Non-Contact Fingerprint Template Protection Using DFT Combined Random Projection

Abstract

In view of the different dangersto which users of contactless biometric systems are exposed, the authors have developed a contactless secure revocable model based on random projection and DFT (Discrete Fourier Transformation) to enhance contactless fingerprint authentication. Two matrices emerge, namely corresponding to the terminations and corresponding to the bifurcations. These matrices are then transformed in a first time thanks to the random projection. In a second time, they apply to them the Discrete Fourier Transformation called the DFT. This proposed non-contact revocable fingerprint model meets the requirements of revocability, diversity, security, and non-reversibility. The evaluation of the model through its results gives the most promising results compared to those existing. The equal error rate (EER) obtained is respectively equal to 0.19% for FVC2002 DB1, 1% for FVC2002 DB2, 4.29% for FVC2002 DB3, and 9.01% for FVC2004 DB2.