Abstract
In this paper, a video scrambling method is proposed. It is realized by separating a video fingerprint when a video is encoded into MPEG-2 stream. A concept of “Video Fingerprint” as the key information to distinguish a video has already been proposed. In the proposed method, the head DC factors of every slice of intra coded frames are manipulated using the value made from a video fingerprint. The MPEG streams those are encoded with the proposed method are based on the MPEG syntax, but they cannot be decoded normally without a video fingerprint. Proposed method reduces PSNR 10.6 dB from 31.9 dB to 20.3 dB applied to nine test sequences that are used MPEG-2 standardization. In proposed method, the amount of information to indicate the method to manipulate DC prediction value is increaded, but the amount of VLC of DC prediction value is decreased, the relationship is also shown in this paper.
