Face Morphing Attack Generation and Detection: A Comprehensive Survey

Abstract

Face recognition has been successfully deployed in real-time applications, including secure applications such as border control. The vulnerability of face recognition systems (FRSs) to various kinds of attacks (both direct and indirect attacks) and face morphing attacks has received great interest from the biometric community. The goal of a morphing attack is to subvert an FRS at an automatic border control (ABC) gate by presenting an electronic machine-readable travel document (eMRTD) or e-passport that is obtained based on a morphed face image. Since the application process for an e-passport in the majority of countries requires a passport photo to be presented by the applicant, a malicious actor and an accomplice can generate a morphed face image to obtain the e-passport. An e-passport with a morphed face image can be used by both the malicious actor and the accomplice to cross a border, as the morphed face image can be verified against both of them. This can result in a significant threat, as a malicious actor can cross the border without revealing the trace of his/her criminal background, while the details of the accomplice are recorded in the log of the access control system. This survey aims to present a systematic overview of the progress made in the area of face morphing in terms of both morph generation and morph detection. In this paper, we describe and illustrate various aspects of face morphing attacks, including different techniques for generating morphed face images and state-of-the-art morph attack detection (MAD) algorithms based on a stringent taxonomy as well as the availability of public databases, which allow us to benchmark new MAD algorithms in a reproducible manner. The outcomes of competitions and benchmarking, vulnerability assessments and performance evaluation metrics are also provided in a comprehensive manner. Furthermore, we discuss the open challenges and potential future areas that need to be addressed in the evolving field of biometrics.

An Introduction to Digital Face Manipulation

“…There has been recently a large amount of research in the field of face morphing. Comprehensive surveys have been published in [70,74] including both morphing techniques and also morphing attack detectors. In general, the following three consecutive stages are considered in the generation process of face morphing images: (i) determining correspondences between the face images of the different subjects.…”

Single Morphing Attack Detection Using Feature Selection and Visualization Based on Mutual Information

“…The resulting images are of high quality without artefact in the background. As we mentioned before, after creation of the morphed images, all the faces were cropped using a modified dlib face detector implementation 6 . Figure 6 shows examples of the FERET cropped face database.…”

“…A morphing attack's success depends on the decision of human observers, especially a passport identification expert. The real-life application for a border police expert who compares the passport reference image of the traveller (digital extracted from the embedded chip) with the facial appearance of the traveller [6] is too hard because of the improvements of the morphing tools and because of the difficulty for the human expert to localise facial areas, in which morphing artefacts are present. This work proposes to add an extra stage of feature selection after feature extraction based on Mutual Information M I to estimate and keep the most relevant and remove the most redundant features from the face images to separate bona fide and morphed images on a S-MAD scenario.…”

MagNet: Detecting Digital Presentation Attacks on Face Recognition

“…Overall, existing works have demonstrated, beyond doubt, that “morphing” is a threat for face recognition systems. Several survey papers have also highlighted the vulnerability of face recognition algorithm against digital manipulation and limitations of existing detection algorithms ( Akhtar et al, 2019 ; Scherhag et al, 2019 ; Singh et al, 2020 ; Tolosana et al, 2020 ; Venkatesh et al, 2020 ). The survey papers bring out the boundaries of existing detection algorithms such as non-generalizability against manipulation types and image resolution and computationally inefficiency.…”