Fingerprint liveness detection using local ridge frequencies and multiresolution texture analysis techniques

It has been demonstrated that simple and inexpensive techniques are sufficient to spoof fingerprint scanners. Previously, effective use of physiological phenomenon of perspiration is shown as a countermeasure against such attacks. These techniques require more than one image for performing the liveness check and hence may not be suited for on-line processing. In this work, a liveness measure based on single image is developed. The inherent texture and density differences between 'live' and 'not live' fingerprint images are exploited. Multiresolution texture analysis techniques are used to minimize the energy associated with phase and orientation maps. Cross ridge frequency analysis of fingerprint images is performed by means of statistical measures and weighted mean phase is calculated. These different features along with ridge reliability or ridge center frequency are given as inputs to a fuzzy c-means classifier. The proposed algorithm was applied to a dataset of approximately 58 live, 50 spoof and 28 cadaver fingerprint images, from three different types of scanners. An error rate of 1.4% is achieved. The algorithm provides a faster technique for doing a liveness test which relies on only one fingerprint image.