A Novel Geometrical Robust Image Data Hiding Scheme

Robustness to geometrical distortions is essential to the data hiding scheme. However, most of existing methods mainly focus on the geometrical invariant watermark, and data hiding techniques with high robustness to geometrical attacks is still a challenge. In this paper, we propose a novel geometrical invariant image data-hiding scheme based on fast correlation attacks of stream ciphers. The scheme can resynchronize the attacked image on the basis of the robustness of feature points and the powerful error-correcting characteristic of fast correlation attacks technique. The paper explicitly analyses the channel noise model, which provides a basis on which the parameters of fast correlation attacks is designed. Simulation results demonstrate that our data-hiding scheme is not only robust to geometrical distortions including RST and general affine transformations, but also robust to common signal processing such as JPEG compression.