结合模糊理论和RBF网络的彩色图像盲水印算法

通过对彩色图像的深入分析,以彩色图像为载体图像,将灰度图像作为原始水印,通过模糊控制理论实现水印嵌入强度的动态设定.最后利用RBF来训练模拟量化的逆过程,借助训练得到的RBF神经网络来完成嵌入和提取水印,进一步提高水印鲁棒性.实验结果表明该算法能够动态地控制水印嵌入强度,较好地平衡了鲁棒性和可感知性的矛盾.

A Color Image Watermarking Algorithm Based on Fuzzy Control and RBF

In this paper, in-depth analysis of color images to color images as a vector image, the grayscale images as the original watermark, through the fuzzy control theory to achieve watermark embedding strength of the dynamic setting. At last, simulation of RBF to train the inverse quantization process, using the training received RBF neural network to complete the watermark embedding and extraction, and further improve the watermark robustness. Experimental results show that the algorithm can dynamically control the watermark embedding strength, better balance of robustness and perceived contradiction.