基于正交扩频码和HVS的DCT域图像数字水印技术

文章基于对数字水印通信模型与通信系统相似性的分析，提出了一种利用正交扩频码构造多个独立信道对同一水印进行重复正交嵌入的想法，并引入HVS模型，把图像分块在DCT域分成三类，然后采用不同的嵌入强度系数嵌入到DCT中频系数中，最后从多个独立信道分别提取水印并做合成。实验结果表明，该方案较一般的扩频水印方法对高斯噪声、JPEG压缩、剪切和中值滤波具有更好的鲁棒性。

Digital Image Watermarking Based on Orthogonal Spread-spectrum Code and HVS Model in DCT Domain

In the light of analysis on the similarity between the digital image watermarking model and spread-spectrum communication system, an idea for embedding one watermark for several times dependently by using orthogonal spreadspectrum code is proposed in the paper. Then, the image blocks are dividied into three types in DCT domain and matched with three coefficients respectively for embedding into the medium frequency coefficients of DCT by introducing the HVS model. Furthermore, the experimental results indicates that the scheme improves the robustness of the image watermarking to Gaussian noise, JPEG attack, cropping and to the median filtering.