Fingerprint Liveness detection using probabality density function
Introduction: Presently, the fingerprint is the most commonly used biometric identifier in authentication systems. It was responsible for more than 50% of the biometric revenue in 2009 . According to Roberts , one way to overtake the security of those systems is providing to the sensor a fake physical biometric. Thus, an efficient technique for spoof detection [3,4] is an essential requirement for any fingerprint based system in operation. Moon et al.  proposed a wavelet analysis of the fingertip surface texture. This approach relies on the fact that commonly used materials in spoof fingerprints consist of large organic molecules which tend to agglomerate at the moment that the forgery is created. As a consequence, asperities are introduced to the surface of the fake fingerprints. In this method, the surface coarseness is modelled as Gaussian white noise added to the image. Moon et al.  achieved significant results for images captured in a high resolution fingerprint scanner (∼1000 dpi). For economic reasons, most of the commercialised scanners nowadays generate images of lower resolution (typically, 500 dpi). In fact, the databases used in the second edition of the Fingerprint Liveness Detection Competition (LivDet), in 2011, comprise only fingerprint images of 500 dpi .