3D Transform Domain Based Robust Watermarking Algorithm For Video

With the rapid development of multimedia technology and computer network, the storage and transmission of digital media becomes more convenient. But the behavior that is infringing the legitimate rights of copyright holder has become increasingly rampant. As an effective method to protect copyright of digital media, digital watermarking technology has been attached great importance to scholars and research institutions. Digital video watermarking becomes a research hotspot in recent years. This paper describes the research background, research status, typical algorithms, basic knowledge, common mathematical transformation, video pre-processing technology of the digital watermarking. This paper proposes robust video watermarking algorithms based on three-dimensional transform domain. The main research contents are as follows:(1) Proposing a robust video multi-watermarking algorithm base on3-D DFT. In this algorithm, using video shot segmentation technique to split the video into different scenes, and then extract the key frames within each lens. The key frames are stacked into video volume date. Then the3-D DFT is performed to the video volume date and we get a feature vector of the video volume date in the transform domain. Using the feature vector to embed and extract the watermarking. The watermarking is extracted without original video, so the watermarking algorithm is a blind extraction algorithm. The watermarking information dose not been embedded into the video, so it is also a zero-watermarking algorithm. The algorithm combines video feature vector, the third party concept and zero-watermarking together. Simulation results show that the video watermarking algorithm can resist conventional attacks and geometric attacks, which has strong robustness.(2) Proposing a3-D DWT-DCT based video watermarking to resist geometric attacks. The algorithm is based on3-D DWT-DCT. Firstly, the3-D DWT is performed to the video volume date and get the approximate coefficient. Secondly, the global3-D DCT is performed to approximate coefficient and get the feature vector form the low-frequency coefficient of3-D DCT. A binary image is selected as the watermark. In order to increase the security of the watermark, we do two-dimensional Arnold scrambling to the binary image. The algorithm combines video feature vector, image scrambling, the third party concept and zero-watermarking together to embed and extract the watermarking. The experimental results show that the proposed watermarking algorithm has strong robustness.