摘要
研究了一种基于DWT与SVD相结合的半盲彩色数字水印算法,鉴于嵌入水印的实用性,原始图像与水印图像采用的都是彩色图像。在置乱变换系数和半盲水印的双重保护下,提高了水印的安全性,奇异值分解对常见的几何失真是稳健的,而小波分解可以把图像的主要能量集中到低频域中,结合它们的优点可以更好地提高水印的抗攻击性。实验结果表明,嵌入水印后的图像在加噪、滤波(如模糊、锐化)、缩放、旋转、剪切和有损压缩等攻击后,其提取水印与原始水印仍有较高的相似度与峰值信噪比,该算法具有较好的鲁棒性。
This paper focused on the study of semi-blind color watermarking algorithm based on the combination DWT with SVD. In consideration of the practicality of embedding watermark, the original image and watermark image all used color pic- tures. In scrambling transformation and semi-blind watermarking double protection, enhanced the security of watermark. Be- cause SVD was steady towards the common geometric distortion, and with using DWT, the main energy of image occupied the low frequency domain. It was better to improve the resistance to attack of watermark by combining with the advantages of DWT and SVD. The result demonstrates that after the watermark image subjected to the attacks of adding noise, filtering ( such as obscuring and sharpening) , scaling, rotating, cutting and loss compression, the extracted watermark also has high similarity and peak signal to noise ratio with the oriainal watermark. The alzorithm is robust.
出处
《计算机应用研究》
CSCD
北大核心
2013年第3期917-920,共4页
Application Research of Computers
关键词
离散小波变换
奇异值分解
半盲水印
置乱变换
鲁棒性
discrete wavelet transform (DWT)
singular value decomposition ( SVD )
semi-bind watermarking
scramblingtransformation
robustness