printer friendly version

Beyond the door

For many years now, the promise of biometrics has not been fully realised in large part because performance in the lab is not representative of performance in the field. That is the reason major programs for Citizen ID and others have never really been fulfilled.

The core problem is that conventional biometric technologies rely on unobstructed and complete contact between the fingerprint and the sensor, a condition that is elusive in the real world, a world that is wet, dry, or dirty and users are not all young office workers with great skin who are experienced at using biometrics.

Wet fingers fail

Wet conditions are notoriously difficult for both semiconductor and conventional optical fingerprint sensors to handle. And, yet, moisture is a fairly common real world condition. Some environments are naturally damp, due to climate (London) or setting (a spa). Some people have moist hands. It is also typical for people going through security to be nervous — and to have sweaty hands.

Conventional optical technologies are often unable to produce images in wet conditions because excess moisture obscures fingerprint ridges, resulting in images of puddles, not fingerprints.

Dryness is a problem

Has any other real world condition caused so much trouble in the biometrics industry? Dry fingertips are common, caused by anything from climate conditions and natural skin characteristics to frequent hand-washing and air travel. For instance, a high desert climate causes dry fingers in an entire population!

Most optical sensors are configured to look for the presence or absence of total internal reflectance (TIR), which is the phenomenon whereby the interface between glass and air acts like a mirror at certain angles. The contact between the skin and the platen defeats the TIR, allowing those points of contact between the finger and the sensor to be imaged. Thus, those points of contact must be complete and unobscured to enable the conventional sensor to collect a fingerprint image. With dry fingers, this is simply not the case. Establishing firm and complete contact with the sensor is very difficult with dry fingers. There is not enough moisture in the skin nor is the skin pliable enough to facilitate the contact necessary for TIR imaging.

It is a dirty world out there

The real world is a rough place and most of us are showing some wear and tear on our hands. Additionally, people do not have time to wash and lotion their hands before they use a fingerprint sensor.

A construction site is an interesting real world case. Construction workers work with their hands and have the cuts and calluses to prove it. Additionally, the construction site is dirty so workers may have grime on their hands when they approach a fingerprint sensor. Altogether, this real world scenario is a nightmare for system administrators whose conventional fingerprint sensors depend on quality contact between the finger and the platen.

We are all different

Many people, both young and adult, have small or fine fingerprint features that can be difficult to image. If the sensor cannot differentiate between these fine characteristics, system performance will suffer.

Age is another physiological characteristic that can affect the ability of a sensor. One effect of ageing is the loss of collagen in the skin; elderly fingers have soft fingerprint ridges that collapse into each other when the finger touches a surface. Because many sensor technologies depend on the quality of contact between the finger and the sensor to collect a good image, soft fingerprint ridges can be difficult.

There are behavioural differences across user populations that can also affect performance. People have different levels of experience with technology and biometrics and this affects how they approach the fingerprint sensor. For example, some people may tend to press hard and others, being more tentative, may barely touch the sensor at all. For technologies that depend on the quality of that touch, this can be a big problem.

You do not want to get spoofed

Being able to discriminate a real finger from an imposter or spoof fingerprint is extremely important. Access – through national borders, into buildings or physical plants, and into electronic devices such as PCs and networks – is increasingly unsupervised. Nonetheless, security, labour costs and convenience often necessitate the use of biometric access control methods such as fingerprint verification.

Ridges are easy to imitate using common household products and ingredients. A variety of materials, from the inexpensive to the very sophisticated, can be used to circumvent traditional fingerprint identification systems. Some of these materials are so thin and colourless that they can even be used, undetected, in access control environments that have trained attendants. For example, a gummy bear candy that costs a few cents can make a very accurate fingerprint that will spoof a traditional fingerprint.

Overcoming these common obstacles

Multispectral imaging is a sophisticated technology developed to overcome the fingerprint capture problems conventional imaging systems have in less-than-ideal conditions. This more effective solution is based on using multiple spectrums of light and advanced polarisation techniques to extract unique fingerprint characteristics from both the surface and subsurface of the skin. The nature of human skin physiology is such that this subsurface information is both relevant to fingerprint capture and unaffected by surface wear and other environmental factors.

The fingerprint ridges that we see on the surface of the finger have their foundation beneath the surface of the skin, in the capillary beds and other sub-dermal structures. The fingerprint ridges we see on our fingertips are merely an echo of the foundational ‘inner fingerprint’.

Unlike the surface fingerprint characteristics that can be obscured by moisture, dirt or wear, the inner fingerprint lies undisturbed and unaltered beneath the surface. When surface fingerprint information is combined with subsurface fingerprint information and reassembled in an intelligent and integrated manner, the results are more consistent, more inclusive and more tamper resistant.

Multispectral imaging technology can also detect living flesh from non-living flesh or other organic or synthetic materials. Liveness detection is built from machine learning algorithms. Using these algorithms and the wealth of information available from multispectral fingerprint images, liveness detection capabilities can be updated if new spoofs are identified.

Unlike any other fingerprint technology, this learning capability allows multispectral imaging sensors to keep up with new threats. The inexpensive and readily-available films and prostheses that easily defeat conventional fingerprint devices are rendered ineffective against this technology.

What does all this mean?

For some time, biometrics have been limited primarily to door types of applications – ingress/egress, time and attendance and similar tasks – because of the problems of traditional biometric technologies. Some organisations have gone so far as to use a PIN bypass if the biometric reader is having problems, defeating the whole purpose of having a biometric.

Contrarily, multispectral fingerprint sensors capture high-quality images because the direct imaging process does not depend on a clean finger/sensor interface. At last, biometrics can provide the same type of reliability as a card but removes all the negatives of the card, including cost of the cards themselves and the more expensive cost of managing cards. After all, nobody leaves their finger at home nor does it wear out.

More importantly, now biometrics, which determines that you are you – not what you carry – can be used in more places and more applications. For example, more than 40 million people are already enrolled on multispectral imaging-based systems at locales ranging from the classic door access control situation to the gates of the world’s favourite theme parks. Such readers are keeping borders secure around the world. Indeed, more than 400 000 people pass through multispectral imaging sensors every day at the Hong Kong border crossing.

Most importantly, though, multispectral imaging provides reliable identity verification in settings going beyond the door. For instance, CareFusion (formally a division of Cardinal Health) provides secure medical drug dispensing equipment. As an experienced user and client of biometrics technology, CareFusion understood these challenges. After an extensive industry wide evaluation of biometrics technology and devices, CareFusion selected multispectral imaging fingerprint sensors and software for integration into its latest generation of dispensing cabinets.

Results have been outstanding. The performance of multispectral imaging in its MedStation 4000 has clearly demonstrated a dramatic decrease in failure rates and other performance problems experienced with conventional biometric devices. With multispectral imaging, the system’s failure to enrol and failure to acquire has been dropped to virtually zero. Where in the past, the solution was only effective about 80% of the time for four out of five users, the new MedStation 4000 performance is near perfect.

Similar applications can now be undertaken for getting on a company’s data network, operating a forklift, signing for merchandise, creating major citizen ID programs and virtually every type of application in which organisations now feel forced to use cards, keys or PINs.

Brand New technologies distributes Lumidigm in South Africa.

Share this article:

Further reading: