基于DWT和RBF神经网络的水印算法研究

The study of watermarking algorithm based on DWT and RBF neural networks

提出了一种在经过四级小波变换的原始图像中嵌入数字水印的算法。根据人类视觉特征来决定嵌入水印的强度,用秘钥来决定水印嵌入的位置,通过使用训练的RBF神经网络来嵌入和提取水印。实验结果表明:该算法简单有效,且对常见的图像处理具有较好的鲁棒性。

This paper proposed a blind watermarking scheme in which a watermark is embedded into the 4-level discrete wavelet transform domain .The watermark insertion strength is decided by the Human Visual System .The watermark location is decided by the secret key .The trained RBF neural networks is used to embed and extract watermark. The experimental results show the effectivity and robustness of the digital watermark.