Improved fingerprint identification using directional filter banks

Abstract

In the finger print area, the majority of approaches are based on bitwise directional analysis. However, several approaches employ filter-based directional analysis (Sarbadhikari, S. et al., 1998; Sherlock, B.G., 1994; Hong, L. et al., 1998). The directional image can be thought of as an image transform, where each pixel of an image represents a direction of the local gray level uniformity. The statistics derived from the directional image are used for segmentation of the original image. The method is suited for simple images like fingerprints and other images that consist of only a background and a foreground. The directional image is partitioned into "homogenous" connected regions according to the fingerprint topology, thus giving a synthetic representation which can be exploited as a basis for classification. We provide a fingerprint enhancement approach based on "directional filter banks". A fingerprint image is decomposed into eight directional images and cleaned. Once the eight directional images are cleaned, the final enhanced image is obtained by combining the cleaned directional images