Title: A new efficient method of fingerprint image enhancement

Authors: Piotr Porwik, Lukasz Wieclaw

Addresses: Institute of Informatics University of Silesia Bedzinska 39, 41-200 Sosnowiec, Poland. ' Institute of Informatics University of Silesia Bedzinska 39, 41-200 Sosnowiec, Poland

Abstract: Fingerprint images frequently have poor characteristics, because these traces are captured in crime places and are left on objects of different structure. For this reason, a lot of fingerprints have to be rejected – they are useless in identification processes. Fortunately, in some cases, low quality fingerprints can be additionally analysed by means of special algorithms. In this paper, a new fingerprint enhancement method has been presented. In the proposed approach, quality of image is enhanced by means of convolution of a given image and Gaussian kernel. Finally, the local binarisation is applied. Results obtained from the experiments have demonstrated and presented the efficiency of the method.

Keywords: biometrics; fingerprint images; image enhancement; Gaussian filtering; image quality.

DOI: 10.1504/IJBM.2008.018662

International Journal of Biometrics, 2008 Vol.1 No.1, pp.36 - 46

Published online: 04 Jun 2008 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article