Fingerprint Verification Using Weighted Zernike Moments and Fast Computation Based on Multi-disc ROI

Abstract

A texture-based method for fingerprint verification is proposed based on weighted Zernike moments. Zernike moment features are calculated from the region of interest (ROI) which consists of several concentric discs, called multi-discs. Zernike moments of multi-disc ROI are weighted according to the regions of discs. In order to reduce the computational complexity of Zernike moments, the fast Zernike moments algorithm is exploited according the structure of multi-discs. The proposed method is rotation and scale invariant, and it is also robust even the reference point locates in the corner of input image. The experiments results show the proposed method has high verification accuracy than most of the state-of-the-art texture-based methods.