【摘要】 针对目前图像水印算法存在的水印可见性与抗攻击鲁棒性的矛盾,为获得理想的图像水印效果,设计一种基于奇异值分解和粒子群优化算法的图像水印算法.首先对原始载体图像进行尺度不变特征变换,选择水印嵌入的区域,并将水印嵌入区域划为多个子块;然后采用奇异值分解算法对子块进行处理,建立奇异值矩阵,并对水印和水印嵌入区域子块进行融合生成水印矩阵;最后采用粒子群优化算法确定水印嵌入的强度.图像水印仿真实验结果表明,该算法可得到理想的水印嵌入效果,水印的不可见性较好,人眼不能感觉出水印嵌入的影响,水印对各种攻击具有较强的鲁棒性,且该水印算法的整体性能明显优于当前其他图像水印算法.更多还原

【Abstract】 Aiming at the contradiction between watermark visibility and anti-attack robustness in the current image watermarking algorithm,we designed an image watermarking algorithm based on singular value decomposition(SVD)and particle swarm optimization(PSO)algorithm to obtain the ideal image watermark effect.Firstly,the image of original carrier was transformed by scale invariant feature,the watermark embedding area was selected,and the watermark embedding region was divided into multiple sub-blocks.Secondly,the singular value decomposition algorithm was used to deal with the sub-blocks,and the singular value matrix was established,and the watermark and the subblock were fused into the watermark matrix.Finally,the particle swarm optimization algorithmwas used to determine the intensity of watermark embedding.The simulation experiment results of image watermarking show that the algorithm can get the ideal watermark embedding effect,the watermark is invisible,the human eye cannot feel the effect of the watermark embedding,the watermark is robust to various attacks,and the overall performance of the watermarking algorithm is much better than other current image watermarking algorithms.更多还原