Analysis and Selection of Features for the Fingerprint Vitality Detection

Abstract

Although fingerprint verification systems have attained a good performance, researchers recently pointed out their weakness under fraudulent attacks by fake fingers. In fact, the acquisition sensor can be deceived by fake fingerprints created with liquid silicon rubber. Among the solutions to this problem, the software-based ones are the cheapest and less intrusive. They use feature vectors made up of measures extracted from one or multiple impressions (static measures) or multiple frames (dynamic measures) of the same finger in order to distinguish live and fake fingers. In this paper, we jointly use both static and dynamic features and report an experimental investigation aimed to compare them and select the most effective ones.

Vitality Detection from Fingerprint Images: A Critical Survey

“…Therefore, making the image processing module more "intelligent", that is, making it able to detect if a fake finger has been submitted is an interesting alternative to the hardware-based approaches. Several approaches aimed to extract vitality features from the fingerprint images directly have been recently proposed [9][10][11][12][13][14][15]. The general rationale behind these approaches is that some peculiarities of "live fingerprints" cannot be hold in artificial reproductions, and they can be detected by a more or less complex analysis of fingerprint images.…”

“…The experimental investigation of Coli et al [14] considers both the elasticdeformation based method and the morphology-based one. The elastic deformation is evaluated by computing the averaged sum of all the distances among the matched minutiae of input and template fingerprints.…”

“…In order to avoid this drawback two new features named DM5 (Dry saturation percentage change) and DM6 (Wet saturation percentage change) are elaborated. With a selection of these measures, Coli et al [14] applies a liveness detection on their extended database.…”

“…This is due to the characteristic of the measure, which requires the comparison of the input impression with the related template client (e.g. elastic distortion or morphological measures [10,14] which require an additional minutiae extraction step), and also to the possibility of evaluating the relationship between fake detection rate and verification performance. In this case the protocol adopted is a little more complex, because the fake detection features can be extracted only by the comparison phase.…”

“…In this case the protocol adopted is a little more complex, because the fake detection features can be extracted only by the comparison phase. As an example, the protocol adopted in [14] is made up of the following steps:…”