2D Gabor filters-based steganalysis of content-adaptive JPEG steganography

Abstract

To improve the detection accuracy for content-adaptive JPEG steganography, which often constrains the embedding changes to complex texture regions, a new steganalysis feature called the GRF (Gabor Rich Feature) based on two-dimensional (2D) Gabor filters is proposed. First, the diverse 2D Gabor filters are generated and used to filter the decompressed JPEG image. Second, five types of statistical features are extracted from the filtered images and these features are merged according to their respective symmetry. Third, all the features are combined and feature selection is performed to reduce dimensionality. Last, an ensemble classifier is used to assemble the steganalysis feature as well as the final steganalyzer. The experimental results show that the proposed steganalysis feature can achieve a performance that is competitive with the state-of-the-art steganalysis features when used for the detection of the latest content-adaptive JPEG steganography algorithms.