Handwriting

At first glance, using handwriting to identify people might not seem like a good idea. After all, many people can learn to copy other people's handwriting with a little time and practice. It seems like it would be easy to get a copy of someone's signature or the required password and learn to forge it.

But biometric systems don't just look at how you shape each letter; they analyze the act of writing. They examine the pressure you use and the speed and rhythm with which you write. They also record the sequence in which you form letters, like whether you add dots and crosses as you go or after you finish the word.

Photo courtesy Softpro This Tablet PC has a signature verification system.

Unlike the simple shapes of the letters, these traits are very difficult to forge. Even if someone else got a copy of your signature and traced it, the system probably wouldn't accept their forgery.

A handwriting recognition system's sensors can include a touch-sensitive writing surface or a pen that contains sensors that detect angle, pressure and direction. The software translates the handwriting into a graph and recognizes the small changes in a person's handwriting from day to day and over time.

Determining Accuracy All biometric systems use human traits that are, to some degree, unique. Which system is best depends on the necessary level of security, the population who will use the system and the system's accuracy. Most manufacturers use measurements like these to describe accuracy: False Accept Rate (FAR): How many imposters the system accepts False Reject Rate (FRR): How many authorized users the system rejects Failure to Enroll Rate (FTE): How many people's traits are of insufficient quality for the system to use Failure to Acquire Rate (FTA): How many times a user must present the trait before the system correctly accepts or rejects them