Abstract: An adaptive digital watermarking algorithm based on GHM and fuzzy clustering theory is presented.The wavelet blocks are classified by utilizing the energy of multiwavelet coefficients and important coefficients.According to the above results and the image local characteristics，a binary watermark is embedded into two correlated detail images adaptively in multiwavelet domain.The embedding process is repeated to improve the robustness of the watermark scheme.The watermark can be extracted without the original image.Experimental results indicate that the performance of the proposed watermarking is invisible and robust to common image processing operation，such as crop，JPEG compression and sharp etc.

Key words: GHM multiwavelet, fuzzy clustering theory, digital watermarking

摘要： 提出一种基于GHM多小波变换域上的模糊自适应水印算法。利用多小波系数块的能量、重要系数对小波块进行分类，结合图像自身的局部特性，在多小波系数中的两个相关的细节子图像中自适应地嵌入一幅二值水印图像；为了增强水印的鲁棒性，采用冗余嵌入方法；水印的提取不需要原始图像。实验结果证明，该算法对剪切、JPEG压缩、锐化等多种图像处理具有很好的鲁棒性。

关键词: GHM多小波, 模糊聚类, 数字水印