CN103976740B - A kind of EEG signals identification system of network-oriented environment and recognition methods - Google Patents
A kind of EEG signals identification system of network-oriented environment and recognition methods Download PDFInfo
- Publication number
- CN103976740B CN103976740B CN201410220371.0A CN201410220371A CN103976740B CN 103976740 B CN103976740 B CN 103976740B CN 201410220371 A CN201410220371 A CN 201410220371A CN 103976740 B CN103976740 B CN 103976740B
- Authority
- CN
- China
- Prior art keywords
- eeg
- module
- data
- eeg data
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 210000004556 brain Anatomy 0.000 claims abstract description 166
- 238000012549 training Methods 0.000 claims abstract description 97
- 239000008280 blood Substances 0.000 claims abstract description 44
- 210000004369 blood Anatomy 0.000 claims abstract description 44
- 238000012546 transfer Methods 0.000 claims abstract description 35
- 238000012545 processing Methods 0.000 claims abstract description 21
- 238000012706 support-vector machine Methods 0.000 claims abstract description 19
- 230000005540 biological transmission Effects 0.000 claims abstract description 15
- 238000005259 measurement Methods 0.000 claims abstract description 14
- 230000005611 electricity Effects 0.000 claims description 40
- 238000012216 screening Methods 0.000 claims description 22
- 230000011218 segmentation Effects 0.000 claims description 16
- 230000003321 amplification Effects 0.000 claims description 15
- 238000001914 filtration Methods 0.000 claims description 15
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 15
- 238000006243 chemical reaction Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 12
- 238000010276 construction Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 7
- 239000012141 concentrate Substances 0.000 claims description 5
- 238000013467 fragmentation Methods 0.000 claims description 3
- 238000006062 fragmentation reaction Methods 0.000 claims description 3
- 230000008447 perception Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 19
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000007635 classification algorithm Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000001027 hydrothermal synthesis Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 208000012880 Finger injury Diseases 0.000 description 1
- 208000003443 Unconsciousness Diseases 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Landscapes
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention relates to a kind of EEG signals identification system and recognition methods of network-oriented environment, recognition system comprises some wearable EEG measuring devices, more than one mobile network's terminal and an eeg data blood processor; The EEG Processing measured is become network eeg data and is processed to eeg data blood processor by mobile network's terminal transmission by EEG measuring device, and eeg data blood processor exports recognition result and transfers to mobile network's terminal and shows.Recognition methods comprises: adopt the EEG signals of the tested individuality of EEG measuring measurement device and be processed into network eeg data, wireless network card through mobile network's terminal transfers to eeg data blood processor, eeg data blood processor carries out segment processing to EEG signals and extracts brain electrical feature information, calculate based on prestige and brain electrical feature vector training sample is screened and determines input amendment and the degree of membership of fuzzy support vector machine, complete the identification to tested individuality and Output rusults.
Description
Technical field
The present invention relates to a kind of identification system and recognition methods, particularly about a kind of EEG signals identification system and recognition methods of network-oriented environment.
Background technology
Informationization and the networking process of society propose new requirement to identity recognizing technology.The a lot of attacks for network identity data base occurred in recent years, cause serious impact to the privacy of user, the rights and interests of enterprise and even the stable of society.Therefore, new identity recognizing technology not only will meet safe and reliable basic demand, more needs to adapt to network environment.Identity recognizing technology common at present can be divided three classes: based on the identity recognizing technology of specific having and the identity recognizing technology based on the biological character such as fingerprint, iris such as identity recognizing technology, identity-based card, passport of the specific knowledge such as password, password.The above two are comparatively common, and technology is relative maturity also, but are more easily usurped by the illegal steal information in the external world, malice.Identity recognizing technology based on biological character becomes study hotspot in recent years with its higher specificity and safety, and such as fingerprint identification technology is widely used in the measurement of driving school in all parts of the country training time, appearance recognition technology identifies for booting computer user.Biological character recognition technology have not easily stolen, availability by force, the feature such as safe ready more.In theory, any physiology of people or behavior characteristics all can be used for identification on the basis met the following conditions: 1) availability, are easy to gather and measure; 2) universality, each possess per capita; 3) uniqueness, everyone feature is different; 4) stability, not in time place change and change.Although by using novel electricity, optics, acoustics biosensor in conjunction with the means such as computer science, statistics, dependable with function based on the identity recognizing technology of biological character progressively improves, but still there is insurmountable problem based on the identity recognizing technology of the traditional biological such as fingerprint, appearance character, such as: appearance identification cannot be distinguished the similar twins of appearance and be subject to the impacts such as ambient lighting; Fingerprint recognition is subject to the impact of finger injuries, and sample fingerprint is easily maliciously stolen; Voice recognition can copy waveform etc. by Technologies of Handling Voice in Computer.
The identification of brain electricity (Electroencephalograph, EEG) signal is a kind of identity recognizing technology based on biological character newly.The identification of brain electricity distinguishes Different Individual by the characteristic information of measurement, extraction, comparison Different Individual EEG signals.Compared with the identity recognizing technology based on other biological character, EEG signals identification, due to the specificity of the cognitive signal of human thinking and disguise, is more difficultly cracked and usurps, and has started to attempt application at safety-security area.The system description at first such as Poulos EEG signals identification mechanism.Huang etc. use 64 passage medical science EEG measuring equipment and complicated EEG measuring task to reach the recognition accuracy of 100%, demonstrate the feasibility in theory of EEG signals identification.Matthias etc. cut down brain wave acquisition port number, use homemade simple measurement device to complete authentication and identifying.The task to EEG measuring such as Marcel simplifies.Along with the development of wearable survey calculation technology, take health as physical support, be worn on tested individuality EEG measuring with it by supporters such as headband, the helmet, earphones and become study hotspot.Corey etc. and Chuang etc. start the identity identifying and authenticating attempting using the wearable EEG measuring helmet realization of consumer level radio communication for safety-security area.Actual application background and technology trends require that the identification of brain electricity does not re-use traditional large-scale wet electrode EEG measuring equipment, but while realizing wearable EEG measuring system structure, simplify EEG measuring task, and adopt highly reliable feature extraction, algorithm for pattern recognition guarantee accuracy of identification.Complex forms, the length consuming time of the EEG measuring task adopted in existing wearable brain electricity identification research, the following demand that network environment hypencephalon electricity identifies cannot be met: 1) reliability: reliability is the basis of identification mechanism, comprise the robustness of front end measuring system accuracy of identification, wireless network transmissions safety and Back end data process and Long-distance Control; 2) real-time: needing to adopt when guaranteeing precision the methods such as sparseness measuring to cut down data volume, shortening EEG measuring task consuming time; 3) wearable: brain electricity identification procedure needs to design hardware structure that is wearable, portable, radio communication, to adapt to heterogeneous networks measurement environment, optimize the practicality of brain electricity identification scheme; 4) sparse and low expense: the authentication of network brain electricity needs in conjunction with network environment, reduces unnecessary measurement links and hardware device, reduces the expense of time, energy and cost of manufacture.
For meeting the demand of above network brain electricity authentication, sparse wearable EEG measuring software and hardware system need be redesigned, simplify EEG measuring task, more needing the existing feature extraction mode recognition methods of supporting improvement.Support vector machine (SupportVectorMachine, SVM) is a kind of machine learning method of Corpus--based Method theory of learning, is suitable for for small sample, non-linear and high dimensional feature vector field homoemorphism formula identification.The EEG signals sample size that wearable EEG measuring equipment obtains is limited, and the characteristic vector dimension extracted is high, so mainly take support vector cassification algorithm to complete brain electricity identification task at present.But, sparse eeg data measurement data amount is few, very easily by the data transmission environments of the features such as outside noise affects, unstable networks and support vector machine itself to the sensitivity of input training sample noise, make traditional support vector machine method also cannot adapt to the needs of network environment hypencephalon electricity data category.
Summary of the invention
For the problems referred to above, the object of this invention is to provide a kind of EEG signals identification system and recognition methods of reliable, the real-time network-oriented environment based on wearable EEG measuring device.
For achieving the above object, the present invention takes following technical scheme: a kind of EEG signals identification system of network-oriented environment, is characterized in that: it comprises some wearable EEG measuring devices, more than one mobile network's terminal and an eeg data blood processor; Described wearable EEG measuring device transfers to described mobile network's terminal after the EEG Processing of the tested individuality measured is become eeg data, the eeg data that described mobile network's end-on receives stores, and the eeg data of storage is transferred to described eeg data blood processor, described eeg data blood processor exports recognition result after processing the network eeg data received, and recognition result is transferred to described mobile network's terminal and show.
Described wearable EEG measuring device comprises the dry electrode of some EEG measuring, an amplification filtering module, an analog-to-digital conversion module and a bluetooth serial ports module; The brain electric analoging signal measured is transferred to described amplification filtering module by wire by the dry electrode of some described EEG measuring, brain electric analoging signal after amplification filtering transfers to described analog-to-digital conversion module, and being converted to brain waiting for transmission electricity digital signal, brain electricity digital signal is through described bluetooth serial ports module transfer extremely described mobile network's terminal.
Described mobile network's terminal comprises Bluetooth receptions module, data memory module, wireless network card and display module; The brain received electricity digital data transmission stores to described data memory module by described Bluetooth receptions module, the eeg data that described data memory module stores transfers to described eeg data blood processor by described wireless network card, recognition result is transferred to display module by described wireless network card and shows after processing the eeg data received by described eeg data blood processor.
Described mobile network's terminal adopts the one in mobile phone, panel computer and notebook computer.
Described eeg data blood processor comprises High_speed NIC, data sectional and characteristic extracting module, prestige computing module and identification module, the eeg data that described mobile network's terminal sends transfers to described data sectional and characteristic extracting module by described High_speed NIC, described data sectional and characteristic extracting module carry out segment processing to the eeg data received and after extracting brain electrical feature information, obtain brain electrical feature vector training sample and transfer to described prestige computing module, described prestige computing module carries out credit value assessment to the brain electrical feature sample received, filter out reliable brain electrical feature sample, the reliable brain electrical feature sample filtered out and credit value thereof are transferred to described identification module by described prestige computing module, the identity of described identification module to tested individuality identifies and exports recognition result, recognition result transfers to described mobile network's terminal by described High_speed NIC and shows.
Adopt an EEG signals personal identification method for the network-oriented environment of identification system, it comprises the following steps: 1) arrange the EEG signals identification system that comprises the network-oriented environment of some wearable EEG measuring devices, more than one mobile network's terminal and an eeg data blood processor; Wherein, wearable EEG measuring device comprises the dry electrode of EEG measuring, amplification filtering module, analog-to-digital conversion module and bluetooth serial ports module; Mobile network's terminal comprises Bluetooth receptions module, data memory module, wireless network card and display module; Eeg data blood processor comprises High_speed NIC, data sectional and characteristic extracting module, prestige computing module and identification module; 2) EEG signals of the tested individuality of the dry electrode pair of EEG measuring is used to carry out perception measurement, the brain electric analoging signal measured exports brain electricity digital signal after amplification filtering module and analog-to-digital conversion module process, and brain electricity digital signal passes through bluetooth serial ports module transfer to mobile network's terminal; 3) in mobile network's terminal, Bluetooth receptions module by receive brain electricity digital data transmission to data memory module, the eeg data of storage is transferred to eeg data blood processor by wireless network card by data memory module; 4) eeg data blood processor processes the EEG signals received, complete the identification of tested individuality and export recognition result, it specifically comprises the following steps: (1) data sectional and characteristic extracting module are carried out segment processing to the network eeg data received by High_speed NIC and after extracting brain electrical feature information, the brain electrical feature obtained vector training sample set transferred to prestige computing module; (2) in prestige computing module, calculate the brain electrical feature vector training sample received and concentrate various kinds credit value originally, and prestige threshold value is set, brain electrical feature vector training sample is screened, and the brain electrical feature vector training sample through screening and credit value thereof are transferred to identification module; (3) according to the vector of the brain electrical feature through the screening training sample and credit value thereof received, the input amendment of identification module determination fuzzy support vector machine and degree of membership parameter, complete the identification of tested individuality and export recognition result; 5) the identification result that identification module exports feeds back to mobile network's terminal by High_speed NIC and shows.
Described step 4) in, data sectional and characteristic extracting module are carried out segment processing to the network eeg data received and are extracted brain electrical feature information, and it specifically comprises the following steps: (I) preset data segmentation and characteristic extracting module receive i-th
sthe network eeg data of individual tested individuality
for:
In formula, N
efor the data length of tested individual brain wave original amplitude sequence, i
s=1,2 ... N
s, N
sfor the number of tested individuality; (II) data sectional and characteristic extracting module are by i-th
sthe network eeg data of individual tested individuality
be divided into N
isection, random selecting starting point in each decile eeg data section, to network eeg data
carry out not decile segmentation, obtain N
dindividual not decile eeg data section, the following principle of demand fulfillment in fragmentation procedure: the end of first the not head end of decile eeg data section and last not decile eeg data section is N with data length respectively
ethe head end of brain wave original amplitude sequence identical with end; The length of each not decile eeg data section is greater than the length of decile eeg data section, and does not exceed the first and last end of brain wave original amplitude sequence; (III) respectively to N
sthe N of individual tested individuality not after decile segmentation
dindividual eeg data section carries out feature extraction, Primary Construction brain electrical feature vector training sample set:
In formula,
represent i-th
si-th of individual tested individuality
dsection brain electrical feature vector training sample,
represent the class label of this training sample, N
dfor the number of not decile eeg data section, N
sfor the number of tested individuality.
Described step 4) in, prestige computing module calculates brain electrical feature vector training sample and concentrates various kinds credit value originally, and by arranging prestige threshold value, brain electrical feature vector training sample is screened, it comprises the following steps: (I) uses correlation computations theoretical, measures identical tested individual i
sthe similarity of different brain electrical features vector training samples; For same tested individual i
s, calculate i-th
d1with i-th
d2the similarity L of individual brain electrical feature vector training sample:
In formula,
with
represent tested individual i respectively
si-th
d1with i-th
d2individual brain electrical feature vector training sample; (II) according to similarity assessment result, i-th is calculated
si-th of individual tested individuality
dsection brain electrical feature vector training sample
credit value
In formula, i
d=1,2...N
d, i
s=1,2...N
s, N
dfor the number of not decile eeg data section, N
sfor the number of tested individuality; (III) thresholding credit value is set in prestige computing module, by comparing credit value and the thresholding credit value of brain electrical feature vector training sample, complete the screening to brain electrical feature vector training sample, it specifically comprises the following steps: 1. the credit value of each brain electrical feature vector training sample and thresholding credit value are compared, remove the brain electrical feature vector training sample of credit value lower than thresholding credit value, and the label of not decile eeg data section corresponding for the brain electrical feature of removal vector training sample is fed back to data sectional and characteristic extracting module; 2. according to the label of the not decile eeg data section received, data sectional and characteristic extracting module choose starting point again from each decile eeg data section, not decile segmentation is carried out to network eeg data, and the neencephalon electrical feature vector training sample utilizing the new not decile eeg data section intercepted to build replaces credit value lower than the brain electrical feature vector training sample of thresholding credit value; 3. repeat step 1. with step 2., until the brain electrical feature vector training sample that all structures obtain meets the requirement of thresholding credit value, complete the screening to brain electrical feature vector training sample, and the brain electrical feature vector training sample through screening and credit value thereof are transferred to identification module.
Described step 4) in, identification module completes the identification of tested individuality and exports recognition result, it specifically comprises the following steps: (I), according to the vector of the brain electrical feature through the screening training sample and credit value thereof received, identification module determination degree of membership also constructs prestige and optimizes two class fuzzy support vector classification devices; (II) according to two class fuzzy support vector classification devices of construction complete, two classification problem is solved; Adopt the multiple two class fuzzy support vector classification devices of 1-a-r rule structure, complete the fuzzy support vector machine Multiclass Classification calculated based on prestige; (III) under a certain network environment, tested individuality is made to complete EEG measuring task, using the eeg data of the tested individuality of wearable EEG measuring measurement device as data to be identified, and by mobile network's terminal transmission to eeg data blood processor, adopt fuzzy support vector machine Multiclass Classification in identification module, the identity of tested individuality is identified.
The present invention is owing to taking above technical scheme, it has the following advantages: 1, the present invention measures owing to adopting the EEG signals identification system comprising the network-oriented environment of some wearable EEG measuring devices, more than one mobile network's terminal and an eeg data blood processor by the EEG signals of side individuality, transmit and identifies, therefore compared with adopting the EEG measuring equipment of complexity in prior art, the present invention has expanded the application scenarios of brain electricity authentication, is applicable to the identification of network environment.When 2, adopting the identity of personal identification method of the present invention to tested individuality to identify, utilize the EEG measuring task of simple, real-time, non-directivity to reduce and measure difficulty, therefore consuming time shorter when the present invention utilizes the wearable EEG signals of EEG measuring device to tested individuality to measure, data volume is less.3, the present invention is due to setting data segmentation and characteristic extracting module and prestige computing module in eeg data blood processor, data sectional and characteristic extracting module carry out segment processing to the eeg data received and after extracting brain electrical feature information, obtain brain electrical feature vector training sample and transfer to prestige computing module, thresholding credit value is set in prestige computing module, by comparing credit value and the thresholding credit value of brain electrical feature vector training sample, complete the screening to brain electrical feature vector training sample, for the brain electrical feature vector training sample not meeting the requirement of thresholding credit value, brain electrical feature vector training sample is rebuild by data sectional and characteristic extracting module, the reliable brain electrical feature sample filtered out and credit value thereof are transferred to identification module and carry out identification by prestige computing module, therefore recognition methods of the present invention is adopted can to provide enough reliable training samples, thus make recognition result more reliable.4, the present invention identifies owing to adopting the identity of fuzzy support vector machine Multiclass Classification to tested individuality based on prestige result of calculation determination degree of membership in identification module, therefore recognition methods of the present invention can improve the accuracy rate of identification, realize in a network environment reliably, fast brain electricity identification.Based on above advantage, the present invention can be widely used in network remote identification.
Accompanying drawing explanation
Fig. 1 is the structural representation of the EEG signals identification system of network-oriented environment of the present invention;
Fig. 2 is the flow chart of the EEG signals personal identification method of network-oriented environment of the present invention;
Fig. 3 is network eeg data section stepwise schematic views;
Fig. 4 is the real-time data acquisition processing platform based on LabVIEW;
Fig. 5 is network eeg data section result schematic diagram; Wherein, horizontal axis representing time, its unit is s, and the longitudinal axis represents EEG signals amplitude, and its unit is μ V;
Fig. 6 is eeg data section prestige result of calculation schematic diagram; Wherein, transverse axis represents eeg data segment number, and the longitudinal axis represents the credit value of eeg data section;
Fig. 7 is the result schematic diagram of the eeg data similarity measurement after prestige is optimized; Wherein, transverse axis represents the numbering of tested individuality, and the longitudinal axis represents the average similarity of eeg data section;
the curve formed represents the average similarity of identical tested individual eeg data section,
the curve formed represents the average similarity of different tested individual eeg data section;
Fig. 8 is different sorting technique brain electricity identification result comparison schematic diagram; Wherein, transverse axis represents the segments of eeg data, and the longitudinal axis represents the accuracy of brain electricity identification;
the curve formed represents the accuracy of the brain electricity identification adopting basic svm classifier algorithm to obtain;
the curve formed represents the accuracy of the brain electricity identification adopting the fuzzy svm classifier algorithm of Gauss distribution to obtain;
the curve formed represents the accuracy of the brain electricity identification adopting prestige Optimization of Fuzzy svm classifier algorithm to obtain.
Detailed description of the invention
Below in conjunction with drawings and Examples, the present invention is described in detail.
As shown in Figure 1, the EEG signals identification system of network-oriented environment of the present invention comprises some wearable EEG measuring device more than 1, one mobile network's terminals 2 and an eeg data blood processor 3.Wearable EEG measuring device 1 transfers to mobile network's terminal 2 after the EEG Processing of the tested individuality measured is become eeg data, mobile network's terminal 2 stores the eeg data received, and the eeg data of storage is transferred to eeg data blood processor 3, eeg data blood processor 3 processes rear output recognition result to the network eeg data received and transfers to mobile network's terminal 2, is shown by mobile network's terminal 2 pairs of recognition results.
In above-described embodiment, wearable EEG measuring device 1 comprises some EEG measuring dry electrode 11, amplification filtering module 12, analog-to-digital conversion module 13 and a bluetooth serial ports module 14.The dry electrode 11 of some EEG measuring transfers to amplification filtering module 12 by wire by measuring the brain electric analoging signal obtained, brain electric analoging signal after amplification filtering transfers to analog-to-digital conversion module 13, brain electric analoging signal is converted to brain waiting for transmission electricity digital signal, and brain electricity digital signal transfers to mobile network's terminal 2 through bluetooth serial ports module 14.
In above-described embodiment, mobile network's terminal 2 adopts the one in mobile phone, panel computer and notebook computer.Mobile network's terminal 2 comprises Bluetooth receptions module 21, data memory module 22, wireless network card 23 and display module 24.The brain received electricity digital data transmission stores to data memory module 22 by Bluetooth receptions module 21, the eeg data that data memory module 22 stores transfers to eeg data blood processor 3 by wireless network card 23, after eeg data blood processor 3 processes the eeg data received, recognition result is transferred to display module 24 by wireless network card 23, by display module 24 Identification display result.
In above-described embodiment, eeg data blood processor 3 comprises High_speed NIC 31, data sectional and characteristic extracting module 32, prestige computing module 33 and identification module 34.The eeg data that mobile network's terminal 2 sends transfers to data sectional and characteristic extracting module 32 by High_speed NIC 31, data sectional and characteristic extracting module 32 carry out segment processing to the eeg data received and after extracting brain electrical feature information, obtain brain electrical feature vector training sample and transfer to prestige computing module 33, prestige computing module 33 carries out credit value assessment to the brain electrical feature sample received, filter out reliable brain electrical feature sample, the reliable brain electrical feature sample filtered out and credit value thereof are transferred to identification module 34 by prestige computing module 33, according to the reliable brain electrical feature sample received and credit value thereof, the identity of identification module 34 to tested individuality identifies and exports recognition result, recognition result transfers to mobile network's terminal 2 by High_speed NIC 31 and shows.
As shown in Figure 2, adopt the EEG signals identification system of network-oriented environment of the present invention tested individuality to be carried out to the method for identification, it comprises the following steps:
1) the EEG signals identification system that comprises the network-oriented environment of some wearable EEG measuring device more than 1, one mobile network's terminals 2 and an eeg data blood processor 3 is set.Wherein, wearable EEG measuring device 1 comprises the dry electrode 11 of EEG measuring, amplification filtering module 12, analog-to-digital conversion module 13 and bluetooth serial ports module 14.Mobile network's terminal 2 comprises Bluetooth receptions module 21, data memory module 22, wireless network card 23 and display module 24.Eeg data blood processor 3 comprises High_speed NIC 31, data sectional and characteristic extracting module 32, prestige computing module 33 and identification module 34.
2) the dry electrode of EEG measuring 11 is used to carry out perception measurement to the EEG signals of tested individuality, the brain electric analoging signal measured exports brain electricity digital signal after amplification filtering module 12 and analog-to-digital conversion module 13 process, brain electricity digital signal transfers to mobile network's terminal 2 by bluetooth serial ports module 14, provides initial data for building brain electricity identity data training set.
3) in mobile network's terminal 2, Bluetooth receptions module 21 by receive brain electricity digital data transmission to data memory module 22, the eeg data of storage is transferred to eeg data blood processor 3 by wireless network card 23 by data memory module 22.
4) eeg data blood processor 3 processes the EEG signals received, and complete the identification to tested individuality, it specifically comprises the following steps:
(1) data sectional and characteristic extracting module 32 receive by High_speed NIC 31 the network eeg data that mobile network's terminal 2 sends, as shown in Figure 3, data sectional and characteristic extracting module 32 carry out segment processing to the network eeg data received and after extracting brain electrical feature information, obtain brain electrical feature vector training sample set and transfer to prestige computing module 33, it specifically comprises:
(I) preset data segmentation and characteristic extracting module 32 receive i-th
sthe network eeg data of individual tested individuality
for:
In formula, N
efor the data length of tested individual brain wave original amplitude sequence, i
s=1,2 ... N
s, N
sfor the number of tested individuality.
(II) data sectional and characteristic extracting module 32 are by the network eeg data of the i-th s tested individuality
be divided into N
isection.Random selecting starting point in each decile eeg data section, to network eeg data
carry out not decile segmentation, obtain N
dindividual not decile eeg data section, the following principle of demand fulfillment in fragmentation procedure:
First not the head end of decile eeg data section and last not decile eeg data section end respectively with step 1. in data length be N
ethe head end of brain wave original amplitude sequence identical with end.
The length of each not decile eeg data section is greater than the length of decile eeg data section, and does not exceed the first and last end of brain wave original amplitude sequence.
(III) respectively to N
sc the eeg data section of individual tested individuality not after decile segmentation carries out feature extraction, Primary Construction brain electrical feature vector training sample set:
In formula,
represent i-th
si-th of individual tested individuality
dsection brain electrical feature vector training sample,
represent the class label of this training sample, N
dfor the number of not decile eeg data section.
(2) in prestige computing module 33, calculate the brain electrical feature vector training sample received and concentrate various kinds credit value originally, and by arranging prestige threshold value, brain electrical feature vector training sample is screened, and the brain electrical feature vector training sample through screening and credit value thereof are transferred to identification module 34, it specifically comprises:
(I) use correlation computations theoretical, measure identical tested individual i
sthe similarity of different brain electrical features vector training samples;
For same tested individual i
s, calculate i-th
d1with i-th
d2the similarity L of individual brain electrical feature vector training sample:
In formula,
with
represent tested individual i respectively
si-th
d1with i-th
d2individual brain electrical feature vector training sample.
(II) according to similarity assessment result, i-th is calculated
si-th of individual tested individuality
dsection brain electrical feature vector training sample
credit value
In formula, i
d=1,2...N
d, i
s=1,2...N
s, N
dfor the number of not decile eeg data section, N
sfor the number of tested individuality.
(III) in prestige computing module 33, arrange thresholding credit value, by comparing credit value and the thresholding credit value of brain electrical feature vector training sample, complete the screening to brain electrical feature vector training sample, it specifically comprises the following steps:
1. the credit value of each brain electrical feature vector training sample and thresholding credit value are compared, remove the brain electrical feature vector training sample of credit value lower than thresholding credit value, and the label of not decile eeg data section corresponding for the brain electrical feature of removal vector training sample is fed back to data sectional and characteristic extracting module 32.
2. according to the label of the not decile eeg data section received, data sectional and characteristic extracting module 32 choose starting point again from each decile eeg data section, not decile segmentation is carried out to network eeg data, and the neencephalon electrical feature vector training sample utilizing the new not decile eeg data section intercepted to build replaces credit value lower than the brain electrical feature vector training sample of thresholding credit value.
3. repeat step 1. with step 2., until the brain electrical feature vector training sample that all structures obtain meets the requirement of thresholding credit value, complete the screening to brain electrical feature vector training sample, and the brain electrical feature vector training sample through screening and credit value thereof are transferred to identification module 34.
(3) according to the vector training sample of the brain electrical feature through screening received and credit value thereof, identification module 34 determines input amendment and the degree of membership parameter of fuzzy support vector machine, complete the identification of tested individuality and export recognition result, it specifically comprises the following steps:
(I) according to the vector training sample of the brain electrical feature through screening received and credit value thereof, identification module 34 is determined degree of membership and constructs prestige to optimize two class fuzzy support vector classification devices;
For i-th
si-th of individual tested individuality
dsection brain electrical feature vector training sample
class label
introduce fuzzy membership
two classification problem representation is:
In formula,
This two classification problem identification brain electrical feature vector training sample is utilized whether to belong to class a.
In formula, degree of membership
represent that the weight affected determined by sample for optimal hyperlane.
I-th
sindividual tested individual i-th
dsection brain electrical feature vector
degree of membership
be expressed as:
In formula, Q is scaling normalized function, i
d=1,2...N
d, i
s=1,2...N
s.
Carry out classification difficulty very greatly owing to finding suitable hyperplane in input amendment space to brain electrical feature vector training sample, therefore adopt map-germ function phi by input amendment spatial mappings is found hyperplane to feature space Z, hyperplane is expressed as:
In formula, w and b is the parameter determining hyperplane, and w represents the vector perpendicular to hyperplane, and b represents the displacement of hyperplane to initial point.
For constructing optimal hyperlane when brain electrical feature vector training sample is linearly inseparable, introduce slack variable
with punishment component C.For brain electrical feature vector training sample be the situation of linearly inseparable, the problem arises solving optimal hyperlane is following quadratic programming problem:
In formula, i
d=1,2 ... N
d, i
s=1,2 ... N
s, slack variable
Punishment component C measures fuzzy support vector machine to all strength of punishments comprising error point, and degree of membership
different punishment weights is given, to reduce the larger input amendment of deviation to the impact of classifying quality to the input amendment in different range of error.Degree of membership
less, i.e. sample
the error comprised is larger, and on determining that the impact of optimal hyperlane is less, vice versa.Slack variable
the tolerance of the inseparable error of metric linear.Punishment component C, slack variable
match with scaling normalized function Q, and be determined by experiment.
Adopt Lagrangian to solve the quadratic programming problem meeting Karush-Kuhn-Tucker (KKT) condition, construction complete prestige optimizes two class fuzzy support vector classification devices.Thus judge whether brain electrical feature vector training sample belongs to class a.
(II) according to two class fuzzy support vector classification devices of construction complete, two classification problem is solved; Adopt the multiple two class fuzzy support vector classification devices of 1-a-r rule structure, complete the fuzzy support vector machine Multiclass Classification calculated based on prestige.
(III) under a certain network environment, tested individuality is made to complete EEG measuring task, the eeg data of the tested individuality measured by wearable EEG measuring device 1 is as data to be identified, and transfer to eeg data blood processor 3 by mobile network's terminal 2, adopt fuzzy support vector machine Multiclass Classification in identification module 34, the identity of tested individuality is identified.
5) the identification result that identification module 34 exports feeds back to mobile network's terminal 2 by High_speed NIC 31 and shows.
Embodiment: the EEG signals identification system that comprises the network-oriented environment of some wearable EEG measuring device 1, mobile network's terminals 2 and an eeg data blood processor 3 is set.Wherein, wearable EEG measuring device 1 adopts the MINDWAVEMOBILE device of Neurosky company, eeg data blood processor 3 to adopt the LabVIEW Data acquisition and issuance platform of NIPXI hardware supported.The EEG signals that MINDWAVEMOBILE equipment collects transfers to mobile network's terminal 2 by Bluetooth2.1, mobile network's terminal 2 stores the EEG signals received, and the eeg data of storage is transferred to LabVIEW Data acquisition and issuance platform, LabVIEW Data acquisition and issuance platform carries out data sectional and feature extraction and classification based training and calculating to the network eeg data received, and obtains identification result and feeds back to mobile network's terminal 2.
Adopt the identity of EEG signals identification system to tested individuality of network-oriented environment of the present invention to identify, its process is:
1) adopt MINDWAVEMOBILE EEG measuring device 1 to gather the EEG signals of 18 tested individualities, it comprises and all gathers the training data of 60s and the data to be identified of 10s to every tested individuality.The eeg data collected transfers to LabVIEW Data acquisition and issuance platform by mobile network's terminal 2.
Adopt MINDWAVEMOBILE EEG measuring device 1 to measure the EEG signals that 18 tested individualities are in unconscious meditation state 60s, and transfer to LabVIEW Data acquisition and issuance platform by mobile network's terminal 2.
2) LabVIEW Data acquisition and issuance platform processes the eeg data received, and it comprises:
First, LabVIEW Data acquisition and issuance platform carries out segmentation to the 60s training data collected, and obtains network eeg data section as shown in Figure 5.Analyzing and processing is carried out to this eeg data section, and extracts the following feature of EEG signals:
(1) each frequency range brain wave Energy distribution: the energy accounting of α ripple (8Hz ~ 13Hz), β ripple (13Hz ~ 30Hz), γ ripple (30Hz ~ 100Hz), θ ripple (4Hz ~ 8Hz), δ ripple (<4Hz).
(2) power spectrum characteristic analysis: the frequency values that peak point is corresponding.
(3) focus of brain wave and the analysis of meditation degree: the absorbed degree of measurand and meditate depth survey etc.
As shown in table 1, for a certain eeg data section of 3 tested individualities of difference, extract the principal character of its EEG signals.
Table 1 EEG signals principal character is extracted
| Characteristic quantity | Tested individual 1 | Tested individual 2 | Tested individual 3 |
| Low frequency α wave energy accounting | 56.69％ | 58.26％ | 53.93％ |
| High frequency alpha waves energy accounting | 10.99％ | 8.13％ | 15.64％ |
| Low frequency β wave energy accounting | 3.79％ | 3.66％ | 11.92％ |
| High frequency β wave energy accounting | 3.31％ | 1.80％ | 9.84％ |
| Low frequency γ wave energy accounting | 1.93％ | 1.50％ | 3.20％ |
| Intermediate frequency γ wave energy accounting | 1.40％ | 1.40％ | 3.15％ |
| θ wave energy accounting | 0.65％ | 0.90％ | 1.57％ |
| δ wave energy accounting | 21.24％ | 24.35％ | 0.75％ |
| Average absorbed index | 27.63 | 65.09 | 43.18 |
| Average meditation index | 38.00 | 44.77 | 41.73 |
Secondly, based on the prestige theory of computation, by the credit value of each eeg data section of Similarity measures that eeg data is intersegmental.As shown in Figure 6, the eeg data of a certain specific tested individuality is divided into 20 sections, and is numbered 1 ~ 20.According to the credit value of each data segment of mean value calculation of the similarity of each data segment and other data segments.
Arranging thresholding credit value is 0.95, remove lower than thresholding credit value 0.95 be numbered 18 data segment.Again choose data segment, then carry out prestige calculating, until each data segment credit value all exceedes thresholding credit value 0.95.
Utilize the theory that prestige calculates equally, the identical tested individual brain electrical feature similarity of comparison and difference tested individual brain electrical feature similarity, as shown in Figure 7, obtain the eeg data similarity measurement after prestige optimization, identical tested individual eeg data similarity is obviously better than different tested individual eeg data similarity, and average prestige result of calculation exceeds about 5%.Therefore, the data segmentation method that the present invention proposes and the data segment accuracy evaluation calculated based on prestige and screening technique can improve the characteristic similarity of same tested individual eeg data, increase the difference of different tested individual EEG signals feature and identical tested individual EEG signals feature, the input amendment of Optimum Classification algorithm.
Wearable EEG signals through collection, segmentation, and is screened, for fuzzy support vector machine provides enough input amendment by the assessment of prestige result of calculation.Selection of kernel function Radial basis kernel function.In conjunction with the result that prestige calculates, determine degree of membership.When using fuzzy support vector machine to classify, normal use Gauss distribution determination degree of membership, distributes fuzzy support vector classification algorithm as reference using basic support vector cassification algorithm and normal Gaussian here.Fig. 8 is than the brain electricity identification result of right support vector machine basic classification algorithm, Gauss distribution fuzzy support vector classification algorithm and prestige Optimization of Fuzzy support vector cassification algorithm.As shown in Figure 8, along with the increase of segments, three kinds of sorting techniques all obtain more sufficient training sample, and accuracy of identification improves, but when segments is greater than after 12, accuracy of identification improves speed and slows down.Use fuzzy support vector classification algorithm that the character of algorithm itself can be utilized to improve the accuracy of classified counting, and result screening input training sample, more scientific determination degree of membership that modified hydrothermal process uses prestige to calculate, further increase accuracy of identification again, be better than two kinds of comparator algorithms, the highest discrimination reaches 91.98%.Especially, when input amendment is less, carried out assessing screening to brain electrical characteristic data section because prestige calculates, under significantly reducing less input amendment situation, error information section is on the impact of recognition result, and therefore the raising of modified hydrothermal process accuracy of identification is more obvious.
Finally, the identity of identification sorting algorithm to tested individuality of EEG signals is adopted to identify.
The various embodiments described above are only for illustration of the present invention; wherein the structure of each parts, connected mode and method step etc. all can change to some extent; every equivalents of carrying out on the basis of technical solution of the present invention and improvement, all should not get rid of outside protection scope of the present invention.
Claims (9)
1. an EEG signals identification system for network-oriented environment, is characterized in that: it comprises some wearable EEG measuring devices, more than one mobile network's terminal and an eeg data blood processor; Described wearable EEG measuring device transfers to described mobile network's terminal after the EEG Processing of the tested individuality measured is become eeg data, the eeg data that described mobile network's end-on receives stores, and the eeg data of storage is transferred to described eeg data blood processor, described eeg data blood processor exports recognition result after processing the network eeg data received, and recognition result is transferred to described mobile network's terminal and show;
Described eeg data blood processor comprises High_speed NIC, data sectional and characteristic extracting module, prestige computing module and identification module, the eeg data that described mobile network's terminal sends transfers to described data sectional and characteristic extracting module by described High_speed NIC, described data sectional and characteristic extracting module carry out segment processing to the eeg data received and after extracting brain electrical feature information, obtain brain electrical feature vector training sample and transfer to described prestige computing module, described prestige computing module carries out credit value assessment to the brain electrical feature sample received, filter out reliable brain electrical feature sample, the reliable brain electrical feature sample filtered out and credit value thereof are transferred to described identification module by described prestige computing module, the identity of described identification module to tested individuality identifies and exports recognition result, recognition result transfers to described mobile network's terminal by described High_speed NIC and shows.
2. the EEG signals identification system of a kind of network-oriented environment as claimed in claim 1, is characterized in that: described wearable EEG measuring device comprises the dry electrode of some EEG measuring, an amplification filtering module, an analog-to-digital conversion module and a bluetooth serial ports module; The brain electric analoging signal measured is transferred to described amplification filtering module by wire by the dry electrode of some described EEG measuring, brain electric analoging signal after amplification filtering transfers to described analog-to-digital conversion module, and being converted to brain waiting for transmission electricity digital signal, brain electricity digital signal is through described bluetooth serial ports module transfer extremely described mobile network's terminal.
3. the EEG signals identification system of a kind of network-oriented environment as claimed in claim 1, is characterized in that: described mobile network's terminal comprises Bluetooth receptions module, data memory module, wireless network card and display module; The brain received electricity digital data transmission stores to described data memory module by described Bluetooth receptions module, the eeg data that described data memory module stores transfers to described eeg data blood processor by described wireless network card, recognition result is transferred to display module by described wireless network card and shows after processing the eeg data received by described eeg data blood processor.
4. the EEG signals identification system of a kind of network-oriented environment as claimed in claim 2, is characterized in that: described mobile network's terminal comprises Bluetooth receptions module, data memory module, wireless network card and display module; The brain received electricity digital data transmission stores to described data memory module by described Bluetooth receptions module, the eeg data that described data memory module stores transfers to described eeg data blood processor by described wireless network card, recognition result is transferred to display module by described wireless network card and shows after processing the eeg data received by described eeg data blood processor.
5. the EEG signals identification system of a kind of network-oriented environment as claimed in claim 1 or 2 or 3 or 4, is characterized in that: described mobile network's terminal adopts the one in mobile phone, panel computer and notebook computer.
6. adopt an EEG signals personal identification method for the network-oriented environment of the identification system as described in any one of Claims 1 to 5, it comprises the following steps:
1) the EEG signals identification system that comprises the network-oriented environment of some wearable EEG measuring devices, more than one mobile network's terminal and an eeg data blood processor is set; Wherein, wearable EEG measuring device comprises the dry electrode of EEG measuring, amplification filtering module, analog-to-digital conversion module and bluetooth serial ports module; Mobile network's terminal comprises Bluetooth receptions module, data memory module, wireless network card and display module; Eeg data blood processor comprises High_speed NIC, data sectional and characteristic extracting module, prestige computing module and identification module;
2) EEG signals of the tested individuality of the dry electrode pair of EEG measuring is used to carry out perception measurement, the brain electric analoging signal measured exports brain electricity digital signal after amplification filtering module and analog-to-digital conversion module process, and brain electricity digital signal passes through bluetooth serial ports module transfer to mobile network's terminal;
3) in mobile network's terminal, Bluetooth receptions module by receive brain electricity digital data transmission to data memory module, the eeg data of storage is transferred to eeg data blood processor by wireless network card by data memory module;
4) eeg data blood processor processes the EEG signals received, and complete the identification of tested individuality and export recognition result, it specifically comprises the following steps:
(1) data sectional and characteristic extracting module are carried out segment processing to the network eeg data received by High_speed NIC and after extracting brain electrical feature information, the brain electrical feature obtained vector training sample set are transferred to prestige computing module;
(2) in prestige computing module, calculate the brain electrical feature vector training sample received and concentrate various kinds credit value originally, and prestige threshold value is set, brain electrical feature vector training sample is screened, and the brain electrical feature vector training sample through screening and credit value thereof are transferred to identification module;
(3) according to the vector of the brain electrical feature through the screening training sample and credit value thereof received, the input amendment of identification module determination fuzzy support vector machine and degree of membership parameter, complete the identification of tested individuality and export recognition result;
5) the identification result that identification module exports feeds back to mobile network's terminal by High_speed NIC and shows.
7. the EEG signals personal identification method of a kind of network-oriented environment as claimed in claim 6, it is characterized in that: described step 4) in, data sectional and characteristic extracting module are carried out segment processing to the network eeg data received and are extracted brain electrical feature information, and it specifically comprises the following steps:
(I) preset data segmentation and characteristic extracting module receive i-th
sthe network eeg data of individual tested individuality
for:
In formula, N
efor the data length of tested individual brain wave original amplitude sequence, i
s=1,2 ... N
s, N
sfor the number of tested individuality;
(II) data sectional and characteristic extracting module are by i-th
sthe network eeg data of individual tested individuality
be divided into N
isection, random selecting starting point in each decile eeg data section, to network eeg data
carry out not decile segmentation, obtain N
dindividual not decile eeg data section, the following principle of demand fulfillment in fragmentation procedure:
The end of first the not head end of decile eeg data section and last not decile eeg data section is N with data length respectively
ethe head end of brain wave original amplitude sequence identical with end;
The length of each not decile eeg data section is greater than the length of decile eeg data section, and does not exceed the first and last end of brain wave original amplitude sequence;
(III) respectively to N
sthe N of individual tested individuality not after decile segmentation
dindividual eeg data section carries out feature extraction, Primary Construction brain electrical feature vector training sample set:
In formula,
represent i-th
si-th of individual tested individuality
dsection brain electrical feature vector training sample,
represent the class label of this training sample, N
dfor the number of not decile eeg data section, N
sfor the number of tested individuality.
8. the EEG signals personal identification method of a kind of network-oriented environment as claimed in claim 6, it is characterized in that: described step 4) in, prestige computing module calculates brain electrical feature vector training sample and concentrates various kinds credit value originally, and by arranging prestige threshold value, screen brain electrical feature vector training sample, it comprises the following steps:
(I) use correlation computations theoretical, measure identical tested individual i
sthe similarity of different brain electrical features vector training samples;
For same tested individual i
s, calculate i-th
d1with i-th
d2the similarity L of individual brain electrical feature vector training sample:
In formula,
with
represent tested individual i respectively
si-th
d1with i-th
d2individual brain electrical feature vector training sample;
(II) according to similarity assessment result, i-th is calculated
si-th of individual tested individuality
dsection brain electrical feature vector training sample
credit value
In formula, i
d=1,2...N
d, i
s=1,2...N
s, N
dfor the number of not decile eeg data section, N
sfor the number of tested individuality;
(III) in prestige computing module, arrange thresholding credit value, by comparing credit value and the thresholding credit value of brain electrical feature vector training sample, complete the screening to brain electrical feature vector training sample, it specifically comprises the following steps:
1. the credit value of each brain electrical feature vector training sample and thresholding credit value are compared, remove the brain electrical feature vector training sample of credit value lower than thresholding credit value, and the label of not decile eeg data section corresponding for the brain electrical feature of removal vector training sample is fed back to data sectional and characteristic extracting module;
2. according to the label of the not decile eeg data section received, data sectional and characteristic extracting module choose starting point again from each decile eeg data section, not decile segmentation is carried out to network eeg data, and the neencephalon electrical feature vector training sample utilizing the new not decile eeg data section intercepted to build replaces credit value lower than the brain electrical feature vector training sample of thresholding credit value;
3. repeat step 1. with step 2., until the brain electrical feature vector training sample that all structures obtain meets the requirement of thresholding credit value, complete the screening to brain electrical feature vector training sample, and the brain electrical feature vector training sample through screening and credit value thereof are transferred to identification module.
9. the EEG signals personal identification method of a kind of network-oriented environment as described in claim 6 or 7 or 8, it is characterized in that: described step 4) in, identification module completes the identification of tested individuality and exports recognition result, and it specifically comprises the following steps:
(I) according to the vector of the brain electrical feature through the screening training sample and credit value thereof received, identification module determination degree of membership also constructs prestige and optimizes two class fuzzy support vector classification devices;
(II) according to two class fuzzy support vector classification devices of construction complete, two classification problem is solved; Adopt the multiple two class fuzzy support vector classification devices of 1-a-r rule structure, complete the fuzzy support vector machine Multiclass Classification calculated based on prestige;
(III) under a certain network environment, tested individuality is made to complete EEG measuring task, using the eeg data of the tested individuality of wearable EEG measuring measurement device as data to be identified, and by mobile network's terminal transmission to eeg data blood processor, adopt fuzzy support vector machine Multiclass Classification in identification module, the identity of tested individuality is identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410220371.0A CN103976740B (en) | 2014-05-23 | 2014-05-23 | A kind of EEG signals identification system of network-oriented environment and recognition methods |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410220371.0A CN103976740B (en) | 2014-05-23 | 2014-05-23 | A kind of EEG signals identification system of network-oriented environment and recognition methods |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103976740A CN103976740A (en) | 2014-08-13 |
| CN103976740B true CN103976740B (en) | 2016-03-30 |
Family
ID=51269091
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410220371.0A Active CN103976740B (en) | 2014-05-23 | 2014-05-23 | A kind of EEG signals identification system of network-oriented environment and recognition methods |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103976740B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105376692B (en) * | 2014-08-19 | 2020-02-28 | 中兴通讯股份有限公司 | Method and device for information interaction between intelligent terminal and wearable device |
| EP3221807B1 (en) * | 2014-11-20 | 2020-07-29 | Widex A/S | Hearing aid user account management |
| CN104523268B (en) * | 2015-01-15 | 2017-02-22 | 江南大学 | Electroencephalogram signal recognition fuzzy system and method with transfer learning ability |
| CN110121711A (en) * | 2016-12-21 | 2019-08-13 | 华为技术有限公司 | A kind of personal identification method and equipment of view-based access control model Evoked ptential |
| CN107411734A (en) * | 2017-03-06 | 2017-12-01 | 华斌 | A kind of device that user characteristics is obtained according to human-body biological electromagnetic wave |
| CN107066103B (en) * | 2017-05-27 | 2021-03-30 | 深圳市宏智力科技有限公司 | Method and device for recognizing living being by brain wave |
| CN108056770A (en) * | 2018-02-02 | 2018-05-22 | 合肥芯福传感器技术有限公司 | A kind of heart rate detection method based on artificial intelligence |
| CN108830865B (en) * | 2018-05-08 | 2021-06-15 | 南京伟思医疗科技股份有限公司 | Method for determining stable upper and lower boundaries of dynamic electroencephalogram image |
| CN111387936B (en) * | 2019-01-02 | 2023-07-21 | 中国移动通信有限公司研究院 | Sleep stage identification method, device and equipment |
| CN109464152B (en) * | 2019-01-09 | 2021-11-26 | 浙江强脑科技有限公司 | Method and device for detecting brain fatigue state and computer readable storage medium |
| CN113178195B (en) * | 2021-03-04 | 2022-08-26 | 杭州电子科技大学 | Speaker identification method based on sound-induced electroencephalogram signals |
| CN113238655A (en) * | 2021-05-19 | 2021-08-10 | 河北酷点科技有限公司 | Method for investigation and data authenticity check based on electric wave neural drive |
| CN113190122A (en) * | 2021-05-31 | 2021-07-30 | 江苏集萃脑机融合智能技术研究所有限公司 | Intelligent device and method based on brain signals, intelligent system and application |
| CN115828208B (en) * | 2022-12-07 | 2023-09-08 | 北京理工大学 | Touch brain electrolytic locking method and system based on cloud edge cooperation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1883378A (en) * | 2006-06-30 | 2006-12-27 | 上海博维康讯信息科技发展有限公司 | A wireless health monitor system |
| CN102349829A (en) * | 2011-10-28 | 2012-02-15 | 重庆大学 | Remote monitoring system of electrocardiosignal |
| CN202365775U (en) * | 2011-12-05 | 2012-08-08 | 冯辉 | Remote electrocardiogram monitoring system |
| CN202589519U (en) * | 2012-05-23 | 2012-12-12 | 上海海事大学 | Wearing type wireless brain signal collecting system based on radio frequency identification |
| CN203122372U (en) * | 2013-03-12 | 2013-08-14 | 李晓龙 | Real-time health monitoring and intelligent warning system based on Android technology and Internet of Things |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030144875A1 (en) * | 1997-09-06 | 2003-07-31 | Suffin Stephen C. | EEG prediction method for medication response |
| US8065529B2 (en) * | 2007-05-21 | 2011-11-22 | Ut-Battelle, Llc | Methods for using a biometric parameter in the identification of persons |
| US9020585B2 (en) * | 2007-06-18 | 2015-04-28 | New York University | Electronic identity card |
| KR101031507B1 (en) * | 2010-07-28 | 2011-04-29 | (주)아이맥스 | A portable measuring instrument of electroencephalograph and control system |
-
2014
- 2014-05-23 CN CN201410220371.0A patent/CN103976740B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1883378A (en) * | 2006-06-30 | 2006-12-27 | 上海博维康讯信息科技发展有限公司 | A wireless health monitor system |
| CN102349829A (en) * | 2011-10-28 | 2012-02-15 | 重庆大学 | Remote monitoring system of electrocardiosignal |
| CN202365775U (en) * | 2011-12-05 | 2012-08-08 | 冯辉 | Remote electrocardiogram monitoring system |
| CN202589519U (en) * | 2012-05-23 | 2012-12-12 | 上海海事大学 | Wearing type wireless brain signal collecting system based on radio frequency identification |
| CN203122372U (en) * | 2013-03-12 | 2013-08-14 | 李晓龙 | Real-time health monitoring and intelligent warning system based on Android technology and Internet of Things |
Non-Patent Citations (3)
| Title |
|---|
| Fuzzy support vector machine for classification of EEG signals using wavelet-based features;Qi Xu 等;《Medical Engineering & Physics》;20090428;全文 * |
| 基于脑电波信号的身份识别技术;夏立文;《中国优秀硕士学位论文全文数据库 信息科技辑》;20110915(第09期);正文第10-11页第2.4节 * |
| 普适环境下基于脑电的身份及上下文状态识别;刘泉影 等;《东南大学学报(自然科学版)》;20101130;第40卷;第264页第6行-倒数第1行 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103976740A (en) | 2014-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103976740B (en) | A kind of EEG signals identification system of network-oriented environment and recognition methods | |
| Omara et al. | A novel geometric feature extraction method for ear recognition | |
| CN105678273B (en) | The starting point detection algorithm of radio-frequency fingerprint identification technology transient signal | |
| CN103605963A (en) | Fingerprint identification method | |
| CN106446828A (en) | User identity identification method based on Wi-Fi signal | |
| CN102231213A (en) | ECG (electrocardiograph) access card identity identification method and system | |
| JP2017123170A (en) | Electrocardiogram authentication method and apparatus, and training method | |
| Li et al. | Fingerprints in the air: Unique identification of wireless devices using RF RSS fingerprints | |
| Al-Naffakh et al. | Unobtrusive gait recognition using smartwatches | |
| CN105184254A (en) | Identity authentication method and system | |
| CN110062379B (en) | Identity authentication method based on channel state information under human behavior scene | |
| CN107315995B (en) | Face recognition method based on Laplace logarithmic face and convolutional neural network | |
| CN108090462A (en) | A kind of Emitter Fingerprint feature extracting method based on box counting dimension | |
| Shang et al. | A usable authentication system using wrist-worn photoplethysmography sensors on smartwatches | |
| CN105654046B (en) | Electrocardiosignal personal identification method and device | |
| Zhang et al. | Human identification using WiFi signal | |
| CN106991773A (en) | A kind of safety-protection system vibration signal recognition based on S-transformation feature extraction | |
| CN103400069A (en) | Pulse wave identification device and method based on power spectrum analysis | |
| Liu et al. | Behavior privacy preserving in RF sensing | |
| CN110123289B (en) | Biological identification method based on pulse wave and related device | |
| CN104679967A (en) | Method for judging reliability of psychological test | |
| CN111652132B (en) | Non-line-of-sight identity recognition method and device based on deep learning and storage medium | |
| Lv et al. | An algorithm of Iris feature-extracting based on 2D Log-Gabor | |
| CN103903017A (en) | Face recognition method based on self-adaption soft histogram local binary patterns | |
| Zhang et al. | Identification of cloned HF RFID proximity cards based on RF fingerprinting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |