一种调色板聚类的信息隐藏算法*

针对目前调色板信息隐藏方法的弱点, 结合聚类分析技术, 提出了一种新的基于调色板聚类的信息隐藏算法ClusterStego 。根据定义的相似性测度和聚类准则函数, 通过对由调色板形成的RGB 空间进行合理聚类来隐藏信息, 具有良好的直观效果。同时由于引入了质量系数的概念, 使得该算法具有隐藏质量可控性, 而且其隐藏容量远大于EzStego 等算法, 可达到每像素87. 5% 。

Palette Clustering Method of Steganography

A new method of steganography is proposed. The method is based on the cluster analysis derived from the tech-nique of pattern cluster analysis. According to the similarity measure and amalgamation rule, the RGB space formed by the palette is analyzed. After that, the better categories are obtained. In each category, all palette items are used to hide some message bits. Because the dynamic bits produced by cluster, and with experiments, the proposed method can avoid the statisti-cal attack in a certain extent. With the quality coefficient, the steganographic quality can be controlled by user choice and the steganographic result has a better intuitional effect. Additional, the steganographic capacity using this method is much larger than that using the EzStego. The maximum capacity can reach 87. 5% per pixel.