Can Foundation Models Generalise the Presentation Attack Detection Capabilities on ID Cards?

TL;DR

This paper investigates how foundation models can enhance the generalisation of presentation attack detection on ID cards across different countries, focusing on zero-shot and fine-tuning approaches with diverse datasets.

Contribution

It benchmarks foundation models for ID card PAD and explores methods to improve cross-country generalisation, emphasizing the importance of bona fide images.

Findings

Bona fide images are crucial for generalisation.

Foundation models can improve PAD performance across diverse ID datasets.

Zero-shot and fine-tuning approaches have different impacts on generalisation.

Abstract

Nowadays, one of the main challenges in presentation attack detection (PAD) on ID cards is obtaining generalisation capabilities for a diversity of countries that are issuing ID cards. Most PAD systems are trained on one, two, or three ID documents because of privacy protection concerns. As a result, they do not obtain competitive results for commercial purposes when tested in an unknown new ID card country. In this scenario, Foundation Models (FM) trained on huge datasets can help to improve generalisation capabilities. This work intends to improve and benchmark the capabilities of FM and how to use them to adapt the generalisation on PAD of ID Documents. Different test protocols were used, considering zero-shot and fine-tuning and two different ID card datasets. One private dataset based on Chilean IDs and one open-set based on three ID countries: Finland, Spain, and Slovakia. Our findings indicate that bona fide images are the key to generalisation.