Research On Video Content-Authentication Watermarking Based On MPEG-2

With the rapid development of Internet and computer technology, people have entered into the digital Information Society. While Information Era supplies many conveniences to create and transmit digital multimedia, it also brings a severe problem —the piracy and distortion of digital multimedia. As an effective implements in the copyright protection and integrity authentication for multimedia, digital watermarking techniques has become an active and flouring research topic. This paper mostly combines with MPEG-2 Video Compression Standard, and makes some research on video content authentication watermarking techniques.The paper first explores and analyzes the existing video authentication techniques, and also gives a general introduction in digital video watermarking techniques. Then departing from the characteristic of video content authentication, it improves the video spatial-domain feature extraction algorithm which is based on differential energy. Moreover, the paper proposes a video content authentication watermarking scheme based on MPEG-2, which can detect and localize the tempered areas accurately in both spatial-domain and temporal-domain of the video. While the scheme achieves the goal of video content authentication, it also fairly meets the essentially technical requirement of video watermarking system,such as real time property, invariable bit rate, blind extraction and imperceptibility.The main work covers:1. The existing video content authentication watermarking algorithms are introduced and analyzed, and this technology is also quite deeply investigated from the objects, fashions and methods of authentication. Furthermore, the paper analyzes and compares the typical fragile watermarking and semi-fragile watermarking algorithms used in image authentication at present, which enlightens the further researches of the paper.2. Through altering the combination mode of energy blocks, it improves the video spatial-domain content feature extraction algorithm based on differential-energy, which enhances the capability of exact detection and localization of tempered areas, and then designs a novel video spatial-domain localization algorithm. The modified algorithm can also be extended to image authentication.