Fingerprint Minutiae Matching Based On Complex Minutiae Vector
XiFeng Tong,Xianglong Tang,Jianhua Huang,Xiao Li
DOI: https://doi.org/10.1109/icmlc.2004.1380466
2004-01-01
Abstract:Because of strict restriction on False Reject Rate, False Accept Rate, and computational time cost, fingerprint minutiae matching is a challenging task. In this paper, we proposed a novel fingerprint feature named Complex Minutiae Vector (CMV) for fingerprint minutiae matching, which consists of a ridge rotation angle associated with a minutia and four ridge counts between the minutia and the four corresponding adjacent points. In the first stage, Complex Minutiae Vector is used to find possible minutiae pairs. Then one minutiae set is rotated and translated to achieve alignment. At last, matching score calculation stage is used to find reliable minutiae matching and get matching score by a novel strategy, which considers not only the number of matched minutiae pairs but also the similarity of corresponding Complex Minutiae Vector pairs. Our promising experimental results show that the proposed methodology is capable to keep a good trade-off between speed and accuracy.