ANALYSIS OF GALTON-HENRY CLASSIFICATION METHOD FOR FINGERPRINT DATABASE FVC 2002 AND 2004

Abstract

According to Galton-Henry, fingerprint classification varies into arch, tented arch, left loop,
right loop, whorl, and twin loop schemes. The approach is depended on the existence of a core and delta of
each fingerprint. If this method is implemented for all fingerprint benchmark databases, then less than 41% of
the fingerprint is being classified, by 37.19% for FVC 2002 and 40.31% for FVC 2004. Therefore, in this
research, three requirements are needed to improve the classification result of the fingerprint, i.e. core point
and its number, ridge frequency and ridge direction, and tented arch as additional requirement. This approach
improves the result which is only 5.94% and 1.56% that is unclassified for FVC 2002 and 2004 consecutively.
Then, to evaluate the time taken in executing the algorithm, this research does the evaluation by offering two
possible conditions of the input of the fingerprint. The first type is without the fingerprint classification,
while the second type is with the classification step along the algorithm process. The latter type requires an
additional step namely RoI selection process to select a desired area of the fingerprint.