Abstract :
The fingerprint orientation field is a widely used feature for developing indexing strategies. Such feature brings stability and decreases the response times during the identification process. The use of attribute relational graphs and dynamic masks can exploit the features provided by the partitioning scheme. The penetration index and the error rate for each of the proposed strategies were determined. To verify the results obtained were used the databases provided by the Fingerprint Verification Competition. Both search schemes reveal the facilities offered by the orientation field for guiding the searching process; reducing the number of comparisons in 53.24%; proving its stability against different fingerprint image qualities. Based on the results, it was proven that the adoption of such indexing strategies will reduce the response time in the identification module proposed by the Identification and Digital Security Center at the University of Informatics Sciences.
Keyword :
attribute relational graphs, dynamic masks, fingerprints indexing, orientation field