An efficient video watermarking scheme with luminance differential DC coefficient modification

Abstract

In this paper, an efficient video watermarking scheme is presented through modifying the third decoded luminance differential DC component in each selected macro block. The modification is implemented by binary dither modulation with adaptive quantization step. The proposed scheme is based on the observation that luminance differential DC components inside one macro block are generally space correlated, so the quantization step can be adjusted according to adjacent differential components, to utilize properties of human visual system (HVS). This method is very robust to gain attacks since amplitude scaling will have the same effect on differential components and the quantization step. Experimental results show that it can be implemented in real time with better visual quality than uniform-quantizing scheme. © Springer-Verlag Berlin Heidelberg 2007 Video watermarking, real-time embedding, luminance.