Reversible Data Hiding scheme for color images based on skewed histograms and cross-channel correlation

Abstract

Reversible data hiding (RDH) is an undetectable communication technology that enables restoration of the original cover image after extracting the hidden secret data from the image. It's being widely used in sevaral fields such as military, medical, etc,. Till date, various RDH have been developed to increase the amount of hidden data while maintaining their quality. However, the problem of traditional trade-off between the embedding capacity (EC) and image quality still persists. For this, we propose a new RDH method for color images based on skewed histograms and cross-channel correlation. The skewed histogram is generated with the help of extreme predictors, aids in achieving less distortion by only incorporating pixels from the peak and short tail. Further, the proposed method splits the secret data (or payload) based on each channel's characteristics so that the most smooth channel can be exploited for embedding to further benefit in reducing the caused distortion and increasing the EC. Moreover, embedding in the complex regions of images is done with the help of more comprehensive predictor which also takes into account various possible edges. Thus, the proposed method achieves greater EC with better quality stego-images than the related and existing RDH methods for color images.