一种基于粗糙集属性约简的图像隐藏信息检测方法

Steganalysis Based on the Rough Set Attribute Reduction

基于统计分析的检测方法是图像隐藏信息检测常用的方法,与特定隐写分析方法相比,它比较灵活,能够快速地适应新的或未知的隐写算法。针对特征属性高维的特点,采用粗糙集属性约简的方法来降低数据规模。实验表明:与采用全部属性的检测系统相比,该方法在不影响分类精度的前提下有效提高了检测速度。

The statistical detection is generally used in steganalysis. Comparing with the special stega- nalysis, it has the advantage of flexing. However, it has the disadvantages in large-scale data,and low calculate speed. For the high dimensions of feature attributes, the attributes reduction based on rough set can delete the non-important knowledge while not alter the classification ability of knowledge. The result of experiment shows that the proposed detection technique achieves a comparable precise detection rate as algorithm based on all feature attributes, however, it evidently decreases detection time.