Fake Finger Detection Using the Fractional Fourier Transform

This paper introduces a new method for detecting fake finger using the fractional Fourier transform (FrFT). The advantage of this method is to require one fingerprint image. The fingerprint is a texture with the interleaving of ridge and valley. When the fingerprint is transformed into the spectral domain, we found energy of fingerprint. Generally, the energy of live fingerprints is larger than the energy of fake fingerprints. Therefore, the energy in the spectral image of a fingerprint can be a feature for detecting of fake fingers. We transformed the fingerprint image into the spatial frequency domain using 2D Fast Fourier transform and detected a specific line in the spectrum image. This lineis transformed into the fractional Fourier domain using the fractional Fourier transform. And, the standard deviation of it is used to discriminate between fake and live fingerprints. For experiment, we made fake fingers of silicone, gelatin, paper and film. And, the fake finger database is created, by which the performance of a fingerprint verification system can be evaluated with higher precision. The experimental results demonstrate that the proposed method can detect fake fingers.