If you’ve tired of hearing over the years that biometrics are about to come in, for staff identification – the devices have come in, the prices have come down, and for guarding companies they are perfectly usable, says a software company.
For several years those selling biometric access products have it’s fair to say been dogged by speed – or lack of it – of devices reading people’s biometric, whether it’s a voice, or hand, finger, some part of the eye, or face. Some biometrics more than others have had problems with false readings. And then price of devices has out buyers off. It’s made biometric technology something more for ‘high’ security uses, in nuclear power plants, for instance; something Hollywood shows when seeking to impress cinema-goers that a place is highly secure and important. No longer, says Ian Lenagan, chairman of WorkPlace Systems, that writes software so that the devices work. He reports that 75 per cent of the workplace terminals being installed for staff access control have a biometric. That means the likes of bakeries. The advantage: an employee can claim to ‘forget’ their badge or access code, but can hardly forget to bring their finger or hand. Speed of identifying someone presenting their biometric has come down to a second or less, which compares with a swipe card reader. Users accept that. <br><br>He predicts a rush of guarding companies to use biometrics, to clock staff in and out: "You will find security companies are using the advantage of biometric control of their own people as a quality recognition factor, a unique selling point, for them selling their services to customers." Besides the contractor knowing the right guard has turned up to the right place, and doing away with one employee clocking on a mate who’s late by keying in his password or swiping his badge, you can use the data for time and attendance. You’ve got staff taken on under TUPE working different hours and pay scales? No biting a pencil and scratching your head to work out who’s entitled to what, software can do it. One guard cannot be bothered to turn up on Saturday night and another has to stand in? The device is only half the story; the software tells you who was the no-show and who should be paid for standing in. As was shown in the May article about contractor Wilson James’ use of WorkPlace Systems software, it saves on admin time and effort and minimises staff queries over pay packets and hours. <br><br>Ian Lenagan separates false accept rate – an unauthorised guy trying to get through the door – and false reject rate – a genuine user of the system, refused entry by error. And he argues that a commercial use of a biometric is different from a security guarding use. A single guard presenting his biometric and having to wait ten seconds? Acceptable – or at any rate more acceptable than at an office or shop, with staff making a queue all wanting to enter for 9am at the last minute. Say a high street retailer has 100 people going in and out of a site, an average four times a day, five times a week. That adds up to 2000 attempted registrations. If you’re being sold a product that claims a reject rate of 0.1 per cent, that may sound fine, but in this case it means on average two genuine people a week will have their biometric rejected. As Ian Lenagan says, that will annoy employees. In another case, say you have a nuclear power station, that wants to protect against unauthorised entry. If a product claims a false accept rate of 0.1 per cent, that again might sound fine, but it means that on average one person in a thousand not authorised to enter, will. The false reject and accept rates depend on how sensitive the device is – is it measuring 100 or 1000 points on your hand or finger? The more, the more accurate it can be. <br><br>And what sort of biometric technology? Ian Lenagan divides them by four:<br>- optical; similar to a photograph, taking an image. <br>- conductive, measuring an electric charge between the points on your finger; current is higher or lower, depending on the troughs and ridges of your finger. <br>- Thermal, rather like swiping, on the basis of heat. Cool and warm readings again depend on ridges and troughs of your finger. <br>- And pressure sensors. You press your finger on a device, and it measures the pressure across say 1000 points; again, the readings are higher or lower depending on your ridges or troughs. Most technology is optical. <br><br>Another factor is how users take to the devices. Having a low-power laser look at the back of your eye – your retina – may well be less liked by staff than having their face scanned by a camera three feet away. To come back to false rejections: we have all been on a telephone to an insurer or other corporation using voice recognition technology and it doesn’t seem to work. Ian Lenagan speaks of people who work with cardboard as having a tendency to having their fingerprints worn away, shallower; the same with handling some chemicals. Dirt, too, can have an impact on the fingers, when you come to present them to an access terminal. Hence some terminals may require you to wipe your hand on a cloth or with some liquid before you place your hand to the device. <br><br>"But what I would say about everyone of these devices is that they are getting better," Ian Lenagan says, " and whereas in 2000 for example you would pay between £1500 and £3000 for a terminal, you are now down into a price range which is generally around £500." And coming down further, he adds. For that money, you can get a device that can read your fingerprint optically, and read it to identify it against a couple of hundred, stored inside the device (to be exact, an algorithm that calculates a numerical representation of your finger); or thousands, if you are verifying. By comparison, if a guarding company is thinking of using biometric recognition where ten or 20 guards are at any site, that, Ian Lenagan suggests, is financially most possible. <br><br>To return to what’s being measured: it depends on the algorithm. Take for example facial recognition – your face is measured from the eyes down to the middle of the chin, including your eyes, nose, mouth and eyebrows. What’s gathered are not so much a snapshot as the shape, the points between your identifying marks, and the algorithm comes up with a mathematical, unique numerical code of you. Similarly, what’s scanned in your iris (the front of your eye) is the geometric pattern, and a number calculated. For a guarding application, if about 20 are ever on a site, it’s possible, Ian Lenagan says, to store up to a couple of hundred fingerprint templates in the device; and within one second, when someone presents their finger at the terminal, the device can compare that finger against the 100 or 200 on the database, until it finds the right one, and lets you in, and clocks you in; that’s where the rostering software comes in, using that biometric data.
