Robust technique for watermark embedding in a video stream based on a block-matching algorithm

As digital multimedia are transmitted more easily and rapidly, pirates can more easily and unlimitedly copy them. Therefore, people are more attentive to intellectual property and copyright protection. Watermarking is a potential technique for copyright protection and authentication of multimedia data on the internet. A watermark, acting like content, is embedded into the original host signal. For digital watermarking of an image, it is often inserted in the spatial or the frequency domain. A watermark can be embedded into the image intensity in the spatial domain, or into the coefficients of the discrete cosine transform in the frequency domain. We propose a novel video-watermarking algorithm based on a block-matching algorithm to achieve copyright protection in the spatial domain. The advantages of our proposed algorithm are robustness against not only signal attacks but also intentional attacks. Moreover, our proposed technique is easy to perform and can provide an adjustable threshold to achieve superior performance in imperceptibility and robustness. The results show that our approach can effectively extract the watermark under various common attacks such as MPEG compression, luminance, noise, filtering, cropping, and blurring, and can thus achieve copyright protection.