一种基于主成分分析的图像信息隐藏质量评价方法

Evaluation method of image information hiding quality based on principal component analysis

评价信息隐藏方法的隐藏质量是信息隐藏研究的一个重要部分。针对传统的主客观评价方法存在的缺点和不足,提出了一种图像信息隐藏质量的评价方法。选取对图像信息隐藏比较敏感的10个图像质量评价量(IQMs),作为评价信息隐藏质量的特征量。利用主成分分析法对各特征量进行分析,排除各特征量之间的线性相关性,选择主成分量实现了对几种图像信息隐藏方法的综合评价。该方法灵活、高效,评价结果符合人眼视觉特性,是比较实用的图像信息隐藏质量评价方法。

Hiding quality evaluation of information hiding method is an important part of information hiding research. An evaluation method of image information hiding quality is proposed to eliminate the drawbacks of traditional subjective and objective evaluation methods. Ten image quality metrics （IQMs） are selected, which are sensitive to image information hiding. These IQMs serve as characteristics to evaluate information hiding quality. The characteristics are analyzed by using principal component analysis method and the linear correlation between characteristics is eliminated. Comprehensive evaluation to some image information hiding methods is realized by choosing principal component characteristics. This method is flexible and efficient. Evaluation results fit human visual feature and it is a practical evaluation method of image information hiding quality.