Latent fingerprints segmentation based on Rearranged Fourier Subbands

In this work, we present a latent fingerprint segmentation algorithm based on spatial-frequency domain analysis. The algorithm arranges the overlapped block-based Fourier coefficients into groups of frequency and orientation subbands, called Rearranged Fourier Subband (RFS). The RFS reveals latent fingerprint spectra in only a limited number of subbands. The algorithm then boosts, sorts, and extracts, from complex background and noise, the latent fingerprint spectra in the RFS subbands. Several experiments are evaluated based on ground truth comparison, feature extraction, and latent matching on the NIST SD27 latent database. Our experimental results show that the proposed algorithm achieves better accuracy compared to those of the published automatic segmentation algorithms.