Calibration based reliable detector for detecting LSB matching steganography

In this paper, a novel detector for detecting LSB matching steganography in gray-scale images is presented. We extend the previously introduced detectors, the one based on the center of mass of the histogram characteristic function (HCF-COM) and its calibrated version (C-HCF-COM), in two aspects. On one hand, to carry out detection, HCF-COM and C-HCF-COM computed from the first and second order difference images are combined together to build a feature set. On the other hand, the calibration used in C-HCF-COM, i.e., the downsample of original image, is changed to the Gaussian filtered image. With these extensions, our method significantly improves HCF-COM and C-HCF-COM. Its superiority over some state-of-the-art works is also demonstrated through extensive experiments. ? 2013 Springer-Verlag.EI