Halftone Image Steganalysis by Reconstructing Grayscale Image

Halftone Image Steganalysis by Reconstructing Grayscale Image

연구 분야: Strategies

학회: International Conference on Artificial Intelligence and Security

Utilizing some special pixels patterns constructing the steganalytic features is popular in binary image steganalysis, which could be also used in halftone image. There is almost no specific image steganalysis for halftone image. In this paper, a halftone image steganalysis scheme is proposed and achieves a satisfactory performance, which is totally different from the previous works focussing on some special pixels patterns. Inspired by the fact that halftoning techniques based on the low-pass characteristic of human visual system model (HVS model), the grayscale image is considered reconstructing with a Gaussian filter. And the distortions caused by embedding secret messages will still exist. After that, some common grayscale image steganalysis can be used for extracting the steganalytic features. Furthermore, a series of experiments are conducted and the experimental results show that the proposed scheme is effective on halftone image steganalysis.