Unknown Presentation Attack Detection against Rational Attackers

TL;DR

This paper introduces a game-theoretic approach with a novel generator-based detection technique and C-marmax loss to improve presentation attack detection, especially against unknown and rational attackers, with enhanced explainability and few-shot learning.

Contribution

It proposes a new optimization framework, a generator-based feature set, and a C-marmax loss function to enhance detection of both known and unknown attacks in adversarial settings.

Findings

Achieves state-of-the-art detection performance for known and unknown attacks.

Provides balanced performance across attack types and settings.

Demonstrates potential for few-shot learning and pixel-level explainability.

Abstract

Despite the impressive progress in the field of presentation attack detection and multimedia forensics over the last decade, these systems are still vulnerable to attacks in real-life settings. Some of the challenges for existing solutions are the detection of unknown attacks, the ability to perform in adversarial settings, few-shot learning, and explainability. In this study, these limitations are approached by reliance on a game-theoretic view for modeling the interactions between the attacker and the detector. Consequently, a new optimization criterion is proposed and a set of requirements are defined for improving the performance of these systems in real-life settings. Furthermore, a novel detection technique is proposed using generator-based feature sets that are not biased towards any specific attack species. To further optimize the performance on known attacks, a new loss function coined categorical margin maximization loss (C-marmax) is proposed which gradually improves the performance against the most powerful attack. The proposed approach provides a more balanced performance across known and unknown attacks and achieves state-of-the-art performance in known and unknown attack detection cases against rational attackers. Lastly, the few-shot learning potential of the proposed approach is studied as well as its ability to provide pixel-level explainability.