Investigating Fingerprint Quality Features for Liveness Detection

Abstract

Fingerprint-based recognition systems are vulnerable to presentation attacks. To identify these attacks one of the solution is fingerprint liveness detection which ensures the presence of a live or fake fingerprint. In this paper, we have investigated the use of quality features for the detection of liveness of given fingerprint image. We have proposed a novel set of features which can be used for liveness detection in fingerprint images. Along with these features, efficacy of other existing quality features is also evaluated for the liveness detection. Based on these quality features fingerprint images are classified into fake and live fingerprints using various classifiers. The robustness of the proposed approach is evaluated on publicly available LivDet 2015 competition database. The advantage of the proposed method is that it utilizes the quality features for liveness detection which are also utilized for the quality analysis of fingerprint images. Therefore, it is possible to combine two different modules, namely, quality analysis and liveness detection of the Automatic Fingerprint Identification System (AFIS) using our proposed approach.