Fingerprint orientation reconstruction from minutiae points

Fingerprint orientation reconstruction from minutiae points is an important procedure in fingerprint reconstruction. In this paper, a novel fingerprint orientation reconstruction approach is proposed. First the initial orientation field is estimated through interpolation strategy. The orientation of central block is predicted by the nearest minutia in eight different sectors. Then a 2-D Fourier series expansion method for orientation field is proposed. Using the coefficients of expansion function, orientation field can be presented and reconstructed effectively. The noise can also be eliminated evidently through this method. Furthermore singular points, which are always needed for other orientation model, are not required by 2-D Fourier series expansion method. The experimental results comparing our method with other methods prove that our proposed method is very effective in representing and reconstructing the fingerprint orientation field.