WO2005109310A1 - Biometric identification system - Google Patents

Biometric identification system Download PDF

Info

Publication number
WO2005109310A1
WO2005109310A1 PCT/GB2005/001739 GB2005001739W WO2005109310A1 WO 2005109310 A1 WO2005109310 A1 WO 2005109310A1 GB 2005001739 W GB2005001739 W GB 2005001739W WO 2005109310 A1 WO2005109310 A1 WO 2005109310A1
Authority
WO
WIPO (PCT)
Prior art keywords
biometric
template
identification
detection unit
central processing
Prior art date
Application number
PCT/GB2005/001739
Other languages
French (fr)
Inventor
Iain Wilkie
Alan Thomson
Original Assignee
Yarg Biometrics Ltd
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 Yarg Biometrics Ltd filed Critical Yarg Biometrics Ltd
Priority to EP05742554A priority Critical patent/EP1784764A1/en
Publication of WO2005109310A1 publication Critical patent/WO2005109310A1/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/12Fingerprints or palmprints
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/94Hardware or software architectures specially adapted for image or video understanding
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Definitions

  • the present invention relates to a system and method for identifying individuals on the basis of biometric characteristics.
  • Preferred embodiments of the invention employ an adaptive re-enrolment process making the invention useful particularly, but not exclusively, in relation to the identification of children; e.g. in a school environment.
  • Biometrics refers to the automatic identification of a person based on his/her physiological or behavioural characteristics. Biometric identification has many advantages over traditional identification methods that typically require the use of passwords and/or PIN numbers. For example, biometric identification requires the person to be identified to be physically present at the point-of- identification. Similarly, biometric identification removes the necessity of remembering a password or PIN etc.
  • biometric systems based on, for example, face, iris, retina, voice and fingerprint matching.
  • a typical biometric identification process involves a sequence of distinct processes, namely biometric data acquisition, feature extraction and matching.
  • biometric data such as an image of a fingerprint, face, retina or iris, or a speech sample
  • biometric reader a suitable sensor or the like
  • biometric reader a suitable sensor or the like
  • unique pattern information referred to herein as a "template”
  • the extracted template is matched against a database of reference templates associated with enrolled users to identify the user corresponding to the extracted template.
  • the matching process is used to determine the specific identity of the user (“identification”) .
  • identity verification an identity claimed by the user
  • the matching process may simply identify the user as a member of a set of individuals ("authentication") .
  • authentication is potentially useful in all of these scenarios.
  • the invention will be described with particular reference to exemplary systems employing fingerprint detection, more particularly to systems employing "fingerswipe" type fingerprint sensors, and most particularly to systems employing thermal fingerswipe sensors.
  • fingerprint detection more particularly to systems employing "fingerswipe” type fingerprint sensors
  • thermal fingerswipe sensors most particularly to systems employing thermal fingerswipe sensors.
  • the invention is applicable to all types of biometrics and biometric readers, particularly "image-based" biometrics . It is envisaged that the invention is particularly also applicable to thermal facial imaging.
  • Biometric identification systems are potentially useful in any application where a user's identity has to be established or verified.
  • the present invention may be employed in connection with any such applications and not only with particular applications that may be described herein by way of example.
  • a biometric system typically comprises at least one biometric reader (e.g. a fingerprint reader) for acquiring raw biometric data, an extraction process for extracting a , template from the raw data, a database of enrolled users and their associated biometric templates, and a matching process for matching an extracted template with an enrolled user.
  • the acquisition, extraction and matching processes are all embedded in the biometric reader together with the database of the enrolled users' biometric records.
  • this fully embedded approach can limit the size of the biometric records database.
  • the biometric records database is stored and individually accessed by each biometric reader, the process of maintaining up to date biometric records may require the continual replication of the current biometric records database to each of the biometric readers.
  • fully embedded biometric systems may prove difficult to integrate with third party applications.
  • simple biometric readers are connected directly to host PCs.
  • the fingerprint sensors acquire the fingerprint image and transmit it to the host PC, Template extraction is performed by the host PC.
  • the matching process is performed either by the host PC or by a separate server connected to the host PC via a network. Whilst this approach makes integration with third party applications easier, it requires the presence of a host PC at each identification location, thereby adding greatly to the system cost, complexity and maintenance Object of the Invention
  • An object of the present invention is to provide improved biometric identification systems and methods.
  • a further object of preferred embodiments of the invention is to provide adaptive biometric identification systems and methods that are particularly suited for identifying children; e.g. in a school environment.
  • a biometric identification system comprising: a central data processing system, and at least one biometric detection unit, each detection unit being operable to acquire biometric data from a user, process the biometric data to create a reduced size template representative of the biometric data, and send the template to the central processing system, the central processing system being operable to compare the received template with one or more of a plurality of reference templates, each associated with an enrolled user, so as to match the received template with a reference template.
  • a biometric identification method comprising: within a biometric detection unit, acquiring biometric data from a user, processing the biometric data to create a reduced size template representative of the biometric data, and sending the template to a central processing system, within the central processing system, comparing the received template with one or more of a plurality of reference templates, each associated with an enrolled user, so as to match the received template with a reference template.
  • the acquisition of raw biometric data and the extraction of a template therefrom are performed locally in an integrated biometric detection unit.
  • the matching process is performed remotely in a separate central data processing system that also hosts the enrolled user database, typically a remote server.
  • This division of functionality in the identification system provides a very different process topology compared with prior art systems that either perform all the operations in a biometric reader (embedded system) or perform the extraction and matching processes in a separate computer (i.e. PC or server).
  • the unique process topology of the present identification system provides a number of advantages over the afore- mentioned prior art systems.
  • An extracted biometric template is generally much smaller than the raw data from which it is derived. Extraction can reduce the size of a thermal fingerprint image from around 180KB to about 500 bytes.
  • extraction is performed locally in the biometric detection unit(s) and only the extracted templates are transmitted to the central matching system. Consequently, the unique process topology of the present identification system effectively reduces the network traffic between a biometric detection unit and the central processing system, especially compared with prior art systems that transmit raw biometric data to a server. Tests have shown that the present identification system can achieve identification speeds of less than two seconds per fingerprint swipe point with up to six swipe points running concurrently using a single remote central server and a database containing in excess of two thousand entries.
  • a further advantage of the present system is that, since a biometric detection unit transmits only an extracted template to a central processing system, complete biometric images are not transmitted during identification, or stored by biometric detection unit (other than transiently until template extraction is complete) or by the central processing system.
  • An extracted template consists of a set of specific points, represented by, for example, a series of numbers.
  • an extracted template could not be used to recreate an actual biometric image that might be used to "spoof" the system.
  • an extracted template cannot be directly compared with a normal fingerprint, an extracted template could not be used to identify a person in a criminal investigation.
  • an extracted template has a secure proprietary format it could not be used to compromise a user' s privacy in the event that another party inappropriately captured the template.
  • the reference templates of enrolled users could be stored in a separate location from the central processing system.
  • the preferred embodiments of the present identification system provide a number of features to improve the flexibility and scalability of the identification process.
  • the present identification system preferably processes the matching algorithm and the enrolled users reference template database in a thread structure, which enables the identification system to handle multiple inputs simultaneously.
  • the preferred embodiments of the present identification system allow a number of pre-defined system thresholds to be modified to adjust the dynamics of the matching operation.
  • the preferred embodiments of the present identification system provide a facility for automatically re-enrolling users as they use the system.
  • This facility is particularly useful for identifying young people, who may still be growing and whose physical attributes may therefore still be developing during the period in which they use the system.
  • the re-enrolment facility automatically maintains valid up-to-date reference template records to accommodate the effects of finger growth over long periods of time.
  • This adaptive re-enrolment process is equally useful for other types of biometric and other age groups, particularly facial biometrics which will change significantly throughout a user's lifetime.
  • a fingerswipe type fingerprint sensor provides a number of advantages over more conventional "pad"-based sensors that are used for reading fingerprints.
  • a pad-based sensor requires a user to place and keep their finger in position on an array of sensors that scan the finger. Accordingly, pad-based sensors must be large enough to accommodate a whole finger and are typically much larger than strip sensors. Consequently, pad-based readers are typically more expensive and prone to the build-up of dirt and resulting damage than strip sensors.
  • a fingerswipe sensor In contrast with pad-based readers, a fingerswipe sensor requires the user to draw their finger across a strip-type sensor array. Accordingly, the process of successive finger movements across the strip, as multiple users employ the strip sensor, ensures that the sensor is self-cleaning and furthermore ensures that no latent prints can be left behind on the strip.
  • Strip-type fingerswipe sensors include thermal imaging types and capacitive sensing types, the former being preferred for the purposes of the present invention. Since the a thermal strip sensor detects differential thermal patterns, it can only be operated with a "living" finger. Consequently, a print copied onto latex could not be used to spoof the user-identification system.
  • the present identification system may also be connected to one or more external device (s) or systems and can send data or control signals to the external devices/systems, for use in any of a wide range of applications.
  • the identification system may be connected to a billing system so that amounts spent by a user can be automatically recorded in his/her account.
  • the identification system would be typically located at a till or a self-service terminal.
  • the identification system could be connected to an access control system controlling access to a computer terminal or network or to a physical location etc.
  • the central processing system also provides an application interface for integration with new and existing third party applications. This provides a simple method of providing a biometric "front-end" to any database driven application without requiring the " integrator to have any in-depth specialist biometrics knowledge.
  • Figure 1 is a block diagram of the hardware/software architecture of one embodiment of a biometric identification system in accordance with the present invention
  • Figure 2 is a flow diagram showing the distribution of processes implemented in preferred embodiments of the invention
  • Figure 3 is a flow diagram showing the 2 distribution of the processes implemented in Figure 3 2 for the control of one or more external devices.
  • 4 5 Detailed Description 6 7
  • preferred embodiments 8 of the invention will now be described, referring 9 firstly to hardware/software architecture of the
  • 25 invention is applicable to biometric data and 26 biometric sensors other than fingerprints and 27 fingerswipe devices.
  • a biometric identification system 5 embodying the present invention comprises an enrolment unit 10 and one or more fingerswipe units 12, each of which can communicate with a remote server 14 via any suitable network 16 (e.g. Ethernet) .
  • the enrolment unit 10 may be located next to the server 14 or to a computer or the like (not shown) connected to the sever 14, to facilitate the input of relevant user details at the time of enrolment.
  • the fingerswipe unit 12 is typically connected to at least one other external device 18A (e.g. a till or a door activation device) through a communications interface 20 (e.g. an RS232 or USB interface) .
  • Different units 12 may be connected to a single device 18 or to separate and/or different types of device and may control different functions within the system. Alternatively or additionally, external devices 18B may be connected directly to the network 16.
  • the unit 12 may also include a contact closure (controllable switch, not shown) enabling the unit to control the operation of a door lock or the like.
  • the enrolment unit 10 is used during enrolment of users onto the system, whereas the fingerswipe units 12 are used for identification of previously enrolled users.
  • the enrolment unit 10 may be identical to the fingerswipe units 12 and may also function as a fingerswipe unit in use of the system. It will be understood that the use of the enrolment unit 10 should result in biometric reference templates being stored in the system that are the same in type and format as those generated by the fingerswipe units 12. It is not essential that the enrolment unit 10 itself performs both data acquisition and feature extraction. For the purposes of enrolment, feature extraction may be performed by the server 14 or another computer associated with the enrolment unit. However, as a matter of practicality, it is convenient for the enrolment unit itself to generate the template in the same way as the other fingerswipe units. Accordingly, the following description of the fingerswipe unit 12 may apply equally to the enrolment unit.
  • the fingerswipe unit 12 includes a thermal strip sensor 22 and a microprocessor, application specific integrated circuit (ASIC) or the like 24, suitably an ARMTM processor.
  • the thermal strip sensor 22 typically is approximately 15mm long by 2mm wide and comprises a strip array of thermally sensitive points of suitable size and density to measure small temperature differences between the ridges and valleys of skin.
  • the thermal strip sensor 22 also includes analogue to digital conversion hardware, which enables the strip array to send signals to the microprocessor 24.
  • the small dimensions of the thermal strip sensor 22 ensure that only a small strip of the finger is in contact with the strip array at any given time. As the finger is swiped over the strip array, the microprocessor 24 is presented with a number of overlapping image slices representing the finger's temperature profile.
  • the associated data acquisition software includes a process for reconstructing a single complete image from these multiple image slices.
  • Fingerswipe sensors of this general type are known in the art (for example the FingerChip (trade mark) sensor produced by Atmel of the United States) , and will not be described in more detail herein.
  • the fingerswipe unit 12 is a compact, integrated unit that performs biometric data acquisition and feature extraction so as to generate a biometric template as output.
  • the software algorithms for these functions are embedded in the unit, encoded in hardware or firmware, preferably in a single chip or as a chipset on a single PCB.
  • suitable network communication software such as a TCP/IP stack, which also serves to uniquely identify the unit 12 within the wider system.
  • acquisition software Si acquires thermal image data from the sensor 22 and reconstructs a complete thermal map of a finger from the thermal data.
  • Feature extraction software S 2 extracts the biometric template from the thermal map. Acquisition and feature extraction software suitable for this purpose is known and will not be described in detail herein.
  • the microprocessor 24 is also connected to a LCD or similar display unit 26 that enables messages to be displayed to a user.
  • the microprocessor 24 may also be connected to a piezo sounder or other audio output device (not shown) to provide audio cues to the user.
  • the server 14 includes : - a matching algorithm 28 ; - a primary database 30 of reference templates; - enrolment software S 3 - adaptive enrolment software S 4 ; and (optionally) - a secondary database 31 of extracted templates for use with the adaptive enrolment software.
  • the enrolment software enables users to be enrolled on the system, by acquiring biometric template data and associating this in the primary database with data (at minimum, a unique ID code) identifying the individual user.
  • data at minimum, a unique ID code
  • the matching algorithm attempts to match this with an enrolled user.
  • Software suitable for this purpose is known and will not be described in detail herein.
  • the secondary database 31 need not be a physically separate database from the primary database 30, but may instead be a portion of the primary database specifically dedicated for adaptive re-enrolment, as discussed further below.
  • the server 14 also provides an application interface 15 for integration with third party applications 17.
  • This provides a simple method of furnishing a biometric "front-end" for any database driven application without requiring the integrator to have any in-depth specialist biometrics knowledge.
  • the application 17 may include more detailed personal information regarding users and details of fingerswipe unit locations, and will determine the actions to be taken in response to a user being identified at a particular fingerswipe unit 12. Accordingly, it is only necessary for the server 14 to transmit to the application 17 a user ID code and a fingerswipe unit ID code.
  • the server itself need not have any detailed knowledge of users, the locations of fingerswipe units, or the use to which transmitted ID and location data will be put.
  • the system architecture described above allows devices or processes to be controlled on the basis of user IDs in a variety of ways.
  • the server 14 may transmit a message on the network 16 indicating that a particular person had been identified at a particular location. Any external network devices 18B listening on the network 16 could respond to the server's message as appropriate.
  • the server 14 may send a message to the appropriate fingerswipe units 12 for display to the user (i.e. through the LCD unit 26) .
  • An external device 18A directly connected to a fingerswipe unit 12 (through the communications interface 20) may be controlled directly from the fingerswipe device 12 in response to a message from the server 14 that a particular user had been identified at a particular location.
  • the application may communicate with devices 18A/B and fingerswipe units 12 via the server 14, or directly with devices 18B and fingerswipe units 12 (and hence devices 18A) via the network 16.
  • Figure 2 shows how the processes implemented in the identification system 5 are distributed amongst its various components.
  • Figure 2 is divided into three sections: the left-most section, middle section and right-most section respectively depict the processes performed in the enrolment unit 10, the fingerswipe unit 12 and the server 14.
  • the processes implemented in the identification system 5 can be broadly divided into two distinct operational phases, namely an enrolment phase 32 and an identification phase 34 (depending on the application this may involve identity verification and/or authentication as mentioned above) .
  • an enrolment phase 32 fingerswipe template data uniquely associated with a user is stored in the server's primary database 30 together with the name and/or other identifier (s) of the user.
  • the user Once the user has successfully enrolled with the enrolment unit 10, he can use any of the fingerswipe units 12 to identify himself (identification phase 34) .
  • enrolment phase 32 and the identification phase 34 are the processes of acquiring 36, reconstructing 38 and extracting 40 information from a user' s fingerswipe on the thermal strip sensor 22 (in the enrolment unit 10 or the fingerswipe unit 12) .
  • the acquisition, reconstruction and extraction processes conducted during the enrolment phase 32 will be designated with the identifier ⁇ e' (i.e. 36e, 38e and 40e) .
  • the acquisition and extraction processes conducted during the identification phase 34 will be designated with the identifier ⁇ v' (i.e. 36v, 38v and 40v) .
  • a series of overlapping slices of the finger's temperature profile are acquired 36e/36v by the acquisition software Si which also reconstructs 38e/38v a complete thermal map of the finger from the multiple overlapping slices acquired by the acquisition software.
  • the thermal map is then processed by the feature extraction software S 2 to extract 40e/40v a reduced data set (referred to henceforth as minutiae) and various other pattern features so as to produce an extracted template that is representative of the thermal map.
  • minutiae a reduced data set
  • any suitable extraction technique can be used for the extraction process 40e/40v, provided that each of the resulting extracted templates can be uniquely associated with a given individual.
  • the enrolment unit/fingerswipe unit transmits the extracted template (e.g. through a TCP/IP stack) to the server 14 for any required further processing and inclusion in the enrolment database.
  • an extracted template consists of a set of specific points, represented by, for example, a series of numbers and/or patterns.
  • an extracted template could not be used to create a fingerprint and/or a false identity.
  • an extracted template could not be used to identify a person in a criminal investigation or to compromise a user's privacy in the event that another party inappropriately captured the template.
  • extraction can reduce a thermal map of around 180KB to a template of about 500 bytes.
  • many conventional biometric systems perform the extraction process in a central server. More particularly, in these conventional biometric systems, a fingerswipe unit acts as a dumb terminal and simply transfers an entire thermal map to a central server.
  • the extraction process is implemented locally in the fingerswipe units and the resulting extracted templates are transmitted to the server 14. Consequently, the specific implementation and distribution of processes between the devices in the identification system 5 effectively reduces the network traffic between a fingerswipe unit and the server 14. This feature becomes particularly important as the identification system 5 scales up to include increasing numbers of fingerswipe units 12.
  • a template is extracted from a thermal map of a user's finger by the enrolment unit 10.
  • the resulting extracted template is transmitted to the server 14 together with the user's personal details. This process is preferably repeated two more times with the same finger.
  • the enrolment software S 3 generalises from the three extraction templates of the user's finger by combining the three extraction templates to create the best composite template therefrom.
  • the enrolment software S 3 then stores 42 the resulting extracted template in the server's primary database 30 as a reference template together with (at least) a unique ID code.
  • the identification system 5 When used for identifying young children who may grow considerably over time, the identification system 5 must be capable of accommodating changes in finger size to avoid the inconvenience of having to repeatedly re-enrol the children. To provide this facility, the identification system 5 implements an automatic adaptive re-enrolment procedure 44.
  • the matching algorithm 28 generates a score representing the match between the user's extracted template and the reference templates stored in the server's primary database 30.
  • the adaptive re-enrolment procedure 44 stores 46 the user's extracted templates in the server's secondary database 31. More specifically, for any given enrolled user, the adaptive re-enrolment procedure 44 only stores 46 those user's templates that closely match the user's reference template stored in the server's primary database 30.
  • the server 14 After a pre-determined time interval (e.g. a month) 48, the server 14 automatically selects a primary subset of the best scoring templates (of the user) from the server's secondary database 31. The server 14 then determines the optimal combination and/or permutation of templates in a further subset from the primary subset. The server 14 then creates 50 a new extracted template from this optimal subset, using the generalisation techniques employed in the normal enrolment phase 32. The new extracted template then replaces the corresponding reference template in the server's primary database 30 and is used for identification until the next automatic adaptive enrolment period.
  • a pre-determined time interval e.g. a month
  • This process gradually increases the separation within the server's primary database 30 by ensuring that reference templates are successively overwritten with templates that: (a) most closely identify the user; and (b) are most strongly differentiated from the other templates in the primary server database 30.
  • the continual adaptation of this combination of features progressively reduces the risk of false identification and thereby gradually improves the performance of the identification system 5.
  • the server 14 determines 52 the user's identity by comparing the received template against all the reference templates stored in the server's primary database 30. More particularly, the server's matching algorithm 28 allocates scores to the received extracted template on the basis of its similarity with each of the stored reference templates and the number of matching minutiae therebetween. In general terms, a large minutiae value and a large similarity score indicate a good match between the received extracted template and a stored reference template. Depending on a number of pre-defined system thresholds, the matching algorithm 28 employs the scores of the received extracted template to generate a positive or negative identification of the user.
  • the identification system 5 provides a number of features to improve the flexibility and scalability of the identity verification phase 34.
  • the identification system 5 employs the matching algorithm 28 and the primary server database 30 in a thread structure, which enables the server 14 to handle multiple inputs simultaneously.
  • the identification system 5 allows the pre-defined system thresholds to be modified to adjust the dynamics of the matching operation. For instance the chance of a "False Acceptance" (i.e. where a person is recognised as being someone else) could be set at 1 in 100,000.
  • Figure 3 focuses on the distribution of processes in the identification system 5 after a user has been identified. More particularly, in a similar fashion to Figure 2, Figure 3 is divided into three sections with the left-most, middle and rightmost sections respectively depicting the processes performed in a fingerswipe unit 12, the server 14 and third party application 17.
  • the server 14 transmits 54 a message to that effect (with time stamp and/or originating fingerswipe unit 12 location) to the third party application 17 which may use the information as appropriate (e.g. include it in a spreadsheet etc.).
  • the third party application 17 may also transmit a response to the server 14, which the server 14 transmits to the external devices 18a/b and/or the fingerswipe unit 12 (as described below) .
  • the server 14 also relays 56 an authentication signal ⁇ auth_sig) to the fingerswipe unit 12.
  • the fingerswipe unit 12 may display 58 an appropriate message to the user and/or transmit a control signal 60 to the external device 18 to take an appropriate action (e.g. the debiting of a user account or, the opening of a door to allow access to a designated area) .
  • the external device 18B is directly connected to the network 16.
  • the server 14 may transmit: (a) a control signal ( ctrl_sig) over the network 16, on receipt of which, the relevant external device 18 takes an appropriate action; and (b) a message ( auth__info) , indicating that a user had been identified, to the fingerswipe unit 12 (on receipt of which, the fingerswipe unit 12 may display an appropriate message to the user on its LCD 26) .
  • Schools within the UK provide a system whereby children whose parents have an income that is below a certain threshold can receive free school dinners.
  • the class teacher collects fees for school dinners in the classroom.
  • receiving free school meals can stigmatise children at an early age, when their confidence and self-esteem can be fragile. This can damage them emotionally and inhibit their social and educational progress.
  • the present identification system provides a viable alternative to the above-mentioned smartcard/magnetic card system.
  • the identification system 5 eliminates the need for ID cards or PIN numbers etc. This is particularly useful for very young users. Furthermore, since the fingerswipe units 12 transmit an extracted template to a server 14 rather than a direct image of a child's finger, the privacy and security of the identification system 5 is enhanced. This is a particularly attractive feature in a school environment, where security and privacy issues are particularly important.
  • the reduction in network traffic provided by the process topology of the identification system 5 and the resulting improved processing speed is particularly beneficial in a school queuing environment, where a large number of impatient children must be processed within a short period of time.
  • the identification system 5 avoids the inconvenience and disruption of repeatedly manually re-enrolling the children in the school setting.
  • the invention is applicable to fingerprint sensors other than thermal fingerswipe types and to biometrics and associated sensors/readers other than fingerprints.
  • FIG. 1 shows only one fingerswipe unit 12, it will be appreciated that the identification system 5 may typically comprise multiple fingerswipe units at one or more locations.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Software Systems (AREA)
  • Collating Specific Patterns (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

A biometric identification system comprises one or more biometric sensor units (12) connected via a network (16) to a server (14) The sensor units are compact integrated units adapted to acquire biometric data, perform feature extraction to generate a biometric template and transmit extracted templates to the server. The server hosts a database (30) of enrolled users and performs matching of received templates against the database to identify users. The server may cooperate with a third party application (17), e.g. to transmit control signals to devices (18A, 18B) when a user is identified. Preferred embodiments employ thermal fingerswipe sensors as the biometric sensors and apply an adative re-enrolment process to automatically update biometric data held in the database for matching purposes.

Description

Biometric Identification System
Field of the Invention
The present invention relates to a system and method for identifying individuals on the basis of biometric characteristics. Preferred embodiments of the invention employ an adaptive re-enrolment process making the invention useful particularly, but not exclusively, in relation to the identification of children; e.g. in a school environment.
Background of the Invention
Biometrics refers to the automatic identification of a person based on his/her physiological or behavioural characteristics. Biometric identification has many advantages over traditional identification methods that typically require the use of passwords and/or PIN numbers. For example, biometric identification requires the person to be identified to be physically present at the point-of- identification. Similarly, biometric identification removes the necessity of remembering a password or PIN etc.
Various types of biometric systems are presently in use, based on, for example, face, iris, retina, voice and fingerprint matching.
A typical biometric identification process involves a sequence of distinct processes, namely biometric data acquisition, feature extraction and matching. During the acquisition phase "raw" biometric data such as an image of a fingerprint, face, retina or iris, or a speech sample, is acquired by a suitable sensor or the like ("biometric reader") . During the feature extraction phase, unique pattern information, referred to herein as a "template", is extracted from the raw data. During the matching phase, the extracted template is matched against a database of reference templates associated with enrolled users to identify the user corresponding to the extracted template. In some scenarios, the matching process is used to determine the specific identity of the user ("identification") . In other scenarios, the matching process is used to verify an identity claimed by the user ("identity verification") . In still other scenarios, the matching process may simply identify the user as a member of a set of individuals ("authentication") . The present invention is potentially useful in all of these scenarios. The invention will be described with particular reference to exemplary systems employing fingerprint detection, more particularly to systems employing "fingerswipe" type fingerprint sensors, and most particularly to systems employing thermal fingerswipe sensors. However it will be understood that, except where features of the invention are specifically related to fingerprint based biometrics etc., the invention is applicable to all types of biometrics and biometric readers, particularly "image-based" biometrics . It is envisaged that the invention is particularly also applicable to thermal facial imaging.
Biometric identification systems are potentially useful in any application where a user's identity has to be established or verified. The present invention may be employed in connection with any such applications and not only with particular applications that may be described herein by way of example.
A biometric system typically comprises at least one biometric reader (e.g. a fingerprint reader) for acquiring raw biometric data, an extraction process for extracting a, template from the raw data, a database of enrolled users and their associated biometric templates, and a matching process for matching an extracted template with an enrolled user. In a first prior art example, the acquisition, extraction and matching processes are all embedded in the biometric reader together with the database of the enrolled users' biometric records. However, this fully embedded approach can limit the size of the biometric records database. Similarly, since the biometric records database is stored and individually accessed by each biometric reader, the process of maintaining up to date biometric records may require the continual replication of the current biometric records database to each of the biometric readers. Furthermore, fully embedded biometric systems may prove difficult to integrate with third party applications.
In a more common prior art example, simple biometric readers (most commonly fingerprint sensors) are connected directly to host PCs. The fingerprint sensors acquire the fingerprint image and transmit it to the host PC, Template extraction is performed by the host PC. The matching process is performed either by the host PC or by a separate server connected to the host PC via a network. Whilst this approach makes integration with third party applications easier, it requires the presence of a host PC at each identification location, thereby adding greatly to the system cost, complexity and maintenance Object of the Invention
An object of the present invention is to provide improved biometric identification systems and methods. A further object of preferred embodiments of the invention is to provide adaptive biometric identification systems and methods that are particularly suited for identifying children; e.g. in a school environment.
Summary of the Invention
In accordance with a first aspect of the invention, there is provided a biometric identification system comprising: a central data processing system, and at least one biometric detection unit, each detection unit being operable to acquire biometric data from a user, process the biometric data to create a reduced size template representative of the biometric data, and send the template to the central processing system, the central processing system being operable to compare the received template with one or more of a plurality of reference templates, each associated with an enrolled user, so as to match the received template with a reference template.
In accordance with a second aspect of the invention, there is provided a biometric identification method comprising: within a biometric detection unit, acquiring biometric data from a user, processing the biometric data to create a reduced size template representative of the biometric data, and sending the template to a central processing system, within the central processing system, comparing the received template with one or more of a plurality of reference templates, each associated with an enrolled user, so as to match the received template with a reference template.
These and further preferred features of the invention are defined in the claims appended hereto.
Advantages provided by the Invention
Advantages provided by the present invention include, but are not limited to, the following.
In the identification system of the present invention, the acquisition of raw biometric data and the extraction of a template therefrom are performed locally in an integrated biometric detection unit. The matching process is performed remotely in a separate central data processing system that also hosts the enrolled user database, typically a remote server. This division of functionality in the identification system provides a very different process topology compared with prior art systems that either perform all the operations in a biometric reader (embedded system) or perform the extraction and matching processes in a separate computer (i.e. PC or server). The unique process topology of the present identification system provides a number of advantages over the afore- mentioned prior art systems.
An extracted biometric template is generally much smaller than the raw data from which it is derived. Extraction can reduce the size of a thermal fingerprint image from around 180KB to about 500 bytes. In the present identification system, extraction is performed locally in the biometric detection unit(s) and only the extracted templates are transmitted to the central matching system. Consequently, the unique process topology of the present identification system effectively reduces the network traffic between a biometric detection unit and the central processing system, especially compared with prior art systems that transmit raw biometric data to a server. Tests have shown that the present identification system can achieve identification speeds of less than two seconds per fingerprint swipe point with up to six swipe points running concurrently using a single remote central server and a database containing in excess of two thousand entries. These advantages of reduced network traffic bedome more significant as the identification system scales up to include increasing numbers of biometric detection units.
A further advantage of the present system is that, since a biometric detection unit transmits only an extracted template to a central processing system, complete biometric images are not transmitted during identification, or stored by biometric detection unit (other than transiently until template extraction is complete) or by the central processing system.
An extracted template consists of a set of specific points, represented by, for example, a series of numbers. Thus, an extracted template could not be used to recreate an actual biometric image that might be used to "spoof" the system. Similarly, since an extracted template cannot be directly compared with a normal fingerprint, an extracted template could not be used to identify a person in a criminal investigation. Also, since an extracted template has a secure proprietary format it could not be used to compromise a user' s privacy in the event that another party inappropriately captured the template. Further, since only the central processing system actually processes the extracted templates (during matching) , the reference templates of enrolled users could be stored in a separate location from the central processing system.
The preferred embodiments of the present identification system provide a number of features to improve the flexibility and scalability of the identification process. In particular, the present identification system preferably processes the matching algorithm and the enrolled users reference template database in a thread structure, which enables the identification system to handle multiple inputs simultaneously. Similarly, the preferred embodiments of the present identification system allow a number of pre-defined system thresholds to be modified to adjust the dynamics of the matching operation.
The preferred embodiments of the present identification system provide a facility for automatically re-enrolling users as they use the system. This facility is particularly useful for identifying young people, who may still be growing and whose physical attributes may therefore still be developing during the period in which they use the system. In particular, the re-enrolment facility automatically maintains valid up-to-date reference template records to accommodate the effects of finger growth over long periods of time. This adaptive re-enrolment process is equally useful for other types of biometric and other age groups, particularly facial biometrics which will change significantly throughout a user's lifetime.
Turning to the hardware aspects of the preferred embodiments of the present identification system, the use of a fingerswipe type fingerprint sensor provides a number of advantages over more conventional "pad"-based sensors that are used for reading fingerprints. In particular, a pad-based sensor requires a user to place and keep their finger in position on an array of sensors that scan the finger. Accordingly, pad-based sensors must be large enough to accommodate a whole finger and are typically much larger than strip sensors. Consequently, pad-based readers are typically more expensive and prone to the build-up of dirt and resulting damage than strip sensors.
In addition, since a finger must be held on the sensor array for a sufficient period of time for the finger to be scanned, there is a risk that a "latent print" could be left behind on a pad-based reader. This creates a security risk, as the latent print could be copied from the pad-based reader.
In contrast with pad-based readers, a fingerswipe sensor requires the user to draw their finger across a strip-type sensor array. Accordingly, the process of successive finger movements across the strip, as multiple users employ the strip sensor, ensures that the sensor is self-cleaning and furthermore ensures that no latent prints can be left behind on the strip. Strip-type fingerswipe sensors include thermal imaging types and capacitive sensing types, the former being preferred for the purposes of the present invention. Since the a thermal strip sensor detects differential thermal patterns, it can only be operated with a "living" finger. Consequently, a print copied onto latex could not be used to spoof the user-identification system.
The present identification system may also be connected to one or more external device (s) or systems and can send data or control signals to the external devices/systems, for use in any of a wide range of applications. For example, the identification system may be connected to a billing system so that amounts spent by a user can be automatically recorded in his/her account. In this case, the identification system would be typically located at a till or a self-service terminal. Alternatively, the identification system could be connected to an access control system controlling access to a computer terminal or network or to a physical location etc.
In preferred embodiments, the central processing system also provides an application interface for integration with new and existing third party applications. This provides a simple method of providing a biometric "front-end" to any database driven application without requiring the" integrator to have any in-depth specialist biometrics knowledge.
Brief Description of the Drawings
Embodiments of the present invention will now be described by way of example only with reference to the accompanying drawings in which: Figure 1 is a block diagram of the hardware/software architecture of one embodiment of a biometric identification system in accordance with the present invention; Figure 2 is a flow diagram showing the distribution of processes implemented in preferred embodiments of the invention; and 1 Figure 3 is a flow diagram showing the 2 distribution of the processes implemented in Figure 3 2 for the control of one or more external devices. 4 5 Detailed Description 6 7 Referring now to the drawings, preferred embodiments 8 of the invention will now be described, referring 9 firstly to hardware/software architecture of the
10 identification system and then turning to a more
11 functional description of the processes performed by
12 the identification system, specifically focussing on
13 the distribution of the processes amongst the
14 various component parts of the system. 15
16 For the sake of brevity and convenience, the central
17 processing system will be referred to in this
18 description as a "server" and the biometric
19 detection unit(s) will be referred to as an
20 "enrolment unit" or a "fingerswipe unit" as
21. appropriate. It will be understood, however, that
22 the functionality of the central processing unit
23 might be distributed between one or more data
24 processing systems other than "servers" and that the
25 invention is applicable to biometric data and 26 biometric sensors other than fingerprints and 27 fingerswipe devices.
28 29 30 31 32 A. Hardware/Software Architecture
Referring to Figure 1, a biometric identification system 5 embodying the present invention comprises an enrolment unit 10 and one or more fingerswipe units 12, each of which can communicate with a remote server 14 via any suitable network 16 (e.g. Ethernet) . The enrolment unit 10 may be located next to the server 14 or to a computer or the like (not shown) connected to the sever 14, to facilitate the input of relevant user details at the time of enrolment. The fingerswipe unit 12 is typically connected to at least one other external device 18A (e.g. a till or a door activation device) through a communications interface 20 (e.g. an RS232 or USB interface) . Different units 12 may be connected to a single device 18 or to separate and/or different types of device and may control different functions within the system. Alternatively or additionally, external devices 18B may be connected directly to the network 16. The unit 12 may also include a contact closure (controllable switch, not shown) enabling the unit to control the operation of a door lock or the like.
The enrolment unit 10 is used during enrolment of users onto the system, whereas the fingerswipe units 12 are used for identification of previously enrolled users. The enrolment unit 10 may be identical to the fingerswipe units 12 and may also function as a fingerswipe unit in use of the system. It will be understood that the use of the enrolment unit 10 should result in biometric reference templates being stored in the system that are the same in type and format as those generated by the fingerswipe units 12. It is not essential that the enrolment unit 10 itself performs both data acquisition and feature extraction. For the purposes of enrolment, feature extraction may be performed by the server 14 or another computer associated with the enrolment unit. However, as a matter of practicality, it is convenient for the enrolment unit itself to generate the template in the same way as the other fingerswipe units. Accordingly, the following description of the fingerswipe unit 12 may apply equally to the enrolment unit.
The fingerswipe unit 12 includes a thermal strip sensor 22 and a microprocessor, application specific integrated circuit (ASIC) or the like 24, suitably an ARM™ processor. The thermal strip sensor 22 typically is approximately 15mm long by 2mm wide and comprises a strip array of thermally sensitive points of suitable size and density to measure small temperature differences between the ridges and valleys of skin. The thermal strip sensor 22 also includes analogue to digital conversion hardware, which enables the strip array to send signals to the microprocessor 24. The small dimensions of the thermal strip sensor 22 ensure that only a small strip of the finger is in contact with the strip array at any given time. As the finger is swiped over the strip array, the microprocessor 24 is presented with a number of overlapping image slices representing the finger's temperature profile. For a sensor of this type, the associated data acquisition software includes a process for reconstructing a single complete image from these multiple image slices. Fingerswipe sensors of this general type are known in the art (for example the FingerChip (trade mark) sensor produced by Atmel of the United States) , and will not be described in more detail herein.
The fingerswipe unit 12 is a compact, integrated unit that performs biometric data acquisition and feature extraction so as to generate a biometric template as output. The software algorithms for these functions are embedded in the unit, encoded in hardware or firmware, preferably in a single chip or as a chipset on a single PCB. Also embedded in the unit in firmware or hardware is suitable network communication software such as a TCP/IP stack, which also serves to uniquely identify the unit 12 within the wider system.
As illustrated schematically in Fig. 1, acquisition software Si acquires thermal image data from the sensor 22 and reconstructs a complete thermal map of a finger from the thermal data. Feature extraction software S2 extracts the biometric template from the thermal map. Acquisition and feature extraction software suitable for this purpose is known and will not be described in detail herein. The microprocessor 24 is also connected to a LCD or similar display unit 26 that enables messages to be displayed to a user. The microprocessor 24 may also be connected to a piezo sounder or other audio output device (not shown) to provide audio cues to the user.
The server 14 includes : - a matching algorithm 28 ; - a primary database 30 of reference templates; - enrolment software S3 - adaptive enrolment software S4; and (optionally) - a secondary database 31 of extracted templates for use with the adaptive enrolment software.
The enrolment software enables users to be enrolled on the system, by acquiring biometric template data and associating this in the primary database with data (at minimum, a unique ID code) identifying the individual user. When a template is received from a fingerswipe unit 12, the matching algorithm attempts to match this with an enrolled user. Software suitable for this purpose is known and will not be described in detail herein.
It should be noted that the secondary database 31 need not be a physically separate database from the primary database 30, but may instead be a portion of the primary database specifically dedicated for adaptive re-enrolment, as discussed further below.
The server 14 also provides an application interface 15 for integration with third party applications 17. This provides a simple method of furnishing a biometric "front-end" for any database driven application without requiring the integrator to have any in-depth specialist biometrics knowledge. In this respect, the application 17 may include more detailed personal information regarding users and details of fingerswipe unit locations, and will determine the actions to be taken in response to a user being identified at a particular fingerswipe unit 12. Accordingly, it is only necessary for the server 14 to transmit to the application 17 a user ID code and a fingerswipe unit ID code. The server itself need not have any detailed knowledge of users, the locations of fingerswipe units, or the use to which transmitted ID and location data will be put.
The system architecture described above allows devices or processes to be controlled on the basis of user IDs in a variety of ways. The server 14 may transmit a message on the network 16 indicating that a particular person had been identified at a particular location. Any external network devices 18B listening on the network 16 could respond to the server's message as appropriate. The server 14 may send a message to the appropriate fingerswipe units 12 for display to the user (i.e. through the LCD unit 26) . An external device 18A directly connected to a fingerswipe unit 12 (through the communications interface 20) may be controlled directly from the fingerswipe device 12 in response to a message from the server 14 that a particular user had been identified at a particular location. Where a third party application 17 is involved, the application may communicate with devices 18A/B and fingerswipe units 12 via the server 14, or directly with devices 18B and fingerswipe units 12 (and hence devices 18A) via the network 16.
B. Processes Performed in the Identification System (5)
1. Overview
Figure 2 shows how the processes implemented in the identification system 5 are distributed amongst its various components. Figure 2 is divided into three sections: the left-most section, middle section and right-most section respectively depict the processes performed in the enrolment unit 10, the fingerswipe unit 12 and the server 14.
Referring to Figures 1 and 2, the processes implemented in the identification system 5 can be broadly divided into two distinct operational phases, namely an enrolment phase 32 and an identification phase 34 (depending on the application this may involve identity verification and/or authentication as mentioned above) . During the enrolment phase 32, fingerswipe template data uniquely associated with a user is stored in the server's primary database 30 together with the name and/or other identifier (s) of the user. Once the user has successfully enrolled with the enrolment unit 10, he can use any of the fingerswipe units 12 to identify himself (identification phase 34) .
Common to both the enrolment phase 32 and the identification phase 34 are the processes of acquiring 36, reconstructing 38 and extracting 40 information from a user' s fingerswipe on the thermal strip sensor 22 (in the enrolment unit 10 or the fingerswipe unit 12) . For the sake of clarity, the acquisition, reconstruction and extraction processes conducted during the enrolment phase 32 will be designated with the identifier λe' (i.e. 36e, 38e and 40e) . Similarly, the acquisition and extraction processes conducted during the identification phase 34 will be designated with the identifier Λv' (i.e. 36v, 38v and 40v) .
On swiping a finger across the thermal strip sensor 22 (of an enrolment unit 10 or a fingerswipe unit 12), a series of overlapping slices of the finger's temperature profile are acquired 36e/36v by the acquisition software Si which also reconstructs 38e/38v a complete thermal map of the finger from the multiple overlapping slices acquired by the acquisition software. The thermal map is then processed by the feature extraction software S2 to extract 40e/40v a reduced data set (referred to henceforth as minutiae) and various other pattern features so as to produce an extracted template that is representative of the thermal map. In practice, any suitable extraction technique can be used for the extraction process 40e/40v, provided that each of the resulting extracted templates can be uniquely associated with a given individual. Once the extraction process 40e/40v is complete, the enrolment unit/fingerswipe unit transmits the extracted template (e.g. through a TCP/IP stack) to the server 14 for any required further processing and inclusion in the enrolment database.
In practice, an extracted template consists of a set of specific points, represented by, for example, a series of numbers and/or patterns. Thus, an extracted template could not be used to create a fingerprint and/or a false identity. Similarly, an extracted template could not be used to identify a person in a criminal investigation or to compromise a user's privacy in the event that another party inappropriately captured the template.
As previously noted, extraction can reduce a thermal map of around 180KB to a template of about 500 bytes. As also discussed above, many conventional biometric systems perform the extraction process in a central server. More particularly, in these conventional biometric systems, a fingerswipe unit acts as a dumb terminal and simply transfers an entire thermal map to a central server. However, in the present identification system 5, the extraction process is implemented locally in the fingerswipe units and the resulting extracted templates are transmitted to the server 14. Consequently, the specific implementation and distribution of processes between the devices in the identification system 5 effectively reduces the network traffic between a fingerswipe unit and the server 14. This feature becomes particularly important as the identification system 5 scales up to include increasing numbers of fingerswipe units 12.
2. Detailed Description of the Enrolment Phase (32)
During the enrolment phase 32, a template is extracted from a thermal map of a user's finger by the enrolment unit 10. The resulting extracted template is transmitted to the server 14 together with the user's personal details. This process is preferably repeated two more times with the same finger. The enrolment software S3 generalises from the three extraction templates of the user's finger by combining the three extraction templates to create the best composite template therefrom. The enrolment software S3 then stores 42 the resulting extracted template in the server's primary database 30 as a reference template together with (at least) a unique ID code. 2 (a) Adaptive Re-enrolment Procedure
When used for identifying young children who may grow considerably over time, the identification system 5 must be capable of accommodating changes in finger size to avoid the inconvenience of having to repeatedly re-enrol the children. To provide this facility, the identification system 5 implements an automatic adaptive re-enrolment procedure 44.
As will be discussed in the following description of the identification phase 34, every time a user uses the identification system 5, the matching algorithm 28 generates a score representing the match between the user's extracted template and the reference templates stored in the server's primary database 30. In parallel with these operations, the adaptive re-enrolment procedure 44 stores 46 the user's extracted templates in the server's secondary database 31. More specifically, for any given enrolled user, the adaptive re-enrolment procedure 44 only stores 46 those user's templates that closely match the user's reference template stored in the server's primary database 30.
After a pre-determined time interval (e.g. a month) 48, the server 14 automatically selects a primary subset of the best scoring templates (of the user) from the server's secondary database 31. The server 14 then determines the optimal combination and/or permutation of templates in a further subset from the primary subset. The server 14 then creates 50 a new extracted template from this optimal subset, using the generalisation techniques employed in the normal enrolment phase 32. The new extracted template then replaces the corresponding reference template in the server's primary database 30 and is used for identification until the next automatic adaptive enrolment period.
This process gradually increases the separation within the server's primary database 30 by ensuring that reference templates are successively overwritten with templates that: (a) most closely identify the user; and (b) are most strongly differentiated from the other templates in the primary server database 30. The continual adaptation of this combination of features progressively reduces the risk of false identification and thereby gradually improves the performance of the identification system 5.
3. Identification Phase (34)
On receipt of an extracted template from a fingerswipe unit 12 the server 14 determines 52 the user's identity by comparing the received template against all the reference templates stored in the server's primary database 30. More particularly, the server's matching algorithm 28 allocates scores to the received extracted template on the basis of its similarity with each of the stored reference templates and the number of matching minutiae therebetween. In general terms, a large minutiae value and a large similarity score indicate a good match between the received extracted template and a stored reference template. Depending on a number of pre-defined system thresholds, the matching algorithm 28 employs the scores of the received extracted template to generate a positive or negative identification of the user.
The identification system 5 provides a number of features to improve the flexibility and scalability of the identity verification phase 34. In particular, the identification system 5 employs the matching algorithm 28 and the primary server database 30 in a thread structure, which enables the server 14 to handle multiple inputs simultaneously. Similarly, the identification system 5 allows the pre-defined system thresholds to be modified to adjust the dynamics of the matching operation. For instance the chance of a "False Acceptance" (i.e. where a person is recognised as being someone else) could be set at 1 in 100,000.
Figure 3 focuses on the distribution of processes in the identification system 5 after a user has been identified. More particularly, in a similar fashion to Figure 2, Figure 3 is divided into three sections with the left-most, middle and rightmost sections respectively depicting the processes performed in a fingerswipe unit 12, the server 14 and third party application 17. Referring to Figure 3, assuming that the user is positively identified 52 (as being an enrolee in the identification system 5) , the server 14 transmits 54 a message to that effect (with time stamp and/or originating fingerswipe unit 12 location) to the third party application 17 which may use the information as appropriate (e.g. include it in a spreadsheet etc.). The third party application 17 may also transmit a response to the server 14, which the server 14 transmits to the external devices 18a/b and/or the fingerswipe unit 12 (as described below) .
In a first embodiment of the identification system 5, in which external devices 18A are directly connected to the originating fingerswipe unit 12, the server 14 also relays 56 an authentication signal { auth_sig) to the fingerswipe unit 12. On receipt of the authentication signal, the fingerswipe unit 12 may display 58 an appropriate message to the user and/or transmit a control signal 60 to the external device 18 to take an appropriate action (e.g. the debiting of a user account or, the opening of a door to allow access to a designated area) .
In a second embodiment of the identification system 5, the external device 18B is directly connected to the network 16. In this case, after transmitting a message to the third party application 17, the server 14 may transmit: (a) a control signal ( ctrl_sig) over the network 16, on receipt of which, the relevant external device 18 takes an appropriate action; and (b) a message ( auth__info) , indicating that a user had been identified, to the fingerswipe unit 12 (on receipt of which, the fingerswipe unit 12 may display an appropriate message to the user on its LCD 26) .
3. Example: School Dinner Payment Scheme
Schools within the UK provide a system whereby children whose parents have an income that is below a certain threshold can receive free school dinners. In many schools in Scotland, the class teacher collects fees for school dinners in the classroom. In this environment, it is very clear to the children who is and who is not entitled to free school dinners. Regrettably, receiving free school meals can stigmatise children at an early age, when their confidence and self-esteem can be fragile. This can damage them emotionally and inhibit their social and educational progress.
To try to protect children from the stigma associated with free school meals, the Scottish Executive has put legislation in place requiring that pupils be identified at the point of sale and fees for school dinners be debited automatically from an appropriate account. In other words, the pupil does not pay at the point of sale but the meal cost is deducted from either their own account or from a subsidised meal account. In this way, children who are entitled to free school meals are not readily identifiable from other children who have parents who can afford to pay. To date, however, the proposed systems have relied on the use of magnetic cards (i.e. smartcards), which the children must carry with them as a means of identification. This can be problematic in schools where the children are very young.
The present identification system provides a viable alternative to the above-mentioned smartcard/magnetic card system.
By using a biometric detection unit, the identification system 5 eliminates the need for ID cards or PIN numbers etc. This is particularly useful for very young users. Furthermore, since the fingerswipe units 12 transmit an extracted template to a server 14 rather than a direct image of a child's finger, the privacy and security of the identification system 5 is enhanced. This is a particularly attractive feature in a school environment, where security and privacy issues are particularly important.
Similarly, the reduction in network traffic provided by the process topology of the identification system 5 and the resulting improved processing speed is particularly beneficial in a school queuing environment, where a large number of impatient children must be processed within a short period of time.
Finally, since the adaptive enrolment procedure automatically accommodates changes in a child' s finger size (caused by the growth of the child over his or her period in school), the identification system 5 avoids the inconvenience and disruption of repeatedly manually re-enrolling the children in the school setting.
Those skilled in the art will appreciate that variations of the disclosed embodiments are possible without departing from the invention. As discussed above, the invention is applicable to fingerprint sensors other than thermal fingerswipe types and to biometrics and associated sensors/readers other than fingerprints.
Similarly, whilst Figure 1 shows only one fingerswipe unit 12, it will be appreciated that the identification system 5 may typically comprise multiple fingerswipe units at one or more locations.
Also as previously discussed, the possible applications of the identification system are not limited to the examples described above.
In view of the above, the skilled person will realize that the above description of the specific embodiments is made by way of example only and not for the purposes of limitation. It will further be clear to the skilled person that modifications may be made without departing from the scope of the invention.

Claims

Claims
1. A biometric identification system comprising: a central data processing system, and at least one biometric detection unit, each detection unit being operable to acquire biometric data from a user, process the biometric data to create a reduced size template representative of the biometric data, and send the template to the central processing system, the central processing system being operable to compare the received template with one or more of a plurality of reference templates, each associated with an enrolled user, so as to match the received template with a reference template.
2. An identification system as claimed in claim 1, wherein each detection unit comprises an integrated unit including a biometric sensor and data processing means for acquiring said biometric data from said sensor, extracting features from said biometric data and generating said template from said extracted features as output for sending to the central processing system.
3. An identification system as claimed in claim 1 or claim 2, wherein the central processing system is further operable to return a signal to the detection unit that is indicative of whether the received template matches a reference template of an enrolled user.
4. An identification system as claimed in any preceding claim, wherein the central processing system is operable to store a plurality of templates received over time and use these automatically to up-date the reference templates.
5. An identification system as claimed in claim 4, wherein the central processing system is operable to automatically up-date the reference template when a pre-determined interval has expired.
6. An identification system as claimed in any preceding claim, wherein the biometric detection unit is a fingerswipe unit operable to capture an image of a finger when a user draws their finger over a sensor.
7. An identification system as claimed in claim 6, wherein the sensor is a thermal imaging sensor.
8. An identification system as claimed in any preceding claim wherein the biometric detection unit is connected to one or more additional devices or systems and operable to receive data or control signals from the central processing system and to send data or control signals to those one or more devices or systems.
9. An identification system as claimed in claim 8, wherein the biometric detection unit is connected to a billing system so that amounts spent by a user can be automatically recorded in his or her account.
10. An identification system as claimed in claim 8, wherein the biometric detection unit is connected to a door activation mechanism, which is operable to open the door or allow the door to be opened, in the event that an enrolled user is identified.
11. An identification system as claimed in any preceding claim, wherein the central processing system is adapted to cooperate with a third party application.
12. A biometric identification method comprising: within a biometric detection unit, acquiring biometric data from a user, processing the biometric data to create a reduced size template representative of the biometric data, and sending the template to a central processing system, within the central processing system, comparing the received template with one or more of a plurality of reference templates, each associated with an enrolled user, so as to match the received template with a reference template.
13. An identification method as claimed in claim 12, wherein the detection unit comprises an integrated unit including a biometric sensor and data processing means, the method further comprising, within the detection unit, acquiring said biometric data from said sensor, extracting features from said biometric data and generating said template from said extracted features as output for sending to the central processing system.
14. An identification method as claimed in claim 12 or claim 13, further comprising returning a signal from the central processing system to the detection unit that is indicative of whether the received template matches a reference template of an enrolled user.
15. An identification method as claimed in any of claims 12 to 14 further comprising, within the central processing system, storing a plurality of templates received over time and using these automatically to up-date the reference templates.
16. An identification method as claimed in claim 15 further comprising up-dating a reference template when a pre-determined interval has expired.
17. An identification method as claimed in any of claims 12 to 16, wherein the biometric detection unit is a fingerswipe unit operable to capture an image of a finger when a user draws their finger over a sensor.
18. An identification method as claimed in claim 17, wherein the sensor is a thermal imaging sensor.
PCT/GB2005/001739 2004-05-07 2005-05-06 Biometric identification system WO2005109310A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP05742554A EP1784764A1 (en) 2004-05-07 2005-05-06 Biometric identification system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GB0410201.8 2004-05-07
GBGB0410201.8A GB0410201D0 (en) 2004-05-07 2004-05-07 User identification system

Publications (1)

Publication Number Publication Date
WO2005109310A1 true WO2005109310A1 (en) 2005-11-17

Family

ID=32482841

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/GB2005/001739 WO2005109310A1 (en) 2004-05-07 2005-05-06 Biometric identification system

Country Status (3)

Country Link
EP (1) EP1784764A1 (en)
GB (1) GB0410201D0 (en)
WO (1) WO2005109310A1 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012059446A1 (en) 2010-11-05 2012-05-10 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Method and x-ray test system for testing identical components by means of x-rays
GB2556625A (en) * 2016-10-27 2018-06-06 Zwipe As Secure enrolment of biometric data
US10216786B2 (en) 2010-05-13 2019-02-26 Iomniscient Pty Ltd. Automatic identity enrolment
CN110852382A (en) * 2019-11-12 2020-02-28 山东大学 Behavior recognition system based on space-time multi-feature extraction and working method thereof
WO2020067990A1 (en) * 2018-09-27 2020-04-02 Coolpay Pte. Ltd. Biometric identification method

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105575046A (en) * 2016-01-13 2016-05-11 广东小天才科技有限公司 Safety monitoring method and apparatus, smart watch, and safety monitoring system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5497430A (en) * 1994-11-07 1996-03-05 Physical Optics Corporation Method and apparatus for image recognition using invariant feature signals
US6061464A (en) * 1996-11-05 2000-05-09 Thomson-Csf Fingerprint-reading system with integrated heating resistors
US6076167A (en) * 1996-12-04 2000-06-13 Dew Engineering And Development Limited Method and system for improving security in network applications
US6535622B1 (en) * 1999-04-26 2003-03-18 Veridicom, Inc. Method for imaging fingerprints and concealing latent fingerprints
US20030123714A1 (en) * 2001-11-06 2003-07-03 O'gorman Lawrence Method and system for capturing fingerprints from multiple swipe images
US20030128240A1 (en) * 1999-08-09 2003-07-10 Martinez Chris J. Method, system, and computer program product for a GUI to fingerprint scanner interface

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5497430A (en) * 1994-11-07 1996-03-05 Physical Optics Corporation Method and apparatus for image recognition using invariant feature signals
US6061464A (en) * 1996-11-05 2000-05-09 Thomson-Csf Fingerprint-reading system with integrated heating resistors
US6076167A (en) * 1996-12-04 2000-06-13 Dew Engineering And Development Limited Method and system for improving security in network applications
US6535622B1 (en) * 1999-04-26 2003-03-18 Veridicom, Inc. Method for imaging fingerprints and concealing latent fingerprints
US20030128240A1 (en) * 1999-08-09 2003-07-10 Martinez Chris J. Method, system, and computer program product for a GUI to fingerprint scanner interface
US20030123714A1 (en) * 2001-11-06 2003-07-03 O'gorman Lawrence Method and system for capturing fingerprints from multiple swipe images

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10216786B2 (en) 2010-05-13 2019-02-26 Iomniscient Pty Ltd. Automatic identity enrolment
WO2012059446A1 (en) 2010-11-05 2012-05-10 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Method and x-ray test system for testing identical components by means of x-rays
GB2556625A (en) * 2016-10-27 2018-06-06 Zwipe As Secure enrolment of biometric data
WO2020067990A1 (en) * 2018-09-27 2020-04-02 Coolpay Pte. Ltd. Biometric identification method
CN110852382A (en) * 2019-11-12 2020-02-28 山东大学 Behavior recognition system based on space-time multi-feature extraction and working method thereof
CN110852382B (en) * 2019-11-12 2023-04-18 山东大学 Behavior recognition system based on space-time multi-feature extraction and working method thereof

Also Published As

Publication number Publication date
GB0410201D0 (en) 2004-06-09
EP1784764A1 (en) 2007-05-16

Similar Documents

Publication Publication Date Title
Pankanti et al. Biometrics: The future of identification [guest eeditors' introduction]
Maltoni et al. Handbook of fingerprint recognition
Wayman et al. An introduction to biometric authentication systems
Jain et al. An introduction to biometric recognition
Wayman et al. Biometric systems: Technology, design and performance evaluation
Schuckers Spoofing and anti-spoofing measures
Schouten et al. Biometrics and their use in e-passports
EP2667329A1 (en) Authorized party authentication template learning system and authorized party authentication template learning method
LV13365B (en) Secure biometric verification of identity
US20160217277A1 (en) One touch two factor biometric system and method for identification of a user utilizing a portion of the person's fingerprint and a vein map of the ub-surface of the finger
EP1784764A1 (en) Biometric identification system
Matyáš et al. Biometric authentication systems
US20040146186A1 (en) Method and system for processing personal characteristics
Perwej An Empirical Investigation of Human Identity Verification Methods
Scheuermann et al. Usability of biometrics in relation to electronic signatures
Dhir et al. Biometric recognition: A modern era for security
Fernandez Biometric sample quality and its application to multimodal authentication systems
EP1302907A1 (en) Method and system for processing personal characteristics
Dorizzi Biometrics at the frontiers, assessing the impact on Society Technical impact of Biometrics
Rejman-Greene Secure authentication using biometric methods
Bohm et al. Biometric systems
KORICHI Biometrics and Information Security for a Secure Person Identi cation
Stanley et al. Biometric authentication: A trustworthy technology for improved authentication
CN208938112U (en) A kind of INVS20M type resident identification card reading equipment
Jain Strengthening of Fingerprint Technology

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Country of ref document: DE

WWE Wipo information: entry into national phase

Ref document number: 2005742554

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 2005742554

Country of ref document: EP

WWW Wipo information: withdrawn in national office

Ref document number: 2005742554

Country of ref document: EP