Face morphing detection in the presence of printing/scanning and heterogeneous image sources
Open Access
- 18 February 2021
- journal article
- research article
- Published by Institution of Engineering and Technology (IET) in IET Biometrics
- Vol. 10 (3) , 290-303
Abstract
Nowadays, face morphing represents 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 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).To solve this problem, the authors propose new approaches to train Deep Neural Networks for morphing attack detection: in particular the generation of simulated printed‐scanned images together with other data augmentation strategies and pre‐training on large face recognition datasets, allowed reaching state‐of‐the‐art accuracy on challenging datasets from heterogeneous image sources.Keywords
All Related Versions
This publication has 40 references indexed in Scilit:
- Automatic Generation and Detection of Visually Faultless Facial Morphs2017
- Detecting morphed face images2016
- Deep Residual Learning for Image Recognition2016
- On the Effects of Image Alterations on Face Recognition Accuracy2016
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Deep Face Recognition2015
- The magic passport2014
- Fingerprint and On-Line Signature Verification Competitions at ICB 20092009
- The FERET evaluation methodology for face-recognition algorithmsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2000
- The FERET database and evaluation procedure for face-recognition algorithmsImage and Vision Computing, 1998