Tatouage video robuste basé sur les régions d'intérêt

In this paper, we propose two new robust and imperceptible video watermarking approaches based on feature regions. The first one adds the signature in the mosaic while the second takes into account the moving objects. The choice of the mosaic and moving objects is based on their robustness against the most important video's attacks such as collusion and MPEG- 4 compression. In addition, to maximize the invisibility, we propose, for each of these two approaches, different methods based on the choice of feature regions and we discuss the robustness and invisibility of each one. Experimental results have shown that methods which add signature in moving objects allow obtaining the best compromise robustness-invisibility.