Face morphing detection in the presence of printing/scanning and heterogeneous image sources
This addresses a security threat in electronic identity documents, though it appears incremental as it builds on existing morphing detection methods.
The paper tackled the problem of detecting face morphing attacks in printed-scanned and cross-database scenarios, achieving state-of-the-art accuracy on challenging datasets through novel training approaches for deep neural networks.
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.