Face anti-spoofing by identity masking using random walk patterns and outlier detection
- PDF / 2,819,558 Bytes
- 20 Pages / 595.276 x 790.866 pts Page_size
- 100 Downloads / 227 Views
INDUSTRIAL AND COMMERCIAL APPLICATION
Face anti‑spoofing by identity masking using random walk patterns and outlier detection Balaji Rao Katika1 · Kannan Karthik1 Received: 25 July 2019 / Accepted: 6 February 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Existing architectures used in face anti-spoofing tend to deploy registered spatial measurements to generate feature vectors for spoof detection. This means that the ordering or sequence in which specific statistics are computed cannot be changed, as one moves from one facial profile to another. While this arrangement works in a person-specific setting, it becomes a major drawback when single-sided training is done based on the natural face class alone. To mitigate subject identity linked content interference within the anti-spoofing frame, we propose a identity-independent architecture based on random correlated scans of natural face images. The same natural face image can be scanned multiple times through independent correlated random walks before deriving simple differential features on the 1D scanned vectors. This proposed frame tends to capture the pixel correlation statistics with minimal content interference and shows great promise, particularly when trained on natural face sets, using a one-class support vector machine and cross-validated on other databases. Performance measured in terms of EER for detection of spoof face is found to be 3.8291% with proposed random scan features, and 2.02% with auto-population samples for inter database. Keywords Face anti-spoofing · Random scan · Auto-population · Planer spoofing · One-class SVM
1 Introduction Generally, access to closed organizations is granted through a multi-biometric identification system, wherein one of the biometrics is most likely to be the face. This is partly so because the face, as a readable and relatable biometric, imparts a distinct social identity to any individual. The other distinct advantage of deploying face is remote authentication and tracking select individuals (if required) who have entered the organization over this internal space. Entry into private and government organizations is usually granted using unmanned face recognition units, wherein an individual presents his/her face to the camera for recognition. The camera takes a snapshot, performs a quick check over a pre-stored image dataset (viz., faces of individuals registered * Balaji Rao Katika [email protected] Kannan Karthik [email protected] 1
Department of Electronics and Electrical Engineering, Indian Institute of Technology Guwahati, Guwahati, Assam 781039, India
with a specific organization) and then grants/denies the individual passage to the interiors. The problem with this unmanned recognition system is that a particular hidden person X may masquerade as another individual Y, either by wearing a prosthetic mask [4] or by deploying planar spoofing [21] by either presenting a printed photograph of Y or by replaying a an old video of Y’s face. Facial recognition systems cannot t
Data Loading...
