Fingerprint classification using crease features

Existing fingerprint classification techniques make use of Level-1 or Level-2 features or their combination for fingerprint classification. On the other hand, Level-3 features have not been explored much for fingerprint classification. Specifically, the use of crease for fingerprint classification does not exist in the literature. In this work, we explore the use of crease features for fingerprint classification. Experiments are performed on the publicly available Hong Kong PolyU high-resolution fingerprint database DBII. Our experimental results are very encouraging and indicate that creases have huge potential to improve the performance of fingerprint classification.