Deep Features for Contactless Fingerprint Presentation Attack Detection: Can They Be Generalized?

TL;DR

This study evaluates the ability of various pre-trained deep learning models, including CNNs and a Vision Transformer, to detect contactless fingerprint presentation attacks across different datasets and attack types, highlighting ResNet50's superior generalization.

Contribution

It provides a comparative analysis of multiple deep learning models' effectiveness in generalizing contactless fingerprint PAD, introducing a benchmark for unseen attack detection.

Findings

ResNet50 achieved the best generalization performance.

Deep features from pre-trained models can effectively detect presentation attacks.

The study offers insights into model selection for contactless fingerprint PAD.

Abstract

The rapid evolution of high-end smartphones with advanced high-resolution cameras has resulted in contactless capture of fingerprint biometrics that are more reliable and suitable for verification. Similar to other biometric systems, contactless fingerprint-verification systems are vulnerable to presentation attacks. In this paper, we present a comparative study on the generalizability of seven different pre-trained Convolutional Neural Networks (CNN) and a Vision Transformer (ViT) to reliably detect presentation attacks. Extensive experiments were carried out on publicly available smartphone-based presentation attack datasets using four different Presentation Attack Instruments (PAI). The detection performance of the eighth deep feature technique was evaluated using the leave-one-out protocol to benchmark the generalization performance for unseen PAI. The obtained results indicated the best generalization performance with the ResNet50 CNN.