A Robust and Blind 3D Mesh Watermarking Driven by Stable Regions

Watermarking 3D meshes is very important in many areas of activity including digital cinematography, virtual reality as well as in CAD design. Due to the complexity of 3D meshes, mesh watermarking schemes are far from the maturity of watermarking algorithms dedicated to audio, image or video watermarking. No existing scheme is robust against all attacks. To verify a good trade-off between invisibility and robustness, we propose in this paper a mesh watermarking technique which combines three different algorithms to profit from their robustness. In plus, we embed three different signatures. This allows to know the type of used attack by detecting which of the signatures resisted. To maximize robustness, we embed these signatures in stable regions detected using an extension of MSER algorithm. Experimental results show that this new approach is robust against numerous attacks including geometric transformations, remeshing, simplification, noise and smoothing. In addition, it presents a blind detection.