Understanding deep face anti-spoofing: from the perspective of data
- PDF / 4,057,104 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 35 Downloads / 170 Views
ORIGINAL ARTICLE
Understanding deep face anti-spoofing: from the perspective of data Yujing Sun1
· Hao Xiong1 · Siu Ming Yiu1
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Face biometrics systems are increasingly used by many business applications, which can be vulnerable to malicious attacks, leading to serious consequences. How to effectively detect spoofing faces is a critical problem. Traditional methods rely on handcraft features to distinguish real faces from fraud ones, but it is difficult for feature descriptors to handle all attack variations. More recently, in order to overcome the limitation of traditional methods, newly emerging CNN-based approaches were proposed, most of which, if not all, carefully design different network architectures. To make CNN-related approaches effective, data and learning strategies are both indispensable. In this paper, instead of focusing on network design, we explore more from the perspective of data. We present that appropriate nonlinear adjustment and hair geometry can amplify the contrast between real faces and attacks. Given our exploration, a simple convolutional neural network can solve the face antispoofing problem under different attack scenarios and achieve state-of-the-art performance on well-known face anti-spoofing benchmarks. Keywords Face anti-spoofing · Biometrics · Image adjustment · Image processing
1 Introduction Nowadays, the usage of facial biometrics in various scenarios of business and industry is dramatically increasing and becoming popular in authenticating user identities. One could protect his privacy in electronic devices using face unlocking techniques, to conveniently open bank account remotely via identity verification with webcam, and even to authenticate payment with facial biometrics. However, it is insecure to use face as a biometric measure for authentication. A recent study [30] on face recognition using commercial matchers shows that face biometric Yujing Sun and Hao Xiong have contributed equally to this work. This project is partially supported by a Collaborative Research Fund (CRF, C1008-16G) and Innovation and Technology Support Programme (ITS/173/18FP) of the Hong Kong Government.
B
Yujing Sun [email protected] Hao Xiong [email protected] Siu Ming Yiu [email protected]
1
systems can be vulnerable to spoofing attacks, such as fraud photographs, videos, or masks that launch against face authentications or recognition systems, and can lead to inestimable privacy leak and property loss, for instance, private photographs, and sensitive bank information. Moreover, given the rapid development and prevalence of social media, people are sharing their facial photographs on the internet intentionally. Malicious people can easily obtain such photographs to attack facial recognition systems. Comparing with other biometrics, such as fingerprint and iris, facial images are much more convenient to acquire. As a consequence, the demand to effectively prevent face spoofing attacks is significantly on the rise. Researchers are
Data Loading...
