A Fingerprint Enhancement Algorithm using a Federated Filter

In the application of the Automatic Fingerprint Identification System (AFIS), some fingerprint images are noisy and corrupted. In such situation, an effectively enhancement algorithm to improve the clarity of the ridge structure is necessary. The available enhancement algorithms are based on either the local orientation field filtering scheme in space domain or the Gabor filtering scheme in the frequency domain. The local orientation could not be correctly estimated for fingerprint images of poor quality, which greatly restricts the applicability of these filtering techniques. The Gabor filters could obtain the reliable orientation estimate even for corrupted images, but it is time consuming. It is unsuitable for an on-line fingerprint recognition system such as AFIS. In this paper, an enhancement algorithm using a space-frequency federated filtering scheme is proposed, which adapts the filtering methods to the input images according to a pre-defined quality factor. The quality factor is calculated on the orientation field filtered image. Only parts of the images of which quality factor do not meet the requirement are filtered by the Gabor filter. Experiments show that the proposed algorithm is computationally efficient, with the same level of the enhancement performance.