基于神经网络的灰度级数字水印嵌入技术

提出一种将灰度级数字水印嵌入到彩色图像中的方法。利用DCT变换，先将灰度水印编码成二值位流信息：用神经网络建立彩色图像中所选择的像素之间关系模型。最后，通过调整被选择像素点与模型输出值之间大小关系来嵌入水印的二值位流信息。采用信息放大技术，加强水印的嵌入强度。利用神经网络、DCT反变换，提取灰度水印。实验结果表明，该算法对目前JPEG图像压缩变换和某些图像处理操作具有极强的鲁棒性。

Embedding technique with gray digital watermark based on neural network

A method of embedding a gray digital watermark into a color image is proposed. By discrete cosine transform technique, the gray digital watermark is encoded into a series of binary numbers for watermarking. The relational model based on neural network is established among those selected pixels of color image. Binary numbers of watermark are embed by adjusting the value between the selected pixels of the color image and the output of the model. The robust performance of the embed watermark is enhanced by the information enlarged technique. The embed watermark can be extracted without original image but by the same model and inversed discrete cosine transform. Experiment results show that the proposed technique is very robust against some image processing operations and joint photographic experts group (JPEG) lossy compression.