Local Distributional Smoothing for Noise-invariant Fingerprint Restoration

Existing fingerprint restoration models fail to generalize on severely noisy fingerprint regions. To achieve noise-invariant fingerprint restoration, this paper proposes to regularize the fingerprint restoration model by enforcing local distributional smoothing by generating similar output for clean and perturbed fingerprints. Notably, the perturbations are learnt by virtual adversarial training so as to generate the most difficult noise patterns for the fingerprint restoration model. Improved generalization on noisy fingerprints is obtained by the proposed method on two publicly available databases of noisy fingerprints.