Reconstruction of fingerprint images

Abstract

Generally, the minutiae sent to the final matching phase are extracted from the skeleton images. The accuracy of minutiae extraction depends on the quality of the skeleton image. This paper presents a novel approach of fully reconstructing the fingerprint images besides using merely the traditional procedures of fingerprint enhancement. The reconstruction would not only improve the quality of the skeleton image but also increase the accuracy of minutiae extraction and thereby improve the recognition results. The proposed algorithm deals with the three quality degrading problems in skeleton images; ridges not being strictly continuous, parallel ridges not being well separated, and finally the ridges not being smooth. The algorithm reconstructs the broken ridges into fine ridges/valleys, increases the uniformity of the ridges/valley width and renders a smooth ridge structure. The reconstruction is an independent process and can be implemented as a separate module between the preprocessing and minutiae extraction phases of a typical Automatic Fingerprint Identification System (AFIS).