摘要
提出一种基于多通道及小波变换的彩色图像公开数字水印技术方案,该方案充分考虑彩色图像中人类视觉(HVS)的掩蔽特性,将彩色图像由RGB彩色空间转换到YCbCr彩色空间,利用YcbCr良好的空间独立性获取3个相对较为独立的颜色通道,,。对每个独立颜色通道中的宿主图像进行充分的内容信息挖掘,并计算其视觉阈值矩阵JND(just noticeable difference)。将宿主图像分块后在离散小波域内进行水印的嵌入。实验结果表明,该算法不仅具有较好的透明性,而且对诸如叠加噪声、JPEG压缩、几何剪切、图像增强、分色攻击等攻击性实验中均具有较好的鲁棒性。
A multi-channel public watermarking scheme is presented to color image in DWT domain. In the scheme, the special masking characters of human visual system are taken into account. First, the original image is transformed from the RGB color space into the YCbCr, which has the excellent independence of every channel. Secondly, the information of the carrier image is deeply mined in every channel furthermore; the embedding tolerance of watermarking is adaptively calculated according to the content of the carrier image. Finally the watermarking is embedded into the carrier image in DWT domain, The experiment results indicated the scheme has good im- perceptibility and robustness to resist the normal attacks of adding noises, JPEG compression, clipping, image enhancement and the attacking every channel of color image.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第3期675-679,共5页
Computer Engineering and Design
基金
贵州省科研基金项目(黔科合20052109)
