3D Facial Geometric Attributes Based Anti-Spoofing Approach against Mask Attacks

3D scanning and 3D printing techniques, as the technical impetus of 3D face recognition, also boost unconsciously the security threat against it from the spoofing attacks via manufactured mask. In order to improve the robustness of 3D face recognition system, several countermeasures against mask attacks based on photometric features have been reported in recent years. However, the anti-spoofing approach involving 3D meshed face scan and the related 3D facial features have not been studied yet. For filling this gap, in this paper, we propose to exploit the anti-spoofing performance of geometric attributes based 3D facial description. It synthesises the advantages of the selected geometric attributes, named principal curvature measures, and the meshSIFT-based feature descriptor. Specifically, the estimation of geometric attributes is coherent to the property of discrete surface, and the feature related to them can accurately describe the shape of facial surface. These characteristics are beneficial to discovering the geometry-based dissimilarity between genuine face and fraud mask. In the experiment part, the baselines of verification and anti-spoofing performance are evaluated on the Morpho database. Furthermore, for simulating a real-world scenario and testing the corresponding anti-spoofing performance, the size of genuine face set is massively extended by uniting the Morpho database and the FRGC v2.0 database to increase the ratio of genuine faces to fraud masks. The evaluation results prove that the proposed 3D face verification system can guarantee competitive verification accuracy for genuine faces and promising anti-spoofing performance against mask attacks.

I. Introduction

Due to the natural, non-intrusive and contactless acquisition pattern of facial data, the face recognition (FR) technology has been extensively accepted by public and applied widely in our daily life [11], [12]. This effective and reliable biometrics recognition patten has spread in numerous practical applications like time attendance, bank service, access control, boarder entry & exit, criminal investigation, etc. In recent years, massive researches and achievements towards face recognition via 2D face images and 3D face scans have been studied and reported for enhancing and improving the recognition performance of a face recognition system [1], [25], [26], [2], [18], [9], [35]. However, a potential menace hiding behind these successful researches, called spoofing attacks, threats insidiously the security of a face recognition system. Spoofing attack is defined as an intrusive act of deceiving a biometric system by presenting a fake evidence or copied biometric trait to obtain valid authentication [28]. Particularly, in face related spoofing attacks, the fake evidence is commonly presented as a photograph or a video of a valid user which are captured at distance or collected via internet. Due to the development of 3D scanning and 3D printing techniques, the convenient manufacture of 3D mask provides an easier way to intruders for masquerading as a registered person in a FR system. Moreover, a social public website “Thats My Face” ¹

http://www.thatsmyface.com

provides the manufacturing service of wearable 3D masks, which only demands to offer one frontal photo (a side-view photo demanded as option). Unfortunately, the convenient face data acquisition and the face model manufacture, which should be the advantage of face recognition, become gradually the jeopardy and the calamity to a face recognition system.