Orientation-improved minutiae for fingerprint matching

For minutiae-based fingerprint matching, alignment-based point pattern matching algorithm has been proposed to reduce the computational cost with a suitable criterion on ridge similarity. However, a large percentage of spurious pairs have rather similar ridges as well, which slows down the matching process and limits the matching accuracy. In this paper, orientation-based ridge patterns are utilized to improve the minutiae feature, which remove more spuriously matched pairs. In addition, to reduce errors caused by broken bifurcation, we propose a new scheme that compares two minutiae regardless their types. Experiments conducted on a large fingerprint database (NIST-4) show that the proposed method produces a much improved performance.