Fingerprint Liveness Detection using Minutiae-Independent Dense Sampling of Local Patches

Abstract

Fingerprint recognition and matching is a common form of user authentication. While a fingerprint is unique to each individual, authentication is vulnerable when an attacker can forge a copy of the fingerprint (spoof). To combat these spoofed fingerprints, spoof detection and liveness detection algorithms are currently being researched as countermeasures to this security vulnerability. This paper introduces a fingerprint anti-spoofing mechanism using machine learning.