摘要： 在DCT域提出了基于模糊和人类视觉系统的新颖的数字水印算法,该算法将图像块按照亮度和纹理进行分类,可以自适应地根据图像特点嵌入水印。水印为二值图像,根据块分类以不同的嵌入强度嵌入到所选择块的中频系数中。该方法允许在一定的视觉距离下嵌入更为鲁棒的水印,达到鲁棒性和不可见性的良好折衷。仿真结果显示了该方法可以抵抗如压缩,加噪等常用攻击。

关键词: DCT, 人类视觉系统, 模糊C均值聚类

Abstract: This paper proposes a DCT-based image adaptive watermarking scheme utilizing HVS and fuzzy clustering theory, through the classification of brightness, texture of image blocks, can adaptively embed the watermark which is a binary image. According to the block classification, it embedded the watermark into the larger coefficients of the mid-frequency of selected blocks. This scheme allows embedding more robust watermark under a certain visual distance so as to achieve a better solution to solve the confliction between implicitly and robustness. Simulation results indicate that the scheme proposed in this paper can survive JPEG compression, noise as well as cropping.

中图分类号: 

• TP309