Fingerprint reference point location based on multi-scale analysis of wavelet

Abstract

It is still difficult to locate reference point for low-quality fingerprints in fingerprint recognition. In this paper, a novel algorithm based on quadratic B-spline wavelet is proposed to deal with this problem. The main idea of the paper is to make a progressive analysis on the gradient-changing of the orientation field and combine the gradient information at different scales together. Due to multi-scale analysis and the adaptive process of improved A'trous algorithm, this method naturally possesses high location rate and simple operation, especially for low-quality fingerprints. The experimental results on the FVC2004 show the high effectiveness and practicability of the approach.