An improved scheme to fingerprint classification

* Final gross prices may vary according to local VAT.

Get Access

Abstract

An improved scheme to fingerprint classification is presented in this paper. It is well known that automatic fingerprint identification system usually needs to access very huge data — thousands of fingerprints stored in database. Thus a good automatic classification algorithm is an important module in such a system. Further more, for identification of fingerprint, it also asks few error occurred in its pre-classification stage. The scheme presented here extracts the singularities of fingerprints. Based on those points, the flow-lines which describe the global pattern of fingerprint are obtained with flow-line tracing algorithm using the direction and skeleton information. The properties of the flowlines are analyzed with respect to fingerprint pattern. Then the scheme performs the overlapped classification algorithm with these information, which reduced the error of classification greatly. Some experimental results are reported.