Abstract:
Fingerprint classification provides an important indexing mechanism in a fingerprint
database. Fingerprint classification as technique to reduce long time for search and
reduce computational complexity of automatic fingerprint recognition system .This
paper presents a technique for fingerprint classification based on detection the global
features of the fingerprints by estimation orientation field of image using sobel
operator then classify fingerprint image based on global features. The proposed
technique was tested on FVC2004 database using 320 images. The highest
classification accuracy of the results achieved was 45.8% for Right Loop class and
the lowest was 2.2% for Whorl class, for Left loop fingerprint is 29.4%, The Arch
fingerprint is 4.10%.