Robust anaglyph 3D video watermarking based on cyan mosaic generation and DCT insertion in Krawtchouk moments

Abstract

The need to protect 3D video content against pirating has become very important due to the widespread expansion of the internet and the proliferation of 3D video content consumption over networks. Several 3D video watermarking approaches have been proposed to resolve this problem, but they have been lacking efficiency to resist malicious attacks, especially the collusion attack. In fact, this attack has not been considered by most existing work. This paper proposes a novel approach for 3D anaglyph watermarking based on cyan mosaic as an embedding target. First, cyan mosaic is generated using the cyan channel extracted from the original 3D anaglyph video. Then, a signature is embedded into the obtained mosaic using Krawtchouk moments and a DCT based technique. This maximizes the compromise between invisibility and robustness against usual and malicious attacks. Finally, the marked 3D anaglyph video is obtained by combining the marked cyan video and the original red video. The experimental results illustrate good robustness against compression, collusion and several other additional attacks, such as geometric manipulation and temporal attacks. Besides, the suggested technique provides high invisibility thanks to DCT based embedding.