Some Studies On Video Watermarking Algorithm Against Camcorder Recording

Because the network has the character of openness and sharing, some digitalmedia products are often tampered with malicious attacks and illegal infringement ofcopyright. A common situation is when a film has just released, some piratedresources of the film have been available for users to download. This phenomenoncauses serious damage to the creator’s creation enthusiasm and interest. Therefore, theprotection of the original video content has become a hot and difficult problem of theacademics, and digital video watermarking technology is one of the major means ofsolving the problem of video content copyright.This article revolves around digital video watermarking technology. In view ofthe attack of high-definition cameras recording videos, we analyze and summarize allkinds of attacks brought by camcorder recording, such as geometric attacks,transcoding, brightness changing, frame structure changing, etc. Then we put forwardthe following digital video watermarking algorithms against camcorder recording.1. With the purpose of copyright protection of digital video, this paper proposesan HDTV watermarking method that is robust to camcorder recording. Becausecontents in the consecutive frames of a video are almost identical, we embed thecopyright information by fine-tuning the luminance relationship of the consecutiveframes. To ensure the quality of the video and the robustness of the algorithm, we usean adaptive watermark pattern to reduce the image area to be modified, and we makethe embedding strength adaptive according to the improved Watson visual model.When detecting watermarks, to reduce detecting errors caused by recording errors andshot changes, we determine the detecting area adaptively according to the size of thewatermarked video and segment the video into shots using the directional empiricalmode decomposition method. Experimental results show that the proposed methodachieves high video quality and is robust to camcorder recording, transcoding,recoding, and other geometric attacks.2. The computational cost of traditional local invariant features such asSIFT(Scale-invariant feature transform) and SURF(Speed Up Robust Feature) is toobig, thus in this paper we propose a robust video zero-watermark scheme based onORB(oriented FAST and rotated BRIEF) feature against camcorder recording andother common video processing attacks. We select suitable feature points based on theresponse value, then we construct local geometric invariant regions according to the orientation of these feature points. Transform these regions to DCT domain andconstruct zero-watermark information based on the low frequency DCT coefficients.In order to resist the synchronization attack, we add a synchronization sequence in thebeginning of the watermark information. The experimental results show that thealgorithm is robust to geometric attacks, common video processing attacks andcamcorder recording attack.