CN109620218A - Brain wave intelligence screening method and system - Google Patents
Brain wave intelligence screening method and system Download PDFInfo
- Publication number
- CN109620218A CN109620218A CN201910087854.0A CN201910087854A CN109620218A CN 109620218 A CN109620218 A CN 109620218A CN 201910087854 A CN201910087854 A CN 201910087854A CN 109620218 A CN109620218 A CN 109620218A
- Authority
- CN
- China
- Prior art keywords
- eeg signals
- brain wave
- screening method
- intelligence screening
- characteristic parameter
- 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
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
- A61B5/7235—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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Molecular Biology (AREA)
- Psychiatry (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Psychology (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention provides a kind of brain wave intelligence screening method and system, this method comprises: obtaining the EEG signals of patient；Dissection process is carried out to EEG signals, obtains the multiple groups characteristic parameter of EEG signals；To parse multiple groups characteristic parameter obtained as input, according to multiple EEG signals samples in sample database, the anomalous event on EEG signals is marked from time domain.
Description
Technical field
The present invention relates to computer fields, and in particular to a kind of based on brain wave intelligence screening method and system.
Background technique
EEG signals (EEG) are overall reflection of the cranial nerve cell bioelectrical activity in cerebral cortex or scalp surface.Brain
A large amount of physiology and disease information are contained in electric signal, in terms of clinical medicine, EEG Processing not only can be certain brains
Disease provides diagnosis basis, but also provides effective treatment means for certain cerebral diseases.The electrical activity of brain is in computer
Wave is rendered as on display, doctor determines the working condition of brain by interpreting wave.
For electroencephalogramsignal signal analyzing mainly by visually observing, this can be regarded as artificial time domain's analysis at present.Artificial time domain's analysis
It is mainly used to directly extract wave character, such as the analysis of zero passage section, histogram analysis, variance analysis, correlation analysis, peak detection
And waveform parameter analysis, coherence average, waveform recognition etc..The detection and analysis of brain electricity is usually by veteran neurology department
Doctor or neural electrophysiology expert manually carry out.But the Chinese brain electricity technician talent is deficient at present, situation of all-level hospitals lacks excellent
The interpreting blueprints doctor of matter, especially vast basic medical unit, brain electro-detection talent's serious loss.
In addition, patient usually requires to carry out for a long time in the detection field (such as epileptic condition disease) of special disease
Brain electro-detection, possible last from days, until capturing brain paradoxical discharge.Therefore, brain electricity technician is frequently necessary to length
Brain text part up to a couple of days is exhaustively interpreted.Not only time-consuming for manual reading of drawings, with high costs and efficiency is lower, it is difficult to have
Effect is intervened or control epileptic condition, manual analysis are also easy to appear mistaken diagnosis and fail to pinpoint a disease in diagnosis.
Summary of the invention
The present invention for overcome the deficiencies in the prior art, provide it is a kind of can Automatic sieve select anomalous event in EEG signals
Brain wave intelligence screening method and system.
To achieve the goals above, the present invention provides a kind of brain wave intelligence screening method, this method comprises:
Obtain the EEG signals of patient；
Dissection process is carried out to EEG signals, obtains the multiple groups characteristic parameter of EEG signals；
To parse multiple groups characteristic parameter obtained as input, according to multiple EEG signals samples in sample database
This, marks the anomalous event on EEG signals from time domain.
An embodiment according to the present invention, carrying out dissection process to EEG signals includes:
The EEG signals of acquisition are pre-processed；
Pretreated EEG signals are divided into multiple segments with certain time length from time domain, obtain each segment
On characteristic parameter.
An embodiment according to the present invention, characteristic parameter includes time domain parameter and frequency domain parameter, when extracting frequency domain parameter
By each fragment segmentation at 2 seconds and have 1 second overlapping small fragment, extract multiple frequency domain character parameters of each small fragment.
An embodiment according to the present invention, EEG signals are multichannel brain electric signal, and pretreatment includes multichannel brain telecommunications
The removal of eye electricity artefact, step include: in number
It is standardized original EEG signals S to obtain SC；
Then " db6 " wavelet function is used to carry out seven layers of wavelet transformation, and the wavelet systems that will be obtained after decomposition to SC signal
Number is together in series, and obtains a wavelet coefficient vector matrix X；
The transposition for seeking matrix X obtains device matrix Y；
Canonical correlation analysis is carried out to vector matrix X and device matrix Y, calculates base vector matrix WxAnd Wy, acquire typical case
Canonical variable after constituent analysis identifies eye electricity artefact ingredient using related coefficient, will using canonical correlation analysis inverse transformation
Each representative vectors after removing eye electricity artefact carry out projective transformation, then carry out the inverse transformation of wavelet transformation, and it is pseudo- to obtain removal eye electricity
EEG signals after mark.
An embodiment according to the present invention, multiple EEG signals samples in sample database are constructed with multiple groups feature ginseng
Number is as input, Random Forest model of the corresponding anomalous event of every group of characteristic parameter as output.
An embodiment according to the present invention, brain wave intelligence screening method is after the EEG signals for obtaining patient by brain electricity
Under attribute directory corresponding to exclusive identification code of the signal associated storage to characterization patient identity.
An embodiment according to the present invention, brain wave intelligence screening method further include:
Request is checked based on terminal browser input, the brain electricity after obtaining label corresponding with the information checked in request
Signal data is simultaneously drawn on terminal browser according to the EEG signals data after the label and shows the brain after corresponding label
Electrical pattern.
An embodiment according to the present invention, brain wave intelligence screening method further include:
Detection modifies operation based on the label of anomalous event for the brain wave figure after the label shown；
In response to the label modification operation detected, the flag data of corresponding EEG signals is corrected.
An embodiment according to the present invention, using correct modified marked EEG signals data as EEG signals sample
This, updates sample database.
Corresponding, the present invention also provides a kind of brain wave intelligence screening systems comprising signal acquisition module, memory
And processor.The EEG signals of signal acquisition module acquisition patient.Memory is stored with computer program.Processor processing is deposited
The computer program stored in reservoir, can be realized following steps when calculation procedure is executed by processor:
Obtain the EEG signals of patient；
Dissection process is carried out to EEG signals, obtains the multiple groups characteristic parameter of EEG signals；
To parse multiple groups characteristic parameter obtained as input, according to multiple EEG signals samples in sample database
This, marks the anomalous event on EEG signals from time domain.
In conclusion brain wave intelligence screening method provided by the invention and system are right after the EEG signals for obtaining patient
EEG signals are parsed, and the multiple groups characteristic parameter in EEG signals is obtained.Multiple EEG signals in sample for reference database
Whether sample is normal to judge EEG signals corresponding to every group of characteristic parameter, when occurring abnormal then in the time domain to the exception
Event is identified.Interpreting blueprints doctor need to only pay close attention to anomalous event, substantially increase the interpreting blueprints efficiency of interpreting blueprints doctor, effectively solve
The problem of inefficiency brought by existing artificial knowledge drawing method of having determined.And a large amount of EEG signals sample in sample database
This then substantially increases the accuracy of intelligent recognition.
Further, brain wave intelligence screening method provided in this embodiment and system are by the electroencephalogram after intelligent recognition
Constantly label modification operation of the detection interpreting blueprints doctor for the electroencephalogram after identification while showing the interpreting blueprints doctor of profession
And modification operation is marked to respond to it, manual confirmation is combined on the basis of intelligently knowing figure, it is ensured that the accuracy of interpreting blueprints.This
Outside, by being included in the high marked EEG signals data of the accuracy rate after doctor's confirmation of interpreting blueprints in sample database constantly
Training identification model, to further increase the accuracy of intelligent knowledge figure.
For above and other objects of the present invention, feature and advantage can be clearer and more comprehensible, preferred embodiment is cited below particularly,
And cooperate attached drawing, it is described in detail below.
Detailed description of the invention
Fig. 1 show the flow chart of the brain wave intelligence screening method of one embodiment of the invention offer.
Fig. 2 show the specific flow chart of step S20 in Fig. 1.
Fig. 3 show the functional block diagram of the brain wave intelligence screening method of one embodiment of the invention offer.
Specific embodiment
As shown in Figure 1, brain wave intelligence screening method provided in this embodiment includes: the EEG signals (step for obtaining patient
Rapid S10).Dissection process is carried out to EEG signals, obtains the multiple groups characteristic parameter (step S20) of EEG signals.It is obtained with parsing
The multiple groups characteristic parameter obtained is as input, and according to multiple EEG signals samples in sample database, brain electricity is marked from time domain
Anomalous event (step S30) on signal.Below with reference to Fig. 1 and Fig. 2 be discussed in detail the present embodiment provides brain wave intelligently sieve
The concrete operating principle of checking method.
Brain wave intelligence screening method provided in this embodiment starts from step S10, in this step brain wave acquisition equipment (packet
Include electroencephalograph, lead sleep detection instrument (PSG) more etc.) acquire patient's with frequency for the sample frequency of 500Hz~4000Hz
Physiological signal obtains EEG signals therein.What brain wave acquisition equipment acquired in the present embodiment is frontal region, central area and occipital region three
The EEG signals in a channel.However, the present invention does not do any restriction to this.In other embodiments, the EEG signals of acquisition
It can be single pass EEG signals.
Electroencephalogramsignal signal collection equipment selects brain telecommunications according to the current network state of equipment after getting EEG signals
Number storage mode.When equipment is under network environment, the EEG signals that electroencephalogramsignal signal collection equipment will acquire are transmitted to
Cloud server；When under no network environment, electroencephalogramsignal signal collection equipment stores collected EEG signals to local
On server, EEG signals are transmitted on cloud server again when equipment is under network environment.The network environment
It can be wifi network or hot spot.After cloud server receives EEG signals, by the EEG signals associated storage to characterize suffer from
Under attribute directory corresponding to the exclusive identification code of person's identity, the centralized management of EEG signals is realized.Characterize patient identity only
One identifier can be medical number etc. of the cell-phone number of patient, identification card number or hospital.
Step S20 is executed after getting EEG signals, and dissection process is carried out to the EEG signals of acquisition, obtains brain telecommunications
Number multiple groups characteristic parameter.By the collected EEG signals meeting of cerebral cortex, there are many noises, also have physiology artefact signal,
These interference signal amplitudes are larger, and useful signal amplitude is less.Therefore the present embodiment includes: step to the dissection process of EEG signals
S201 pre-processes the EEG signals of acquisition, removes the physiology artefact signal in EEG signals；Step S202, from time domain
On pretreated EEG signals are divided into multiple segments with certain time length, obtain the characteristic parameter in each segment.
Eye electricity artefact is removed in step S201, and specific step is as follows: original EEG signals S is standardized
To SC.Then " db6 " wavelet function is used to carry out seven layers of wavelet transformation, and the wavelet coefficient string that will be obtained after decomposition to SC signal
Connection gets up, and obtains a wavelet coefficient vector matrix X；The transposition for seeking matrix X obtains device matrix Y.To vector matrix X and dress
It sets matrix Y and carries out canonical correlation analysis, calculate base vector matrix WxAnd Wy, acquire the canonical variable after typical composition is analyzed, benefit
Identify eye electricity artefact ingredient with related coefficient, after eye electricity artefact will be removed using canonical correlation analysis inverse transformation it is each it is typical to
Amount carries out projective transformation, then carries out the inverse transformation of wavelet transformation, the EEG signals after obtaining removal eye electricity artefact.Further to subtract
The interference of few garbage signal, in this present embodiment, the pretreatment in step S201 further include filtering out 50Hz using bandpass filter
More than, 0.5Hz frequency content below.However, the present invention is not limited in any way this.
Step S202 is executed after obtaining pretreated EEG signals, divides pretreated EEG signals from time domain
At it is multiple when a length of 30 seconds segments, the extraction of characteristic parameter is carried out for each segment.However, the present invention is to each segment
Length is not limited in any way.In this present embodiment, the characteristic parameter of extraction include time domain charactreristic parameter, frequency domain character parameter and
Nonlinear characteristic parameters will be described in detail the extraction of three kinds of characteristic parameters below.Time domain charactreristic parameter includes the peak of EEG signals
Value, variance and Hjorth parameter.Wherein variance indicates the variation range of different sleep stage EEG signals.Assuming that one
Fragment signal is X (i), i=1,2 ..., N, and N is the length of segment, then its variance V are as follows:
Wherein,It is the mean value of signal X (i).
For for Hjorth parameter, Hjorth parameter includes Hjorth mobility and Hjorth complexity:
1. Hjorth mobility HmAre as follows:
2. Hjorth complexity HcFor
Wherein,difi=X (i)-X (i-
1), X (i) is fragment signal, and i=1,2 ..., N, N are the length of segment.
When carrying out frequency domain character extraction, because each rhythm and pace of moving things wave is no longer than 2 seconds, in order to which the feature of extraction is more complete
Face includes sleep info, therefore each fragment segmentation at 2 seconds and is had 1 second small fragment being overlapped, and extracts the more of each small fragment
A frequency domain character parameter, frequency domain character parameter include: the frequency band energy E (kc) of K complex wave in each 2 seconds small fragments, energy
Than ratio (kc)；The frequency band energy E (δ) of δ wave, energy ratio ratio (δ)；The frequency band energy E (θ) of θ wave, energy ratio ratio
(θ)；The frequency band energy E (α) of α wave；Energy ratio ratio (α)；The frequency band energy E (β) of β wave, energy ratio ratio (β), and extract
The statistical natures such as minimum value, maximum value, average value, the variance of each frequency band energy.
Each its frequency domain character of 2 seconds small fragments is extracted as follows:
Firstly, carrying out M layers of decomposition to EEG signals X (i) using Mallat algorithm, corresponding wavelet coefficient is as follows:
Wherein, Aj,kAnd Dj,k(j=1,2 ..., M) is respectively the approximation coefficient and detail coefficients of j scale space, h0, h1Point
Not Wei low frequency and high-frequency decomposition filter, m-2k indicate scale displacement, Z indicate integer set.
In this present embodiment, it carries out seven layers to EEG signals X (i) using " db6 " wavelet function to decompose, i.e. M=7, wherein
A1Represent K complex wave, A1+D1Represent δ wave, D2Represent θ wave, D3Represent α wave, D4Represent β wave.
The ENERGY E (δ) of δ wave is obtained by formula 5:
Wherein, A1(i) the K complex wave after small echo signal decomposition in i-th layer, D are indicated1(i) it indicates through small echo signal decomposition
δ wave in i-th layer afterwards.
The ENERGY E (θ) of θ wave is obtained by formula 6:
Wherein, D2(i) the θ wave after small echo signal decomposition in i-th layer is indicated.
The ENERGY E (α) of α wave is obtained by formula 7:
Wherein, D3(i) the α wave after small echo signal decomposition in i-th layer is indicated.
The ENERGY E (β) of β wave is obtained by formula 8:
Wherein, D3(i) the β wave after small echo signal decomposition in i-th layer is indicated.
The ENERGY E (kc) of K complex wave is obtained by formula 9:
Gross energy and Es=E (δ)+E (θ)+E (α)+E (β)+E (kc).
And then the energy ratio of each rhythm and pace of moving things wave, ratio (kc)=E (kc)/E can be calculateds, ratio (δ)=E (δ)/Es,
Ratio (θ)=E (θ)/Es, ratio (α)=E (α)/Es, ratio (β)=E (β)/Es。
Nonlinear characteristic parameters include approximate entropy, renyi ' s entropy and correlation dimension.
1. approximate entropy algorithm is as follows:
For original input signal X (i)=[x1,x2,…,xN] the new subsequence of construction, X (i, l)=[xi,xi+1,…,
xi+l-1], 1≤i≤N-l, wherein l is the length of subsequence, takes 1,2 or 3；
R is defined as signal noise grade, r=kSD, and wherein SD is the standard deviation of signal X (i), k=0, and 0.1,
0.2,…,0.9；
Building space submatrix X (j, l)=X (j, l) | j ∈ [1,2 ..., N-l] }, element each in matrix is calculated:
C (i, l) represents the ratio of number and sum N-l in matrix X (j, l) less than r, and calculation formula is as follows:
Therefore approximate entropy can be calculated by formula 12:
2. Renyi ' s entropy is defined as:
Wherein, q is the flexible strategy of Renyi ' s entropy, when q level off to 1 when, HqConverge to Shannon entropy, when q level off to 0 when, HqIt receives
Minimum entropy is held back,X (i) is fragment signal, and i=1,2 ..., N, N are the length of the segment.
3. correlation dimension algorithm:
Correlation dimension illustrates the complexity of system, and correlation dimension is higher, shows that system is more complicated, correlation integral is by public affairs
Formula 14 indicates:
WhereinFor Heaviside function.
There are following relationships between correlation integral C (r) and scale r:
Wherein, D indicates desired correlation dimension.
By the available correlation dimension D of formula 15:
It extracted through above-mentioned time domain charactreristic parameter, obtain each after frequency domain character parameter extraction and nonlinear characteristic parameters
One group of characteristic parameter corresponding to section, executes step S30 later, is input with multiple groups characteristic parameter, according in sample database
Multiple EEG signals samples, from time domain mark EEG signals on anomalous event.In this present embodiment, in sample database
Multiple EEG signals samples construct using multiple groups characteristic parameter as input, the corresponding anomalous event conduct of every group of characteristic parameter
The Random Forest model of output.Therefore input step S20 multiple groups characteristic parameter obtained in Random Forest model, after training
Random Forest model is by successively label has the segment of anomalous event from time domain.In this present embodiment, it is obtained using step S20
The multiple groups characteristic parameter of each EEG signals sample in database is sampled, trains random forest point with these characteristic parameter groups
Multiple decision trees in class device form Random Forest model.However, the present invention is not limited in any way this.In other embodiments
In, other models of mind can also be constructed in sample database to realize the label of anomalous event.
In this present embodiment, brain wave intelligence screening method further includes step S40, based on checking for terminal browser input
Request obtains the EEG signals data after checking the corresponding label of information in requesting with this and according to the brain telecommunications after the label
Number draws on terminal browser and shows the brain wave figure after corresponding label.Specifically, when interpreting blueprints doctor is logical
Crossing browser input includes when the checking request of exclusive identification code for characterizing patient identity, and system is according to unique identification code
Obtaining step S30 mark after EEG signals data and be transmitted to terminal browser after carrying out coding compression to it.Terminal is clear
Device of looking at is decoded after receiving data, and the eeg data after reduction label is simultaneously drawn on terminal browser according to the eeg data
It makes and shows the brain wave figure after label.Brain wave figure is shown in the form of catalogue for the anomalous event of label
Side, interpreting blueprints doctor need to only click a certain anomalous event in catalogue, and brain wave figure will the automatic Display anomalous event institute
The waveform position at place.Or in other embodiments, when carrying out the displaying of waveform, since the waveform at initial anomalous event
It shows, then jumps to waveform position locating for subsequent anomalous event one by one.
For the accuracy for further increasing anomalous event label, in this present embodiment, brain wave intelligence screening method is also wrapped
It includes: label modification operation (step S50) of the detection for the brain wave figure after the label shown based on anomalous event；In response to
The label modification operation detected, corrects the flag data (step S60) of corresponding EEG signals.Mark of the interpreting blueprints doctor to displaying
Brain wave figure after note carries out confirmation feedback, is corrected manually if the place for marked erroneous occur.Based on what is detected
Correct information, the flag state in eeg data do corresponding adjustment.For example, normal condition is corrected as from abnormality.Due to
The accuracy of the smart tags of step S30 depends on Random Forest model, and the number of random forest deep learning is more, brain telecommunications
The label of number sample is more accurate, then recognition accuracy of its label will be higher.For the accuracy for improving smart tags identification, Yu Ben
In embodiment, the correct modified marked EEG signals data of doctor that will be interpreted blueprints in step S60 are as EEG signals sample
This, updates sample database (step S70).
Corresponding with above-mentioned brain wave intelligence screening method, the present embodiment also provides a kind of brain wave intelligence screening system
System, which includes signal acquisition module 1, memory 2, processor 3, brain wave drafting module 4, display module 5, detection module
6, update module 7 and cloud server 8.The EEG signals of the acquisition patient of signal acquisition module 1.Memory 2 is stored with calculating
Machine program.Processor 3 handles the computer program stored in memory 2 to realize the present embodiment step S10 to step S70.
In this present embodiment, EEG signals acquired in signal acquisition module 1 are uploaded to cloud service in a network environment
Device 8 is simultaneously managed concentratedly according to the unique identifier of characterization patient identity.Processor 3 obtains the brain electricity in cloud server 8
Signal simultaneously calls the computer program stored in memory 2 to realize the dissection process of EEG signals and from time domain to parsing place
EEG signals after reason carry out anomalous event label.Eeg data after label is stored in the exclusive identification code of characterization patient identity
Under corresponding attribute directory.When processor 3 receive from terminal browser after checking request, according to identification code calling deposit
Storage is to the EEG signals data after the label under respective attributes catalogue and will be transmitted to terminal browser after data encoding compression
10.Brain wave drafting module 4 on terminal browser 10 decompresses the EEG signals data after marking and draws corresponding brain
Electrical pattern.Display module 5 shows brain wave figure and anomalous event catalogue on terminal browser.Detection module 6 will be real-time
Detection modifies operation based on the label of anomalous event for the brain wave figure after the label shown, and the label that will test is repaired
Change operational feedback to processor 3, processor 3 is according to the reference numerals for correcting EEG signals based on the state of anomalous event of feedback
According to, while the EEG signals data after corrigendum label are transmitted to cloud server 8 and are stored.Update module 7 marks corrigendum
EEG signals data after note are included in sample database, the learning sample as Random Forest model.
In conclusion brain wave intelligence screening method provided by the invention and system are right after the EEG signals for obtaining patient
EEG signals are parsed, and the multiple groups characteristic parameter in EEG signals is obtained.Multiple EEG signals in sample for reference database
Whether sample is normal to judge EEG signals corresponding to every group of characteristic parameter, when occurring abnormal then in the time domain to the exception
Event is identified.Interpreting blueprints doctor need to only pay close attention to anomalous event, substantially increase the interpreting blueprints efficiency of interpreting blueprints doctor.Effectively solution
The problem of inefficiency brought by existing artificial knowledge drawing method of having determined.And a large amount of EEG signals sample in sample database
This then substantially increases the accuracy of intelligent recognition.
Further, brain wave intelligence screening method provided in this embodiment and system are by the electroencephalogram after intelligent recognition
Constantly label modification operation of the detection interpreting blueprints doctor for the electroencephalogram after identification while showing the interpreting blueprints doctor of profession
And modification operation is marked to respond to it, manual confirmation is combined on the basis of intelligently knowing figure, it is ensured that the accuracy of interpreting blueprints.This
Outside, by being included in the high marked EEG signals data of the accuracy rate after doctor's confirmation of interpreting blueprints in sample database constantly
Training identification model, to further increase the accuracy of intelligent knowledge figure.
Although the present invention is disclosed above by preferred embodiment, however, it is not intended to limit the invention, this any known skill
Skill person can make some changes and embellishment without departing from the spirit and scope of the present invention, therefore protection scope of the present invention is worked as
Subject to claims range claimed.
Claims (10)
1. a kind of brain wave intelligence screening method characterized by comprising
Obtain the EEG signals of patient；
Dissection process is carried out to EEG signals, obtains the multiple groups characteristic parameter of EEG signals；
To parse multiple groups characteristic parameter obtained as input, according to multiple EEG signals samples in sample database, from
The anomalous event on EEG signals is marked in time domain.
2. brain wave intelligence screening method according to claim 1, which is characterized in that carry out dissection process to EEG signals
Include:
The EEG signals of acquisition are pre-processed；
Pretreated EEG signals are divided into multiple segments with certain time length from time domain, are obtained in each segment
Characteristic parameter.
3. brain wave intelligence screening method according to claim 2, which is characterized in that the characteristic parameter includes time domain ginseng
Each fragment segmentation at 2 seconds and is had the small fragment of overlapping in 1 second when extracting frequency domain parameter, extracted each by several and frequency domain parameter
Multiple frequency domain character parameters of small fragment.
4. brain wave intelligence screening method according to claim 2, which is characterized in that the EEG signals are multichannel brain
Electric signal, the pretreatment include the removal of eye electricity artefact in multichannel brain electric signal, and step includes:
It is standardized original EEG signals S to obtain SC；
Then " db6 " wavelet function is used to carry out seven layers of wavelet transformation, and the wavelet coefficient string that will be obtained after decomposition to SC signal
Connection gets up, and obtains a wavelet coefficient vector matrix X；
The transposition for seeking matrix X obtains device matrix Y；
Canonical correlation analysis is carried out to vector matrix X and device matrix Y, calculates base vector matrix WxAnd Wy, acquire typical composition
Canonical variable after analysis identifies eye electricity artefact ingredient using related coefficient, will be removed using canonical correlation analysis inverse transformation
Each representative vectors after the electric artefact of eye carry out projective transformation, then carry out the inverse transformation of wavelet transformation, after obtaining removal eye electricity artefact
EEG signals.
5. brain wave intelligence screening method according to claim 1, which is characterized in that multiple brains electricity in sample database
Sample of signal is constructed using multiple groups characteristic parameter as input, and the corresponding anomalous event of every group of characteristic parameter is random as output
Forest model.
6. brain wave intelligence screening method according to claim 1, which is characterized in that the brain wave intelligence screening method
Corresponding to exclusive identification code after the EEG signals for obtaining patient by the EEG signals associated storage to characterization patient identity
Under attribute directory.
7. brain wave intelligence screening method according to claim 1, which is characterized in that the brain wave intelligence screening method
Further include:
Request is checked based on terminal browser input, the EEG signals after obtaining label corresponding with the information checked in request
Data simultaneously draw on terminal browser according to the EEG signals data after the label and show the brain wave after corresponding label
Figure.
8. brain wave intelligence screening method according to claim 7, which is characterized in that the brain wave intelligence screening method
Further include:
Detection modifies operation based on the label of anomalous event for the brain wave figure after the label shown；
In response to the label modification operation detected, the flag data of corresponding EEG signals is corrected.
9. brain wave intelligence screening method according to claim 8, which is characterized in that will be correct modified marked
EEG signals data update sample database as EEG signals sample.
10. a kind of brain wave intelligence screening system characterized by comprising
Signal acquisition module obtains the EEG signals of patient；
Memory is stored with computer program；
Processor handles the computer program stored in the memory, can be real when the calculation procedure is executed by processor
Existing following steps:
Obtain the EEG signals of patient；
Dissection process is carried out to EEG signals, obtains the multiple groups characteristic parameter of EEG signals；
To parse multiple groups characteristic parameter obtained as input, according to multiple EEG signals samples in sample database, from
The anomalous event on EEG signals is marked in time domain.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910087854.0A CN109620218A (en) | 2019-01-29 | 2019-01-29 | Brain wave intelligence screening method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910087854.0A CN109620218A (en) | 2019-01-29 | 2019-01-29 | Brain wave intelligence screening method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109620218A true CN109620218A (en) | 2019-04-16 |
Family
ID=66064192
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910087854.0A Pending CN109620218A (en) | 2019-01-29 | 2019-01-29 | Brain wave intelligence screening method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109620218A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111214226A (en) * | 2020-01-21 | 2020-06-02 | 苏州小蓝医疗科技有限公司 | Electroencephalogram feature extraction and selection method |
| CN111508578A (en) * | 2020-05-19 | 2020-08-07 | 中国电子科技集团公司第三十八研究所 | Brain wave checking device and method based on artificial intelligence |
| CN111543984A (en) * | 2020-04-13 | 2020-08-18 | 重庆邮电大学 | Method for removing ocular artifacts of electroencephalogram signals based on SSDA (steady state data acquisition) |
| CN112089414A (en) * | 2019-06-17 | 2020-12-18 | 阿里健康信息技术有限公司 | Method and device for detecting abnormal discharge of brain |
| CN113712571A (en) * | 2021-06-18 | 2021-11-30 | 陕西师范大学 | Abnormal electroencephalogram signal detection method based on Rinyi phase transfer entropy and lightweight convolutional neural network |
| CN116602691A (en) * | 2023-07-14 | 2023-08-18 | 北京元纽科技有限公司 | Denoising method and device for electroencephalogram signals, electronic equipment and storage medium |
| CN117481667A (en) * | 2023-10-24 | 2024-02-02 | 沈阳工业大学 | Electroencephalogram signal acquisition system |
| CN117547286A (en) * | 2023-12-29 | 2024-02-13 | 中国人民解放军东部战区总医院 | Electroencephalogram signal data analysis management system based on intelligent repair material |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102743157A (en) * | 2012-07-14 | 2012-10-24 | 深圳市邦健电子有限公司 | Monitor and method for displaying long time physiological parameter data |
| US20170202518A1 (en) * | 2016-01-14 | 2017-07-20 | Technion Research And Development Foundation Ltd. | System and method for brain state classification |
| CN107260166A (en) * | 2017-05-26 | 2017-10-20 | 昆明理工大学 | A kind of electric artefact elimination method of practical online brain |
| CN109247936A (en) * | 2018-10-31 | 2019-01-22 | 山东大学 | A kind of abnormal brain electricity behavior monitoring system and method for full night sleep monitor |
-
2019
- 2019-01-29 CN CN201910087854.0A patent/CN109620218A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102743157A (en) * | 2012-07-14 | 2012-10-24 | 深圳市邦健电子有限公司 | Monitor and method for displaying long time physiological parameter data |
| US20170202518A1 (en) * | 2016-01-14 | 2017-07-20 | Technion Research And Development Foundation Ltd. | System and method for brain state classification |
| CN107260166A (en) * | 2017-05-26 | 2017-10-20 | 昆明理工大学 | A kind of electric artefact elimination method of practical online brain |
| CN109247936A (en) * | 2018-10-31 | 2019-01-22 | 山东大学 | A kind of abnormal brain electricity behavior monitoring system and method for full night sleep monitor |
Non-Patent Citations (6)
| Title |
|---|
| HEALY, P.D.O"REILLY, R.D.BOYLAN, G.B.MORRISON, J.P.: "Web-based remote monitoring of live EEG", 《2010 12TH IEEE INTERNATIONAL CONFERENCE ON E-HEALTH NETWORKING, APPLICATIONS AND SERVICES》 * |
| RAMACHANDRAN 等: "Design and Implementation of an Open-Source Browser-based Laboratory Platform for EEG Data Analysis", 《INTERNATIONAL CONFERENCE ON ADVANCES IN COMPUTING, COMMUNICATIONS AND INFORMATICS》 * |
| 刘红 等: "《心血管内科临床诊断与综合治疗》", 30 April 2016, 长春：吉林科学技术出版社 * |
| 彭芳: "《生理学实验指导》", 31 March 2013, 西安：第四军医大学出版社 * |
| 李明爱 等: "眼电伪迹自动去除方法的研究与分析", 《仪器仪表学报》 * |
| 郑逢斌: "《计算机科学导论》", 31 August 2016 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112089414A (en) * | 2019-06-17 | 2020-12-18 | 阿里健康信息技术有限公司 | Method and device for detecting abnormal discharge of brain |
| CN111214226A (en) * | 2020-01-21 | 2020-06-02 | 苏州小蓝医疗科技有限公司 | Electroencephalogram feature extraction and selection method |
| CN111543984A (en) * | 2020-04-13 | 2020-08-18 | 重庆邮电大学 | Method for removing ocular artifacts of electroencephalogram signals based on SSDA (steady state data acquisition) |
| CN111508578A (en) * | 2020-05-19 | 2020-08-07 | 中国电子科技集团公司第三十八研究所 | Brain wave checking device and method based on artificial intelligence |
| CN113712571A (en) * | 2021-06-18 | 2021-11-30 | 陕西师范大学 | Abnormal electroencephalogram signal detection method based on Rinyi phase transfer entropy and lightweight convolutional neural network |
| CN116602691A (en) * | 2023-07-14 | 2023-08-18 | 北京元纽科技有限公司 | Denoising method and device for electroencephalogram signals, electronic equipment and storage medium |
| CN116602691B (en) * | 2023-07-14 | 2023-10-10 | 北京元纽科技有限公司 | Denoising method and device for electroencephalogram signals, electronic equipment and storage medium |
| CN117481667A (en) * | 2023-10-24 | 2024-02-02 | 沈阳工业大学 | Electroencephalogram signal acquisition system |
| CN117547286A (en) * | 2023-12-29 | 2024-02-13 | 中国人民解放军东部战区总医院 | Electroencephalogram signal data analysis management system based on intelligent repair material |
| CN117547286B (en) * | 2023-12-29 | 2024-05-28 | 中国人民解放军东部战区总医院 | Electroencephalogram signal data analysis management system based on intelligent repair material |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109620218A (en) | Brain wave intelligence screening method and system | |
| Fatemian et al. | A new ECG feature extractor for biometric recognition | |
| CN109602417A (en) | Sleep stage method and system based on random forest | |
| EP2428160A2 (en) | Monitoring physiological activity using partial state space reconstruction | |
| CN110811609A (en) | Intelligent epileptic spike detection method based on fusion of adaptive template matching and machine learning algorithm | |
| CN103610456B (en) | Brain pressure non-invasive monitoring analysis system based on physiological signal characteristic parameters | |
| CN105595990A (en) | Intelligent terminal device for evaluating and distinguishing quality of electrocardiosignal | |
| DE102008002933A1 (en) | Data recording for patient status analysis | |
| CN109497997A (en) | Based on majority according to the seizure detection and early warning system of acquisition | |
| CN103767702A (en) | Electrocerebral non-linear dual-measure feature extraction and fusion processing method | |
| CN103637795B (en) | Automatic diagnosis function detection method for electrocardiogram instrument | |
| CN104173046A (en) | Method for extracting color marked amplitude-integrated electroencephalogram | |
| DE102015218948A1 (en) | Apparatus and method for determining a medical health parameter of a subject by means of voice analysis | |
| CN110234374B (en) | Anesthetic drug identification method, and method and device for processing anesthetic electroencephalogram signals | |
| Queiroz et al. | Support system for classification of beat-to-beat arrhythmia based on variability and morphology of electrocardiogram | |
| DE102016204397A1 (en) | CLASSIFICATION OF A TIME SERIES SIGNAL AS VENTRICULAR EXTRASYSTOLE | |
| CN107837083B (en) | J wave automatic testing method based on least square method supporting vector machine | |
| CN108968950A (en) | A kind of digitlization electrocardiogram acquisition method, apparatus and system | |
| CN116092641B (en) | Method, system and device for dynamically adjusting sleep sensory stress level | |
| CN117281479A (en) | Human lower limb chronic pain distinguishing method, storage medium and device based on surface electromyographic signal multi-dimensional feature fusion | |
| CN109998530A (en) | Portable brain pyroelectric monitor system based on VR glasses | |
| Sun et al. | A practical cross-domain ecg biometric identification method | |
| Yuan et al. | Diagnosing cardiac abnormalities from 12-lead electrocardiograms using enhanced deep convolutional neural networks | |
| DE102016204398A1 (en) | CLASSIFICATION OF A TIME SERIES SIGNAL AS VENTRICULAR EXTRASYSTOLE AND VENTRICULAR TACHYCARDIA | |
| CN116369853A (en) | Olfactory function standardized evaluation device and method based on brain-computer interaction technology |
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 |