Cryptography Reference
In-Depth Information
Nevertheless, signature recognition has considerable problems. First, the
reject rate (FRR) is high if the system is to be moderately secure. You
will have noticed that your handwriting changes, for example, after phys-
ical efforts or under stress. Another problem is that though you obtain a
huge amount of data from a Smartpen, you have to know how to evaluate
them: there is an obvious lack of theory. On the other hand, due to data pri-
vacy in some countries, including Germany, it is not possible to use arbitrary
words instead of signatures since graphologists can learn pretty much about
a personality from the handwriting.
•
Hand geometry:
This is a relatively old method. The geometry of hands
differs individually. There are forging possibilities similar to fingerprints (but
the hand geometry is easier to 'steal'), and I can imagine that swollen hands
cause problems. Nevertheless, the method is the second most frequently used
system after fingerprint recognition, particularly for admission control, e.g., in
nuclear power stations, or in the students' dining hall at Georgia University,
to reduce counterfeiting.
•
Vein patterns:
Similarly to retinal scan systems, these systems acquire the
extremely individual arrangement of veins in a hand. The method appears to
be robust and pretty secure against forgery, but it is not widely used yet.
•
Combined methods:
At the CeBIT 1998 trade fair, a system by the name of
BioID
(
www.bioid.com
) was introduced, which acquires facial traits, voice,
and lip dynamics concurrently. It was remarkably insensitive to eyeglasses,
beards, heavy tongues, and new face wrinkles. While achieving a good FAR
with this system, however, you have to put up with a very high FRR. The
system has meanwhile disappeared.
•
Methods of the future:
In addition to ongoing research in voice recognition,
which is currently still unreliable (at least for civilian uses), intensive research
work has been put into acquiring other individual characteristics. A particu-
larly original system was introduced in the
New Scientist
dated December 12,
1999: every human has a different
gait
. This research work was motivated by
the desire to identify masked bank robbers on video clips. Also, shoplifters
pretending to be pregnant while smuggling stolen merchandise out of stores
could be identified by their gait (genuinely pregnant women walk differently).
These are promising prospects. However, I could also imagine that there is
a potential for surveillance of people from considerable distances, and I'm
afraid this won't remain utopia. More about this in Chapter 8.
Figure 6.11:
(
continued
)
