Research On Second Generation Video Watermarking Algorithms Based On Content Features

With the development of information technology and the growing popularization of network, multimedia information (image, video, audio, etc) has become an indispensable way of communication. People could download and spread digital products through the Internet very conveniently, but simultaneously, various behaviors like copyright violation, illegal copying, tampering are becoming easier too. So the security of multimedia data has become a very important issue.Digital watermarking is a technology that watermark information is embedded into digital multimedia then carries out the function such as identification, copyright safeguarding, content authentication etc. At present, digital video watermarking is becoming a powerful tool for video copyright and security authentication.The second generation video watermarking schemes are studied in this paper. Based on video content features, the watermark is embedded into the perceptually significant features of video. The watermarking system is becoming more robust and stable by making the watermark embedding and significant features of video linked closely. The main researches which have been done in this thesis are described as the following:(1)In order to embed the watermark into the significant features of video, the content features and how to detect these features are studied firstly in this thesis. The important features which are explored include:motion objects of video, feature points and video objects.(2)In the premise of extracting the moving objects of video effectively, a motion location based video watermarking scheme is proposed, in which we embed the watermark into the located motion region. Experimental results show that the proposed algorithm is robust to usual processing attacks, especially to MPEG2 compression. On this basis, another novel self-embedded video watermarking scheme is proposed. The simulations show that the scheme has a good performance to resist common signal processing attacks, and is also effective for video authentication. The results demonstrate the algorithm is robust to resist local changes and could recover the major content when the watermarked video is partly tampered so as to protect the completeness of the video.(3)In the premise of detecting the stable feature points, a robust watermarking scheme is proposed based on local feature regions. Experimental results show that the proposed scheme is robust against common processing operations, furthermore, it can resist the attacks of geometric distortions like cropping, scaling and rotating.(4)Based on detecting the accurate video object region, an object-based video watermarking algorithm using features for correcting synchronizing information is proposed. Experiments demonstrated that the algorithm can well protect the video object, and be robust to usual signal processing attacks and geometric attacks like scaling and rotating. Furthermore, it is also robust to special video attacks such as frame dropping, MPEG4 compression etc.