Page 75 - Automated Fingerprint Identification Systems (AFIS)

60 AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS

Figure 3.3
Later Photo

change in lighting, the direction of shadows, background color or activity, or
even the size of the face in relation to the background affect the accuracy of
photographic recognition devices.
The field of facial recognition has not yet matured to the level where the
reality meets all the claims of its marketing staff. The hype that followed facial
recognition testing in airports, for example, was less than spectacular. Still,
facial recognition software is often promoted as an important method of iden-
tification. Movies and television programs, in one example, take viewers inside
the secure rooms of Las Vegas casinos, where walls of monitors show live images
of the players on the casino floor taken by hidden cameras. In a typical sce-
nario, security personnel look at a screen, and then suddenly magnify the face
of a gambler. “Wasn’t he barred from this casino?” they ask. Did facial recog-
nition software find this person? Not really. The first thing that drew the atten-
tion of the security personnel was the action of the player, not his or her face.
Only once they noticed the unusual activity did they focus on the player’s face
and capture the image. Following this, the facial recognition software they were
using made a comparison and presented choices; the casino personnel made
the final determination of identification. This is a quantum leap from the facial
identification hype that claims that any person can be found in any group
purely through electronic means.
The images that are captured in photographs are subject to changes in light-
ing, shadows, background, “noise” from other lighting sources, etc. For an iden-
tifier to be truly unique, however, it cannot be changed either by the owner of
the identifier or governmental or cultural differences; it must remain unique
for perpetuity. Fingerprints remain constant. Fingerprints are unique.