A New Method of Fingerprints Singular Points Detection

Abstract

This paper proposes a new method of fingerprint singular points searching algorithm based on fingerprint orientation maps mainly. In these orientation maps, mathematical morphology theory and operations are used to search the fingerprint’s singular points in view of the phenomenon that the change of ridges’ direction at the points is the fastest in the whole fingerprint image and the directions of finger ridges which surround the points are symmetric with each other. For obtaining more accuracy result, before obtain the orientation maps, we use db3 wavelet technology to remove the most of fingerprint images’ noises in this paper. For the same reason, before the main mathematical morphology operations, we first obtain fingerprint images’ binary image models based on the db3-denoise image, and use these models to remove the disturb directions of the image background secondly. Experiments indicate this method is easy to understand, achieve and has a good robustness for the fingerprint’s translation and deformation.