改进的基于压缩感知的鲁棒性水印算法

An Improved Watermarking Algorithm with Strong Robustness Based on CS

目的为了提高水印算法的抗攻击能力,提出一种改进的基于压缩感知的鲁棒性水印算法。方法为了确保嵌入的水印信息在载体中分布均匀,对彩色载体图像的R,G,B分量进行N×N不重复分块后得到Rij,Gij,Bij,按照下标值不同的原则,随机选择3个颜色分块构成嵌入块;为了提高水印的鲁棒性和检测的成功率,对水印信息采用循环嵌入的方式;详细论述水印的嵌入和提取流程,并通过仿真与其他3种算法进行实验结果对比。结果水印提取不需要原始图像的参与,具有盲检测性,水印的循环嵌入策略极大地提高了水印鲁棒性和水印检测成功率。仿真实验结果表明,重建的水印图像具有较小的失真度,算法能够很好地抵抗一系列的几何攻击。结论算法具有较好的鲁棒性和不可感知性,实现较简单,在版权保护及信息安全领域具有一定的参考价值。

The work aims to put forward an improved watermarking algorithm with strong robustness based on CS to improve the anti-attack capability of watermarking algorithm.In order to ensure the uniform distribution of embedded watermark information in the carrier,the Rij,Gij and Bij were obtained from N×N non-repetitive blocking of R,G and B of color carrier image.According to the principle of different subscript values,the embedded blocks were randomly selected from each of the three color categories.In order to improve the watermark robustness and the success rate of detection,the strategy of cyclic embedding of watermark information was adopted.The watermark embedding and extracting process were described in details.The proposed algorithm was compared with other three algorithms through simulation.The process of extracting the embedded watermark without the original image achieved a blind detection.The strategy of cyclic embedding of watermark greatly improved the watermark robustness and the success rate of watermark detection.The experimental results of simulation demonstrated that,the reconstructed watermark image had less distortion and the proposed algorithm could resist a series of geometric attacks.With better robustness and visual insensitivity,the proposed watermarking algorithm is easy to realize,which has certain reference value in copyright protection,information security and other fields.