Steganalysis of a PVD-based content adaptive image steganography
Signal Processing
| Hi-index | 0.00 |
Pixel-value-differencing (PVD) based steganographic methods, which take into account the characteristic of human visual system, usually lead to large embedding capacity and good visual imperceptibility. In this paper, a recently proposed PVD-based steganography is further studied. First, by deriving a sufficient condition for data embedding/extraction, we show that the method can be generalized, and we have in fact a much wider choice of embedding parameters. Then, the security of this PVD-based steganography is analyzed. A targeted steganalyzer for this method is designed, and the experiments show that the method can be easily detected. That is to say, like other PVD-based steganographic methods, this most recent one can not resist statistical detection and it is also not secure. Therefore, as a conclusion, more investigation should be done on PVD-based steganographic methods, to enhance their statistical undetectability.