An Improved Interpolation based Reversible Data Hiding for Medical Images

Despite having a good embedding performance for natural images, performance of the Interpolation based Reversible Data Hiding (IRDH) schemes is yet to be verified for medical images. An IRDH scheme embeds only in the up-sampled pixels leaving the original pixels intact in an input image, which is strictly required in many medical and military image applications. This paper therefore presents an improved IRDH scheme and analysis of its embedding performance for medical images. Our proposed scheme utilizes local dispersion based adjustment to improve the quality of embedded images further. For higher embedding capacity, correlation between each interpolated pixel and its embedded versions is used. Experiments have been carried out for embedding data into a set of medical images, and our early results are evident of significant improvement in embedding performance.