基于子采样方法的鲁棒性水印算法

提出了一种新的基于子采样的图像数字水印算法．随机产生的水印信号被分别嵌入到离散傅里叶变换后的各子采样图像中，利用各子采样图像间的相似性特征，能够有效地提高水印算法的检测性能．理论分析表明，本算法能够在不降低宿主图像质量前提下，使得实际嵌入水印强度达到扩频水印方法的4倍，从而有效地提高算法的鲁棒性．数值试验结果说明，同传统的扩频水印算法相比，本文水印算法对多种攻击具有更好的鲁棒性．

The Robust Watermarking Scheme Based on Subsampling

A novel subsampling based robust image watermarking scheme was proposed.Randomly generated watermark sequences are modulated into the discrete Fourier transform amplitude of those subsampled images, respectively. Taking advantages of the high similarities among those subsampled images, the detection performance of the proposed watermarking scheme can be effectively improved. The theoretical analysis shows that the actual embedding strength of the proposed method is four times that of the traditional spread spectrum watermarking scheme under the same image quality. The experimental results under different attacks show that the proposed watermarking scheme can greatly improve the robustness of the spread spectrum watermarking scheme.