Anzeige
Erschienen in:
2015 | OriginalPaper | Buchkapitel
90. Ridge-Slope-Valley Feature for Fingerprint Liveness Detection
verfasst von : Feng Wang, Jian Cheng, Yan Jiang
Erschienen in: The Proceedings of the Third International Conference on Communications, Signal Processing, and Systems
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
Attacking fingerprint-based biometric systems by presenting fake fingers is a serious threat for unattended devices. In this work, we introduce a novel algorithm, by extracting features along the fingerprint curves, to discriminate between fake fingers and real ones on static images. Pairs of mean value and standard deviation are sampled from the ridge, slope and valley of the curves. Then bag-of-words model is used to select cluster centers and form a 128-dimension feature of words’ frequency. We test our method on a dataset collected by Chinese Academy of Science, which contains 960 live fingerprints and 960 fake ones made by silicon. Though the fake fingerprints is too verisimilar to be distinguished by naked eyes, we still get an accuracy of 98.85 %. Because our method is based on single static fingerprint image, it can be freely embedded into existing fingerprint-based biometric systems.