WO2007125478A2 - Procédé et appareil d'identification biométrique d'un individu - Google Patents

Procédé et appareil d'identification biométrique d'un individu Download PDF

Info

Publication number
WO2007125478A2
WO2007125478A2 PCT/IB2007/051508 IB2007051508W WO2007125478A2 WO 2007125478 A2 WO2007125478 A2 WO 2007125478A2 IB 2007051508 W IB2007051508 W IB 2007051508W WO 2007125478 A2 WO2007125478 A2 WO 2007125478A2
Authority
WO
WIPO (PCT)
Prior art keywords
person
model
feature vector
point
curve
Prior art date
Application number
PCT/IB2007/051508
Other languages
English (en)
Other versions
WO2007125478A3 (fr
Inventor
Marijn C. Damstra
Alphons A. M. L. Bruekers
Sieglinde Neerken
Original Assignee
Koninklijke Philips Electronics N.V.
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Koninklijke Philips Electronics N.V. filed Critical Koninklijke Philips Electronics N.V.
Publication of WO2007125478A2 publication Critical patent/WO2007125478A2/fr
Publication of WO2007125478A3 publication Critical patent/WO2007125478A3/fr

Links

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C9/00—Individual registration on entry or exit
    • G07C9/30—Individual registration on entry or exit not involving the use of a pass
    • G07C9/32—Individual registration on entry or exit not involving the use of a pass in combination with an identity check
    • G07C9/37—Individual registration on entry or exit not involving the use of a pass in combination with an identity check using biometric data, e.g. fingerprints, iris scans or voice recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C2209/00—Indexing scheme relating to groups G07C9/00 - G07C9/38
    • G07C2209/60—Indexing scheme relating to groups G07C9/00174 - G07C9/00944
    • G07C2209/63—Comprising locating means for detecting the position of the data carrier, i.e. within the vehicle or within a certain distance from the vehicle
    • G07C2209/65—Comprising locating means for detecting the position of the data carrier, i.e. within the vehicle or within a certain distance from the vehicle using means for sensing the user's hand
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C9/00—Individual registration on entry or exit
    • G07C9/20—Individual registration on entry or exit involving the use of a pass
    • G07C9/22—Individual registration on entry or exit involving the use of a pass in combination with an identity check of the pass holder
    • G07C9/25—Individual registration on entry or exit involving the use of a pass in combination with an identity check of the pass holder using biometric data, e.g. fingerprints, iris scans or voice recognition
    • G07C9/257—Individual registration on entry or exit involving the use of a pass in combination with an identity check of the pass holder using biometric data, e.g. fingerprints, iris scans or voice recognition electronically

Definitions

  • the present invention relates to a method of biometric authentication/identification of a person.
  • the invention further relates to an apparatus and a system for biometric authentication/identification of a person, as well as to a computer program product.
  • Biometric authentication and identification are related concepts.
  • biometric authentication an alleged identity of a person being authenticated is validated using a feature vector derived from a physiological characteristic. This feature vector is established based on a physiological characteristic of the person such as e.g. a feature vector derived from a fingerprint of a finger of the person being authenticated.
  • the person also supplies an alleged identity, such as her name.
  • the feature vector is subsequently matched with an enrolled feature vector that was obtained from the person associated with the alleged identity. Based on this comparison the person is either authenticated as having the alleged identity or not.
  • identification also involves establishing a feature vector for the person, in this case the person being identified.
  • This feature vector is matched with enrolled feature vectors stored in a database.
  • identification this involves, at least one, but typically many, matches with enrolled feature vectors of different reference persons, until a (sufficient) match is found, or the database is exhausted.
  • authentication and identification involve a comparison of an acquired feature vector with an enrolled feature vector, authentication is often described as being a one-to-one comparison, whereas identification is a one-to-many comparison.
  • Biometric authentication/identification has gained popularity due to the nature of biometrics. Biometric characteristics are closely coupled to a single person and generally cannot be transferred to another person. Moreover biometric characteristics are generally difficult to duplicate/forge. Consequently biometric authentication/identification are generally considered to be more secure than e.g. the use of tokens for authentication/identification.
  • Biometric authentication/identification are also attracting the interest of counterfeiters, and other malicious parties that aim to thwart security systems. And while many people are under the impression that biometric systems are infallible, research has repeatedly shown that this is not the case, such as illustrated by e.g. "Impact of Artificial "Gummy” Fingers on Fingerprint Systems” by Tsutomu Matsumoto, et al. published in the Proceedings of SPIE Vol. #4677, Optical Security and Counterfeit Deterrence Techniques IV, 24-25 January 2002. This document illustrates the impact of imposter attacks using "gummy" fingers.
  • the present invention provides a method of biometric authentication/identification that comprises the step of using a predictable time- variant first physiological characteristic.
  • feature vectors may be derived from the first physiological characteristic that are representative of the first physiological characteristic of the person at a particular timestamp.
  • a feature vector representative of the height of the junction between epidermis and dermis may be derived from a finger of a person being authenticated/identified at the time of authentication/identification.
  • Such a model can be a generic model valid for all individuals, or a model that is specific to a particular reference person and was established during enrolment. This model can be used to establish an expected feature vector of the reference person at a particular timestamp, or alternatively can be used to establish a timestamp when the feature vector has a particular value.
  • the present invention proposes to use such a model to establish consistency of a first reference point comprising a first timestamp being the time of authentication/identification of the person and a first feature vector being a feature vector representative of the first physiological characteristic of the person at the first timestamp with the model of the reference person. It is important to note that the first reference point is associated with the person being authenticated/identified, whereas the model is associated with the reference person. After establishing consistency, the outcome thereof can be used in the decision making for the authentication/identification.
  • the present invention thus provides an enhanced method of authentication/identification in that the present invention can be used to reduce the number of false positives.
  • the present invention can be used to rule out that the person authenticating/identifying himself/herself is the reference person when the first feature vector is not consistent. Alternatively, when consistency is established this corroborates that the person can be the reference person.
  • the present invention may be used to counter imposter attacks using a gummy finger by determining the height of the junction between the epidermis and the dermis of a finger during authentication/identification and matching this with the model of the reference person. Not only will such a gummy finger generally not be consistent with the model at the time of authentication/identification, but moreover it will not show the same feature vector development over time as the actual finger. As a result this particular imposter attack can be thwarted using the present invention, thereby preventing a false positive authentication/identification decision.
  • establishing consistency of the first reference point with the model of the reference person comprises establishing a difference measure indicative of the extent to which the first reference point is inconsistent with the model of the reference person. Once the difference measure is established it is verified whether the difference measure is within a predetermined range. If the difference measure is within the predetermined range the first reference point is deemed consistent, if the difference measure is outside the predetermined range the first reference point is deemed inconsistent. Provided the underlying first physiological characteristic exhibits the same or a sufficiently similar development over time for all individuals, a single model of the feature vector of the reference person over time can be used for all reference persons. However this model will require personalization, by means of temporal synchronization to the specific individual.
  • the model is adapted for each individual reference person, using a second reference point comprising a second timestamp corresponding to a time of enrolment of the reference person and a second feature vector representative of the first physiological characteristic of the reference person at the time of enrolment.
  • the model may be adapted during enrolment and stored together with other enrolment data for the reference person.
  • the second reference point may be stored instead, allowing the generic model to be adapted on-the-fly during authentication/identification.
  • the difference measure corresponds to the distance between a first curve based on the model through the first reference point and a second curve based on the model passing through a second reference point associated with the reference person.
  • the second curve corresponds with the actual model over time
  • the first curve can be derived from the second curve by translating the second curve either along the time axis or along the feature vector axis/axes until the first curve passes through the first reference point.
  • the difference measure is established by selecting a first point located on one of the first curve and the second curve, and predicting a second point by projecting the first point along the time axis or along the feature vector axis/axes onto the other one of the first curve and the second curve. Consequently the second point has either the timestamp or the feature vector value in common with the first point. Subsequently the difference measure can be computed by computing the distance between the first and the second point.
  • the advantage of this manner of establishing a difference measure is its simplicity; it requires the projection of a single point onto a curve, and the computation of a single distance.
  • establishing the difference measure comprises a summation of the absolute value of the differences between the first and the second curve.
  • the first physiological characteristic is cyclic claim 8 provides a more preferable embodiment wherein the summation is a summation over one period of the first physiological characteristic, thereby simplifying the summation.
  • the method of authentication/identification is a multimodal authentication/identification method, wherein the first physiological characteristic is used as a modality for validating the outcome of the authentication/identification of a further modality.
  • This embodiment enables the combination of a stable primary modality that has sufficient discerning capabilities for authentication/identification, with a time- variant, possibly less discerning, secondary modality.
  • multiple secondary time- variant modalities are used to verify time consistency. This embodiment is particularly advantageous when using time based difference measures, as it involves combining scalar time based difference measures.
  • the method further comprises a step for storing the first reference point, the storing in part dependent on a positive authentication/identification decision. In this manner it is possible to acquire up to date enrolment data based on the first reference point of the person being authenticated/identified, provided the authentication/identification was (sufficiently) successful.
  • the method comprises a step for updating the model using the first reference point, provided that authentication/identification was positive.
  • This embodiment allows the present invention to track changes over time that are a-typical for the development of the first physiological characteristic of a particular reference person.
  • a good example of such an a-typical change is an increased drop in bone mineral density as a result of prolonged exposure to a particular medication.
  • Fig. 1 shows a graph of a generic model M ref of a time- variant first physiological characteristic over time.
  • Fig. 2 shows a graph of two models, Mi and M 2 for two individuals based on a generic model M re f and enrolment data of a first person and a second person respectively.
  • Fig. 3 illustrates the determination of a difference measure using the first reference point as the first point for predicting a second point.
  • Fig. 4 illustrates the determination of a difference measure using the second reference point as the first point for predicting a second point.
  • Fig. 5 illustrates an alternate determination of a difference measure using a first point for predicting a second point.
  • Fig. 6 illustrates two issues related to the modeling of cyclic first physiological characteristics.
  • Fig. 7 illustrates the determination of a further difference measure for a non- cyclic first physiological characteristic.
  • Fig. 8 illustrates the determination of a further difference measure for a cyclic first physiological characteristic.
  • Fig. 9 depicts a block diagram of a device for biometric authentication of a person according to the present invention.
  • Fig. 10 depicts a block diagram of a system for biometric identification of a person.
  • the human physiology offers a variety of time- variant physiological characteristics. Some of these characteristics exhibit a gradual change throughout the natural life of a person, such as the bone mineral density. Other processes exhibit a much faster rate of change, such as Cortisol levels in human blood.
  • Time-variant first physiological characteristics suitable for use by the present invention are not limited to e.g. monotonously increasing/decreasing time-variant characteristics.
  • transient or cyclic changing physiological characteristics having either a slow or fast rate of change.
  • any one of the following physiological characteristics may be used as the first physiological characteristic: the height of the junction between epidermis and dermis, - the height of a person, the dimensions of the ear, the dimensions of the nose, the collagen density in the dermis, the bone mineral density, and - the Cortisol levels in human blood.
  • a model may be generated of a feature vector representative of the first physiological characteristic over time.
  • Techniques for creating a model of the first physiological characteristic are well known to those skilled in the art, and may involve curve-fitting techniques, or piece-wise linear modeling techniques.
  • the advantage of using a piecewise linear model is that it not only allows fast forward calculations, i.e. from timestamp to feature vector, but moreover also enables fast reverse operation, i.e. from feature vector to timestamp.
  • a simple, yet practical example of a suitable time- variant physiological characteristic is the bone mineral density.
  • a model may be generated of a feature vector representative of the bone mineral density as a function of the age of a person. This type of model, having a time related input, is particularly useful when determining predictions of a feature vector. Alternatively a model of the age as a function of a feature vector representative of the bone mineral density may be generated, the latter being particularly useful when determining predictions of timestamps. Either model may be used individually or together in a method according to the present invention.
  • Fig. 1 shows a graph of a generic model M re f of a feature vector representative of a slowly changing time- variant first physiological characteristic over age.
  • the value Q of the time- variant first physiological characteristic decreases monotonously over time, until it is close to 0 at the age of 80.
  • the model depicted here involves a feature vector comprising a single component Q
  • the present invention is not limited to single component models. Models that involve multiple components may be simplified by considering each individual component as a separate physiological characteristic that is modeled over time. Consequently a physiological characteristic with multiple components can be interpreted as multiple independent single component physiological characteristics that are combined in the decision for authentication/identification. Although a straightforward simplification this way of working may not yield the best possible results.
  • a model of a feature vector representative of a physiological characteristic comprising N-components over time can also be interpreted as a curve in an N-dimensional space that represents the N-components over time.
  • a distance measure that considers all components simultaneously, e.g. by calculating an N-norm for verifying whether the first point is within a predetermined distance from the second point, that corresponds to a point on the curve described by the model. This added accuracy/reliability comes at the cost of added complexity.
  • a person In order to be able to authenticate/identify a person using biometrics, persons need to be enrolled. During enrolment the first physiological characteristic is measured resulting in a second feature vector as well as a second timestamp corresponding with the time of enrolment. The second feature vector and the time of enrolment form a second reference point. The second reference point for each reference person is stored together with a unique identifier for later reference. This unique identifier could be the identity of the reference person or a reference for establishing that identity, such as a reference-number or a (hyper)link.
  • enrolment data is generated in a similar manner as feature vectors during authentication/identification. More often the enrolment procedure involves multiple measurements and additional processing in order to obtain a better quality enrolment data.
  • enrolment data of (a) reference person(s) it is possible to adapt a generic model of the first physiological characteristic to create a model for each individual reference person, provided the model of the first physiological characteristic so requires.
  • Fig. 1 shows a general model representative of the first physiological characteristic M re f. Assume that the shape of the monotonous decrease shown is representative for humans in general. Further assume that the height of the curve is different for different individuals. During enrolment the general model will have to be scaled to compensate for differences between individuals.
  • Fig. 2 illustrates how the model of a first physiological characteristic M re f may be adapted for two individuals.
  • a first person is enrolled, age 36.
  • the value of the second feature vector acquired during the enrolment is represented by Qi.
  • the model Mi for this first person is established using this measurement and the general model representative of the first physiological characteristic M re f.
  • the general shape of the curve Mi is identical to that of the curve M re f but it is scaled such that the second reference point for the first person (36 (age), Qi) is on the curve, yielding the curve Mi representative of the first person.
  • age 17 a value Q 2 was measured for the second feature vector.
  • Fig. 2 shows an example that involves scaling along the axis of the feature vector, other first physiological characteristics may require scaling along the time- scale axis or other operations.
  • the above is merely an example of how a general model may be modified according to enrolment data. In case the model requires adaptation for each enrolled person, such model parameters or the model itself will have to be included with the enrolment data. In case no such customization is required a single model can be used that need not be part of the enrolment data.
  • a model of a feature vector representative of a suitable first physiological characteristic of a person it is possible to utilize the first physiological characteristic as well as the model of that characteristic in a method of biometric authentication/identification according to the present invention.
  • a model can be established, e.g. during enrolment of each reference person.
  • a first reference point comprising a first timestamp and a first feature vector of the person being authenticated/identified is used to establish consistency of that first reference point with at least one model of a reference person.
  • the first reference point is a point in a multi-dimensional space with as dimensions, time and the feature vector dimensions.
  • the model in turn, represents a curve in the same space.
  • noise is accounted for by defining that the point is consistent if it is sufficiently close, that is within a pre-determined range from the curve.
  • a preferred manner to establish consistency is to establish a point on the curve corresponding to the first reference point, for example by selecting a point on the curve with the same time-stamp, or with the same feature vector value. Subsequently the distance between the reference point and the selected point can be calculated, and used as a difference measure.
  • the advantage of this particular manner of determining a difference measure is that it does not involve complex computations. Next several manners for determining a difference measure will be presented in more detail.
  • a first reference point (T acq ,Q acq ) is acquired as depicted in Fig. 3.
  • the first reference point comprises a first timestamp T acq corresponding to the time of authentication/identification of a person and a first feature vector Q aC q representative of the first physiological characteristic of the person at the first timestamp T acq .
  • a second reference point is acquired (T erl ,Q erl ) as depicted in Fig. 3.
  • the second reference point comprises a second timestamp T er i corresponding to the time of enrolment of a reference person and a second feature vector Q acq representative of the first physiological characteristic of the reference person at the second timestamp T erl .
  • a first curve M a may be defined based on the model of the first physiological characteristic that passes through the first reference point.
  • a second curve M e can be defined that passes through the second reference point, both curves identical in shape, the first curve having an offset with respect to one of the timestamp values and the feature vector values compared to the second curve.
  • a first manner to establish a difference measure, in accordance with claim 6, is presented in Fig. 3.
  • the first reference point (T acq ,Q acq ) is used as a first point for predicting a second point located on the second curve M e .
  • the value is determined of the second curve M e at the first timestamp T acq , corresponding to the feature vector of the second point S 1 .
  • the difference measure for this particular case is indicated by difference measure di, representing the difference in the respective feature vector of the first reference point (T acq ,Q acq ) and the second point S 1 .
  • a second manner to establish a difference measure in accordance with claim 6, is also presented in Fig. 3.
  • This second approach also starts with the selection of the first reference point, as the first point for predicting a second point.
  • a second point S 2 is established, by determining the timestamp when the curve M e equals the first feature vector value Q acq .
  • the difference measure for this particular case is indicated by difference measure d 2 , representing the difference in the respective timestamps of the first reference point (T acq ,Q acq ) and the second point S 2 .
  • Fig. 4 also depicts a further difference measure d 3 representative of the distance between the first curve M a and the second curve M e .
  • first and the second curve are plotted in order to visualize the difference measure calculation, it is not necessary to plot or calculate all points on the curves in order to establish the difference measure.
  • Fig. 5 illustrates an embodiment according to claim 5. After acquiring a first reference point (T acq ,Q acq ) and a second reference point (T erl ,Q erl ), a difference measure representative of the distance between a first curve M a through the first reference point and a second curve M e through the second reference point can be determined.
  • a first point Pi is selected.
  • the first point Pi is located on one of the first or the second curve, in the example in Fig. 5 on the second curve M e .
  • a second point is predicted using the model M of the first physiological characteristic.
  • the second point being a point on the other of the first and second curve, in the example in Fig. 5 the first curve M a .
  • a first manner to compute a distance measure di involves determining the value of the feature vector of the first curve M a at the timestamp corresponding to the first point Pi, in this case this corresponds to the feature vector of the second point S 1 .
  • the difference measure for this particular case is indicated by difference measure di, representing the difference in the respective bio metric values of the first point Pi and the second point S 1 .
  • a second manner to compute a distance measure d 2 can be determined by determining the value of the timestamp when the first curve M a has a feature vector corresponding to that of the first point P 1 , in this case this results in the second point S 2 .
  • the difference measure for this particular case is indicated by difference measure d 2 , representing the difference in the respective timestamps of the first point Pi and the second point S 2 .
  • establishing the difference measure comprises using the model to establish an alternate reference point comprising a feature vector and a timestamp representative of either the person or the reference person.
  • a difference measure may be used for verifying whether the difference measure is within a predetermined range, typically this predetermined range will correspond with the absolute value of the difference measure being below a predetermined threshold. If this is not the case, then this may be seen as a proof that the person being authenticated/identified (associated with the first reference point) is not the same person as the reference person (associated with the second reference point). On the other hand when the difference measure is within range the person being authenticated/identified can, but need not, be the reference person.
  • Fig. 6 illustrates two issues associated with modeling a cyclic time- variant first physiological characteristic.
  • Fig. 6 illustrates such by means of the points E 1 , E 2 , E3, and E 4 .
  • the generic model may be adapted, such that the enrolment data are on the generic model, here the curve M e .
  • a cyclic first physiological characteristic it is important to prevent sub-sampling
  • the curve M ss is presented to illustrate a further shape-wise similar curve that passes through the same enrolment points as M e , but has approximately a six times higher frequency than that of M e .
  • This example illustrates that the generation of a model requires knowledge of the first physiological characteristic, in order to prevent sub-sampling (Nyquist sampling theorem).
  • Fig. 6 further highlights a problem that occurs when using a cyclic model to predict a second point. Due to the cyclic nature of the model there typically will be multiple timestamps when the model evaluates to a particular feature vector.
  • a similar problem may occur with an a-cyclic model wherein multiple timestamps have the same feature vector value.
  • the first reference point (T acq ,Q acq ) When predicting a second point on the curve M e with the same biometric value as the first point, here the first reference point (T acq ,Q acq ), multiple points will be found with a feature vector value identical to that of the first feature vector Q acq .
  • the first timestamp T acq may be used to select the timestamp predictions that are closest to the acquired first or second timestamp respectively.
  • Fig. 6 illustrates the selection of relevant candidate timestamps using T acq resulting in the timestamps T pacq 2, and T pacq 3, the timestamps closest to the first timestamp T acq .
  • Fig. 7 illustrates a difference measure determination in accordance with claim 7.
  • First a first reference point (T acq ,Q acq ) and a second reference point (T erl ,Q erl ) are acquired.
  • M the model of the first physiological characteristic
  • the first curve, and the second curve can be established M a and M e respectively.
  • the difference measure representative of the distance between the first and the second curve can be established.
  • establishing the difference measure dAi comprises the summation of the absolute differences between the first and the second curve for an a- cyclic first physiological characteristic.
  • the area Al is established by summation of the absolute differences between the first and the second curve over the overlapping period, here 1964-2014.
  • the difference measure dAi is derived from Al by dividing Al by the overlap in time, here 40 years.
  • Fig. 8 illustrates a difference measure determination in accordance with claim
  • the area A2 is established by means of the summation of the absolute differences between the first and the second curve over one period of the cycle.
  • the area A2 corresponds to the hatched areas.
  • the area A2 can be used as a difference measure, it is possible to factor out the period and use A2/T p as a difference measure instead.
  • Fig. 9 shows an apparatus 805 for biometric authentication of a person according to the present invention.
  • the apparatus 805 represents an apparatus that is arranged to use two modalities for authentication: a fingerprint and a time-variant first physiological characteristic.
  • the first modality, the fingerprint 801 is used to authenticate the individual, and will not be discussed here.
  • the second modality 802, the time-variant first physiological characteristic is used to establish consistency of a feature vector acquired during authentication from the person being authenticated with a model of a reference person comprised in the enrollment data of that reference person.
  • the apparatus 805 comprises an acquisition means 810, an establishing means
  • the apparatus 805 further comprises an on-board timer 850, and storage means 860.
  • the storage means 860 comprises a biometric database.
  • the biometric database comprises entries that correspond to enrolment data of previously enrolled persons.
  • An enrolment data entry for a reference person here comprises: an identity and a model of a feature vector representative of the first physiological characteristic of the reference person over time.
  • the enrolment data also comprises fingerprint enrolment data this is not considered here for clarity sake.
  • the apparatus 805 is arranged to authenticate a person, it may also be used to enroll a person, as the general process of feature vector acquisition is similar, but may involve repeated acquisition of feature vector and further processing thereof.
  • the acquisition means 810 When a person attempts to authenticate herself, she presents the appropriate finger 890 to the apparatus 805.
  • the acquisition means 810 subsequently acquires her fingerprint 801, and in addition acquires information 806 related to the first time- variant physiological characteristic. This information 806 is used by the acquisition means to establish a first feature vector 802.
  • the acquisition means 810 acquires a first timestamp 803 from the on-board timer 850, corresponding to the time of acquisition.
  • the depicted embodiment features an on-board timer the present invention is not limited thereto, as an external timer can be used instead.
  • the person also provides an ID-card 880 to the apparatus 805.
  • the ID-card 880 comprises an alleged identity 809 of the person.
  • the alleged identity 809 is used to retrieve an enrolment data 811 associated with the alleged identity 809 from the bio metric data 860.
  • the enrolment data 811 comprises the alleged identity 809 and a model of a feature vector representative of the first physiological characteristic of the person having the alleged identity 809 over time.
  • the first establishing means 825 is arranged to establish consistency of the first reference point with the model of the reference person.
  • the first establishing means 825 comprises a second establishing means 820 arranged to establish a difference measure 812 indicative of the extent to which the first reference point is inconsistent with the model of the reference person, and a verifying means 830 arranged to verify whether the difference measure 812 is within a predetermined range.
  • a verifying means 830 arranged to verify whether the difference measure 812 is within a predetermined range.
  • the verification means 830 receives the difference measure 812 and verifies whether the difference measure 812 is within a predetermined range.
  • the outcome 814 of this verification is presented to the decision means 840.
  • the decision means 840 combines the outcome 814 of the verification with an outcome 816 of the fingerprint authentication into an authentication decision 817.
  • the authentication outcome 816 is accepted when the first feature vector 802 at the first timestamp 803 is sufficiently consistent with the enrolment data of the reference person, otherwise the authentication outcome 816 is rejected. This result is presented on the output of the decision means in the form of the authentication decision 817.
  • the verification means 830 and decision means 840 may be combined such that the verification means 830 does not provide an outcome 814 in the form of a hard decision, consistent or inconsistent, but provides a probability instead. Provided the fingerprint authentication method also provides a probability, a more accurate soft decision can be made.
  • the present invention can be applied in an advantageous manner for authentication purposes or for identification purposes alike. This can be illustrated using the apparatus 805.
  • the apparatus 805 is an apparatus for authentication of a person, a slightly modified version of the apparatus may be used for identification.
  • the first feature vector 802 is matched repeatedly with feature vector from various candidate reference persons from the biometric data. This process may continue until a match is found or there are no more candidate reference persons left in the database.
  • an apparatus for identification of a person compares the first timestamp and the first feature vector with enrolment data of at least one, but potentially many candidate reference persons until a sufficient match is found, or the database is exhausted.
  • Fig. 10 shows a system 900 for authentication/identification of a person the system comprises two devices 915 according to claim 16 suitable for identification. These respective devices are part of a system comprising a central biometric database 905, wherein the devices 915 acquire enrolment data from the central biometric database 905 over a network 910 rather than from an internal biometric database 860 as shown in Fig. 9.
  • Each device 915 may comprise a computer program product reader, arranged to read a computer program product comprising program code means stored on a computer readable medium 920, such as an optical disc, or a solid state memory device.
  • the first physiological characteristic is considered to be a predictable time-variant physiological characteristic.
  • variations/deviations may occur in the actual first physiological characteristic as a result of external influences.
  • This information may be used to update the parameters of the model to compensate for such trends and provide a better match between the model and the actual first physiological characteristic of the individual.
  • Such updates are particularly useful in case the first physiological characteristic is cyclic, as these updates allow "synchronization" of the model with the actual first physiological characteristic.
  • tracking, and synchronization of the adapted model are highly desirable, tracking changes in the first physiological characteristic by means of the acquired feature vectors embodies an inherent risk.
  • tracking changes in the feature vectors rather than the actual physiological characteristic noise and structural errors in the feature vector generation may lead to erroneous updates that may make the system less robust. This however may be circumvented by means of a periodic re-enrolment.
  • any reference signs placed between parentheses shall not be construed as limiting the claim.
  • the word “comprising” does not exclude the presence of elements or steps other than those listed in a claim.
  • the word “a” or “an” preceding an element does not exclude the presence of a plurality of such elements.
  • the invention can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer.
  • the device claim enumerating several means several of these means can be embodied by one and the same item of hardware.
  • the mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

Abstract

L'invention concerne un procédé d'authentification/d'identification biométrique d'un individu qui consiste à utiliser une première caractéristique physiologique prévisible variable dans le temps pour établir une consistance entre un premier point de référence comprenant une première estampille (803) et une premier vecteur d'attribut (802) de l'individu authentifié/identifié avec un modèle (Mi) d'un individu de référence et d'authentification/identification de l'individu en tant qu'individu de référence en fonction de l'établissement, la première estampille (803) étant le temps d'authentification/identification de l'individu, le premier vecteur d'attribut (802) étant un vecteur d'attribut représentatif de la première caractéristique physiologique de l'individu au niveau de la première estampille, le modèle (Mi) étant un modèle d'un vecteur d'attribut représentatif de la première caractéristique physiologique de l'individu de référence dans le temps. L'invention concerne également un appareil et un système destinés à l'authentification/identification d'un individu.
PCT/IB2007/051508 2006-05-01 2007-04-24 Procédé et appareil d'identification biométrique d'un individu WO2007125478A2 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
EP06113341.9 2006-05-01
EP06113341 2006-05-01

Publications (2)

Publication Number Publication Date
WO2007125478A2 true WO2007125478A2 (fr) 2007-11-08
WO2007125478A3 WO2007125478A3 (fr) 2008-01-10

Family

ID=38534079

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2007/051508 WO2007125478A2 (fr) 2006-05-01 2007-04-24 Procédé et appareil d'identification biométrique d'un individu

Country Status (1)

Country Link
WO (1) WO2007125478A2 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9158906B2 (en) 2008-09-26 2015-10-13 Koninklijke Philips N.V. Authenticating a device and a user

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020138743A1 (en) * 2000-06-08 2002-09-26 Murakami Rick V. Method and apparatus for calibration over time of histological and physiological biometric markers for authentication
US6657538B1 (en) * 1997-11-07 2003-12-02 Swisscom Mobile Ag Method, system and devices for authenticating persons

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6657538B1 (en) * 1997-11-07 2003-12-02 Swisscom Mobile Ag Method, system and devices for authenticating persons
US20020138743A1 (en) * 2000-06-08 2002-09-26 Murakami Rick V. Method and apparatus for calibration over time of histological and physiological biometric markers for authentication

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
COOTES T F ET AL: "Toward automatic simulation of aging effects on face images" IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, IEEE, NEW YORK, NY, US, vol. 24, no. 4, April 2002 (2002-04), pages 442-455, XP011094222 ISSN: 0162-8828 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9158906B2 (en) 2008-09-26 2015-10-13 Koninklijke Philips N.V. Authenticating a device and a user

Also Published As

Publication number Publication date
WO2007125478A3 (fr) 2008-01-10

Similar Documents

Publication Publication Date Title
Dinca et al. The fall of one, the rise of many: a survey on multi-biometric fusion methods
US9268990B2 (en) Apparatus and method for producing an identification device
US8150108B2 (en) Systems and methods of identification based on biometric parameters
JP5531963B2 (ja) 認証システム、装置、認証方法及びプログラム
Nanni et al. Combining local, regional and global matchers for a template protected on-line signature verification system
Lumini et al. A clustering method for automatic biometric template selection
Radha et al. Rank level fusion using fingerprint and iris biometrics
JP5710748B2 (ja) 生体認証システム
Wang et al. Face based biometric authentication with changeable and privacy preservable templates
US20100014655A1 (en) Method and apparatus for generating cryptographic key using biometric data
WO2013084617A1 (fr) Système d'authentification biométrique
Toli et al. A survey on multimodal biometrics and the protection of their templates
Soltane et al. Multi-modal biometric authentications: concept issues and applications strategies
Ashraf et al. Iris and foot based sustainable biometric identification approach
JP2012238256A (ja) 生体認証装置、生体認証方法および生体認証プログラム
Meharia et al. The human key: Identification and authentication in wearable devices using gait
Choras Ear Biometrics.
Sadhya et al. Construction of a Bayesian decision theory‐based secure multimodal fusion framework for soft biometric traits
US11711216B1 (en) Systems and methods for privacy-secured biometric identification and verification
WO2007125478A2 (fr) Procédé et appareil d'identification biométrique d'un individu
KR20070019363A (ko) 타인수락율을 감소시키기 위한 생체정보 인증방법
Findling et al. Mobile gait match-on-card authentication from acceleration data with offline-simplified models
Kanagalakshmi et al. A cancellable and irrevocable approach for fingerprint template protection using optimal iterative solubility algorithm and secure point BASE
Spanakis et al. Developing a Context-Dependent Tuning Framework of Multi-channel Biometrics that Combine Audio-Visual Characteristics for Secure Access of an eHealth Platform
Pathak et al. Performance of multimodal biometric system based on level and method of fusion

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: 07735632

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 07735632

Country of ref document: EP

Kind code of ref document: A2