CN111466909A - Target detection method and system based on electroencephalogram characteristics - Google Patents
Target detection method and system based on electroencephalogram characteristics Download PDFInfo
- Publication number
- CN111466909A CN111466909A CN202010292815.7A CN202010292815A CN111466909A CN 111466909 A CN111466909 A CN 111466909A CN 202010292815 A CN202010292815 A CN 202010292815A CN 111466909 A CN111466909 A CN 111466909A
- Authority
- CN
- China
- Prior art keywords
- electroencephalogram
- signal
- target
- detected
- image
- 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
- A61B5/24—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
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- 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
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- 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
- A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2576/00—Medical imaging apparatus involving image processing or analysis
- A61B2576/02—Medical imaging apparatus involving image processing or analysis specially adapted for a particular organ or body part
- A61B2576/026—Medical imaging apparatus involving image processing or analysis specially adapted for a particular organ or body part for the brain
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Surgery (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Veterinary Medicine (AREA)
- Psychiatry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Psychology (AREA)
- Power Engineering (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention provides a target detection method and a system based on electroencephalogram characteristics, which comprises the following steps: acquiring an electroencephalogram signal of a tested image to be detected; analyzing the event-related potential components of the electroencephalogram signals, and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature; detecting the image to be detected by using a trained preset classifier based on the target characteristic signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains the target to be detected, and the image to be detected does not contain the target to be detected. The method and the device solve the technical problem that the target detection precision of the remote sensing image is not high in the prior art.
Description
Technical Field
The invention relates to the technical field of image detection, in particular to a target detection method and system based on electroencephalogram characteristics.
Background
Image media always contain more information than other information transmission media. The method is one of important means for information processing by detecting an interested target in an image and analyzing related information thereof. The remote sensing image is mainly an image generated by scanning the ground in a full wave band from a satellite, and contains more abundant information compared with a common image. The remote sensing image target detection has wider application in military fields such as enemy investigation of airport and port, search of specific military targets, monitoring of sailing ships and airplanes in wide area and civil fields such as weather forecast, agricultural survey, geographic exploration, sea police forecast and hydrological monitoring, and is an important branch of remote sensing image analysis. An important application in remote sensing image target detection is the detection of specific artificial targets in the visible light band. Although the target detection is in a visible light wave band, compared with the common image target detection, the remote sensing image target detection has the characteristics of complex and variable imaging quality, small target scale, few remote sensing image samples and the like, and has certain difficulty. However, due to the characteristics of small target size and changeable shape in the remote sensing image, the target detection precision is not high when the remote sensing image is detected by using the traditional image detection method.
Disclosure of Invention
In view of this, the present invention provides a target detection method and system based on electroencephalogram characteristics, so as to alleviate the technical problem in the prior art that the target detection accuracy of a remote sensing image is not high.
In a first aspect, an embodiment of the present invention provides a target detection method based on electroencephalogram characteristics, including: acquiring an electroencephalogram signal of a tested image to be detected; analyzing the event-related potential components of the electroencephalogram signals, and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature; detecting the image to be detected by utilizing a trained preset classifier based on the target characteristic signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains a target to be detected, and the image to be detected does not contain the target to be detected.
Further, the event-related potential component analysis is performed on the electroencephalogram signal, and the extraction of the target characteristic signal of the electroencephalogram signal comprises the following steps: analyzing the components of the event-related potential of the electroencephalogram signal, and extracting an initial characteristic signal of the electroencephalogram signal; and carrying out filtering operation on the initial characteristic signal by using an xDAWN spatial filtering algorithm to obtain a target characteristic signal.
Further, the method further comprises: and training the preset classifier to obtain the trained preset classifier.
Further, training a preset classifier to obtain the trained preset classifier, including: acquiring initial electroencephalogram signals of a plurality of stimulation materials to be tested; removing artifacts in the initial electroencephalogram signal, and filtering the initial electroencephalogram signal to obtain an electroencephalogram signal subjected to noise reduction; analyzing event-related potential components of the de-noised electroencephalogram signal, and extracting a sample characteristic signal of the de-noised electroencephalogram signal; the sample characteristic signal is any one of: p3 signature, N2 signature; and training a preset classifier by using the sample characteristic signal to obtain the trained preset classifier.
Further, the preset classifier is a linear discriminant classifier; based on the target characteristic signal, the trained preset classifier is utilized to detect the image to be detected, and a detection result is obtained, wherein the detection result comprises the following steps: acquiring a projection vector of the linear discrimination classifier; calculating electrode channel values of the target feature signals based on the projection vectors; comparing the electrode channel value with a classification threshold value of the trained preset classifier to obtain a comparison result; and detecting the image to be detected based on the comparison result to obtain a detection result.
In a second aspect, an embodiment of the present invention further provides a target detection system based on electroencephalogram characteristics, including: the device comprises an acquisition module, a feature extraction module and a detection module, wherein the acquisition module is used for acquiring an electroencephalogram signal of a tested image to be detected; the characteristic extraction module is used for analyzing the event-related potential components of the electroencephalogram signals and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature; the detection module is used for detecting the image to be detected by utilizing a trained preset classifier based on the target characteristic signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains a target to be detected, and the image to be detected does not contain the target to be detected.
Further, the feature extraction module comprises: the extraction unit is used for analyzing the event-related potential components of the electroencephalogram signals and extracting initial characteristic signals of the electroencephalogram signals; and the filtering unit is used for carrying out filtering operation on the initial characteristic signal by using an xDAWN spatial filtering algorithm to obtain a target characteristic signal.
Further, the system further comprises: and the training module is used for training the preset classifier to obtain the trained preset classifier.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method according to the first aspect when executing the computer program.
In a fourth aspect, the present invention further provides a computer-readable medium having non-volatile program code executable by a processor, where the program code causes the processor to execute the method according to the first aspect.
The invention provides a target detection method and a system based on electroencephalogram characteristics, which comprises the following steps: acquiring an electroencephalogram signal of a tested image to be detected; analyzing the event-related potential components of the electroencephalogram signals, and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature; and detecting the image to be detected by using the trained preset classifier based on the target characteristic signal to obtain a detection result. According to the method, the detection of the weak and small targets in the image containing the complex scene is realized by analyzing the event-related potential components of the electroencephalogram signal of the tested image to be detected, extracting the P3 characteristic signal or the N2 characteristic signal and detecting the image to be detected by utilizing the P3 characteristic signal or the N2 characteristic signal, so that the accuracy of the detection of the weak and small targets is greatly improved, and the technical problem of low target detection accuracy of the remote sensing image in the prior art is solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flowchart of a target detection method based on electroencephalogram characteristics according to an embodiment of the present invention;
fig. 2 is a flowchart of a method for training a preset classifier according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating a method for testing a subject according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a target detection system based on electroencephalogram characteristics according to an embodiment of the present invention;
fig. 5 is a schematic diagram of another target detection system based on electroencephalogram characteristics according to an embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. 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.
The first embodiment is as follows:
fig. 1 is a flowchart of a target detection method based on electroencephalogram characteristics according to an embodiment of the present invention. As shown in fig. 1, the method specifically includes the following steps:
and S102, acquiring the electroencephalogram signal of the image to be detected. Alternatively, the image to be detected may be a remotely sensed image.
Step S104, analyzing the event-related potential components of the electroencephalogram signals, and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature.
An Event-Related Potential (ERP) is a specific brain-evoked Potential, and by intentionally imparting a specific psychological significance to a stimulus, a Potential of the brain caused by a plurality of or various lateral notes is utilized. The classic ERP comprises the following main components: p1, N1, P2, N2 and P3, wherein the first three are exogenous components, and the last two are endogenous components. In the embodiment of the invention, the P3 characteristic signal or the N2 characteristic signal in the ERP component of the electroencephalogram signal is mainly extracted.
Optionally, in the embodiment of the present invention, a 61-dimensional channel signal in the range of 460ms to 500ms in the electroencephalogram signal is used as the P3 feature signal, and a 61-dimensional channel signal in the range of 268ms to 320ms in the electroencephalogram signal is used as the N2 feature signal.
Step S106, detecting the image to be detected by using a trained preset classifier based on the target characteristic signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains the target to be detected, and the image to be detected does not contain the target to be detected.
Alternatively, the preset classifier may be any one of a linear Discriminant classifier (L initial classifier Analysis, L DA), a Support Vector Machine (SVM) classifier, or other classifier.
According to the target detection method based on the electroencephalogram characteristics, provided by the embodiment of the invention, the detection of the dim targets in the images containing complex scenes is realized by carrying out ERP component analysis on the electroencephalogram signals of the tested images to be detected, extracting P3 characteristic signals or N2 characteristic signals and detecting the images to be detected by utilizing the P3 characteristic signals or N2 characteristic signals, so that the accuracy of the detection of the dim targets is greatly improved, and the technical problem of low target detection accuracy of the remote sensing images in the prior art is solved.
Optionally, step S104 includes the steps of:
step S1041, analyzing the components of the event-related potential of the electroencephalogram signal, and extracting an initial characteristic signal of the electroencephalogram signal;
step S1042, carrying out filtering operation on the initial characteristic signal by using an xDAWN spatial filtering algorithm to obtain a target characteristic signal.
ERP components, such as N2 signature or P3 signature, directly characterized by the electrode channel are often contaminated by interfering components, and therefore, it is desirable to improve the signal-to-noise ratio by appropriate linear transformations. In the embodiment of the invention, the spatial filter is designed by using the xDAWN spatial filtering algorithm, and then the spatial filter is used for carrying out filtering operation on the initial characteristic signal to obtain the target characteristic signal, so that the technical effect of improving the signal-to-noise ratio of the characteristic signal can be achieved.
Optionally, the method provided in the embodiment of the present invention further includes: and training the preset classifier to obtain the trained preset classifier. Specifically, as shown in fig. 2, the method for training the preset classifier includes the following steps:
step S21, obtaining initial electroencephalogram signals of a subject about a plurality of stimulation materials. The stimulation material is a remote sensing image set. The method specifically comprises the following substeps:
step S211, making stimulation materials, wherein the experimental stimulation materials are from a public real remote sensing image data set acquired from a network, 2700 remote sensing images are obtained in total, the number of samples containing the target to be detected is 270, the number of samples not containing the target to be detected is 2430, the number ratio of positive samples to negative samples is 1:9, the stimulation materials are divided into 27 groups, 100 samples in each group comprise 10 target stimulation samples, and 90 interference stimulation samples.
In step S212, the experiment was recruited, and a total of 11 subjects (7 males and 4 females) not participating in the similar experiment were recruited for compensation. The age is 16-25 years old, the vision is normal or corrected, and the recent mental state is good. They have not previously known the experimental conditions, but have previously been informed of the relevant conditions, such as the purpose of research, experimental procedures, experimental operations, rights and interests, etc.
Step S213, designing an experiment process, controlling Presentation of stimulation materials by using Presentation experiment software, wherein the flashing frequency of the stimulation materials is 4Hz, 4 stimulation materials are presented per second (namely the Presentation time of each remote sensing picture is 250ms), the stimulation Presentation uses a 24' display (PHI L IPS 242G5 DJB) with the resolution of 1920 × 1080 and the refresh rate of 144Hz, each group of experiments is divided into 4 parts, as shown in FIG. 3, ① is an experiment guide and 5S, the experiment guide part guides a tested key operation and starts an experiment group by displaying a guide language on a screen, ② is a stimulation material Presentation and 25S, 100 remote sensing image materials are provided, 10 stimulation materials containing targets to be detected are randomly inserted into the 100 stimulation materials with equal probability, ③ is a tested input answer and 5S, after Presentation of each group of stimulation materials is finished, counting results of the stimulation materials containing the targets to be tested are input through a screen prompt, ④ is a formal rest time after each group of experiments is finished, the rest time of each group is 25S, after each group of experiments is finished, the experiments are 60S, and the test operation needs to be performed in each experiment group.
Step S214, recording initial EEG signals, and acquiring EEG signals by using 64 channels of EEG signal amplifiers (such as BrainAmps, Brain Products, Germany) and matching software (such as BrainVision Recoder, Brain Products, Germany), wherein the sampling rate is 500Hz, the EEG cap used in the experiment has 64 active Cap electrodes (Brain Products, Germany) which are extended systems by a 10-5 standard system, and the electrodes used in practice are Fp1-2, AFz, AF3-4, AF7-8, Fz, F1-8, FC1-6, FT7-10, C1-6, T7-8, CPz1-6, TP7-10, Pz1-8, POz, PO 3-7374, PO 3-7378, PO 463-7378, Oz 4642 and Oz 462-42.
And step S22, removing artifacts in the initial electroencephalogram signal, and filtering the initial electroencephalogram signal to obtain the electroencephalogram signal subjected to noise reduction. Specifically, the method comprises the following substeps:
step S221, removing artifacts in the initial electroencephalogram signal by an Artifact Subspace Reconstruction (ASR) method, specifically including the following 4 steps: 1) the sampling frequency is reduced to 250Hz, so that the memory required by storage and subsequent calculation is reduced; 2) recalibrating the reference potential, wherein the reference potential is the average potential of TP9 and TP 10; 3) keeping data from 5s before the beginning to 5s after the end of each group of experiments, and deleting other redundant data; 4) and processing the continuous data by using an ASR method.
Step S222, the initial electroencephalogram signal collected by the electroencephalogram cap can be regarded as a mixed signal of multiple independent source signals, and an artifact signal which is statistically independent from electroencephalogram components representing brain neuron electric signals can be effectively separated; by subtracting these independent artifact components, the artifact signals that are mixed in the EEG signal can be removed, which is done in 4 steps as follows: 1) performing low-pass and high-pass filtering on the data by using an FIR (finite Impulse response) filter, wherein the cut-off frequency of the low-pass filtering is 30Hz, and the cut-off frequency of the high-pass filtering is 1 Hz; 2) dividing the filtered continuous data into 27 test runs according to the experimental components, and performing Independent Component Analysis (ICA); 3) dividing the EEG data into 2700 epochs, wherein each epoch is 0.2s before the start of stimulation to 1s after the start of stimulation; 4) determining and removing typical artifact components such as eye movement, eye drift, head movement and the like by referring to prompts given by an ADJUST plug-in; 5) taking the data 200ms before stimulation as reference, removing the baseline, and labeling the preprocessed EEG data segment according to the stimulation event, thereby obtaining two types of EEG sample data: target stimulation-like samples and interferential stimulation-like samples. The number of stimulation samples per tested target is 270, and the number of interference stimulation samples is 2430.
Step S23, analyzing the event-related potential components of the de-noised electroencephalogram signal, and extracting a sample characteristic signal of the de-noised electroencephalogram signal; the sample characteristic signal is any one of: p3 signature, N2 signature. Specifically, the method comprises the following substeps:
step S231, calculating the average ERP of the EEG signals of each tested experimental group, dividing 2700 EEG segment data of each tested group into two types, namely an interference stimulation type and a target stimulation type, and aligning each type of sample data by taking the stimulation occurrence time as a time zero point. Each sample data included 1000ms after stimulation, taking into account the appearance period of the ERP component. All sample data can be used by the data matrix
A description will be given. Where i-1, 2 denote target and interferential stimulation samples, respectively, k-1, 2, …, EiDenotes the data number in each sample, c is 1,2, …, NcDenotes the electrode channel number, t1, 2, …, NtRepresenting the sampling time. In the embodiment of the invention, Nc is 61, Nt is 270, E1＝270,E22430. Therefore, the average ERP of the brain electrical signals of each experimental group to be tested can be expressed as follows:
where mean () is an averaging function, and g 1,2, and … 27 are the numbers of the experimental groups, that is, the average is performed according to the experimental group to which the sample data belongs. There were 100 sample stimuli in each set of experiments, 90 interfering stimuli and 10 target stimuli, thus
10 times of superposition averaging are carried out, and
90 fold averages were performed. The average ERP per tested can be expressed as:
and step S232, extracting a sample characteristic signal of the noise-reduced electroencephalogram signal. By counting all the tested
The average value and the difference value of the measured values are found to be the largest, the two types of sample data recorded by a plurality of electrodes have significant difference, and although the amplitude and the latency of the tested target stimulation ERP have difference, the most tested difference ERP can observe a significant positive peak at about 400ms, namely a P3 characteristic signal; half of the tests were able to observe a negative peak at about 200ms, i.e., the N2 signature.
And step S24, training the preset classifier by using the sample characteristic signal to obtain the trained preset classifier.
In particular, the projection vector of L DA can be expressed as:
wherein mu1And mu2The estimation can be carried out by two types of characteristic sample signals; swIt is estimated that as follows,
wherein, X1Characteristic signals corresponding to target stimulation, X2Characteristic signals corresponding to target stimulation, SwInverse matrix of
The projection vector w of L DA calculated by the above formula can project the matrix of the sample characteristic signal into a virtual electrode channel value:
wherein, tauiFor sampled segments of the selected characteristic signal, τ1＝[268ms,320ms]，τ2＝[460ms,500ms],
Is the N2 characteristic signal and is,
is the P3 signature. The classification threshold is the median of the signal class centers after two types of projection:
wherein ThP3Classification threshold, Th, for P3 signatureN2Is the classification threshold of the N2 signature.
Optionally, in the embodiment of the present invention, if the preset classifier is a linear discriminant classifier, step S106 includes the following steps:
step S1061, acquiring a projection vector of the linear discrimination classifier;
step S1062, calculating an electrode channel value of the target characteristic signal based on the projection vector;
step S1063, comparing the electrode channel value with a classification threshold value of a trained preset classifier to obtain a comparison result;
and step S1064, detecting the image to be detected based on the comparison result to obtain a detection result.
The embodiment of the invention provides a target detection method based on electroencephalogram characteristics, which realizes the detection of a small target in an image containing a complex scene by carrying out ERP component analysis on an electroencephalogram signal of a tested image to be detected, extracting a P3 characteristic signal or an N2 characteristic signal and utilizing a P3 characteristic signal or an N2 characteristic signal to detect the image to be detected, wherein in the embodiment of the invention, a 61-dimensional channel signal in a 460-500 ms interval in the electroencephalogram signal is used as the P3 characteristic signal, a 61-dimensional signal in an 268-320 ms interval in the electroencephalogram signal is used as the N2 characteristic signal, and through experimental comparison, the specific P3 characteristic signal and the specific N2 characteristic signal are found to detect the small target in the image to be detected, so that the technical effect of improving the accuracy can be achieved. Specifically, the method provided by the embodiment of the invention can achieve the following technical effects:
1. compared with the traditional method, the method provided by the embodiment of the invention finds a potential physiological method detection target, and realizes the detection of the weak and small target in a complex scene.
2. The method provided by the embodiment of the invention discovers that the separability of the target is proportional to the separability of the P3 characteristic and the N2 characteristic, and greatly improves the accuracy of the detection of the weak and small targets.
Example two:
FIG. 4 is a schematic diagram of a target detection system based on electroencephalogram characteristics according to an embodiment of the present invention. As shown in fig. 4, the system includes: the system comprises an acquisition module 10, a feature extraction module 20 and a detection module 30.
Specifically, the obtaining module 10 is configured to obtain an electroencephalogram signal of the image to be detected. Optionally, the image to be detected is a remote sensing image.
The characteristic extraction module 20 is used for analyzing the event-related potential components of the electroencephalogram signals and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature.
Optionally, in the embodiment of the present invention, a 61-dimensional channel signal in the range of 460ms to 500ms in the electroencephalogram signal is used as the P3 feature signal, and a 61-dimensional channel signal in the range of 268ms to 320ms in the electroencephalogram signal is used as the N2 feature signal.
The detection module 30 is configured to detect an image to be detected by using a trained preset classifier based on the target feature signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains the target to be detected, and the image to be detected does not contain the target to be detected.
Alternatively, the preset classifier may be any of L DA, SVM, or other classifiers.
According to the target detection system based on the electroencephalogram characteristics, provided by the embodiment of the invention, the detection of weak and small targets in an image containing a complex scene is realized by carrying out ERP component analysis on an electroencephalogram signal of a tested image to be detected, extracting a P3 characteristic signal or an N2 characteristic signal and detecting the image to be detected by utilizing the P3 characteristic signal or the N2 characteristic signal, so that the accuracy of the detection of the weak and small targets is greatly improved, and the technical problem of low target detection accuracy of a remote sensing image in the prior art is solved.
Optionally, fig. 5 is a schematic diagram of another target detection system based on electroencephalogram characteristics according to an embodiment of the present invention. As shown in fig. 5, the feature extraction module 20 includes: an extraction unit 21 and a filtering unit 22.
Specifically, the extracting unit 21 is configured to perform event-related potential component analysis on the electroencephalogram signal, and extract an initial characteristic signal of the electroencephalogram signal.
And the filtering unit 22 is configured to perform a filtering operation on the initial characteristic signal by using an xDAWN spatial filtering algorithm to obtain a target characteristic signal.
Optionally, as shown in fig. 5, the system further includes: and the training module 40 is used for training the preset classifier to obtain the trained preset classifier.
Specifically, the training module 40 is further configured to: acquiring initial electroencephalogram signals of a plurality of stimulation materials to be tested; removing artifacts in the initial electroencephalogram signal, and filtering the initial electroencephalogram signal to obtain an electroencephalogram signal subjected to noise reduction; analyzing the event-related potential components of the de-noised electroencephalogram signal, and extracting a sample characteristic signal of the de-noised electroencephalogram signal; the sample characteristic signal is any one of: p3 signature, N2 signature; and training the preset classifier by using the sample characteristic signal to obtain the trained preset classifier.
Optionally, if the preset classifier is L DA, the detection module 30 is further configured to obtain a projection vector of the linear discriminant classifier, calculate an electrode channel value of the target feature signal based on the projection vector, compare the electrode channel value with a classification threshold of the trained preset classifier to obtain a comparison result, and detect the image to be detected based on the comparison result to obtain a detection result.
The embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and when the processor executes the computer program, the steps of the method in the first embodiment are implemented.
The embodiment of the invention also provides a computer readable medium with a non-volatile program code executable by a processor, wherein the program code causes the processor to execute the method in the first embodiment.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A target detection method based on electroencephalogram features is characterized by comprising the following steps:
acquiring an electroencephalogram signal of a tested image to be detected;
analyzing the event-related potential components of the electroencephalogram signals, and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature;
detecting the image to be detected by utilizing a trained preset classifier based on the target characteristic signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains a target to be detected, and the image to be detected does not contain the target to be detected.
2. The method of claim 1, wherein performing event-related potential component analysis on the electroencephalogram signal to extract a target characteristic signal of the electroencephalogram signal comprises:
analyzing the components of the event-related potential of the electroencephalogram signal, and extracting an initial characteristic signal of the electroencephalogram signal;
and carrying out filtering operation on the initial characteristic signal by using an xDAWN spatial filtering algorithm to obtain a target characteristic signal.
3. The method of claim 1, further comprising: and training the preset classifier to obtain the trained preset classifier.
4. The method of claim 3, wherein training the pre-set classifier to obtain the trained pre-set classifier comprises:
acquiring initial electroencephalogram signals of a plurality of stimulation materials to be tested;
removing artifacts in the initial electroencephalogram signal, and filtering the initial electroencephalogram signal to obtain an electroencephalogram signal subjected to noise reduction;
analyzing event-related potential components of the de-noised electroencephalogram signal, and extracting a sample characteristic signal of the de-noised electroencephalogram signal; the sample characteristic signal is any one of: p3 signature, N2 signature;
and training a preset classifier by using the sample characteristic signal to obtain the trained preset classifier.
5. The method of claim 1, wherein the pre-set classifier is a linear discriminant classifier;
based on the target characteristic signal, the trained preset classifier is utilized to detect the image to be detected, and a detection result is obtained, wherein the detection result comprises the following steps:
acquiring a projection vector of the linear discrimination classifier;
calculating electrode channel values of the target feature signals based on the projection vectors;
comparing the electrode channel value with a classification threshold value of the trained preset classifier to obtain a comparison result;
and detecting the image to be detected based on the comparison result to obtain a detection result.
6. A target detection system based on electroencephalogram characteristics, comprising: an acquisition module, a feature extraction module and a detection module, wherein,
the acquisition module is used for acquiring the electroencephalogram signal of the tested image to be detected;
the characteristic extraction module is used for analyzing the event-related potential components of the electroencephalogram signals and extracting target characteristic signals of the electroencephalogram signals; the target characteristic signal is any one of: p3 signature, N2 signature;
the detection module is used for detecting the image to be detected by utilizing a trained preset classifier based on the target characteristic signal to obtain a detection result; the detection result comprises any one of the following items: the image to be detected contains a target to be detected, and the image to be detected does not contain the target to be detected.
7. The system of claim 6, wherein the feature extraction module comprises: an extraction unit and a filtering unit, wherein,
the extraction unit is used for analyzing the event-related potential components of the electroencephalogram signals and extracting initial characteristic signals of the electroencephalogram signals;
and the filtering unit is used for carrying out filtering operation on the initial characteristic signal by using an xDAWN spatial filtering algorithm to obtain a target characteristic signal.
8. The system of claim 6, further comprising: and the training module is used for training the preset classifier to obtain the trained preset classifier.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the steps of the method of any of the preceding claims 1 to 5 are implemented when the computer program is executed by the processor.
10. A computer-readable medium having non-volatile program code executable by a processor, wherein the program code causes the processor to perform the method of any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010292815.7A CN111466909A (en) | 2020-04-14 | 2020-04-14 | Target detection method and system based on electroencephalogram characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010292815.7A CN111466909A (en) | 2020-04-14 | 2020-04-14 | Target detection method and system based on electroencephalogram characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111466909A true CN111466909A (en) | 2020-07-31 |
Family
ID=71752449
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010292815.7A Pending CN111466909A (en) | 2020-04-14 | 2020-04-14 | Target detection method and system based on electroencephalogram characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111466909A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113598794A (en) * | 2021-08-12 | 2021-11-05 | 中南民族大学 | Training method and system for detection model of ice drug addict |
| CN114288520A (en) * | 2021-12-31 | 2022-04-08 | 广州酷狗计算机科技有限公司 | Sleep assisting method, device, equipment and storage medium based on brain waves |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006026548A1 (en) * | 2004-08-30 | 2006-03-09 | Southern Illinois University Carbondale | Biopotential waveform data fusion analysis and classification method |
| CN101464728A (en) * | 2009-01-05 | 2009-06-24 | 清华大学 | Human-machine interaction method with vision movement related neural signal as carrier |
| CN106527732A (en) * | 2016-11-30 | 2017-03-22 | 中国医学科学院生物医学工程研究所 | Method for selecting and optimizing feature signals in somatosensory electric stimulation brain computer interface |
| CN109255309A (en) * | 2018-08-28 | 2019-01-22 | 中国人民解放军战略支援部队信息工程大学 | Brain electricity and eye movement fusion method and device towards Remote Sensing Target detection |
| CN110781945A (en) * | 2019-10-22 | 2020-02-11 | 太原理工大学 | Electroencephalogram signal emotion recognition method and system integrating multiple features |
| CN110991406A (en) * | 2019-12-19 | 2020-04-10 | 燕山大学 | RSVP electroencephalogram characteristic-based small target detection method and system |
-
2020
- 2020-04-14 CN CN202010292815.7A patent/CN111466909A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006026548A1 (en) * | 2004-08-30 | 2006-03-09 | Southern Illinois University Carbondale | Biopotential waveform data fusion analysis and classification method |
| CN101464728A (en) * | 2009-01-05 | 2009-06-24 | 清华大学 | Human-machine interaction method with vision movement related neural signal as carrier |
| CN106527732A (en) * | 2016-11-30 | 2017-03-22 | 中国医学科学院生物医学工程研究所 | Method for selecting and optimizing feature signals in somatosensory electric stimulation brain computer interface |
| CN109255309A (en) * | 2018-08-28 | 2019-01-22 | 中国人民解放军战略支援部队信息工程大学 | Brain electricity and eye movement fusion method and device towards Remote Sensing Target detection |
| CN110781945A (en) * | 2019-10-22 | 2020-02-11 | 太原理工大学 | Electroencephalogram signal emotion recognition method and system integrating multiple features |
| CN110991406A (en) * | 2019-12-19 | 2020-04-10 | 燕山大学 | RSVP electroencephalogram characteristic-based small target detection method and system |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113598794A (en) * | 2021-08-12 | 2021-11-05 | 中南民族大学 | Training method and system for detection model of ice drug addict |
| CN114288520A (en) * | 2021-12-31 | 2022-04-08 | 广州酷狗计算机科技有限公司 | Sleep assisting method, device, equipment and storage medium based on brain waves |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Groen et al. | From image statistics to scene gist: evoked neural activity reveals transition from low-level natural image structure to scene category | |
| Luckhoo et al. | Inferring task-related networks using independent component analysis in magnetoencephalography | |
| Parra et al. | Recipes for the linear analysis of EEG | |
| CN109255309B (en) | Electroencephalogram and eye movement fusion method and device for remote sensing image target detection | |
| Dellert et al. | Dissociating the neural correlates of consciousness and task relevance in face perception using simultaneous EEG-fMRI | |
| CN111714118B (en) | Brain cognition model fusion method based on ensemble learning | |
| CN110991406A (en) | RSVP electroencephalogram characteristic-based small target detection method and system | |
| CN107480716B (en) | Method and system for identifying saccade signal by combining EOG and video | |
| CN111466909A (en) | Target detection method and system based on electroencephalogram characteristics | |
| Simbolon et al. | An experiment of lie detection based EEG-P300 classified by SVM algorithm | |
| CN111209806B (en) | Video target asynchronous detection method based on single-test electroencephalogram signal | |
| CN111568412A (en) | Method and device for reconstructing visual image by utilizing electroencephalogram signal | |
| KR101714708B1 (en) | Brain-computer interface apparatus using movement-related cortical potential and method thereof | |
| Lv et al. | Design and implementation of an eye gesture perception system based on electrooculography | |
| CN108958489A (en) | A kind of interesting image regions Rapid Detection method based on brain electricity and eye tracker | |
| CN116211305A (en) | Dynamic real-time emotion detection method and system | |
| CN111616702A (en) | Lie detection analysis system based on cognitive load enhancement | |
| Petras et al. | Information redundancy across spatial scales modulates early visual cortical processing | |
| Ouyang et al. | Handling EEG artifacts and searching individually optimal experimental parameter in real time: a system development and demonstration | |
| CN112450949A (en) | Electroencephalogram signal processing method and system for cognitive rehabilitation training | |
| Xie et al. | Detection of weak small image target based on brain-computer interface | |
| CN111671421A (en) | Electroencephalogram-based children demand sensing method | |
| CN110516711B (en) | Training set quality evaluation method of MI-BCI system and optimization method of single training sample | |
| CN112861629B (en) | Multi-window distinguishing typical pattern matching method and brain-computer interface application | |
| CN114601474A (en) | Source domain sample screening method for motor imagery transfer learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200731 |