Research On Video Watermarking Technology Against Resolution Downsampling Transcoding

Video watermarking is an important part of digital watermarking technology which embeds the invisible information into the video signal to implements the funtion of data protection and hidden communication. Video transcoding is a very common video operator, it has considerable influence on video watermarking detection and extraction.Since it is not necessary to completely decode and then encode, the compressed domain video transcoding technology is getting more and more interest. As a result, the practical application of transcode-resistant video watermarking technology becomes a demand.This thesis focuses on the video watermarking principles, algorithms and also the key techniques of video watermarking against transcoding. The contributions of the paper are as follows:(1) According to the development and general method of video watermarking in recent years, we systematically summarize the characteristics, the research status and the classification of video watermarking algorithms and principles. Research on the key techniques of compressed domain video watermarking is emphasized, as well as the conception and classification of transcoding, including the conversion of bit-rate, spital resolution, frame-rate and code syntax. The key techniques of DCT domain downsampling transcoding are discussed, which consist of macroblock re-encode mode selection, motion vector re-estimation and so on. The effect of video transcoding for compressed video watermarking is analyzed and related experiments are performed.(2) In this paper we present a compressed domain video watermarking technique against video transcoding. The watermarking approach is based on a downsampling algorithm under DCT-domain, with the Watson visual model to control the watermark embedding strength. We choose the middle and low DCT coefficients as the position of watermark sequence to be embedded, so as to achieve the balance between watermark robustness and visual transparency. Experimental results show that the proposed strategies are useful in solving the transcoding attack, especially the downsampling transcoding.In this paper, the research is verified and analyzed by Visual C++ 2005 Tool.