Data Hiding in Independent Components of Video

摘要

Independent component analysis (IA) is a recently developed statistical technique which often characterizes the data in a natural way. Digital watermarking is the main technique for copyright protection of multimedia digital products. In this paper, a novel blind video watermarking scheme is proposed, in which ICA is applied to extract video independent components (ICs), and a watermark is embedded into the ICA domain by using a 4-neighboring-mean-based method. The simulation shows that the scheme is robust to MPEG-2 compression and able to temporally resynchronize.