Fingerprint Image Sharpening and Denoising Method Based on Nonlinear Anisotropic Reverse Diffusion Equation

Considering the feature of the fingerprint, a nonlinear anisotropic reverse diffusion algorithm is proposed from the reverse diffusion phenomena in physics. The method is effective for both noise removing and edge preserving, mainly because it prevents the blurring effect on the degenerate diffusion equation for the points with big gray level variations. Compared with nonlinear diffusion methods, the algorithm has better performances for preprocessing noisy fingerprint images.