Finally, image quality was assessed using the National Institute
of Standards and Technology (NIST) Fingerprint Image Quality
(NFIQ) metric (version 4.0.1). The NFIQ is a public (export
controlled) software distribution that defines quality as a
‘‘predictor of matcher performance’’ (i.e., high quality fingerprint
images generate high matcher performance) [7]. The premise is
that a high quality fingerprint image will generate a high match
score when compared to a known-match ‘‘mate’’, and that this
score will be well separated from the scores associated with the
known non-match distribution [7]. In actual practice, computation
of a normalized match score is relatively straightforward and
based on the difference in the known-match score for a questioned
print, versus the average of all known non-match scores
normalized by the standard deviation of the KNMs. Based on this,
a high quality image will generate known-match and known nonmatch
scores that are well separated, and the non-matching scores
will be tightly clustered. As a result, the high quality image will