Fingerprint Liveness Detection Adapted to Different Fingerprint Sensors Based on Multiscale Wavelet Transform and Rotation-Invarient Local Binary Pattern

Abstract

In recent years, fingerprint authentication systems are convenient for us to verify the identity of the user by extracting and analysing these biometric features, so they have been rapidly developed in our daily life. However, current existing problem is that fingerprint authentication systems are vulnerable to spoofing attacks, such as artificial fake fingerprints. Moreover, the classification accuracy of traditional liveness detection methods for different sensors is not satisfactory. Therefore, in order to solve these spoofing attacks and enhance the classification performance for samples of different fingerprint sensors, a new software-based fingerprint liveness detection method, which is based on the multiscale wavelet transform and the rotaion-invarient local binary pattern (RILBP), was proposed in this paper. The fingerprint samples are derived from four different fingerprint sensors in LivDet 2011. Experimental results demonstrate that our method can detect the fingerprint liveness with higher classification performance compared with other methods of fingerprint liveness detection.