The Research Of Stereoscopic Video Watermarking Technology And Its Application

Stereoscopic video technology has been under consistent development, as it can provide stereoscopic perception with a strongly sense of realism and immersion to end users. Nevertheless, it is a great challenge to ensure the security of stereoscopic video. Watermarking technology is regarded as an effective approach to resolve information security. Many researchers have begun to turn their attention to the stereo-video watermarking. Because of the huge size of video data, video usually storage and transmission in compressed format. Hence, the research of compressed domain watermarking is more important. Besides that, stereoscopic video usually be encrypted before upload to cloude. Thus, the research of watermarking into stereoscopic video encrypted domain is very urgent. Three stereoscopic video watermarking algorithms are proposed as following: 1. To solve the capacity problem caused by distortion drift we propose a larger capacity watermarking algorithm without distortion drift. Firstly, the mechanism of error-drift is deeply researching to analyze the seriousness of error-drift to any frame types. The frames which drift-level are higher are selected to watermark embedding. Secondly, in order to increase watermarking capacity, an exact statistical analysis in the compressed domain is made to find more locations which are appreciated to embedding. Thirdly, the copyright information is preprocessed by secret sharing method, and in order to maintain the perceptual quality and embedding efficiency, the matrix embedding method based on Error-correcting codes is adopted. Experimental results show that the propose algorithm has higher embedding capacity and efficiency than current algorithms.2. At present, the stereoscopic video watermarking mostly has sole function. To resolve this problem, a stereoscopic video dual-watermarking algorithm with controllable error-drift is proposed. Firstly, by full analyzing the error-drift problem, an embedding method with limited error-drift is proposed. Secondly, the copyright information which preprocessing by Error-correcting encoding method, is embed into the textual region in I-frame by correlation modulating method. Thirdly, the authentication information which generate from I frame, the most important frame in a group pictures(GOP), is embedded as fragile watermark. Experimental results show that the proposed watermarking algorithm has dual effect: copyright protecting and authentication.3. To protect the security of the data in cloud, a watermarking embedding method that vacating the embedding room before encryption is proposed. Firstly, the motion/disparity vector difference and integer discrete cosine coefficients from the non-critical frames are reserved before encrypted with a suitable amount of watermark capacity and controllable error-drift. Secondly, encrypt the encoding modes, motion/disparity vector difference and integer discrete cosine transform coefficients from the important frames by equal length code word substitution. Thirdly, embed watermark into the encrypted stereoscopic video that has reserved redundant room before encrypted. Experimental results show that the encrypted stereoscopic video is invisible enough to get non-effective information; the proposed watermarking algorithm that performs in the encrypted stereoscopic video not only with controllable error-drift but also maintains the bit-rates unchangeable, can meet the reality practical needs to ensure information security.