CN113243924A - Identity recognition method based on electroencephalogram signal channel attention convolution neural network - Google Patents
Identity recognition method based on electroencephalogram signal channel attention convolution neural network Download PDFInfo
- Publication number
- CN113243924A CN113243924A CN202110544789.7A CN202110544789A CN113243924A CN 113243924 A CN113243924 A CN 113243924A CN 202110544789 A CN202110544789 A CN 202110544789A CN 113243924 A CN113243924 A CN 113243924A
- Authority
- CN
- China
- Prior art keywords
- signals
- identity recognition
- convolution
- electroencephalogram
- layer
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/117—Identification of persons
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Pathology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an identification method based on an electroencephalogram signal channel attention convolution neural network, which comprises the following steps: s1, selecting EEG signals of different channels from an emotion electroencephalogram database as original signals; s2, removing the ocular artifact signals and the power frequency interference signals in the original signals by using a band-pass filter to obtain pure emotion electroencephalogram signals; and S3, inputting the preprocessed emotion electroencephalogram signals into a deep learning identity recognition model, and performing identity recognition on the emotion electroencephalogram signals by using a deep learning algorithm. The invention selects the emotion EEG signal for identity recognition, the emotion EEG is easy to obtain, and the identity recognition method has universality and generalization. The invention shortens the number of neurons connected between the front layer and the rear layer, alleviates the problem of gradient disappearance, enhances the characteristic propagation and reduces the network parameters, and more effectively utilizes the EEG signal characteristics of different emotional states, thereby effectively carrying out the identity recognition of the emotional EEG signal.
Description
Technical Field
The invention relates to the technical field of communication electronics and biological feature recognition, in particular to an identity recognition method based on a biological feature extraction algorithm of emotion electroencephalogram signals.
Background
In recent years, the internet and smart cities offer opportunities for economic development and also bring many potential safety hazards, some of which are related to identification and verification of identity information. The traditional identity recognition modes such as face recognition, iris recognition, fingerprint recognition and the like all have the defects of being easily tampered, copied and used under duress, so that certain potential safety hazards exist. With the development of technology, the continuous intellectualization and the technological process of criminal means make the biometric identification technology face new challenges. The traditional biological feature recognition method mainly comprises face recognition, fingerprint recognition and iris recognition. For example, the identity recognition system can be broken through a simulation mask which is manufactured by printing a human face image through a 3D technology in the human face recognition; fingerprint identification by using a fake finger made of gelatin, the fingerprint identification system can be easily cheated; the false iris characteristics etched on the contact lenses are used in iris recognition, so that an iris recognition system is true and false and is difficult to distinguish. Therefore, the traditional identity identification method has certain potential safety hazard.
At present, a novel identity recognition method, namely an identity recognition research based on electroencephalogram signals, appears, because the electroencephalogram signals are signals acquired by electrical signals released by the activity of neuron cells of the brain in the cerebral cortex, namely the electroencephalogram signals. Therefore, the electroencephalogram signal has the unique advantages of invisibility, non-steatability, non-imitability, non-coercibility, necessity of living bodies and the like, and can make up the defects of the traditional identity recognition means when being applied to the identity recognition.
In recent years, algorithms associated with deep learning have been increasingly applied to the classification of EEG signals. And EEG signals are time-varying and scalp biopotential recording locations, researchers tend to use deep learning architectures to capture features in EEG signals for classification tasks. The deep learning model can extract important electroencephalogram characteristic information from input original data for classification tasks. In these studies, CNN was used directly to learn features of EEG signals from changes in amplitude of EEG time series. The CNN algorithm performs convolution operations using filters that can automatically learn key features of the EEG signal, and these filters can combine the automatically learned feature stacks of multiple CNN layers into local patterns of more complex features. The pooling layer sub-samples the output of the convolutional layer by outputting only the maximum value for each small region. Subsampling allows the convolutional layer after the pooling layer to operate at a different ratio than the layer before it. Thereby making the CNN model adaptive.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide the EEG signal-based identity recognition method based on the EEG signal channel attention convolutional neural network, which selects the emotion EEG signal for identity recognition, is easy to acquire the emotion EEG, reduces the limitation on a user when the EEG signal is acquired, and ensures that the EEG signal-based identity recognition method has universality and generalization.
The purpose of the invention is realized by the following technical scheme: an identification method based on an electroencephalogram signal channel attention convolution neural network comprises the following steps:
s1, selecting data, and selecting EEG signals of different channels from an emotion electroencephalogram database as original signals;
s2, preprocessing signals, namely removing ocular artifact signals and power frequency interference signals in the original signals by using a band-pass filter to obtain pure emotion electroencephalogram signals;
and S3, inputting the preprocessed emotion electroencephalogram signals into a deep learning identity recognition model, performing identity recognition on the emotion electroencephalogram signals by using a deep learning algorithm, and outputting identity IDs.
Further, the deep learning identity recognition model comprises a channel attention mechanism model, a dense layer 1, a transition layer 1, a dense layer 2, a transition layer 2 and a full connection layer which are connected in sequence.
Further, the step S3 specifically includes the following sub-steps:
s31, dividing the continuous emotion electroencephalogram signals into 1S electroencephalogram data segments of each segment by using a sliding window, then inputting the depth S31, dividing the continuous emotion electroencephalogram signals into 1S electroencephalogram data segments of each segment by using the sliding window, and then inputting a channel attention mechanism model of the deep learning identity recognition model;
s32, the channel attention mechanism model performs convolution operation on the input signal by using a 2D-CNN convolution network, and then inputs the signal into a dense layer;
s33, the dense layer performs convolution by using 1 2D CNN convolution layer, the size of a convolution kernel is 3, the size of the convolution kernel is 6, normalization operation is performed after convolution, an activation function LeakyRelu is input, then a signal input into the dense layer is spliced with an output signal of the dense layer to form a large signal matrix, and then the large signal matrix is input into the transition layer;
s34, the transition layer carries out normalization operation on the input signal, and then inputs the normalized input signal into a nonlinear activation function LeakyReLU, a 3 x 3 convolution network and a maximization pooling layer in sequence;
and S35, repeating the steps S34 and S35, and finally outputting through the full connection layer.
Further, the internal structure of the channel attention mechanism model is as follows: the 2D-CNN convolution operation is two-dimensional convolution operation and is suitable for two-dimensional input data, and the definition of the convolution operation is defined for an input two-dimensional signal S and an inner core W;
in this equation, the operators represent discrete convolution operations, the output of the convolution process is called a signature map, and n represents the dimension of the input signal S.
Further, a dropout layer is added in the dense layer, and the dropout is 0.5 and is used for preventing overfitting.
Further, the activation function in the fully connected layer is softmax.
Further, in the deep learning classification model, loss is calculated by using a categorical _ cross entropy function, and an optimization function of the model optimizes model parameters by using adm.
The invention has the beneficial effects that:
1) the invention selects the emotion EEG signal for identity recognition, the emotion EEG is easy to obtain, and the limitation to the user when the EEG signal is obtained is reduced, so that the identity recognition method based on the EEG signal has universality and generalization.
2) The invention designs a new algorithm with high accuracy as an identification algorithm of the electroencephalogram signal. The algorithm avoids the problem of complex process and large calculation amount of the feature extraction and screening steps, and effectively saves the operation time. The algorithm shortens the number of neurons connected between the front layer and the rear layer, alleviates the problem of gradient disappearance, enhances the characteristic propagation and reduces network parameters, and more effectively utilizes EEG signal characteristics of different emotional states, thereby effectively carrying out identity recognition of emotional electroencephalogram signals.
3) The deep learning classification model designed by the invention is added with a channel attention mechanism on the basis of CNN. The channel attention mechanism improves the classification efficiency of the model by calculating the weight of each channel EEG signal and then adding to facilitate deep learning for feature learning.
Drawings
FIG. 1 is a flow chart of an identification method based on an electroencephalogram signal channel attention convolutional neural network of the present invention;
FIG. 2 is a block diagram of a deep learning identity recognition model of the present invention;
FIG. 3 is a schematic diagram of a dense layer structure according to the present invention.
Detailed Description
The technical scheme of the invention is further explained by combining the attached drawings.
As shown in FIG. 1, the identification method based on the electroencephalogram signal channel attention convolution neural network of the invention comprises the following steps:
s1, selecting data, and selecting EEG signals of different channels from an emotion electroencephalogram database as original signals;
the database of electroencephalographic signals is from the published HeadIT emotional EEG dataset, which contains both positive and negative emotional tasks. This example had a total of 20 healthy volunteers participating in the experiment. Stimulating the subject to be tested by directed language narration to induce a realistic emotion; a total of 15 guide image narratives, each narrative describing different emotions and scenes separated by a speech-guided relaxation episode; the subject uses the image as an expression to evoke a suggested emotion. The method divides each evoked emotion EEG into 1sEEG signal segments for identity recognition. 265 affective EEG experiments were performed for each test, for a total of 8480 samples (265 subsamples x 32 tests); the ID number of each participant acts as a tag in the identification. The data and tags are described as follows:
1. data: number of participants 265 subsamples 256EEG data points (1s,256HZ adoption rate)
2. Labeling: identification number 265 subsample 1(ID)
TABLE 1, 15 kinds of emotion classification table of emotion electroencephalogram
| Positive emotion | Negative emotions |
| Love | Anger and anger |
| Joyous | Jealousy |
| Happiness | Aversion to |
| Relief device | Depression and depression |
| In the same situation | Fear of |
| Satisfy the requirement of | Sadness and sorrow |
| Excitement | Sadness and pain |
| (awe) |
The feasibility of the emotion electroencephalogram on identity recognition is explored: the identification based on the electroencephalogram signals does not depend on feature extraction and feature selection any more, but utilizes the electroencephalogram signals as the main basis of identification. The method has the advantages that the dense convolutional neural network method is utilized to carry out identity recognition on the learning characteristics of the electroencephalogram signals, a good effect can be obtained, and the visible deep learning method can automatically learn the characteristics of the electroencephalogram signals for identity recognition, so that the method can be used for finishing the identity recognition of the electroencephalogram signals.
S2, preprocessing signals, namely removing ocular artifact signals and power frequency interference signals in the original signals by using a band-pass filter to obtain pure emotion electroencephalogram signals;
and S3, inputting the preprocessed emotion electroencephalogram signals into a deep learning identity recognition model, performing identity recognition on the emotion electroencephalogram signals by using a deep learning algorithm, and outputting identity IDs.
The invention adopts a mode of combining channel attention and a dense convolutional neural network to carry out identity recognition research on the multi-channel electroencephalogram signal, thereby realizing the identity recognition method based on the electroencephalogram signal. The dense convolutional neural network is used for learning electroencephalogram characteristics and identifying identities, shortens the number of neurons connected between a front layer and a rear layer, alleviates the problem of gradient disappearance, enhances characteristic propagation, reduces network parameters, and more effectively utilizes EEG signal characteristics of different emotional states, thereby effectively identifying identities of emotional electroencephalogram signals. As shown in fig. 2, the deep learning identity recognition model includes a channel attention mechanism model, a dense layer 1, a transition layer 1, a dense layer 2, a transition layer 2, and a fully connected layer, which are connected in sequence.
The step S3 specifically includes the following sub-steps:
s31, dividing the continuous emotion electroencephalogram signals into 1S electroencephalogram data segments of each segment by using a sliding window, then inputting the depth S31, dividing the continuous emotion electroencephalogram signals into 1S electroencephalogram data segments of each segment by using the sliding window, and then inputting a channel attention mechanism model of the deep learning identity recognition model; since the sampling frequency of the signal is 256Hz, 256 data are contained in 1s, that is, each sample has 256 sample points; the channel attention mechanism model is connected and input by full-connection layers of 256 neurons; assuming that the total number of samples is N and the number of samples in the training set is N1, the format of the input data during training is N1 × 256 × 1; the input format at the time of the test was (NN1) × 256 × 1; the activation function, leak relu, was then used to calculate the weight for each channel to add to the raw EEG signal (the number of samples in the training set was 4240, so the format of the input data was 4240 × 256 × 32 × 1, and the number of samples in the test set was 1060).
S32, the channel attention mechanism model performs convolution operation on the input signal by using a 2D-CNN convolution network, and then inputs the signal into a dense layer;
s33, performing convolution on the dense layer by using 1 2D CNN convolution layer, wherein the size of a convolution kernel is 3, the size of the convolution kernel is 6, performing normalization operation after convolution and inputting an activation function LeakyRelu, splicing a signal input into the dense layer and an output signal of the dense layer to form a large signal matrix, and inputting the signal into a transition layer, as shown in FIG. 3;
s34, outputting through a dense layer, and selecting the maximum pooling by a transition layer in order to reduce the redundancy of signals; the transition layer carries out normalization operation on input signals and then inputs the normalized input signals into a nonlinear activation function LeakyReLU, a 3 x 3 convolution (Conv) network and a maximization pooling layer in sequence;
and S35, repeating the steps S34 and S35, and finally outputting the result through the full connection layer, wherein the number of the people identified by the method is 32, so that the number of the units of the full connection layer is 32.
Further, the internal structure of the channel attention mechanism model is as follows: the 2D-CNN convolution operation is two-dimensional convolution operation and is suitable for two-dimensional input data, and the definition of the convolution operation is defined for an input two-dimensional signal S and an inner core W;
in this equation, the operators represent discrete convolution operations, the output of the convolution process is called a signature map, and n represents the dimension of the input signal S.
Further, a dropout layer is added in the dense layer, and the dropout is 0.5 and is used for preventing overfitting.
Further, the activation function in the fully connected layer is softmax.
Further, in the deep learning classification model, loss is calculated by using a categorical _ cross entropy function, and an optimization function of the model optimizes model parameters by using adm.
After the electroencephalogram signals are subjected to identity recognition by using the deep learning classification model, the identity recognition result needs to be evaluated to judge the quality of the staging algorithm. A commonly used evaluation parameter is mainly the classification accuracy. This example uses the electroencephalogram signals of 20 persons for identification. The identity recognition result based on the emotion electroencephalogram signal is established and is shown in the following table 2.
TABLE 2 identification results of the respective tests tested
As can be seen from table 2 above, in the case of the deep learning identity recognition model, the highest identity recognition accuracy rate can reach 100%, and the average accuracy rate can reach 95.27% in the identity recognition of 20 users, but the accuracy rate of the deep learning identity recognition model is low because the corresponding emotion electroencephalogram is excited in the process of narratively guiding and inducing electroencephalogram, which causes the accuracy rate of the identity recognition to be low. The overall accuracy of the electroencephalogram signal identification is higher and reaches 95.27%. In addition, a kappa coefficient may be calculated from the overall accuracy, which may be used to measure how well the identity ID identified by the algorithm matches the actual identity ID.
Where ACC represents overall classification accuracy and ACC0 represents the level of opportunity. ACC (adaptive cruise control)o＝1/Ny，NyIs the number of classifications. When the kappa coefficient is 0, it indicates that the classification accuracy is an opportunity level degree, and a kappa coefficient of 1 means the best classification effect. The kappa coefficient of the method is 0.95, which shows that the staging effect has high consistency. Therefore, the method can realize high-efficiency and accurate identification based on the electroencephalogram signals, and has certain advantages compared with a general algorithm for deep learning identification.
It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (7)
1. An identification method based on an electroencephalogram signal channel attention convolution neural network is characterized by comprising the following steps:
s1, selecting data, and selecting EEG signals of different channels from an emotion electroencephalogram database as original signals;
s2, preprocessing signals, namely removing ocular artifact signals and power frequency interference signals in the original signals by using a band-pass filter to obtain pure emotion electroencephalogram signals;
and S3, inputting the preprocessed emotion electroencephalogram signals into a deep learning identity recognition model, performing identity recognition on the emotion electroencephalogram signals by using a deep learning algorithm, and outputting identity IDs.
2. The EEG channel attention convolutional neural network-based identity recognition method of claim 1, wherein the deep learning identity recognition model comprises a channel attention mechanism model, a dense layer 1, a transition layer 1, a dense layer 2, a transition layer 2 and a fully connected layer, which are connected in sequence.
3. The identification method based on the electroencephalogram signal channel attention convolution neural network as claimed in claim 2, wherein the step S3 specifically comprises the following sub-steps:
s31, dividing the continuous emotion electroencephalogram signals into 1S electroencephalogram data segments, and inputting a channel attention mechanism model of the deep learning identity recognition model;
s32, the channel attention mechanism model performs convolution operation on the input signal by using a 2D-CNN convolution network, and then inputs the signal into a dense layer;
s33, the dense layer performs convolution by using 1 2D CNN convolution layer, the size of a convolution kernel is 3, the size of the convolution kernel is 6, normalization operation is performed after convolution, an activation function LeakyRelu is input, then a signal input into the dense layer is spliced with an output signal of the dense layer to form a large signal matrix, and then the large signal matrix is input into the transition layer;
s34, the transition layer carries out normalization operation on the input signal, and then inputs the normalized input signal into a nonlinear activation function LeakyReLU, a 3 x 3 convolution network and a maximization pooling layer in sequence;
and S35, repeating the steps S34 and S35, and finally outputting through the full connection layer.
4. The EEG channel attention convolutional neural network-based identity recognition method of claim 3, wherein the internal structure of the channel attention mechanism model is as follows: the 2D-CNN convolution operation is two-dimensional convolution operation and is suitable for two-dimensional input data, and the definition of the convolution operation is defined for an input two-dimensional signal S and an inner core W;
in this equation, the operators represent discrete convolution operations, the output of the convolution process is called a signature map, and n represents the dimension of the input signal S.
5. The EEG channel attention convolution neural network-based identity recognition method of claim 3, characterized in that a dropout layer is added in the dense layer, and the dropout is 0.5 for preventing overfitting.
6. The method for identity recognition based on an electroencephalogram signal channel attention convolutional neural network of claim 3, wherein the activation function in the fully-connected layer is softmax.
7. The method for identity recognition based on the electroencephalogram signal channel attention convolution neural network, wherein in the deep learning classification model, losses are calculated by using a categorical _ cross entropy multi-classification cross entropy function, and an optimization function of the model uses adm to optimize model parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110544789.7A CN113243924A (en) | 2021-05-19 | 2021-05-19 | Identity recognition method based on electroencephalogram signal channel attention convolution neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110544789.7A CN113243924A (en) | 2021-05-19 | 2021-05-19 | Identity recognition method based on electroencephalogram signal channel attention convolution neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113243924A true CN113243924A (en) | 2021-08-13 |
Family
ID=77182742
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110544789.7A Pending CN113243924A (en) | 2021-05-19 | 2021-05-19 | Identity recognition method based on electroencephalogram signal channel attention convolution neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113243924A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113723247A (en) * | 2021-08-20 | 2021-11-30 | 西安交通大学 | Electroencephalogram identity recognition method and system |
| CN114081505A (en) * | 2021-12-23 | 2022-02-25 | 成都信息工程大学 | Electroencephalogram signal identification method based on Pearson correlation coefficient and convolutional neural network |
| CN116070141A (en) * | 2023-04-06 | 2023-05-05 | 博睿康科技（常州）股份有限公司 | Signal detection method, detection model, detection equipment and application |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886792A (en) * | 2017-01-22 | 2017-06-23 | 北京工业大学 | A kind of brain electricity emotion identification method that Multiple Classifiers Combination Model Based is built based on layering |
| CN108776788A (en) * | 2018-06-05 | 2018-11-09 | 电子科技大学 | A kind of recognition methods based on brain wave |
| CN111329474A (en) * | 2020-03-04 | 2020-06-26 | 西安电子科技大学 | Electroencephalogram identity recognition method and system based on deep learning and information updating method |
-
2021
- 2021-05-19 CN CN202110544789.7A patent/CN113243924A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886792A (en) * | 2017-01-22 | 2017-06-23 | 北京工业大学 | A kind of brain electricity emotion identification method that Multiple Classifiers Combination Model Based is built based on layering |
| CN108776788A (en) * | 2018-06-05 | 2018-11-09 | 电子科技大学 | A kind of recognition methods based on brain wave |
| CN111329474A (en) * | 2020-03-04 | 2020-06-26 | 西安电子科技大学 | Electroencephalogram identity recognition method and system based on deep learning and information updating method |
Non-Patent Citations (2)
| Title |
|---|
| 周婧: "基于循环神经网络方法的脑电信号身份识别", 《中国优秀博硕士学位论文全文数据库(硕士)医药卫生科技辑 》 * |
| 陶威: "基于注意力机制的脑电情绪识别方法研究", 《中国优秀博硕士学位论文全文数据库(硕士) 医药卫生科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113723247A (en) * | 2021-08-20 | 2021-11-30 | 西安交通大学 | Electroencephalogram identity recognition method and system |
| CN113723247B (en) * | 2021-08-20 | 2024-04-02 | 西安交通大学 | Electroencephalogram identity recognition method and system |
| CN114081505A (en) * | 2021-12-23 | 2022-02-25 | 成都信息工程大学 | Electroencephalogram signal identification method based on Pearson correlation coefficient and convolutional neural network |
| CN116070141A (en) * | 2023-04-06 | 2023-05-05 | 博睿康科技（常州）股份有限公司 | Signal detection method, detection model, detection equipment and application |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108776788B (en) | Brain wave-based identification method | |
| CN112784798B (en) | Multi-modal emotion recognition method based on feature-time attention mechanism | |
| CN108596039B (en) | Bimodal emotion recognition method and system based on 3D convolutional neural network | |
| CN110069958B (en) | Electroencephalogram signal rapid identification method of dense deep convolutional neural network | |
| CN111461176B (en) | Multi-mode fusion method, device, medium and equipment based on normalized mutual information | |
| CN113243924A (en) | Identity recognition method based on electroencephalogram signal channel attention convolution neural network | |
| CN114052735B (en) | Deep field self-adaption-based electroencephalogram emotion recognition method and system | |
| CN111134666A (en) | Emotion recognition method of multi-channel electroencephalogram data and electronic device | |
| CN110135244B (en) | Expression recognition method based on brain-computer collaborative intelligence | |
| CN112800998A (en) | Multi-mode emotion recognition method and system integrating attention mechanism and DMCCA | |
| CN114176607B (en) | Electroencephalogram signal classification method based on vision transducer | |
| An et al. | Electroencephalogram emotion recognition based on 3D feature fusion and convolutional autoencoder | |
| CN115238731A (en) | Emotion identification method based on convolution recurrent neural network and multi-head self-attention | |
| CN116230234A (en) | Multi-mode feature consistency psychological health abnormality identification method and system | |
| CN112465069A (en) | Electroencephalogram emotion classification method based on multi-scale convolution kernel CNN | |
| Rayatdoost et al. | Subject-invariant EEG representation learning for emotion recognition | |
| CN113180659A (en) | Electroencephalogram emotion recognition system based on three-dimensional features and cavity full convolution network | |
| CN117198468A (en) | Intervention scheme intelligent management system based on behavior recognition and data analysis | |
| CN114081505A (en) | Electroencephalogram signal identification method based on Pearson correlation coefficient and convolutional neural network | |
| CN113974627A (en) | Emotion recognition method based on brain-computer generated confrontation | |
| CN116763324A (en) | Single-channel electroencephalogram signal sleep stage method based on multiple scales and multiple attentions | |
| Dia et al. | A novel stochastic transformer-based approach for post-traumatic stress disorder detection using audio recording of clinical interviews | |
| CN115969392A (en) | Cross-period brainprint recognition method based on tensor frequency space attention domain adaptive network | |
| Castro et al. | Development of a deep learning-based brain-computer interface for visual imagery recognition | |
| CN114081492A (en) | Electroencephalogram emotion recognition system based on learnable adjacency matrix |
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 |