一种基于图象融合的数字水印方法

提出了一种基于多分辨率图象融合的数字水印算法.该算法基本思想源自公开钥加密系统与Torus自同构映射置乱的概念,并根据视觉系统HVS的照度掩蔽特性及组合规则,以原始图象的子波分解级数为依据,对原始图象及水印的子波变换系数进行融合,以实现水印在原始图象空间位置和频率方向上的嵌入.实验结果表明,该算法具有良好的鲁棒性和可行性.

IMAGE FUSION-BASED WATERMARKING

It presents an approach for embedding and extraction watermarking based on multiresolution image fusion.This method transforms both image and watermark into wavelet domain where their coefficient are fused according to combination rules that take into account characteristics of the HVS.The watermarking processing is adaptive and depends on the local host image characteristics at each resolution level,repeatedly merges the watermark coefficient at the various resolution levels of the host signal which provides simultaneous spatial localization and frequency spread of the watermark.The experimental results show that the imperceptibility and robustness of the method can be guaranteed.