High-fidelity reversible data hiding scheme based on multi-predictor sorting and selecting mechanism

Abstract

Reversible data hiding can completely recover the cover image without any distortion after the secret data is retrieved. In this paper, a new high-fidelity reversible data hiding scheme is proposed using the prediction-error histogram modification technique. A multi-predictor sorting and selecting mechanism is proposed to select an optimum predictor from a set of predictors. Then a smaller prediction-error is obtained to embed the secret data using the prediction-error histogram modification technique. Since the multi-predictor sorting and selecting mechanism can utilize the advantage of multi-predictor and information of current predicted pixel to obtain a more accurate prediction value, our proposed scheme can introduce less image distortion at the same embedding payload. Experimental results demonstrate that our proposed scheme outperforms the similar reversible data hiding schemes.