Generating Cancellable Fingerprint Template Using Local-Similar Image

Fingerprint template is considered as a promising solution to solve the problem of fingerprint revocation, also is an indispensable key technology in the field of identification. In this context, a novel fingerprint template based on local-similar image (LSIT) is proposed. The generation of the LSIT mainly consists of two parts: replacing and hiding the original minutiae through expansion; mapping and abstracting fingerprint features according to the constructed local-similar image. On this basis, a cancellable fingerprint template characterized by color features transformed from minutiae can be obtained. The security and cancellability of the proposed LSIT are analyzed and discussed, and the recognition performance is experimented using the FVC2002 and FVC2004 databases. The experimental results indicate that the LSIT has promising performances, which provide a new way for fingerprint protection and application.