It’s something of a puzzle why even people working in the security industry don’t know this, but all fingerprint readers aren’t perfect, and some of them aren’t even adequate for their task.

Unlike RFID tags or magnetic cards, fingerprints can’t be lost, stolen or shared. This has obvious advantages for time & attendance and access control systems, but not understanding the technology can lead to poor purchase decisions.

Sensors

As any camerophile will tell you, resolution isn’t everything, but it’s a pretty big deal. The features of a fingerprint are tiny, so paying extra for a higher dpi makes sense. Most sensors you’ll encounter come in one of two flavors: optical and capacitive. The difference is that an optical sensor takes a two-dimensional picture of the finger, while a capacitive chip produces a map in three dimensions. Obviously, this is much harder to fake, especially without the active cooperation of the person whose finger it is – some optical sensors, and no kidding here, will respond positively to an ordinary photocopy of the original print.

If you need an extremely high level of certainty about when someone was where, the more expensive multi-spectral sensors might be what you need. It’s also a type of optical sensors, but actually records fingerprints in several different wavelengths. They actually “see” the pattern of blood vessels below the surface, so they’re extremely difficult to subvert.

Software

Unless you’re an OEM, you’ll not work directly with the embedded software that determines whether a print is valid or not. However, you should understand the concept of false positives (letting someone in who isn’t in the system) and false negatives (not recognizing an enrolled user). The manufacturer will give rates for both, and you can generally rely on these figures.

Many buyers will focus only on the false positive percentage, but this makes little sense when you think about it. A random stranger won’t even try getting into the building if he only has a one in 10,000, or even one in a hundred chance of this working. However, if regular users start getting frustrated, they’ll start to subvert the system by doing things like blocking open controlled doors, and your security will suffer.

On another note, you might want to seamlessly integrate access control, time and attendance, and your payroll software – find out how much this will cost, right at the outset.

Knowing What You Want

A lot of improvements in either business processes or security measures fail in their purpose, just because that purpose was never explicitly defined. Rather than immediately spending money on a biometric system, think about other steps you can take. Putting a $200 lock on a $50 door is a waste of $200.

That’s not to say that biometric technology doesn’t have a role to play; the only time I’ve seen it not to be appropriate was in a factory that made steel wool – many of the workers didn’t have fingerprints. If you do detailed planning before even phoning suppliers, you’ll get the maximum benefit. If you wait until after the installation has been done to make changes, these will be far more expensive.

LEAVE A REPLY