Attack using reconstructed fingerprint

Most fingerprint recognition systems store minutiae-based fingerprint template in a database. As the minutiae template is very compact, many take it for granted that these minutiae points do not contain sufficient information for reconstructing the original fingerprint. This paper proposes a scheme to reconstruct the fingerprint from minutiae points based on the amplitude and frequency modulated (AM-FM) fingerprint model to fool a system that requires a full print. We first generate a binary ridge pattern which has a similar ridge flow to that of the original fingerprint. The continuous phase is intuitively reconstructed by removing the spirals in the phase image estimated from the ridge pattern. We further introduce a phase refinement process to reduce the artifacts created due to the discontinuity in the reconstructed phase image, which is the combination of the continuous phase and the spiral phase (computed from the minutiae points). Compared with previous works, our reconstructed fingerprint matches better against the original fingerprint and the other impressions. In addition, it contains fewer artifacts.