Fingerprint Composition-Feature Extraction Using Directional Filter Banks

Abstract

A method for fingerprint composition-feature extraction by using directional filter banks is proposed. the proposed method decomposes a fingerprint image into eight directional sub bands using directional filter banks and then computes the composition-feature from each sector, corresponding to each directional sub band, segmented from the interested region. The composition-feature consists of mean feature and deviation feature and it can well represent the local feature of a sector. Experimental results show the effectiveness of the proposed method.