Curvature-based singular points detection

Abstract

Singular Points Detection or more commonly known as 'Core Points Detection', is an important process in most fingerprint verification and identification algorithms for locating reference points for minutiae matching and classification. In this paper, we propose a new algorithm for singular points detection, which is based on scale-space analysis and curvature properties of the flow patterns of the fingerprint. The algorithm starts by examining the curvature of the fingerprint image at the coarsest scale and zoom in until the actual resolution of the image is reached. Experimental results show that the proposed algorithm is able to locate singular points in fingerprint with high accuracy. © Springer-Verlag 2001.