CN114027855B - Electroencephalogram signal decoding method and system for recognizing head movement intention - Google Patents
Electroencephalogram signal decoding method and system for recognizing head movement intention Download PDFInfo
- Publication number
- CN114027855B CN114027855B CN202111521319.5A CN202111521319A CN114027855B CN 114027855 B CN114027855 B CN 114027855B CN 202111521319 A CN202111521319 A CN 202111521319A CN 114027855 B CN114027855 B CN 114027855B
- Authority
- CN
- China
- Prior art keywords
- electroencephalogram signal
- signal data
- intention
- head
- map
- 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.)
- Active
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
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
-
- 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
- A61B5/7253—Details of waveform analysis characterised by using transforms
-
- 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
- A61B5/7253—Details of waveform analysis characterised by using transforms
- A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
-
- 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)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Surgery (AREA)
- Veterinary Medicine (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Psychiatry (AREA)
- Animal Behavior & Ethology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Signal Processing (AREA)
- Psychology (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses an electroencephalogram signal decoding method and system for recognizing head movement intention, which relate to the technical field of electroencephalogram signal decoding and comprise the following steps: acquiring an electroencephalogram signal data segment of a target user in a head movement process; processing the electroencephalogram signal data segment to obtain a transformation map corresponding to the target user; the transformation map is a map with time-frequency information; and processing the transformation map by adopting a machine learning algorithm to identify the head movement intention of the target user. The method can achieve the purpose of improving the classification accuracy of the head movement intention.
Description
Technical Field
The invention relates to the technical field of electroencephalogram signal decoding, in particular to an electroencephalogram signal decoding method and system for recognizing head movement intention.
Background
The existing electroencephalogram signal decoding technology solves the problem of limb motor imagery or motor execution. As the head moves close to the electroencephalogram cap, the electroencephalogram signal acquisition is greatly interfered, so that the signal to noise ratio is low, and the classification accuracy is low.
In addition, the head motion paradigm has few applications, so the study of the electroencephalogram decoding technology is blank.
Disclosure of Invention
The invention aims to provide an electroencephalogram signal decoding method and system for identifying head movement intentions so as to achieve the purpose of improving the classification accuracy of the head movement intentions.
In order to achieve the purpose, the invention provides the following scheme:
a brain electrical signal decoding method for recognizing head movement intention comprises the following steps:
acquiring an electroencephalogram data segment of a target user in the head movement process;
processing the electroencephalogram signal data segment to obtain a transformation map corresponding to the target user; the transformation map is a map with time-frequency information;
processing the transformation map by adopting a machine learning algorithm, and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
Optionally, the processing the electroencephalogram signal data segment to obtain a transform map corresponding to the target user specifically includes:
preprocessing the electroencephalogram signal data segment; the preprocessing comprises electrode positioning, useless electrode elimination, re-reference, filtering, segmentation and baseline correction;
processing the preprocessed electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user;
the continuous transform map is a continuous wavelet transform map or a short-time Fourier transform map.
Optionally, the processing the preprocessed electroencephalogram signal data segment by using a short-time fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user specifically includes:
screening the preprocessed electroencephalogram signal data segment to determine electroencephalogram signal data near a motor cortex;
processing the marked electroencephalogram signal data by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user;
the marked electroencephalogram signal data are determined electroencephalogram signal data near the motor cortex.
Optionally, processing the transform map by using a machine learning algorithm, and identifying the head movement intention of the target user specifically includes:
inputting the transform map into a head movement intention decoding model to identify a head movement intention of the target user;
the head movement intent decoding model is determined from training data and a convolutional neural network; the training data includes a transform map having an idle state, a transform map having an intention of head left turning, and a transform map having an intention of head right turning.
Optionally, the process of determining the training data is as follows:
acquiring a first electroencephalogram data segment and a second electroencephalogram data segment of a user through electroencephalogram cap equipment; the first electroencephalogram signal data segment is electroencephalogram signal data before head turning movement; the second electroencephalogram signal data section is electroencephalogram signal data after left-right turning movement; the second electroencephalogram signal data section comprises an electroencephalogram signal data section after the head turns left and an electroencephalogram signal data section during the head turns right;
preprocessing the first electroencephalogram signal data segment and the second electroencephalogram signal data segment;
screening the preprocessed first electroencephalogram signal data segment and the preprocessed second electroencephalogram signal data segment, and determining a first marked electroencephalogram signal data segment and a second marked electroencephalogram signal data segment; the first marked electroencephalogram signal data segment is electroencephalogram signal data of the preprocessed first electroencephalogram signal data segment near a motor cortex; the second marked electroencephalogram signal data segment is the electroencephalogram signal data of the preprocessed second electroencephalogram signal data segment near the motor cortex;
respectively processing the first marked electroencephalogram signal data segment and the second marked electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm to obtain a transform map with an idle state, a transform map with a left head turning intention and a transform map with a right head turning intention; or respectively processing the first marked electroencephalogram signal data segment and the second marked electroencephalogram signal data segment by adopting a continuous wavelet transform algorithm to obtain a transform map with an idle state, a transform map with a head left-turn intention and a transform map with a head right-turn intention.
Optionally, the inputting the transform map into a head movement intention decoding model to identify the head movement intention of the target user specifically includes:
inputting the transformation map into a head movement intention decoding model to obtain a first classification result and a second classification result;
the first classification result is in an idle state or a head movement intention state, the second classification result represents a head left-turning intention or a head right-turning intention, and when the first classification result is in the head movement intention state, the second classification result is the head left-turning intention or the head right-turning intention.
A brain electrical signal decoding system that recognizes head movement intent, comprising:
the data acquisition module is used for acquiring an electroencephalogram signal data segment of a target user in the head movement process;
the data processing module is used for processing the electroencephalogram signal data segment to obtain a transformation map corresponding to the target user; the transformation map is a map with time-frequency information;
the first head movement intention recognition module is used for processing the transformation map by adopting a machine learning algorithm and recognizing the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides an electroencephalogram signal decoding method and system for recognizing head movement intention. According to the method, the electroencephalogram signal data section is processed to obtain the map with time-frequency information, then the head movement intention of the target user is identified by utilizing the effectiveness of map extraction information and the characteristics of strong learning capability and fitting capability of a machine learning algorithm, and the classification result of classification accuracy is obtained.
A brain electrical signal decoding method for recognizing head movement intention comprises the following steps:
acquiring and preprocessing an electroencephalogram signal data segment of a target user in the head movement process;
processing the preprocessed electroencephalogram signal data segment by adopting a time-sequence neural network, and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
Optionally, the time-sequence neural network is a recurrent neural network RNN or a long-short term memory network LSTM.
An electroencephalogram signal decoding system for recognizing an intention of head movement, comprising:
the data acquisition processing module is used for acquiring and preprocessing an electroencephalogram signal data section of a target user in the head movement process;
the second head movement intention identification module is used for processing the preprocessed electroencephalogram signal data section by adopting a time-sequence neural network and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides an electroencephalogram signal decoding method and system for recognizing head movement intention. According to the method, the preprocessed electroencephalogram signal data segments are processed by adopting a time-sequence neural network, the head movement intention of the target user is identified, and a classification result of classification accuracy is obtained.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic flow chart of an electroencephalogram signal decoding method for recognizing head movement intention according to a first embodiment of the present invention;
FIG. 2 is a schematic diagram of the convolutional neural network of the present invention;
FIG. 3 is a schematic structural diagram of an electroencephalogram signal decoding system for recognizing an intention of head movement according to a second embodiment of the present invention;
FIG. 4 is a schematic flowchart of a method for decoding an electroencephalogram signal for identifying an intention of head movement according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electroencephalogram signal decoding system for recognizing an intention of head movement according to a fourth embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
At present, no special electroencephalogram decoding technology is used for solving the problem of head movement, and meanwhile, the decoding technology related to movement intention identification is relatively few. In view of the problem of low classification accuracy in decoding the head movement intention by using the means in the prior art, the invention provides an electroencephalogram signal decoding method and system for identifying the head movement intention.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Example one
The electroencephalogram signal decoding method for recognizing the head movement intention provided by the embodiment includes the steps of firstly sequentially preprocessing an electroencephalogram signal (EEG) in the head movement process, dividing a data segment and carrying out frequency domain transformation, and then carrying out secondary processing on the processed electroencephalogram signal by means of a Convolutional Neural Network (CNN), so that classification of an idle state and a head movement intention state, a head left-turn intention state and a head right-turn intention state is achieved, and recognition of the head movement intention with high accuracy and low algorithm time consumption is achieved.
Referring to fig. 1, the electroencephalogram signal decoding method for recognizing the intention of head movement provided by the embodiment includes the following steps.
Step 101: acquiring an electroencephalogram signal data segment of a target user in the head movement process.
Step 102: processing the electroencephalogram signal data segment to obtain a transformation map corresponding to the target user; the transformation map is a map with time-frequency information.
Step 103: processing the transformation map by adopting a machine learning algorithm, and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
Preferably, the step 102 described in this embodiment specifically includes:
preprocessing the electroencephalogram signal data segment; the preprocessing includes electrode positioning, rejection of unused electrodes, re-referencing, filtering, segmentation, and baseline correction.
And processing the preprocessed electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user.
The continuous transform map is a continuous wavelet transform map or a short-time Fourier transform map.
The method comprises the following steps of processing a preprocessed electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to a target user, and specifically comprises the following steps:
screening the preprocessed electroencephalogram signal data segment, and determining electroencephalogram signal data near a motor cortex; processing the marked electroencephalogram signal data by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user; the marked electroencephalogram signal data are determined electroencephalogram signal data near the motor cortex.
Preferably, step 103 described in this embodiment specifically includes:
inputting the transform map into a head movement intent decoding model to identify a head movement intent of the target user; the head movement intent decoding model is determined from training data and a convolutional neural network; the training data includes a transform map having an idle state, a transform map having a head left turn intention, and a transform map having a head right turn intention.
The determination process of the training data comprises the following steps:
step A: acquiring a first electroencephalogram signal data segment and a second electroencephalogram signal data segment of a user through electroencephalogram cap equipment; the first electroencephalogram signal data segment is electroencephalogram signal data before head turning movement; the second electroencephalogram signal data section is electroencephalogram signal data after left-right rotary head movement; the second electroencephalogram signal data section comprises an electroencephalogram signal data section after the head turns left and an electroencephalogram signal data section during the head turns right; wherein, the number of channels of the electroencephalogram signals is 21, and the number of channels near the motor cortex is 19.
The collection experiment is as follows: a left-right turning direction prompt appears on a computer screen randomly, a user performs left-right turning movement according to the prompt, and then corresponding electroencephalogram signals are collected through electroencephalogram cap equipment.
And B: and preprocessing the first electroencephalogram signal data segment and the second electroencephalogram signal data segment. The preprocessing specifically includes electrode positioning, rejection of unused electrodes, re-referencing, filtering, segmentation, baseline correction, and the like.
The electrode positioning is to correspond the collected channel signals (namely the collected electroencephalogram signals) to the actual three-dimensional head electrode position to obtain accurate positioning information.
And the step of eliminating useless electrodes is to delete the unselected electrode channels according to the positioning information to obtain the required channel signals.
Re-referencing is to compare the required channel signal with the reference electrode signal to obtain the corrected electroencephalogram signal.
The filtering is to perform band-pass filtering processing on the corrected electroencephalogram signal in a set frequency band (1-50 Hz), and the set frequency band comprises main information of neural activity.
The segmentation is to segment the electroencephalogram signal of each turn (the electroencephalogram signal is the electroencephalogram signal after the band-pass filtering processing), so as to obtain an electroencephalogram sample before each turn and an electroencephalogram sample after each turn.
The baseline correction method is to subtract a baseline value from each value in the electroencephalogram sample to obtain an electroencephalogram signal after secondary correction, namely a final preprocessed electroencephalogram signal. Wherein, the data segment of idle state is selected from 300ms to 0ms before the left-right turn-around prompt, and the data segment of head movement intention state is selected from 50ms to 350ms after the left-right turn-around prompt appears. The idle state and head movement intention state data segments obtained in this way are closer to the actual state time segments.
And C: screening the preprocessed first electroencephalogram signal data segment and the preprocessed second electroencephalogram signal data segment, and determining a first marked electroencephalogram signal data segment and a second marked electroencephalogram signal data segment; the first marked electroencephalogram signal data segment is electroencephalogram signal data of the preprocessed first electroencephalogram signal data segment near a motor cortex; the second marked electroencephalogram signal data segment is the electroencephalogram signal data of the preprocessed second electroencephalogram signal data segment near the motor cortex;
since the motor cortex is the region that controls head motion, 3-channel signals near the motor cortex are screened in this example: FCz, FC3, FC4 as labeled brain electrical signal data segments.
Step D: and respectively processing the first marked electroencephalogram signal data section and the second marked electroencephalogram signal data section by adopting a short-time Fourier transform algorithm to obtain a transform map with an idle state, a transform map with a head left-turning intention and a transform map with a head right-turning intention.
Or respectively processing the first marked electroencephalogram signal data segment and the second marked electroencephalogram signal data segment by adopting a continuous wavelet transform algorithm to obtain a transform map with an idle state, a transform map with a head left-turn intention and a transform map with a head right-turn intention.
Because the electroencephalogram signal is a time sequence signal, namely a voltage value which changes along with time, Continuous Wavelet Transform (CWT) or short-time Fourier transform is needed to convert the time sequence signal into a transform map which has time-frequency information and the depth of 3, and the transform map has time-frequency characteristics of the signal and can extract more classification information.
In addition, the embodiment can select other brain electrical signals besides the 3-channel signals near the motor cortex for subsequent processing. Besides the time-frequency information obtained through wavelet transformation or short-time Fourier transformation, the time-frequency information can also be obtained through other time-frequency analysis methods.
The training process is as follows:
step 1: a Convolutional Neural Network (CNN) as shown in fig. 2 was constructed. The convolutional neural network comprises three convolutional layers and three max pooling layers, two fully-connected layers ( parameters 600 and 300, respectively).
Since the convolutional neural network has strong learning ability and fitting ability to the features, the Convolutional Neural Network (CNN) as shown in fig. 2 may be used to process the transform map.
Step 2: firstly, inputting a transformation map with an idle state and a head movement intention state into a convolutional neural network for two-classification training to obtain a first classification network parameter model for identifying the idle state and the head movement intention; secondly, inputting the transformation map with the head left turning intention state and the head right turning intention state into a convolutional neural network for two-classification training to obtain a second classification network parameter model for identifying the head left turning intention and the head right turning intention; wherein the head movement intent decoding model includes a first classification network parameter model and a second classification network parameter model.
And step 3: and putting the transform map with the idle state, the head left turning intention and the head right turning intention into the trained head movement intention decoding model, and verifying the experimental result to further obtain the final head movement intention decoding model.
Wherein the inputting the transform map into a head movement intention decoding model to identify the head movement intention of the target user specifically comprises:
inputting the transformation map into a head movement intention decoding model to obtain a first classification result and a second classification result; the first classification result is in an idle state or a head movement intention state, the second classification result represents a head left-turn intention or a head right-turn intention, and when the first classification result is in the head movement intention state, the second classification result is the head left-turn intention or the head right-turn intention.
In the embodiment, the electroencephalogram signals are subjected to frequency domain transformation to obtain data with time-frequency characteristics, and then the data with the time-frequency characteristics are trained by means of the strong learning capacity of the convolutional neural network to obtain a classification network parameter model which can classify the characteristics and inhibit noise characteristics. In addition, compared with the existing electroencephalogram decoding technology, the method has the advantage of high classification accuracy, and meanwhile, the established head movement intention decoding model can be applied to the identification of the subsequent head movement intention, so that the method has the advantage of short time consumption.
Example two
Referring to fig. 3, the electroencephalogram signal decoding system for recognizing the intention of head movement provided by the present embodiment includes:
the data acquisition module 301 is configured to acquire an electroencephalogram signal data segment of a target user in a head movement process.
The data processing module 302 is configured to process the electroencephalogram signal data segment to obtain a transform map corresponding to the target user; the transformation map is a map with time-frequency information.
A first head movement intention identifying module 303, configured to process the transform map by using a machine learning algorithm, and identify a head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
EXAMPLE III
Referring to fig. 4, the electroencephalogram signal decoding method for identifying an intention of head movement provided by the embodiment includes:
step 401: acquiring and preprocessing an electroencephalogram signal data segment of a target user in the head movement process.
Step 402: processing the preprocessed electroencephalogram signal data segment by adopting a time-sequence neural network, and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
The time-sequence neural network is a Recurrent Neural Network (RNN) or a long-short term memory network (LSTM).
Example four
Referring to fig. 5, the electroencephalogram signal decoding system for recognizing the intention of head movement provided by the present embodiment includes:
and the data acquisition and processing module 501 is configured to acquire and pre-process an electroencephalogram signal data segment of the target user in the head movement process.
A second head movement intention identifying module 502, configured to process the preprocessed electroencephalogram signal data segment by using a time-sequential neural network, and identify a head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principle and the embodiment of the present invention are explained by applying specific examples, and the above description of the embodiments is only used to help understanding the method and the core idea of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (9)
1. An electroencephalogram signal decoding method for recognizing an intention of head movement, comprising:
acquiring an electroencephalogram signal data segment of a target user in a head movement process;
processing the electroencephalogram signal data segment to obtain a transformation map corresponding to the target user; the transformation map is a map with time-frequency information;
processing the transformation map by adopting a machine learning algorithm, and identifying the head movement intention of the target user; the head movement intent comprises an idle state, a head left turn intent, and a head right turn intent;
the processing the electroencephalogram signal data segment to obtain a transform map corresponding to the target user specifically includes:
preprocessing the electroencephalogram signal data segment; the preprocessing comprises electrode positioning, useless electrode elimination, re-reference, filtering, segmentation and baseline correction;
processing the preprocessed electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user;
the transform map is a continuous wavelet transform map or a short-time Fourier transform map; the transformation maps include a transformation map having an idle state, a transformation map having an intention of head left turning, and a transformation map having an intention of head right turning.
2. The method for decoding electroencephalogram signals for identifying head movement intentions, according to claim 1, is characterized in that the method for processing the preprocessed electroencephalogram signal data segments by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user specifically comprises the following steps:
screening the preprocessed electroencephalogram signal data segment to determine electroencephalogram signal data near a motor cortex;
processing the marked electroencephalogram signal data by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user;
the marked electroencephalogram signal data are determined electroencephalogram signal data near the motor cortex.
3. The method for decoding an electroencephalogram signal capable of recognizing the intention of head movement according to claim 1, wherein the transformation map is processed by a machine learning algorithm to recognize the intention of the head movement of the target user, and specifically comprises:
inputting the transform map into a head movement intent decoding model to identify a head movement intent of the target user;
the head movement intent decoding model is determined from training data and a convolutional neural network; the training data includes a transform map having an idle state, a transform map having a head left turn intention, and a transform map having a head right turn intention.
4. The method for decoding brain electrical signals with intention of head movement according to claim 3, wherein the training data is determined by:
acquiring a first electroencephalogram signal data segment and a second electroencephalogram signal data segment of a user through electroencephalogram cap equipment; the first electroencephalogram signal data segment is electroencephalogram signal data before head turning movement; the second electroencephalogram signal data section is electroencephalogram signal data after left-right turning movement; the second electroencephalogram signal data section comprises an electroencephalogram signal data section after the head turns left and an electroencephalogram signal data section during the head turns right;
preprocessing the first electroencephalogram signal data segment and the second electroencephalogram signal data segment;
screening the preprocessed first electroencephalogram signal data segment and the preprocessed second electroencephalogram signal data segment, and determining a first marked electroencephalogram signal data segment and a second marked electroencephalogram signal data segment; the first marked electroencephalogram signal data segment is electroencephalogram signal data of the preprocessed first electroencephalogram signal data segment near a motor cortex; the second marked electroencephalogram signal data section is the electroencephalogram signal data of the preprocessed second electroencephalogram signal data section near the motor cortex;
respectively processing the first marked electroencephalogram signal data segment and the second marked electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm to obtain a transform map with an idle state, a transform map with a left head turning intention and a transform map with a right head turning intention; or respectively processing the first marked electroencephalogram signal data segment and the second marked electroencephalogram signal data segment by adopting a continuous wavelet transform algorithm to obtain a transform map with an idle state, a transform map with a head left-turn intention and a transform map with a head right-turn intention.
5. The method for decoding brain electrical signals for identifying head movement intention according to claim 3, wherein the inputting the transform map into a head movement intention decoding model to identify the head movement intention of the target user specifically comprises:
inputting the transformation map into a head movement intention decoding model to obtain a first classification result and a second classification result;
the first classification result is in an idle state or a head movement intention state, the second classification result represents a head left-turning intention or a head right-turning intention, and when the first classification result is in the head movement intention state, the second classification result is the head left-turning intention or the head right-turning intention.
6. An electroencephalogram signal decoding system for recognizing an intention of head movement, comprising:
the data acquisition module is used for acquiring an electroencephalogram signal data segment of a target user in the head movement process;
the data processing module is used for processing the electroencephalogram signal data segment to obtain a transformation map corresponding to the target user; the transformation map is a map with time-frequency information;
the first head movement intention recognition module is used for processing the transformation map by adopting a machine learning algorithm and recognizing the head movement intention of the target user; the head movement intent comprises an idle state, a head left turn intent, and a head right turn intent;
the data processing module specifically comprises:
preprocessing the electroencephalogram signal data segment; the preprocessing comprises electrode positioning, useless electrode elimination, re-reference, filtering, segmentation and baseline correction;
processing the preprocessed electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map corresponding to the target user;
the transform map is a continuous wavelet transform map or a short-time Fourier transform map; the transformation maps include a transformation map having an idle state, a transformation map having an intention of head left turning, and a transformation map having an intention of head right turning.
7. A method for decoding brain electrical signals for recognizing head movement intention, comprising:
acquiring and preprocessing an electroencephalogram signal data segment of a target user in the head movement process; processing the electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map with an idle state, a transform map with a head left-turn intention and a transform map with a head right-turn intention;
processing the preprocessed electroencephalogram signal data segment by adopting a time-sequence neural network, and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
8. The method for decoding brain electrical signals with intention of head movement according to claim 7, wherein the time-sequence neural network is a Recurrent Neural Network (RNN) or a long-short term memory network (LSTM).
9. An electroencephalogram signal decoding system for recognizing an intention of head movement, comprising:
the data acquisition processing module is used for acquiring and preprocessing an electroencephalogram signal data section of a target user in the head movement process; processing the electroencephalogram signal data segment by adopting a short-time Fourier transform algorithm or a continuous wavelet transform algorithm to obtain a transform map with an idle state, a transform map with a head left-turn intention and a transform map with a head right-turn intention;
the second head movement intention identification module is used for processing the preprocessed electroencephalogram signal data section by adopting a time-sequence neural network and identifying the head movement intention of the target user; the head movement intent includes an idle state, a head left turn intent, and a head right turn intent.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111521319.5A CN114027855B (en) | 2021-12-13 | 2021-12-13 | Electroencephalogram signal decoding method and system for recognizing head movement intention |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111521319.5A CN114027855B (en) | 2021-12-13 | 2021-12-13 | Electroencephalogram signal decoding method and system for recognizing head movement intention |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114027855A CN114027855A (en) | 2022-02-11 |
| CN114027855B true CN114027855B (en) | 2022-09-23 |
Family
ID=80140623
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111521319.5A Active CN114027855B (en) | 2021-12-13 | 2021-12-13 | Electroencephalogram signal decoding method and system for recognizing head movement intention |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114027855B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116035592B (en) * | 2023-01-10 | 2024-06-14 | 北京航空航天大学 | Method, system, equipment and medium for identifying turning intention based on deep learning |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018094720A1 (en) * | 2016-11-24 | 2018-05-31 | 浙江大学 | Clinical electroencephalogram signal-based brain-machine interface system for controlling robotic hand movement and application thereof |
| CN108852349A (en) * | 2018-05-17 | 2018-11-23 | 浙江大学 | A kind of moving decoding method using Cortical ECoG signal |
| JP2019076712A (en) * | 2017-10-20 | 2019-05-23 | パナソニック株式会社 | Electroencephalograph and electroencephalogram measurement system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110307079A1 (en) * | 2010-04-29 | 2011-12-15 | Board Of Trustees Of Michigan State University, The | Multiscale intra-cortical neural interface system |
| US9717440B2 (en) * | 2013-05-03 | 2017-08-01 | The Florida International University Board Of Trustees | Systems and methods for decoding intended motor commands from recorded neural signals for the control of external devices or to interact in virtual environments |
-
2021
- 2021-12-13 CN CN202111521319.5A patent/CN114027855B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018094720A1 (en) * | 2016-11-24 | 2018-05-31 | 浙江大学 | Clinical electroencephalogram signal-based brain-machine interface system for controlling robotic hand movement and application thereof |
| JP2019076712A (en) * | 2017-10-20 | 2019-05-23 | パナソニック株式会社 | Electroencephalograph and electroencephalogram measurement system |
| CN108852349A (en) * | 2018-05-17 | 2018-11-23 | 浙江大学 | A kind of moving decoding method using Cortical ECoG signal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114027855A (en) | 2022-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ko et al. | Multi-scale neural network for EEG representation learning in BCI | |
| CN111062250B (en) | Multi-subject motor imagery electroencephalogram signal identification method based on deep feature learning | |
| CN106899968B (en) | Active non-contact identity authentication method based on WiFi channel state information | |
| CN110991406B (en) | RSVP electroencephalogram characteristic-based small target detection method and system | |
| KR20190035368A (en) | Method and Apparatus for Recognizing Emotion Based on Image Converted from Brain Signal | |
| CN110495893B (en) | System and method for multi-level dynamic fusion recognition of continuous brain and muscle electricity of motor intention | |
| CN114027855B (en) | Electroencephalogram signal decoding method and system for recognizing head movement intention | |
| CN109375776B (en) | Electroencephalogram action intention recognition method based on multi-task RNN model | |
| CN113609977B (en) | Pedestrian gait recognition method based on channel state information quotient distance | |
| CN112426160A (en) | Electrocardiosignal type identification method and device | |
| CN113486752B (en) | Emotion recognition method and system based on electrocardiosignal | |
| CN109009098B (en) | Electroencephalogram signal feature identification method under motor imagery state | |
| Ramos-Aguilar et al. | Analysis of EEG signal processing techniques based on spectrograms | |
| CN110543831A (en) | brain print identification method based on convolutional neural network | |
| CN104331892A (en) | Morphology-based neuron recognizing and analyzing method | |
| CN114578967A (en) | Emotion recognition method and system based on electroencephalogram signals | |
| Mao et al. | Classification of non-time-locked rapid serial visual presentation events for brain-computer interaction using deep learning | |
| Hejazi et al. | Feature level fusion for biometric verification with two-lead ECG signals | |
| CN116531001A (en) | Method and device for generating multi-listener electroencephalogram signals and identifying emotion of cross-listener | |
| CN114578963B (en) | Electroencephalogram identity recognition method based on feature visualization and multi-mode fusion | |
| CN115469749A (en) | Target positioning method based on auditory brain-computer interface | |
| Luczak et al. | Spectral representation—analyzing single-unit activity in extracellularly recorded neuronal data without spike sorting | |
| CN116035592B (en) | Method, system, equipment and medium for identifying turning intention based on deep learning | |
| CN115081471B (en) | Peak potential detection and classification method based on multi-band composite waveform and application | |
| CN117251737B (en) | Lightning waveform processing model training method, classification method, device and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |