中文摘要：

提出一种无监督的通用隐写检测框架,根据待测图像的固有统计特性,检索与之相似的载体图像从而构造辅助图像集;对由待测图像和辅助载体图像集构成的测试图像集进行异常检测,将载密图像看作异常点,判断待测图像是否嵌密。实验表明,在对原始单次采样和重采样图像组成的混合异构图像进行隐写检测时,该框架取得了良好的检测效果和检测效率。

英文摘要：

We propose an unsupervised universal steganalysis framework. First, cover images with statistical properties similar to those of the given test image are searched from a massive cover image database to establish an aided sample set. Second, outlier detection is performed on a test set com- posed of the given test image and its aided sample set to determine the type of the given test image. The experimental results illustrate that the proposed framework could achieve high detection perform- anee and efficiency on heterogeneous dataset composed of raw single-sampled and resampled images.