Face Spoofing Detection Based on Combining Different Color Space Models
Junqin He, Jun Luo · 2019
At present, with the widespread application of face recognition for identity authentication technology, spoofing attack has become the biggest security threat. For in the RGB color space model, the real face and the spoofing attack image are difficult to distinguish, a method based on the combination of different color space models is proposed. The main idea is to transform RGB image into YCbCr and Luv color space model to extract LBP (Local Binary Patterns) features, then transform RGB image into HSV color space model to extract CM (Color Moment) features, and finally cascade the extracted features into SVM (Support Vector Machine) for decision classification. Compared with the existing methods on two public datasets of Replay-Attack and CASIA-FASD, the experimental results show that the proposed method is superior to single color space and single feature face spoofing detection method, and the calculation process is very simple, and it has higher discrimination ability.