Myope Models -- Are face presentation attack detection models short-sighted?

TL;DR

This study investigates the impact of background information on face presentation attack detection models, showing that including background improves performance and models selectively utilize background cues, with a novel multi-task approach outperforming state-of-the-art.

Contribution

The paper introduces a comparative analysis of face PAD models with and without background, proposing a multi-task learning method that surpasses existing results and analyzing model focus via Grad-CAM++.

Findings

Background presence improves detection accuracy.

Multi-task learning achieves state-of-the-art EER of 0.2%.

Models selectively use background cues depending on attack type.

Abstract

Presentation attacks are recurrent threats to biometric systems, where impostors attempt to bypass these systems. Humans often use background information as contextual cues for their visual system. Yet, regarding face-based systems, the background is often discarded, since face presentation attack detection (PAD) models are mostly trained with face crops. This work presents a comparative study of face PAD models (including multi-task learning, adversarial training and dynamic frame selection) in two settings: with and without crops. The results show that the performance is consistently better when the background is present in the images. The proposed multi-task methodology beats the state-of-the-art results on the ROSE-Youtu dataset by a large margin with an equal error rate of 0.2%. Furthermore, we analyze the models' predictions with Grad-CAM++ with the aim to investigate to what extent the models focus on background elements that are known to be useful for human inspection. From this analysis we can conclude that the background cues are not relevant across all the attacks. Thus, showing the capability of the model to leverage the background information only when necessary.