WO2016080366A1 - 脳波による認証装置、認証方法、認証システム及びプログラム - Google Patents
脳波による認証装置、認証方法、認証システム及びプログラム Download PDFInfo
- Publication number
- WO2016080366A1 WO2016080366A1 PCT/JP2015/082178 JP2015082178W WO2016080366A1 WO 2016080366 A1 WO2016080366 A1 WO 2016080366A1 JP 2015082178 W JP2015082178 W JP 2015082178W WO 2016080366 A1 WO2016080366 A1 WO 2016080366A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- authentication
- discrimination
- electroencephalogram
- stimulus
- target
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 210000004556 brain Anatomy 0.000 claims abstract description 58
- 230000000638 stimulation Effects 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 22
- 230000001186 cumulative effect Effects 0.000 claims description 20
- 230000036992 cognitive tasks Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 230000019771 cognition Effects 0.000 claims description 2
- 230000000007 visual effect Effects 0.000 description 23
- 230000006870 function Effects 0.000 description 12
- 238000005259 measurement Methods 0.000 description 12
- 238000004458 analytical method Methods 0.000 description 10
- 230000009471 action Effects 0.000 description 9
- 210000003128 head Anatomy 0.000 description 9
- 230000007704 transition Effects 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 7
- 230000004044 response Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 6
- 230000007177 brain activity Effects 0.000 description 5
- 230000001965 increasing effect Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 238000009825 accumulation Methods 0.000 description 4
- 210000004761 scalp Anatomy 0.000 description 4
- 230000008685 targeting Effects 0.000 description 4
- 230000006399 behavior Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000001149 cognitive effect Effects 0.000 description 2
- 230000003920 cognitive function Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000894007 species Species 0.000 description 2
- 241000282693 Cercopithecidae Species 0.000 description 1
- 241000016401 Hatena Species 0.000 description 1
- 241000219793 Trifolium Species 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000006735 deficit Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004070 electrodeposition Methods 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 210000000887 face Anatomy 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 210000000554 iris Anatomy 0.000 description 1
- 230000007794 irritation Effects 0.000 description 1
- 230000006996 mental state Effects 0.000 description 1
- 238000000491 multivariate analysis Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 210000001747 pupil Anatomy 0.000 description 1
- 230000010076 replication Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000035900 sweating Effects 0.000 description 1
- 230000008448 thought Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 210000003462 vein Anatomy 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/117—Identification of persons
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/377—Electroencephalography [EEG] using evoked responses
- A61B5/378—Visual stimuli
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6813—Specially adapted to be attached to a specific body part
- A61B5/6814—Head
Definitions
- the present invention relates to biometric authentication using brain waves, and relates to an authentication device, authentication method, authentication system, and program using brain waves.
- biometric authentication examples include fingerprints and irises in the pupil.
- vein authentication that reads the shape of the palm or finger blood vessels is used in ATMs of banks.
- voiceprints, face shapes, handwriting, etc. has been put into practical use.
- biometric authentication can be broken by replication.
- Non-Patent Document 1 shows a behavior prediction method that can be selected from neural activity taking single neuron activity as an example.
- Patent Documents 1 and 2 have proposed a communication support apparatus and method capable of transmitting intention by analyzing brain activity (see Patent Documents 1 and 2).
- Patent Documents 1 and 2 for example, it is possible to support communication for people with movement disabilities who are difficult to speak and write, and people with severe movement disabilities who are difficult to input various devices using their hands and feet.
- the present inventor has proposed a method for showing information representation in the brain in a map by electroencephalogram analysis for general subjects including healthy subjects (see Patent Document 3). Further, the present inventor has proposed an apparatus and a method for ordering survey objects by electroencephalogram analysis (see Patent Document 4).
- Patent Documents 5 and 6 an authentication method using an electroencephalogram has been proposed (see Patent Documents 5 and 6).
- Patent Document 5 the basis based on the subject's brain wave and the basis based on the brain wave of the target candidate are obtained by using the frequency component obtained by Fourier transform from the brain wave and the time domain component obtained by correlation dimension analysis from the brain wave. It has been proposed to perform personal authentication based on distance.
- Patent Document 6 in addition to the examination of the identity by the existing biometrics authentication, as a personal authentication method for detecting an unauthorized operation by the user himself / herself, an electroencephalogram, a heart rate, a sweating amount, etc. It has been proposed to use this information as biological information representing emotional characteristics for determination of a mental state.
- JP 2012-053656 A JP 2012-073329 A JP 2010-274035 A JP 2013-178601 A JP 2004-248714 A JP 2005-293209 A
- Patent Documents 1 and 2 of the prior art the problem to be solved is to be able to decode at high speed and with high accuracy in order to decode the decision making in the brain by analyzing the electroencephalogram data.
- Patent document 1 is a device that has been proposed in research on communication before that, in order to measure biological information such as brain waves, there is a problem that the noise is large, the probability of correct answer is low, and it takes time to determine, It is a technology to solve.
- Patent Document 1 discloses that the intent in the brain is determined in a short time without erroneous determination, the device is operated directly in real time by the operator thinking in the brain, and a patient or elderly person with speech impairment is basically It is a technology that makes it easier and more direct to tell caregivers about caregiving and feelings around us.
- Patent Document 1 is realized by a technique of quantifying a brain process of decision making called “virtual decision making function” that has been researched and developed by the present inventor.
- the concept of a virtual decision-making function is greatly expanded and utilized as an in-brain intention decoding technique for an intention transmission device based on electroencephalogram measurement.
- the communication support apparatus of Patent Literature 1 includes a device that presents a stimulus and a processing device that processes brain wave data from an electroencephalograph that measures an electroencephalogram after the stimulus presentation by the device.
- the processing device includes the electroencephalogram. Based on the discriminant function obtained by analyzing the data and the success rate, it is determined that a specific decision is made in the brain. Also, based on the product of the cumulative discriminant score by the discriminant function obtained by analyzing the electroencephalogram data and the success rate, when the product exceeds a threshold, it is determined that a specific decision is made in the brain, The judgment result is output to the device.
- a function obtained by analyzing data measured by an electroencephalograph is a function of multivariate analysis, such as a logistic function or a linear discriminant analysis function. It was shown that it was to be set for each subsequent elapsed time.
- Patent Document 1 is a technology that can decode intentions in the brain in real time and support communication, and the success rate has been used as a numerical value for creating an improved virtual decision-making function.
- the devices of Patent Documents 1 and 2 are based on real-time decoding of event-related potentials (ERP) related to cognitive functions, among other electroencephalograms that reflect various brain information.
- ERP event-related potentials
- the present invention is intended to solve these problems, and an object of the present invention is to provide a personal authentication device, a personal authentication method, a system, and a program having high confidentiality and high authentication accuracy as personal authentication. .
- the present invention has the following features in order to achieve the above object.
- the apparatus of the present invention is an electroencephalogram authentication apparatus that estimates brain information for each authentication candidate stored in advance with respect to an electroencephalograph and electroencephalograms for a plurality of stimulation events obtained by the electroencephalograph. And a processing unit that obtains a discrimination score by a discrimination model, and identifies and authenticates an authentication candidate who has provided the model based on comparison between discrimination models of the discrimination score. The identification of the authentication candidate based on the comparison between discrimination models in the processing unit is based on one or both of the discrimination score, or the decoding accuracy obtained from the discrimination score.
- the electroencephalogram is an electroencephalogram associated with cognitive processing for the stimulation event.
- the discriminant model for estimating the brain information is characterized by optimizing the weighting coefficient of the discriminant model so that the target can be decoded from the electroencephalogram in the target selection task.
- the authentication apparatus further includes a stimulus presentation unit that presents a plurality of stimulation events and a target presentation unit that presents which one of the plurality of stimulation events is a target, and the electroencephalogram obtained by the electroencephalograph is A brain wave related to a cognitive task of selecting a target.
- the stimulus presenting unit presents a stimulus event having a different figure or a stimulus event having a different position only.
- the identification of the authentication candidate based on the comparison between the discrimination models of the processing unit is obtained from the rank of the average value for the target in the standardized data of the cumulative discrimination score for each stimulus type, or the rank by the comparison between the discrimination models, or the discrimination score It is characterized in that the highest rank is specified by the rank of the decoding accuracy between the discrimination models or the average value of the ranks of both the discrimination score and the decoding accuracy.
- the apparatus of the present invention is an authentication apparatus using an electroencephalogram, in which an electroencephalograph and a deciphering model for estimating brain information for each authentication candidate stored in advance for electroencephalograms for a plurality of stimulation events obtained by the electroencephalograph And a processing unit that obtains the decryption accuracy, obtains the decryption accuracy, and identifies and authenticates the authentication candidate who has provided the model formula based on the decryption accuracy.
- the method of the present invention is an authentication method using an electroencephalogram, which measures an electroencephalogram of an authentication target person caused by a plurality of stimulation events, and stores brain information for each authentication candidate stored in advance with respect to the electroencephalogram.
- a discrimination score is obtained by a discrimination model for estimating the discrimination score, and an authentication candidate who provides the discrimination model is identified and authenticated based on comparison between discrimination models of the discrimination score.
- the system of the present invention is an electroencephalogram authentication system, which is an electroencephalogram authentication system comprising a stimulus presentation device, an electroencephalograph, and a processing device that processes electroencephalogram data from the electroencephalograph, and the stimulus presentation device Presents a plurality of stimulation events each consisting of a target and a non-target each time, the electroencephalograph measures the electroencephalogram of the person to be authenticated caused by the plurality of stimulation events, and the processing device On the other hand, for each authentication candidate stored in advance, a discrimination score is obtained by a discrimination model for estimating brain information, and the authentication candidate who provided the model is identified and authenticated based on the discrimination score. And
- the program of the present invention estimates a brain information for each authentication candidate stored in advance for a computer, a stimulus presentation means for presenting a plurality of stimulus events each time, and an electroencephalogram data immediately after the stimulus presentation.
- This is a program for obtaining a discrimination score from a discrimination model to be used, and functioning as a processing means for identifying and authenticating an authentication candidate who has provided the model based on the discrimination score and a presentation means for presenting an authentication result.
- the present invention by analyzing the electroencephalogram, it is possible to perform authentication with confidentiality and high reliability.
- the present invention it was possible to achieve 95% authentication accuracy in a short time (5 blocks), which could not be realized by the conventional authentication using an electroencephalogram.
- the accuracy can be further improved by appropriately using the kind of stimulus type, the figure discrimination action task, and the position discrimination task.
- the person to be authenticated and the authentication candidate do not need to memorize the password, so that it is effective not only for healthy persons but also for persons with disabilities. Moreover, since it is authentication by an electroencephalogram, there is no possibility of duplication.
- the present invention relates to biometric authentication based on electroencephalogram analysis, and is a personal authentication technology that enables individual identification using only electroencephalogram.
- the present invention is to identify an individual by analyzing the brain response to the presentation of a stimulus event, focusing on the brain activity, particularly the event-related potential, which is the type of brain wave recorded on the scalp. Specifically, it mainly includes elements of stimulus event presentation, electroencephalogram measurement, electroencephalogram data analysis, personal identification, and identification result presentation.
- the apparatus of the present invention includes an electroencephalogram measurement headgear, a computer for experiment control and data analysis, and a subject sub-monitor (display screen (for stimulus presentation)).
- FIG. 1 is a schematic diagram for explaining personal identification by electroencephalogram decoding for the personal authentication method of the present invention.
- personal authentication by brain waves is performed in the following steps (A), (B), and (C).
- A The brain wave of a person whose ID is unknown is measured.
- B Based on the discrimination model of each person registered in the database by pre-registration, the discrimination score or further decoding process is advanced to calculate the decoding accuracy.
- C The person who provided the maximum model is specified based on one or both of the discrimination score and the decoding accuracy.
- FIG. 1 will be described in detail.
- the hatena mark indicates that the person to be authenticated is a person whose ID is unknown. Faces with different contours indicate persons registered in the database by pre-registration.
- the decoding accuracy is calculated by the discrimination model formula (also referred to as a discrimination model) of these persons, the decoding accuracy becomes 0.50, 0.75, 0.98, 0.43, 0.28, and a maximum value of 0.2.
- the person who provided the 98 model round outline
- a two-class pattern identification model can be used, for example, a model expression based on linear discriminant analysis, a support vector machine, or the like.
- step (A) what is provided by the user (person for personal authentication) for personal authentication is a “target (hereinafter also referred to as target) selection task” that requires cognitive judgment regarding graphic discrimination or position discrimination. This is the EEG data during the performance.
- FIG. 2 shows an example of graphic discrimination of an action task for personal identification by electroencephalogram in the present embodiment.
- the target selection task related to figure discrimination when a predetermined target appears among a plurality of types of picture cards presented in time series, it is only conscious of that.
- FIG. 3 shows an example of the position determination of the action task for personal identification by the electroencephalogram in the present embodiment.
- ⁇ It is better to change the target multiple times to raise the security level.
- the security level can be increased by providing that both types of tasks succeed in decoding with high accuracy.
- the contents of the brain activity are different between the graphics discrimination and the location discrimination, the models required for decoding the electroencephalogram are different. Therefore, the combined use further increases the security level.
- the personal authentication method of the present invention characteristic data of an individual's electroencephalogram for personal identification is registered in advance. That is, the personal authentication method of the present invention includes a “pre-registration for personal identification” step.
- FIG. 4 is a diagram schematically showing an apparatus and method according to this embodiment.
- the stimulus presentation display screen is shown to the individual authentication target person, and the scalp electroencephalogram of the personal authentication target person is measured by an electroencephalograph (in the figure, electroencephalogram). Measurement is recorded by the amplifier 4).
- An individual subject to personal authentication wears an electroencephalograph electrode 3 for measuring an electroencephalogram on the head.
- an electroencephalograph electrode 3 for measuring an electroencephalogram on the head.
- a head cap to which an electroencephalograph electrode is fixed is used.
- Various visual stimuli are presented on the display screen (monitor). EEG data is obtained with an electroencephalograph.
- the raw electroencephalogram data is analyzed by a processing device such as the computer 6 and the result of personal authentication is displayed on a display screen or the like.
- a processing device such as the computer 6
- the result of personal authentication is displayed on a display screen or the like.
- thick arrows are shown from the head where the electroencephalogram electrode is located to the electroencephalogram amplifier 4, and from the electroencephalogram amplifier 4 to the computer 6, but this schematically shows that signals are transmitted by wire or wirelessly. It is illustrated.
- FIG. 2 is a diagram schematically showing the presentation of the stimulus event for figure discrimination and the response of the brain wave of the person for personal authentication in response to the passage of time in the present embodiment.
- a stimulus event also called an alerting event or a test stimulus event
- a simple graphic is presented to the person to be personally authenticated one event (one sheet) at a time.
- the brain wave of the person who is subject to personal authentication who saw this is measured, and the brain wave is analyzed by a brain wave analysis processing device such as a computer.
- the stimulus event is a symbol, an illustration, a picture, a photograph, or the like. This is a figure discrimination problem common to both processes of pre-registration and personal authentication described later.
- Each subject is taught a cognitive task that teaches subjects with one of a plurality of stimulation events, for example, eight figures, and counts the number of presentations in the head if the sequentially presented stimulation events are the target.
- the brain wave at that time is measured.
- Electroencephalograms from single or multiple electrodes placed on the scalp around the top of the head are measured. The measurement is performed according to the following procedure. The electrode positions were arranged at 8 locations based on the standard electrode placement method (10% method).
- Visual stimuli For example, visual stimuli (simple figures in FIG. 2) are continuously presented on a computer display screen or the like in a pseudo-random order like a picture-story show (see FIG. 2).
- the present invention is not limited to visual stimuli, and can be implemented by auditory stimuli (sound, voice, music, etc.), tactile stimuli, odor stimuli, and the like.
- one of a plurality of visual stimuli (in FIG. 2, a plurality of geometric figures (triangle, rhombus, star, bi-ellipse, quadrangle, circle, heart, clover, etc.)) Tell the person to be personally identified as a “target”.
- the person subject to personal authentication teaches and counts the number of presentations in the head only for the target stimulus.
- a series of stimulus presentations targeting a specific visual stimulus is expressed as one game.
- each pseudo-stimulation presentation of all visual stimuli is one block, 5 blocks are presented continuously in each game. In this case, every visual stimulus is presented five times.
- the presentation time for each visual stimulus is 750 milliseconds, and after the blank of 250 milliseconds, the next visual stimulus is presented.
- the electroencephalogram data shown in the lower part of FIG. 2 is an example of electroencephalogram data corresponding to each visual stimulus when “target” is taught as a star shape and the subject is presented with the visual stimulus and counted.
- the electroencephalogram data for the visual stimulus of the target (star shape) has a greater response to the electroencephalogram than the electroencephalogram data for the visual stimulus of the non-target (triangle, rhombus, and double ellipse).
- the electroencephalogram data for the target visual stimulus out of the electroencephalogram data is larger in many cases when the response of the electroencephalogram is larger than the electroencephalogram data for the non-target stimulus. Yes.
- the data acquired in the target selection task includes a response immediately after the target presentation and a response immediately after the non-target presentation.
- EEG for stimuli visual stimuli, auditory stimuli, olfactory stimuli, tactile stimuli, etc.
- an electroencephalogram potential called an event-related potential (or event-related electroencephalogram) is used for stimulation.
- the event-related potential is a transient electroencephalogram that occurs in conjunction with the timing of the occurrence of an external or internal event that affects the cognitive process, and P300 (positive potential change 300 ms after stimulus presentation). and so on.
- Linear discriminant analysis is performed on the two types of tagged data of the electroencephalogram data immediately after the target presentation and the electroencephalogram data immediately after the non-target presentation to generate a discriminant model formula.
- the weighting coefficient of the model formula is optimized so that an output can be made such that a value is positive for the target electroencephalogram data and a value is negative for the non-target electroencephalogram data.
- the weighting coefficient of the discriminant model formula is stored in the storage unit of the apparatus according to the present embodiment for each individual and for each stimulation event. The same data of a plurality of persons subject to personal authentication is accumulated in the storage unit and used as a database.
- the discrimination score (y) for one presentation of each visual stimulus is calculated by a linear discriminant function expressed by the following equation.
- x is a value of electroencephalogram data (voltage) at a certain time of a certain channel.
- the type of x includes a type (n) obtained by multiplying the number of channels (the number of channels corresponding to the number of measurement points to obtain brain wave data at a plurality of measurement points on the scalp of the subject's head) and the data point.
- the weighting coefficient w and the constant term c for each electroencephalogram data can be obtained by linear discriminant analysis.
- electroencephalogram measurement is performed in the same manner as described in [(a) Stimulus event presentation and pre-registration electroencephalogram measurement in pre-registration]. That is, the subject is taught to a subject by targeting one of a plurality of stimulus events, for example, eight figures, and the number of presentations in the head is counted if the stimulus event presented sequentially is the target, The test is performed on each subject, and the electroencephalogram at that time is measured.
- the “decoding accuracy” is calculated according to how many games are successfully decoded in all games (for example, 8 games). Decoding accuracy varies from 0/8 to 8/8. In general, the decoding accuracy increases with an increase in the number of repetitions of stimulus presentation (number of blocks). However, even if the final result is the same, the subject may be high from the beginning or high in the middle depending on the subject, so the process of increasing the decoding accuracy with the increase in the number of blocks is different. Therefore, as a personal authentication system, it is preferable to use a block average of decoding accuracy as detailed data of decoding accuracy.
- FIG. 5 shows a discrimination model formula for each registrant registered in the database for each stimulus ((1) to (8) in the figure) of the person to be authenticated (subject). It is a graph which shows the transition of the discrimination
- the vertical axis in FIG. 5 indicates the accumulated discrimination score, and the horizontal axis indicates the time elapsed by the number of blocks (from 1 block to 5 blocks).
- the cumulative discrimination score of the electroencephalogram data for each stimulus in each block when accumulating the cognitive task of counting the target is accumulated, the cumulative discrimination score is 6 for the specific stimulus event (4) in the accumulation of 5 blocks. It can be seen that the maximum value is exceeded. It can be seen that the cumulative discriminant scores of the other stimulation events (1), (2), (3), (5),... (8) are low and stable as the blocks are accumulated.
- FIG. 6 is a diagram showing a radar chart showing discrimination scores of eight stimulation events in the first game in which the target (target) is taught as ID-1 (for example, a star figure).
- FIG. 6 shows an example of successful decoding in which the discriminated score of ID-1 is larger and maximum than the other discriminant scores, and therefore the taught target and the discriminant score of the maximum match.
- FIG. 7 is a radar chart showing the discrimination scores of the eight stimulation events in the second game in which the target (target) is taught as ID-2 (for example, a triangle).
- FIG. 7 shows an example of decoding failure in which the discriminated score of ID-4 is larger and the maximum than the other discriminant scores, and thus the taught target and the discriminant score maximum do not match.
- the number of successful decryptions in which the taught target and the largest discrimination score match is the total number (the number of decryption failures and the number of successful decryptions in which the taught target and the largest discrimination score do not match).
- the value divided by (sum of) was taken as the decoding accuracy.
- the decoding accuracy is 0/8 to 8/8.
- FIG. 8 is an example of a graph showing a transition of decoding accuracy in the present embodiment.
- the vertical axis in FIG. 8 indicates the decoding accuracy, and the horizontal axis indicates the time passage by the number of blocks (1 to 5 blocks).
- FIG. 8 shows an example in which the decoding accuracy is calculated from the cumulative discrimination score, but the decoding accuracy can also be obtained from the discrimination score of the electroencephalogram data for each block when performing the cognitive task of counting the target. .
- the decoding accuracy in the block can be calculated from the cumulative discriminant score of the electroencephalogram data when the block is executed.
- As a method for digitization including the transition of the decoding accuracy for example, it is possible to digitize by using an integral value of a curve formed by the decoding accuracy. In the first to third embodiments, a discrimination model using the height of this value as an index. Compare between.
- the decoding accuracy generally increases as the number of blocks increases.
- the number of blocks can be appropriately set in the electroencephalogram measurement for authentication.
- the discriminant score is calculated using the discriminant model formula for each of the plurality of registrants registered in the database, and then the discriminant score is determined for each stimulus type. Accumulation over time (number of blocks), and when the stimulus type with the maximum cumulative discrimination score obtained from the subject's brain wave data in each block matches the stimulus type set as the target, it is counted as "decoding success" Then, “decoding accuracy” is calculated according to how many games are successfully decoded in all games (for example, 8 games).
- step (B) Identifies the registrant who provided the discriminant model formula with the highest decoding accuracy among the decoding accuracy obtained in step (B). It authenticates that the person to be authenticated is the specified registrant among the authentication candidates registered in advance.
- the subject of authentication asks the subject of authentication to input the ID of the target candidate that has been pre-registered, such as alphabets and numbers, and determines that the ID matches the registrant identified with the highest decoding accuracy. And can be authenticated.
- the stimulus is mainly described for the visual stimulus.
- an auditory stimulus instead of the visual stimulus, an auditory stimulus, a contact stimulus, an olfactory stimulus, or the like may be given to measure and analyze the corresponding electroencephalogram.
- the target selection task related to position determination some of a plurality of (8 in the example) positions on the screen are randomly flashed, and one of them is a target. Even in the position determination selection task, the target is executed multiple times at different positions.
- the cognitive function of the brain differs depending on whether the action task is a figure discrimination or a position discrimination. Unlike the discrimination of different figures and pictures, the position discrimination is expected to be faster because the brain has the function of discrimination at high speed.
- 9 and 10 are examples of results of personal authentication experiments using brain waves.
- FIG. 9 is a diagram showing how the decoding accuracy using the discriminant model formula of the candidate subject changes according to the accumulation of blocks. It can be seen that when the discriminant model formula of No. 17 candidate subject is applied, the value of the decoding accuracy is the maximum as compared with the case where discriminant model formulas of other candidate subjects are applied. Looking at the transition of the decoding accuracy, it can be seen that when the number of blocks is increased and the decoding accuracy is accumulated, the decoding accuracy becomes the maximum in the case of No. 17 with the accumulation of 3 blocks or more.
- FIG. 10 is a diagram showing the decoding accuracy for discrimination model providers No. 1 to No. 20 based on the result of FIG.
- the personal authentication focused on the decryption accuracy has been described.
- the discrimination score (cumulative) is calculated for each stimulus type. By utilizing this, in the present embodiment, the height of the discriminant score and cumulative discriminant score value related to each stimulus type is compared and authenticated among a plurality of discriminant models.
- a single target teaching game in the target stimulation task (one game in the first embodiment) will be specifically described as one decoding attempt.
- a specific stimulus type is targeted by measuring the brain waves of the person who is seeking authentication and using each discrimination model of multiple registrants registered in the database for the brain wave data
- a discrimination score obtained by accumulating all blocks is calculated for each stimulus type.
- the cumulative discriminant score for the stimulus species thus obtained is standardized so that the average value is 0 and the standard deviation is 1, and the standard cumulative discriminant score is calculated.
- the average value of the standard cumulative discrimination score for the target is determined. This operation is performed for each discrimination model registered in the database. From the result, the person who has provided the discrimination model with the maximum standard cumulative discrimination score for the target is estimated as the person and authenticated.
- a ranking model that ranks from the highest cumulative discrimination score for each decoding trial and compares the average value of ranking data for all decoding trials is provided to provide the highest (1st) discrimination model for authentication.
- the authentication work can be performed even with a decoding attempt targeting one kind of stimulation species or a part thereof without performing a decoding attempt targeting all stimulation types.
- the method of determining the most suitable discrimination model using the average value after obtaining the standard cumulative discrimination score for the decoding attempts (eight decoding attempts) targeted at each of the maximum 8 types of stimuli is as follows: , The authentication accuracy becomes higher.
- the block that is the number of repetitions of the stimulation group can be shortened, and authentication can be performed from at least one block (8 types of stimulation including the target are presented once).
- the highest rank (first place) is determined by calculating the average of the ranks in which the average of the transitions in decoding accuracy is arranged in descending order and the rank in which the cumulative discrimination score is arranged in descending order. Can also be authenticated.
- the person to be authenticated is one of those registered in the database
- authentication focusing on the electroencephalogram in order to identify who the person is A method was presented.
- a person who is not registered in the database may be subject to authentication.
- it is considered effective to pay attention to the decoding accuracy as a technique for eliminating unregistered persons.
- the accuracy of decoding the cognitive task should be very high using his own model (90% or more based on previous knowledge). .
- the decoding accuracy threshold is set to 0.7 (70%) and the decoding accuracy falls below this threshold, it is denied, that is, the person is not registered in the database. it can. If this threshold is set too low, there is an increased risk that a non-registrant in the database will be mistakenly authenticated as someone in the registrant. Conversely, if the threshold is set too high, the database registrant will be The risk of accidentally denying you as a non-registered person increases. Therefore, the threshold should be set according to which risk is desired to be eliminated. As a method of authentication / denial by such threshold setting, it is also possible to use the cumulative discrimination score (average value) for the target as an index instead of the decoding accuracy. It is also possible to set a threshold based on both the index and the decoding accuracy.
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Psychiatry (AREA)
- Psychology (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
Description
(A)ID不明の人物の脳波を計測する。
(B)事前登録によりデータベース登録された各人物の判別モデルによって判別得点又はさらに解読処理を進めて解読精度を算出する。
(C)判別得点又は解読精度のいずれか一方、又は両方に基づき、最大のモデルを提供した人物を特定する。
本実施の形態を図を参照して以下説明する。
図4は、本実施の形態による装置及び方法を模式的に示す図である。図4の個人認証対象者への刺激提示1で図示されるように、刺激提示用の表示画面を個人認証対象者に見せて、個人認証対象者の頭皮上脳波を脳波計(図中、脳波アンプ4)により計測記録する。個人認証対象者は、脳波を測定するための脳波計電極3を頭に装着する。例えば、脳波計電極を固定したヘッドキャップを用いる。表示画面(モニター)に様々な視覚刺激を提示する。脳波計により脳波生波形のデータを得る。脳波生波形のデータをコンピューター6等の処理装置で解析処理して、個人認証の判定結果を表示画面等に示す。図4において、脳波電極の位置する頭部から脳波アンプ4に、そして、脳波アンプ4からコンピューター6に、太い矢印を図示したが、これは有線又は無線により信号が伝達されることを模式的に図示したものである。
判別モデル式の重み付け係数を、個人毎に、かつ刺激事象毎に、本実施の形態の装置の記憶部に、記憶しておく。複数の個人認証対象者の同様のデータを記憶部に蓄積しデータベースとする。
第1の実施の形態では、図2に示すような、脳波による個人識別用行動課題が図形判別である例で説明したが、本実施の形態では、図3に示すような、脳波による個人識別用行動課題として位置判別を用いる。本実施の形態は、位置判別に関する視覚刺激を扱う以外は、第1の実施の形態と、同様である。
第1及び第2の実施の形態では、脳波による個人識別用行動課題が図形判別である例と位置判別である例を説明したが、本実施の形態では、図形判別の課題と位置判別の課題の両方について、判別モデル式の事前登録と、個人認証を行う。また、個人認証の際には、最初に行った課題において2位以下と解読精度の差が生じている場合には、もう一方の課題を実施しないというように、適宜、図形判別の課題と位置判別の課題のいずれか一方を省略してもよい。
第1~3の実施の形態では、解読精度に着目した個人認証について説明したが、本実施の形態は、解読精度に換えて、解読精度を求める際に算出した判別得点自体に着目した個人認証を行うものである。図5に示すように、判別得点(累積)は各刺激種単位で算出される。これを利用して、本実施の形態では、各刺激種に係る判別得点や累積判別得点の値の高さを、複数の認証候補者複数の判別モデル間で比較して認証する。
3 被験者の脳波計電極
4 脳波アンプ
6 コンピューター
Claims (11)
- 脳波による認証装置であって、
脳波計と、
前記脳波計により得られる複数の刺激事象に対する脳波に対して、予め蓄積されている認証候補者毎の、脳情報を推定する判別モデルによって判別得点を求め、前記判別得点の判別モデル間比較に基づき、前記モデルを提供した認証候補者を特定して認証する処理部と、
を備えることを特徴とする脳波による認証装置。 - 前記処理部における、判別モデル間比較に基づく認証候補者の前記特定は、前記判別得点、もしくは前記判別得点から求めた解読精度の、いずれか一方又は両方に基づくことを特徴とする請求項1記載の脳波による認証装置。
- 前記脳波は、前記刺激事象に対する認知的処理に関連する脳波であることを特徴とする請求項1記載の脳波による認証装置。
- 前記脳情報を推定する判別モデルは、標的選択課題における脳波から標的を解読できるように、判別モデルの重み付け係数を最適化したものであることを特徴とする請求項1記載の脳波による認証装置。
- 前記認証装置は、さらに、複数の刺激事象を提示する刺激提示部及び前記複数の刺激事象のうちのいずれが標的であるかを提示する標的提示部を備え、
前記脳波計により得られる前記脳波は、標的を選択する認知課題に関連する脳波であることを特徴とする請求項1記載の脳波による認証装置。 - 前記刺激提示部は、図形の異なる刺激事象、又は位置のみ異なる刺激事象を提示することを特徴とする請求項5記載の脳波による認証装置。
- 前記処理部の判別モデル間比較に基づく認証候補者の前記特定とは、各刺激種に対する累積判別得点の標準化データにおける標的に対する平均値の、判別モデル間比較による順位、もしくは前記判別得点から求めた解読精度の、判別モデル間比較による順位、もしくは前記判別得点と前記解読精度の両方の累積値の順位の平均値により、最上位を特定することであることを特徴とする請求項1記載の脳波による認証装置。
- 脳波による認証装置であって、
脳波計と、
前記脳波計により得られる複数の刺激事象に対する脳波を、予め蓄積されている認証候補者毎の、脳情報を推定する解読モデルによって解読処理して、解読精度を求め、前記解読精度に基づき前記モデル式を提供した認証候補者を特定して認証する処理部と、
を備えることを特徴とする脳波による認証装置。 - 複数の刺激事象により生起される認証対象者の脳波を計測し、
前記脳波に対して、予め蓄積されている認証候補者毎の、脳情報を推定する判別モデルによって判別得点を求め、前記判別得点の判別モデル間比較に基づき、前記判別モデルを提供した認証候補者を特定して認証することを特徴とする脳波による認証方法。 - 刺激提示装置と、脳波計と、該脳波計からの脳波データを処理する処理装置とを備える脳波による認証システムであって、
前記刺激提示装置は、標的及び非標的からなる複数の刺激事象を、それぞれ複数回提示し、
前記脳波計は、複数の刺激事象により生起される認証対象者の脳波を計測し、
前記処理装置は、前記脳波に対して、予め蓄積されている認証候補者毎の、脳情報を推定する判別モデルによって判別得点を求め、前記判別得点に基づき、前記モデルを提供した認証候補者を特定して認証することを特徴とする脳波による認証システム。 - コンピューターを、
複数の刺激事象をそれぞれ複数回提示する刺激提示手段と、
該刺激提示直後の脳波データに対して、予め蓄積されている認証候補者毎の、脳情報を推定する判別モデルによって判別得点を求め、前記判別得点に基づき、前記モデルを提供した認証候補者を特定して認証する処理手段と、
認証結果を提示する提示手段として機能させるためのプログラム。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US15/527,387 US10470690B2 (en) | 2014-11-21 | 2015-11-17 | Authentication device using brainwaves, authentication method, authentication system, and program |
JP2016560223A JP6373402B2 (ja) | 2014-11-21 | 2015-11-17 | 脳波による認証装置、認証方法、認証システム及びプログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2014-236574 | 2014-11-21 | ||
JP2014236574 | 2014-11-21 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016080366A1 true WO2016080366A1 (ja) | 2016-05-26 |
Family
ID=56013902
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2015/082178 WO2016080366A1 (ja) | 2014-11-21 | 2015-11-17 | 脳波による認証装置、認証方法、認証システム及びプログラム |
Country Status (3)
Country | Link |
---|---|
US (1) | US10470690B2 (ja) |
JP (1) | JP6373402B2 (ja) |
WO (1) | WO2016080366A1 (ja) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107196809A (zh) * | 2017-07-07 | 2017-09-22 | 南京邮电大学 | 基于脑电特征的身份认证方法及认证系统 |
US10482227B2 (en) * | 2017-03-28 | 2019-11-19 | International Business Machines Corporation | Electroencephalography (EEG) based authentication |
CN111295673A (zh) * | 2017-11-08 | 2020-06-16 | 国际商业机器公司 | 神经反应检测器 |
JP2021121930A (ja) * | 2016-07-11 | 2021-08-26 | アークトップ リミテッド | ブレイン・コンピュータ・インターフェースを提供するための方法およびシステム |
WO2024058045A1 (ja) * | 2022-09-15 | 2024-03-21 | 株式会社Jvcケンウッド | 情報処理装置および方法、プログラム、認証装置 |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10169560B2 (en) * | 2016-02-04 | 2019-01-01 | Lenovo Enterprise Solutions (Singapore) Pte. Ltd. | Stimuli-based authentication |
CN108108603A (zh) * | 2017-12-04 | 2018-06-01 | 阿里巴巴集团控股有限公司 | 登录方法及装置和电子设备 |
EP3960094A4 (en) | 2019-04-26 | 2023-01-18 | Osaka University | REFERENCE STIMULUS |
CN114707132B (zh) * | 2021-05-20 | 2023-04-18 | 广州番禺职业技术学院 | 基于情感语音的脑波加解密方法和系统 |
CN116421202B (zh) * | 2023-02-13 | 2024-04-02 | 华南师范大学 | 基于脑电快速周期性视觉刺激奇异范式的大脑视觉功能快速检测方法、装置及存储介质 |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004248714A (ja) * | 2003-02-18 | 2004-09-09 | Kazuo Tanaka | 生体信号を用いた認証方法および認証装置 |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5363858A (en) * | 1993-02-11 | 1994-11-15 | Francis Luca Conte | Method and apparatus for multifaceted electroencephalographic response analysis (MERA) |
US7689272B2 (en) * | 2001-06-07 | 2010-03-30 | Lawrence Farwell | Method for brain fingerprinting, measurement, assessment and analysis of brain function |
JP2005293209A (ja) | 2004-03-31 | 2005-10-20 | Ntt Data Corp | 個人認証装置、情報端末、個人認証方法、およびプログラム |
CN101558997A (zh) * | 2009-05-27 | 2009-10-21 | 江西蓝天学院 | 基于二阶盲辨识的脑电信号识别方法 |
JP5414039B2 (ja) | 2009-06-01 | 2014-02-12 | 独立行政法人産業技術総合研究所 | 脳情報の表示方法及び装置 |
JP5544620B2 (ja) | 2010-09-01 | 2014-07-09 | 独立行政法人産業技術総合研究所 | 意思伝達支援装置及び方法 |
JP5472746B2 (ja) | 2010-09-28 | 2014-04-16 | 独立行政法人産業技術総合研究所 | 意思伝達支援装置及び方法 |
JP6146760B2 (ja) | 2012-02-28 | 2017-06-14 | 国立研究開発法人産業技術総合研究所 | 序列化装置、序列化方法及びプログラム |
US20140020089A1 (en) * | 2012-07-13 | 2014-01-16 | II Remo Peter Perini | Access Control System using Stimulus Evoked Cognitive Response |
CN105051647B (zh) * | 2013-03-15 | 2018-04-13 | 英特尔公司 | 基于生物物理信号的搜集时间和空间模式的大脑计算机接口(bci)系统 |
-
2015
- 2015-11-17 JP JP2016560223A patent/JP6373402B2/ja active Active
- 2015-11-17 WO PCT/JP2015/082178 patent/WO2016080366A1/ja active Application Filing
- 2015-11-17 US US15/527,387 patent/US10470690B2/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004248714A (ja) * | 2003-02-18 | 2004-09-09 | Kazuo Tanaka | 生体信号を用いた認証方法および認証装置 |
Non-Patent Citations (1)
Title |
---|
YOSHIKO NAKAMURA ET AL.: "Ninchi Kino ni Chakumoku shita Noha BMI Gijutsu no Kaihatsu", THE JAPANESE SOCIETY FOR COGNITIVE PSYCHOLOGY DAI 11 KAI TAIKAI HAPPYO RONBUNSHU, 11TH, 29 June 2013 (2013-06-29), pages 32 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2021121930A (ja) * | 2016-07-11 | 2021-08-26 | アークトップ リミテッド | ブレイン・コンピュータ・インターフェースを提供するための方法およびシステム |
JP7365373B2 (ja) | 2016-07-11 | 2023-10-19 | アークトップ リミテッド | ブレイン・コンピュータ・インターフェースを提供するための方法およびシステム |
US10482227B2 (en) * | 2017-03-28 | 2019-11-19 | International Business Machines Corporation | Electroencephalography (EEG) based authentication |
US11500973B2 (en) | 2017-03-28 | 2022-11-15 | International Business Machines Corporation | Electroencephalography (EEG) based authentication |
CN107196809A (zh) * | 2017-07-07 | 2017-09-22 | 南京邮电大学 | 基于脑电特征的身份认证方法及认证系统 |
CN107196809B (zh) * | 2017-07-07 | 2020-04-14 | 南京邮电大学 | 基于脑电特征的身份认证方法及认证系统 |
CN111295673A (zh) * | 2017-11-08 | 2020-06-16 | 国际商业机器公司 | 神经反应检测器 |
JP2021502135A (ja) * | 2017-11-08 | 2021-01-28 | インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Machines Corporation | ニューロン応答に基づいた人間検出の方法、システム及びコンピュータ・プログラム |
JP7179059B2 (ja) | 2017-11-08 | 2022-11-28 | キンドリル・インク | ニューロン応答に基づいた人間検出の方法、システム及びコンピュータ・プログラム |
CN111295673B (zh) * | 2017-11-08 | 2024-01-12 | 勤达睿公司 | 神经反应检测器 |
WO2024058045A1 (ja) * | 2022-09-15 | 2024-03-21 | 株式会社Jvcケンウッド | 情報処理装置および方法、プログラム、認証装置 |
Also Published As
Publication number | Publication date |
---|---|
US20170325720A1 (en) | 2017-11-16 |
US10470690B2 (en) | 2019-11-12 |
JPWO2016080366A1 (ja) | 2017-09-14 |
JP6373402B2 (ja) | 2018-08-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6373402B2 (ja) | 脳波による認証装置、認証方法、認証システム及びプログラム | |
Lin et al. | Brain password: A secure and truly cancelable brain biometrics for smart headwear | |
Nakamura et al. | In-ear EEG biometrics for feasible and readily collectable real-world person authentication | |
US9811730B2 (en) | Person identification using ocular biometrics with liveness detection | |
Makowski et al. | DeepEyedentificationLive: Oculomotoric biometric identification and presentation-attack detection using deep neural networks | |
Kinnunen et al. | Towards task-independent person authentication using eye movement signals | |
Juhola et al. | Biometric verification of a subject through eye movements | |
CN107437011A (zh) | 基于脑电信号的身份识别的方法和装置 | |
Debie et al. | Session invariant EEG signatures using elicitation protocol fusion and convolutional neural network | |
Rathi et al. | A novel approach for designing authentication system using a picture based P300 speller | |
Arias-Cabarcos et al. | Inexpensive brainwave authentication: new techniques and insights on user acceptance | |
Casanova et al. | User recognition based on periocular biometrics and touch dynamics | |
Serwadda et al. | fNIRS: A new modality for brain activity-based biometric authentication | |
EP2831810A1 (en) | Person identification using ocular biometrics | |
Rathi et al. | Towards a role-based authentication system based on SSVEP-P300 hybrid brain–computer interfacing | |
Arias-Cabarcos et al. | Performance and usability evaluation of brainwave authentication techniques with consumer devices | |
Farid et al. | Biometric authentication for dementia patients with recurrent neural network | |
Kanaga et al. | An experimental investigations on classifiers for Brain Computer Interface (BCI) based authentication | |
Rathi et al. | Authentication framework for security application developed using a pictorial P300 speller | |
Rathi et al. | A comparative study of classification methods for designing a pictorial P300-based authentication system | |
Awad et al. | Cognitive biometrics for user authentication | |
Chandler et al. | Biometric measurement of human emotions | |
Kaur et al. | A study of EEG for enterprise multimedia security | |
Pham | EEG-based person authentication for security systems | |
Cherepovskaya et al. | An evaluation of biometrie identification approach on low-frequency eye tracking data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15860307 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2016560223 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 15527387 Country of ref document: US |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 15860307 Country of ref document: EP Kind code of ref document: A1 |