Abstract
In this paper, a face spoofing detection method called the Fully Convolutional Network with Domain Adaptation and Lossless Size Adaptation (FCN-DA-LSA) is proposed. As its name suggests, the FCN-DA-LSA includes a lossless size adaptation preprocessor followed by an FCN based pixel-level classifier embedded with a domain adaptation layer. The FCN local classifier makes full use of the basic properties of face spoof distortion namely ubiquitous and repetitive. The domain adaptation (DA) layer improves generalization across different domains. The lossless size adaptation (LSA) preserves the high-frequent spoof clues caused by the face recapture process. The ablation study shows that both DA and the LSA are necessary for high-accuracy face spoofing detection. The FCN-LSA obtains competitive performance among the state-of-the-art methods. With the help of small-sample external data in the target domain (2/50, 2/50, and 1/20 subjects for CASIA-FASD, Replay-Attack, and OULU-NPU respectively), the FCN-DA-LSA further improves the performance and outperforms the existing methods.
Highlights
Faces can be captured conveniently by digital cameras, web cameras, smart phones, etc
With the fast development of face recognition, the modern face recognition algorithms [1]–[3], especially deep networks trained on large scale datasets, can surpass human performance, but they may be fooled by face spoofing attacks which can be launched by inexperienced attackers
As its name suggests, the Fully Convolutional Networks (FCNs)-DA-LSA includes a lossless size adaptation preprocessor followed by an FCN-based pixel-level classifier embedded with a domain adaptation layer
Summary
Faces can be captured conveniently by digital cameras, web cameras, smart phones, etc. For photo/video attacks, faces are first captured by camera, printed/displayed on papers/screens are recaptured by another camera This process is a special cases of the image recapture which has a wide variety of defending methods. Atoum et al [8] designed a patch-based CNN to detect the spoof patterns in extracted face patches of 96 × 96 pixels This is the basic version of the local supervision which is the key to the high performance. The local supervised FCN-based face spoofing detection methods will be reviewed. They can avoid the problem of data inefficiency. The FCN-based face spoofing detection methods are generally superior to the CNN-based ones since they make full use of the basic properties of ubiquitous and repetitive. Based on our previous FCN-based study [10], we are going to proposed a new method called Fully Convolutional Network with Domain Adaptation and Lossless Size Adaptation (FCN-DA-LSA)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
