Premier Electronics is a distributor of CCTV, network video recorders, digital video recorders and biometric access controls; with so many strings to their bow, it's fair to call them experts in the field of casino security. The latest technology in their field revolves around facial recognition, which has been implemented in the world's largest casino, in Macau, for identifying staff; around 100 readers are in place at The Venetian. MD Darryl Coffey talked to us about the system, and how it can help a casino's security.

Casino International: What – looking particularly at the biometric side – are the key benefits for installation of the facial recognition system?
Darryl Coffey: Speed of access. The biggest problem of most biometrics is time to get through; you can go to Heathrow and see their 'fast track' technology, you could try that and just time the people going through, I’ve seen times from 30 seconds to 40 seconds. If you’re working at trying to get your staff in, it may involve putting through several thousand staff at the start of each shift at the Venetian, so you’re going to have queues around the block with times like that to get an individual through. There's a bank in Switzerland where they bring in 1200 people at the start of the day, and they’ve installed A4 Vision (aka A4 Bioscript), facial readers, because of the speed of access.

CI: So what is the speed?
DC: Recognition is under a second.

CI: Why is it so much quicker?
DC: Ease of use and the way it matches, it just works incredibly fast. If you’re using for instance a fingerprint, you put your finger down, it looks at your fingerprint, decides if it’s correct and gives a match, and if not you go back through the process again. The A4 Bioscript facial recognition system works in real time, well virtually real time, it’s doing 15 shots a second, and it will get you in one of those if you’re on the list.

CI: Is it sacrificing any accuracy for speed if it’s under a second for recognition?
DC: Of the systems that have been tested by Government agencies, it’s the only system that hasn’t been spoofed, so the answer is no.

CI: So why isn’t everybody using it?
DC: It’s still a growing industry. The majority of sales are into the States mainly, because there’s numbers out there. It’s a technology that people sometimes don’t trust, mainly I think because fingerprints are pushed very hard, and have that connotation that you’re a criminal because you had your fingerprints taken, and I don’t think people like that. Another reason is that you have to physically work with it, and we’ve heard objections where people don’t want to touch where someone else has touched. If someone’s been eating for instance a bacon sandwich and touches the sensor, and some one else comes along who couldn’t touch bacon, we have a major problem going on.

CI: One of the major obstacles to the adoption of technology is cost, so how affordable is this, is it early days for the technology so it’s still quite expensive?
DC: It’s not the cheapest technology out there, but it has been around for about seven years, and been sold for about five of those, so it’s still maturing. Prices are more expensive than for fingerprint readers because you can pick those up very cheaply these days, but there are inherent problems, and the major thing it’s going to save you is time and accuracy; you have to put that in the equation when you’re asking how much something costs. The main questions are does it do the job, does it work effectively, is it reliable? So you can buy something at lower cost that just spends most of its time down or not recognising people and that’s going to prove more expensive in a revenue sense than buying the equipment that will actually do the job first time.

CI: What does the actual kit comprise?
DC: There’s a couple of parts to it; one is the enrolment station, and your details are stored on to a computer. It is a system that measures ground-based points on the face, so this is not a system that looks at a picture of someone and looks at light and shade and does a comparison against that. That kind of system has a one inherent weakness and that’s all you're looking at is light and shade, so if the lighting changes from when you took the enrolment picture to when you want to use it in a real-life situation, you’re going to have a problem recognising that. Those systems don't like glasses, hats, beards or anything else, so that technology’s not brilliant in that way.
A4 Vision/Bioscript 3D Facial Recognition measures in sub-millimetre patches on the screen; it fires a grid of invisible light at the face so it can even work in total darkness. The grids distort according to the contours on the face and the system reads the change of those angles and that is how it measures you; it takes a number of hard points across the face so the nose, the top of the eyes, the chin, the lips, areas that don’t change dramatically over the life span of the person after about the age of 11. Those minutiae are then stored as the comparison, you can either run it then as a pure one-to-one so either against a swipe card if people are using those, tracking cards or whatever, or one-to-many so it can search databases.

Printer friendly version

Email this article to a friend