Based On Mpeg-2 Video Watermarking Technology

As a beneficial complement to traditional cryptography technology, digital watermarking has received high attention, both from academia and industry. In this thesis, two typical watermarking systems respectively focused on copyright protection and video authentication are studied.The principle of digital watermarking technology especially video watermarking is first introduced as well as a survey on status quo of video watermarking research is also given. Attacks against watermarking schemes are important approaches to assess the performance of variants of algorithms. A new way of classifying watermarking attack algorithms is introduced, in order to make the assessment of watermarking schemes more rapidly.Robustness is one of the crucial requirements for watermarking algorithms, especially in scenario of copyright protection use. Thanks to the excellent anti-jamming ability of spread spectrum technique, spread spectrum (SS) watermarking achieves high robustness in addition to other good performances. To further improve the security and the real-time processing capability, this thesis has refined the parameters of one existing SS watermarking scheme and the pre-treatment process in the extraction of watermarks.The main purpose of video authentication is to guarantee the validity of video content. In this thesis, a new video authentication algorithm based on the spatial-temporal information extracted from the video bit streams is proposed. The proposed scheme works efficiently together with the MPEG-2 encoding/decoding process. Spatial authentication mainly aims at the malicious tamper under frame level, while temporal authentication can deal with the tamper methods in or above frame level. The combination of spatial and temporal characteristics provide perfect protection to the video content.