Intensity histogram steganalysis in BPCS-steganography

BPCS-Steganography is a steganographic method that hides secret messages in digital images. BPCS-Steganography extracts local regions of the image to embed using image segmentation based on a complexity measure that separates the image into ``informative'' and ``noise-like'' regions. The human visual system will be unable to perceive any difference by the replacement of noise regions with random binary data. This property allows us to embed secret data into such noise-like regions if the secret data is a random pattern. To avoid suspicion, an image should look innocent after embedding with secret information, not only visually, but also by analysis. A complexity histogram represents the relative frequency of occurrence of the various complexities. In previous work, we studied the complexity histogram of an image when embedded with secret data using BPCS-Steganography, and pointed out an anomaly in its shape. In this paper, we analyze other features of the image theoretically and practically. We consider the intensity of pixels in color components and luminance and analyze the shape of those histograms. As the result of the experiments, we show a more secure method to embed by BPCS-Steganography.