A facial image authentication method by embedding PLT-PCA features

Facial image authentication is a special and important field in media security for its wide use in many conditions related with security or privacy. In this paper, a facial image authentication scheme is proposed based on a novel robust watermarking algorithm by embedding the statistic features of the face region. In the proposed Robust watermarking algorithm, the block of LL sub-band is selected as the embedding object and the quantized statistic feature of face region is embedded and extracted by the given criterion. And the statistic feature is extracted by involving the sample expanding method named power law transformation (PLT) and principle component analysis. The Experimental results demonstrated the robustness of the proposed watermarking algorithm and feasibility of the authentication scheme.