GENDER CLASSIFICATION BASED ON ANN WITH USING FINGERPRINT FEATURE VECTORS


Ceyhan E. B., SAĞIROĞLU Ş., Akyil E.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, cilt.29, sa.1, ss.201-207, 2014 (SCI-Expanded) identifier identifier

Özet

There are statistical analysis studies on gender classification by analyzing a particular area of fingerprint in literature. In these studies it has been determined that ridge counts of fingerprint were used. The studies was performed with limited data. The data used depend on race or country. Also the processes were performed manually. In this study for the first time in the literature, an intelligent model have been identified using all feature vectors of fingerprint with artificial neural network models for determining the relationship between gender and fingerprint. Then results that have been obtained are presented. In our preliminary analysis studies, it has been observed that fingerprints contain a changeable amount of usable data. The quantity of data used in this study is determined by considering this amounts. Success rate of the developed intelligent system is obtained 72% in our tests. The results show that the relationship between fingerprint and gender is high.