Combining weighted linear project analysis with orientation diffusion for fingerprint orientation field reconstruction

This paper presents a novel algorithm for reconstructing the fingerprint orientation field (FOF). The basic idea of the algorithm is to reconstruct the FOF by combining weighted linear project analysis with orientation diffusion. We first compute the weight values of point gradients according to the similarity of point orientations. In the second procedure, the qualities of blocks are assessed by the coherence of point orientations, and then the block orientation with high quality are estimated by the weighted linear projection analysis based on the vector set of point gradients. The fingerprint blocks that will be reconstructed is selected by the priority of them in the third procedure. In the end, the FOF is reconstructed by orientation diffusion based on the reconstruction priorities of blocks. To validate the performance, the proposed method has been applied to fingerprint enhancement, fingerprint singularity detection, fingerprint minutiae extraction and fingerprint matching using the FVC2000 and FVC2004 databases. The experiments show that the proposed method is more accurate and more reliable, and it is more robust against noise.