The following is a list of areas where biometrics can make a major contribution, and where ABI can advise and assist.

Law Enforcement
Airport security
Counter-terrorism
Child recovery
Access control
ID cards/visas/passports
Drivers licenses
Smartcards

Law enforcement

Historically the major user of biometrics, police agencies have used fingerprinting as a means of identifying criminals for well over a hundred years. Police gain the most benefit because a criminal’s biometric information such as fingerprints, mugshot, DNA, etc, may already be held in a database. This enables forensic information collected at a crime scene to be matched against it.

An Automated Fingerprint Identification System, or AFIS, is designed to enable a fingerprint to be matched extremely quickly against a large number of records in a criminal database. To do this effectively it will almost always hold encodings of all ten fingers.

Law enforcement agencies have achieved significant success with facial recognition, matching the mugshot (or even composite drawing) of a suspect against a database of offenders. This is particularly useful where the individual has refused to give his name, or has given a false name.

back to top

Airport security

Post 9/11 a real need emerged to identify terrorists trying to board planes. As in many cases the only information available on suspected terrorists was a mugshot or surveillance photo, facial recognition was thrust to center stage as the biometric which could help identify them before they board the plane. While much work has been done in this area, the practical and logistical issues which have to be overcome have meant that so far, implementation has not been as fast as originally anticipated.

back to top

Counter-terrorism

The airport security need outlined above is a subset of the wider fight against terrorism. In general, counter-terrorism involves trying to identify someone in a public area who is on a watchlist, where frequently the only description of the individual comes from a low-resolution long-distance photo. This remains challenging from both the technical and public acceptance viewpoints.

back to top

Child recovery

The proliferation of the Web in the second half of the 90’s lead to an explosion in the exchange of CSAM imagery. Imaging and associated face recognition technologies have achieved notable success in tracking down child abusers and identifying missing children.  ABI has worked directly with law enforcement agencies on a system to process these images, including the largest such system in the world at the time.

back to top

Access control – physical

The use of biometrics such as fingerprints or face to verify that an individual is who they claim to be is well-established. While a photo on a government-issued ID such as a driver’s license or passport enables a visual check to detect when the card is being presented by someone other than the individual to whom it was issued, this is far from infallible. Requiring the photo to be matched against the record in a central database strengthens this dramatically. Other biometrics such as iris recognition have increased the confidence level of the matches to approaching 100%, although the recognition process can take more time.

Where speed of identity validation is particularly important, for example at the entrance to a secure building, where many people arrive at the same time in the morning, the use of RFID-based ID cards together with face recognition can enable a valid employee’s identity to be confirmed without them even slowing down.

back to top

Access control – online

The need for online identity validation has increased dramatically over the last few years, and this trend is set to continue. Major application areas include financial transactions, health services, call centres, access to centralised personal records, etc. While the core functions are similar to those involved in physical access control, there are important differences. For example, it’s unrealistic to believe that individuals accessing central systems from their home or home office will pay for additional hardware such as hand or iris scanners to confirm their identity.  This constrains the biometrics that can be used to those where the hardware already exists in a personal computer or smartphone, such as the microphone for speaker recognition or the camera (if available) for face recognition.  Where the person is using a computer, typing style and mouse movement can also be effective.

Issues to be addressed include onboarding/registration, i.e. the establishment of a database record for the individual where their identity has been confirmed. This is generally handled by matching a live on-screen shot of their face against their image on a government-issued ID document, or using speaker recognition to match their style of talking.  The key factor influencing the effectiveness of this is the quality of the image on that document.

In some instances a match can be borderline. When this happens additional approaches can be used, including other biometrics such as fingerprints, and also non-biometric approaches including MAC/Device ID checks, location, out-of-range authentication, knowledge-based authentication, etc.

A key issue with any biometric-based approach is to check for liveness.  With any of the above biometrics there is a risk that someone will try to pass themselves off as someone else in order to gain access by using a previously-recorded copy of a person’s face, voice, fingerprint, etc.  While this can be surprisingly difficult to address, we can advise on the various ways to detect it.

back to top

ID cards/visas/ePassports

The issue here is identity fraud, where the agency needs to know primarily three things –

  • before issuing a document, has a similar document already been issued to this individual under a different name?
  • on subsequent presentation of the document, say at a Point of Entry, is the individual presenting it the same individual to whom it was issued?
  • If not the same individual, who are they?

Various biometrics can be used for these purposes. In 2003 the International Civil Aviation Organization (ICAO), stated a preference for fingerprints and facial recognition to be used on travel documents.

ICAO has mandated that all member countries will move to electronic passports by 2010.  ePassports hold the facial and fingerprint images on an RFID chip in the ePassport.  When the traveller presents their ePassport on arrival at a Point of Entry, their fingerprints are read and a photo is taken of their face.  These, and the images on the chip, are encoded and compared.  If they match within a redefined level of tolerance, they are deemed to be of the same person.  Holding the images, rather than the encoded strings, on the chip gives each country the flexibility to use whatever face and fingerprint recognition systems they wish.

back to top

Drivers licenses

While drivers licenses is in many ways a similar application to ID cards, it is still the case in most countries that ID cards are not used and therefore the drivers license database is by far the largest database available of adult citizens. Drivers licenses are an ideal application for facial recognition because the individuals cooperate in having their photo taken, while the environment has been set up with optimal lighting, camera resolution, distance from the subject, and neutral background. While this offers a potentially invaluable tool in identifying individuals whose details are in the database, because of privacy concerns it cannot be used for this purpose in certain jurisdictions.

back to top

Smartcards

Smartcards are not a different application, but a particularly secure means of providing an individual with an identity card. They are especially appropriate for biometrics because sufficient memory can be made available to hold the individual’s facial image and a number of encode arrays. While these will always be held in a central database as well, having them on the card itself enables it to be used in locations where there may be no network access.

The use of Radio Frequency Identification (RFID) enabled smartcards minimizes the time taken to verify a person’s identity, by allowing data on the card to be read without direct contact.

back to top