Fingerprint Matching Relying on Minutiae Hough Clusters

In this paper, we present an algorithm for fingerprint verification based on global alignment and minutiae-centered circular regions. Minutiaecentered circular matching ensures higher matching speed and robustness to non-linear distortion. This matching is based on finding the first pair of corresponding regions and then finding other pairs of corresponding minutiae in the images. Our approach consists of finding the best alignment between input minutiae and template minutiae using Hough-Matching [1], and then the sets of matched minutiae will be used as input to the minutiae-centered circular regions matching algorithm [2]. The experimental results show higher performance using the Hough matching and the matching circular regions algorithms on the public Database FVC2002 db2 [3].