时空特征分析结合随机密钥的压缩域数字视频 水印嵌入和提取方法.

Aiming at the problems that the existing watermark embedding and extraction algorithms are sensitive to the filtering, compression and noise conditions, this paper proposed a robust digital-domain video watermark embedding and extraction method based on space-time features. The proposed framework consisted of a public key and a private key to prevent self-conspiracy attacks. The algorithm performed spatio-temporal analysis on the video and extracted the public key from the spatiotemporal feature of the compressed video, which was inherently robust. First, it used a random key to select a candidate block from a pre-selected block set to ensure the security of the watermark frame. Then it selected the 4 x 4 sub-block suitable for embedding watermark based on the spatio-temporal features of compressed video. Finally, it embedded the watermark bits using nonzero quantized coefficients. The watermarking framework allowed a limited increase in the video bit rate and reduced the computational overhead . The experimental results show that compared with the contrast method, the proposed method has strong robustness and safety. [ABSTRACT FROM AUTHOR]