Associating Minutiae between Distorted Fingerprints Using Minimal Spanning Tree

摘要

This paper proposes to associate minutiae between distorted fingerprints by locating two matched minimal spanning trees, and then apply affine transformation to one of the two images. This process is iterated until correspondences are stable. A radial basis function -based warping model is established from the final correspondences, and is applied to nonlinearly align the two images. The experimental results show that the proposed method leads to an improved matching performance.