Information Technology Reference
In-Depth Information
one who matches all of his or her own characteristics. Even in the case
of identical twins who share the same DNA pattern, developmental dif-
ferences result in variances in fingerprint, hand geometry, motion kine-
matic, iris, retina, pore, and blood vessel patterns. Effective measurement
and analysis of these characteristics may provide the best means for per-
sonal identification, once the technologies involved have matured from
their current state.
Tests of “what you are” are generally referred to as biometric tests,
in that they measure specific biological characteristics against a pre-
viously identified value. Because biology changes, biometric measure-
ments must be updated regularly in order to remain at peak efficiency.
Performance of biometric identification systems is measured in terms of
two primary metrics:
•
False rejection rate (FRR):
A measure of the number of times a
valid identification fails, also referred to as the type I failure rate.
•
False
acceptance rate (FAR):
A measure of the number of times an
invalid identification succeeds, also referred to as the type II fail-
ure rate.
Biometric identification systems may use direct identification or
behavioral l tests, depend ing on t he t y pe of biologic a l a ssessment employed.
Examples of direct identification systems include the following:
•
Palm-print systems.
Palm-based biometric systems employ a scan-
ner able to measure and identify the patterns of ridges and valleys of
the palm. Like an enlarged fingerprint reader, this solution identifies
unique characteristics on the palm; however, it can generate false
rejections because of injury or mis-positioning of the hand within
the scanner. Like all solutions that involve direct contact with the
reader, palm-based security systems may also be opposed by some
users fearful of contagion.
•
Hand geometry.
A second full-hand biometric system employs the
unique patterns of growth in the bones to provide identification. By
placing the fingers against pegs at specific locations in the reader,
finger length and joint configuration can be identified. This solution
is subject to an elevated potential for false response because finger
length may be matched through accident or surgical manipulation,
