Abstract:
Lots of existing image watermark algorithms based on Contourlet transformation cannot resist brightness and contrast attack.In order to solve the problem,a blind robust quantization-based watermark algorithm against brightness and contrast is presented.The original image is transformed by Contourlet transformation and gets low frequency band and high frequency sub-bands,and the low frequency band is split into non-overlapping blocks.Each block is decomposed with Singular Value Decomposition(SVD) and a bit of watermark is embedded into the Component of S through odd-even quantization.The inverse Contourlet transformation and SVD are applied to obtain the carried image with the watermark.At the detection end,revision for resistance against brightness and contrast is used to the attacked watermarked image and a watermark is blindly extracted through odd-even judgment.Experimental results show that the proposed algorithm has good transparency for the embedded watermark and it is strongly robust against brightness and contrast attack.
Key words:
Contourlet transformation,
digital watermark,
brightness attack,
contrast attack,
Singular Value Decomposition(SVD)
摘要： 针对目前部分基于Contourlet变换域图像数字水印算法无法抵抗亮度和对比度攻击的缺点,提出一种抗亮度和对比度攻击的水印方法。发送端对源图像进行Contourlet变换,得到一个低频分量和多个不同方向的高频子带,将低频子带分成互不重叠的子块,对每一子块进行信号奇异值分解(SVD),利用奇偶量化的方法将经Arnold置乱后的1 bit水印信息嵌入到S的第一分量中,通过SVD和Contourlet逆变换得到含水印图像。检测端对攻击后的含水印图像进行抗亮度和对比度调整修正,根据奇偶性提取出盲水印。实验结果表明,该算法具有较好的水印透明性,当亮度和对比度受到较大攻击时也能够完整地提取出水印信息。
关键词:
Contourlet变换,
数字水印,
亮度攻击,
对比度攻击,
奇异值分解
CLC Number: