Face Liveness Setection
Biometrics offers a powerful and practical solution to authentication-required applications. Due to the breakthrough of biometrics authentication via deep learning and its better security capability compared with traditional authentication methods (e.g., password, secret question, token code), more and more attention has been attracted from both academia and industry nowadays. Typical biometric modalities include fingerprint, iris, face and voice print, among which ``face'' is the most popular one as it does not require any additional hardware infrastructure and almost all mobile phones are equipped with a front-facing camera. Despite the success of face recognition, it is still vulnerable to the presentation attacks due to the popularity of social media from which facial images are easy to acquire. For instance, a presentation attack can record the face information of a person by printing (printing attack), replaying on screen (replay attack) or even counterfeiting the face via 3D masking and VR, which brings extremely challenging security issues.
The diverse facial capturing devices and unexpected capturing environment make face liveness detection problem difficult.
Our approach leverages the advantage of deep learning and domain generalization by extracting generalized feature representation for cross-domain face liveness detection task.
In particular, 3D convolutional neural networks (3D CNN), which have been proved to be efficient for action recognition task, are employed to learn spoofing-specific information based on typical printed and replay video attacks. The solution incorporates 2D and 3D features related to the presentation attack problem, and learns not only spatial variations associated with attacks but also artifacts that take place over time.
More specifically, we employ the 3D CNN architecture with a data augmentation strategy for the spoofing detection task. To obtain a more robust and generalized 3D CNN model, the lack of generalization is dealt with by introducing a regularization mechanism, which focuses on improving classification accuracy during training as well as generalizing to unknown conditions by minimizing the feature distribution dissimilarity across domains. These capabilities allow us to make a further step regarding the detection of attacks under unknown or different conditions.
We also propose a large-scale dataset for face liveness detection, Rose-Youtu Face Liveness Detection dataset (Rose-Youtu). In Rose-Youtu database, there are 3350 videos with 20 subjects for public-research purpose. Five mobile devices (Hasee Smart-Phone, Huawei Smart-Phone, iPad 4, iPhone 5s and ZTE Smart-Phone) were employed for face video acquisition. For spoofing medium, printed paper attack, video display attack, mask attack and video replay attack were considered.
Face Liveness Detection Video