| Error Rate of Fingerprint
Presented by: Xuejun Tan
ABSTRACT: Fingerprint has long been used for person authentication
and people believe that every person's fingerprint is unique.
However, there is not enough scientific research to explain
what is the probability that two fingerprints, which are impressions
of different fingers, may be taken as the same one. In this
paper, we proposed a formal framework to estimate the error
rate of fingerprint matching. Not like a previous work, which
only takes into account of the positions of minutiae, in our
model, we also take into account of the relations between
different minutiae. The results of our work show that the
error rate of fingerprint matching is not only decided by
the number of minutiae in the template and query fingerprints
and the uncertainty area of a minutia, but also the ridge
counts between two different minutiae.
The comparison between the theoretical result and the real-data
result, which uses NIST-4 fingerprint database, shows that
it still needs much work on fingerprint matching in real-world
applications, so that the performance of the existing fingerprint
matching algorithm can be close to the theoretical result.