一种基于变长位平面的自适应视频水印算法研究*

针对视频水印嵌入强度在平衡水印系统的鲁棒性和不可见性时存在的问题，提出了一种变长位平面分解的视频水印算法。该算法根据8位灰度水印图像在位平面分解后各位平面在重构水印图像时的权值不同，将不同的位平面嵌入在不同数量的DCT中低频系数中，并利用HVS和Watson视觉模型的亮度掩蔽特性自适应的调整水印嵌入的强度从而实现了自适应地嵌入水印。实验表明，该视频水印系统在满足鲁棒性的同时具有高度的透明性。

Adaptive video watermarking based on variable bit planes

This paper proposed a new adaptive video watermarking based on variable bit planes which satisfied the problem of balancing robustness and invisibility during video watermarking embedding. According to different weights in image reconfiguration, embedded bit planes of 8 bit gray image in different quantity of DCT coefficients. At the same time, the embedding strength was adaptively adjusted depending on luminance masking of HVS and Watson visual model. Watermark was adaptively embedded in video carrier. Experimental results show that the presented video watermark scheme is not only robust but also highly transparent.