Overlapping pixel value ordering predictor for high-capacity reversible data hiding
- 45 Downloads
In recent years, many reversible information hiding methods have been proposed. Among them, the pixel value ordering (PVO) method can be used to create high-fidelity camouflage images under good embedding capacity. The original PVO method adopted the block-by-block manner, and each block could embed only 2 bits. In this study, we propose an overlapping PVO (OPVO) method. In the PPVO method, secret data are embedded in the maximum and minimum pixel values to improve the situation in which only up to 1 bit can be embedded in a sliding window. In the pixel-by-pixel manner, each pixel can embed data multiple times. We make full use of the correlation between adjacent pixels in the natural image to increase the embedding capacity. An advantage of the overlapping PVO method is that its embedding capacity is more than twice as high as that of previous PVO series methods where an acceptable image quality is maintained. In our experiment, the embedding capacity of the smooth image is up to 130,000 bits, whereas that of the complex image is higher. However, the image quality decrease rate of the proposed method tends to be lower than that of the other methods. Therefore, any kind of image is suitable for our method.
KeywordsReversible data hiding (RDH) Difference expansion (DE) Histogram shifting (HS) Prediction-error expansion (PEE) Pixel-value-ordering (PVO)
This research was partially supported by the Ministry of Science and Technology of the Republic of China under the Grants MOST 106-2221-E-324-006-MY2.
- 15.D. Coltuc, I. C. Dragoi, Context embedding for raster-scan rhombus based reversible watermarking. In: Proceedings of the First ACM Workshop on Information Hiding and Multimedia Security (ACMIHMMSEC), ACM, Montpellier, France, June 17–19, pp. 215–220 (2013)Google Scholar
- 23.Chen, X., Sun, X., Sun, H., Xiang, L., Yang, B.: Histogram shifting based reversible data hiding method using directed-prediction scheme. Multimed. Tools Appl. 74, 1–19 (2014)Google Scholar