Improved pixel value differencing steganography using logistic chaotic maps

The pixel-value differencing (PVD) method is one of the steganographic methods that provide higher embedding capacity without very noticeable artifacts in the cover image to human eyes. However, stego-images generated by PVD can be detected using histogram analysis of the pixel-pair differences. Also once the image is known to have embedded data, the extraction algorithm of PVD can be applied to retrieve the embedded message. In this paper, we propose a new modification to the original PVD based on logistic-chaotic maps. The aim is not only to increase its security against the histogram analysis, but also to add another level of challenge for extracting the secret message by the steganalyzer. The proposed method is empirically evaluated and compared with the original PVD on several images.