Reversible watermarking using median based prediction and Dynamic prediction error histogram shifting

Reversible digital watermarking is used to successfully recover the cover image as well as hidden information with minimum distortion. Reversible watermarking techniques have very sensitive application domains such medical and military fields where integrity of cover image is very important. In this paper, we are using median based predictor to predict the cover image which does not have any effect of watermark insertion. Instead of expanding prediction errors, we are using Dynamic prediction error histogram shifting to embed information. Our algorithm provides high embedding capacity with minimum distortion.