; 4, calculating an intra-class space covariance matrix and an intra-class frequency covariance matrix; and 5, splitting the inter-class scatter matrix. 6, establishing a projection matrix. 7, calculating a characteristic number pair; and 8, obtaining d-dimensional features for training.9, training an SVM model. And 10, detecting and identifying the motor imagery of the detected person. The conventional processing method ignores the spatial information in the electroencephalogram signal. Matrix dimension reduction processing is used, the thought of a matrix variable Gaussian model is introduced, and the utilization rate of space information is further increased.

Embodiments of the present invention provide an EEG identity authority control system and method. The system consists of an interconnected brain-computer Interface Device, an Authority Processor and aPeripheral Device, wherein, the Brain-computer interface device is used for acquiring the EEG signal in the process of the user manipulating the external equipment, and converting the EEG signal intothe manipulation instruction which can be recognized by the external equipment; the computer interface device is used for acquiring the EEG signal in the process of the user manipulating the externalequipment, the Authority processor is used for extracting EEG features from EEG signals and determining user manipulation privileges based on the extracted EEG features, and the Peripheral Device isused for determining whether to execute a manipulation behavior with a permission request indicated by the manipulation instruction according to the manipulation authority. Thus, the EEG features areextracted from the EEG signals obtained in the process of the user manipulating the external device to perform the identification right confirmation, thereby realizing the corresponding identificationright control for the manipulation of the external device with the permission requirement, without the need of the additional identification right confirmation time, and thereby improving the efficiency of the identification right confirmation.

The invention discloses a feature extraction method of motion imaginary EEG signal based on weighted compound multi-scale fuzzy entropy. At first, the composite multi-scale fuzzy entropy time sequenceof MI-EEG is calculated; According to the difference and variation of CMFE entropy in different leads of each motion imagination task T, the optimal time period is determined, and the MI-EEG signal is further used for feature extraction; Then, the weighting factors are introduced to obtain tao weighted coarse-grained sequences, and the fuzzy entropy of each coarse-grained sequence is calculated,and the average value of each coarse-grained sequence is defined as WCMFE; WCMFE at a single scale tao is computed for any lead Ci of various motion imagination tasks T; The variation range of scale factor tao is determined, WCMFE at multiple scales is computed, The eigenvectors (shown in the description) and the eigenvectors (FT) of all kinds of motion imagination tasks T are constructed and fused into MI-EEG feature vector F, further improving the classification accuracy.

The invention provides a motor imagery electroencephalogram signal classification method based on a hybrid model. The motor imagery electroencephalogram signal classification method comprises the following steps: S1, data extraction: collecting two groups of data from a database as an electroencephalogram database; s2, data reconstruction: normalizing electroencephalogram data in an electroencephalogram database, and storing the normalized electroencephalogram data in a four-dimensional tensor structure; s3, feature optimization: optimizing the electroencephalogram data by using an ODV-CSSD algorithm to obtain an optimal electroencephalogram feature; and S4, forming an interpretable cluster by using the prior information, establishing a decision boundary, and obtaining a final category attribution result. The invention provides a novel motor imagery electroencephalogram signal classification method aiming at the problem that a common space mode is too sensitive to abnormal data so that overfitting is prone to occurring, and a novel feature optimization algorithm ODV-CSSD is provided and researched. According to the method, the optimal electroencephalogram feature can be obtained, and meanwhile, the dimension of the optimal electroencephalogram feature can be achieved, so that extraction and classification of the left-hand and right-hand motor imagery electroencephalogram signals are realized, and the recognition accuracy is improved.

The invention discloses an EEG real-time interaction analysis system and real-time interaction analysis method based on a web visualization platform, and the analysis method comprises the following steps: 1, receiving an EEG signal transmitted by an EEG signal collection device through a cloud server; 2, enabling a data preprocessing module to filter abnormal values in the electroencephalogram signal data, then carrying out low-pass filtering processing, and storing the processed data into a database; step 3, performing basic statistical feature extraction and user-defined advanced feature extraction on the EEG time sequence in the database by the server side module; and step 4, enabling a web visualization analysis module to monitor and visualize the data acquisition condition and the data feature condition. Research or experimenters can analyze data conditions and EEG feature conditions in real time, so that doctors or experimenters can know information acquisition conditions and adjust acquisition schemes in real time through low-cost equipment, and meanwhile, effective cloud management is performed on EEG information.

The invention discloses an upper limb movement direction neural decoding method and device based on Riemannian metric. The method comprises the following steps: extracting electroencephalogram signal feature information from a neural signal in an upper limb movement direction in a movement task process through Riemannian metric; and recognizing the movement direction of the upper limbs by using the electroencephalogram features based on Riemannian metrics. According to the method, the upper limb motion intention analysis model is established through the electroencephalogram signal characteristic information based on Riemannian metrics, so that the analysis model can obtain stable motion intention analysis performance in cognitive concentration and cognitive distraction states, which is very important for rehabilitation and assistance. Meanwhile, a new thought is provided for research on upper limb motion decoding, and a basis is provided for further research on man-machine collaborative interaction.