An Enhanced Generative Adversarial Network Model for Fingerprint Presentation Attack Detection

Abstract

Automatic fingerprint recognition systems (AFRS) have played a significant role in biometric security in recent years. However, it is vulnerable to several threats which can put the AFRS at substantial risk. Presentation attack or spoofing is one of these attacks which utilizes a spoof fingerprint created with a fabrication material by an intruder to fool the authentication system. The development of new fabrication materials makes this spoof detection more challenging for cross-materials. In this paper, a new approach for liveness detection is proposed to encounter the presentation attack. An enhanced Generative Adversarial Network is utilized for this purpose. The performance of the proposed method is evaluated in an open-set paradigm on publicly accessible LivDet Competition datasets and the proposed methodology achieves an average accuracy of 98.52 %, 92.02 %, 80.89% and 86.95% for the LivDet 2013, LivDet 2015, LivDet 2017 and LivDet 2019 datasets, respectively, which outperforms the state-of-the-art methods. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.