Geoffrey Hoffman
Steve Weiss
For our project, we propose to research the effects of watermarking on MPEG video. Currently, most watermarking research seems to focus on images, and claims that the concepts apply to audio and video as well. We will focus on MPEG video. Our goal is to only watermark I-frames, while using the original image for prediction in P and B frames, so that every frame in the sequence will appear to be watermarked.
Ingemar Cox's paper, Secure Spread Spectrum Watermarking for Multimedia, presents a high-level watermarking algorithm, which appears to be very effective. We will implement several variations of this algorithm, and compare their effects against both tampering, and the various features of MPEG 1 and 2 (such a prediction, interlaced images, and scalability).
Last Modified: