New steganalysis methods based on camera source identification and outlier detection

Abstract

In order to solve the cover source mismatch problem, a new blind universal image steganalysis algorithm combining camera source identification and outlier detection is proposed in this paper. Cover images from the same source of the test image are retrieved from the reference image database as aided samples. The test image and its aided samples form a whole test set. As the stego image can be regarded as an outlier, the outlier detection is performed on the test set to judge the test image as cover or stego. This method has removed the need for training and mitigated the impact of cover source mismatch. Experiment results show that our method is effective on several JPEG steganography algorithm.