Research On Digital Video Watermarking Techniques For AVS

With the development of Internet and information technology, video information has become the main part of on-line resources, in stead of text and static image. Nevertheless, the transparent digital information is vulnerable to reproduction, tampering and illegal propagation. In order to solve these problems, digital watermark technology, as a new protection method of copyright, has become the key research area in the field of signal and information processing. And digital video watermark technology has attracted special attention for its large amount of potential commercial value. At the same time, the standard of copyright protection in AVS video coding standards is still under construction. AVS video coding standards is a new generation of digital video coding standards established by the audio video coding standard workgroup of China with autonomous intellectual property rights. The digital video watermarking algorithm based copyright protection of AVS will become the main technical method. Based on a deep analysis of technical characteristics of video watermark technology and AVS video coding standards, this paper presents three new digital video watermark algorithms:(1) A subjective tests based adaptive video watermarking algorithm is proposed. By introducing the subjective tests, an adaptive watermarking algorithm is presented which is based on measured visibility thresholds of the impairments brought by watermarking. The visibility thresholds are determined by subjective tests based on comparisons of original and watermarked video frames. Experimental results show that the presented watermarking scheme achieves comparatively high detection rate than spatial non-adaptive watermarking scheme and the simple human visual system based adaptive watermarking scheme.(2) A novel video watermarking algorithm based on motion vectors of AVS is proposed. During the inter prediction step of AVS video encoding system, watermark embedding is implemented by modifying the resolution of the motion vectors of partition blocks in different macroblock partitions. The modulation is based on the mapping rules between the motion vector resolution and the watermark bits. At last, the prediction modes of current macroblock is determined using standard AVS coding algorithm which will ensure that the best prediction mode is used for current watermarked macroblock. The experimental results show that the presented watermarking method achieves comparatively high embedding rate without complicated computation, and has almost no impact on video bit rate and video quality subjectively as well as objectively.(3) A compressed domain based video watermarking algorithm for copy control of AVS bit stream is proposed. In the bit stream of AVS video compression standard, watermarking was implemented by modulating the corresponding Exp-Golomb code, which is based on the mapping rules between these codes and the information bits. The extraction of information needs no original media, and merely requires decoding the Exp-Golomb code from bit stream rather than decoding the whole video. The experimental results show that the presented watermarking method achieves comparatively high embedding rate without increasing the bit rate of data, and has almost no impact on video quality subjectively as well as objectively. In addition, a selective video encryption algorithm is proposed to enhance the security of the copy control system by encrypting the motion vectors in video stream.