Face morphing detection in the presence of printing/scanning and heterogeneous image sources

TL;DR

This paper introduces novel deep learning methods for face morphing detection that effectively handle printed, scanned, and heterogeneous image sources, addressing a critical security challenge in identity verification.

Contribution

The study proposes new training strategies including simulated printed-scanned images and large-scale pre-training, achieving state-of-the-art accuracy on diverse datasets.

Findings

Achieved high accuracy on heterogeneous image datasets

Effective training with simulated printed-scanned images

Pre-training on large face recognition datasets enhances detection performance

Abstract

Face morphing represents nowadays a big security threat in the context of electronic identity documents as well as an interesting challenge for researchers in the field of face recognition. Despite of the good performance obtained by state-of-the-art approaches on digital images, no satisfactory solutions have been identified so far to deal with cross-database testing and printed-scanned images (typically used in many countries for document issuing). In this work, novel approaches are proposed to train Deep Neural Networks for morphing detection: in particular generation of simulated printed-scanned images together with other data augmentation strategies and pre-training on large face recognition datasets, allowed to reach state-of-the-art accuracy on challenging datasets from heterogeneous image sources.