Authors : A. Rasmi and M. Mohanapriya
Abstract: Reversible data hiding with high capacity is a challenging area of digital image steganography. Prediction error histogram of images have large frequency of zero values. The modification of prediction error histograms can be utilized to store secret digits. This study presents a method called to encode a secret image into digits in base and these are called trits which hides it by modifying the prediction error histogram of the cover image. This improves the embedding capacity while allowing the original cover image to be recovered with zero errors. Experimental results indicate that the secret image can be recovered with little loss. The peak signal to noise ratio of the recovered secret image is above 30 decibels. The embedding capacity is high enough to allow the hiding of 80×80 secret grayscale images inside 256×256 cover grayscale images. The method is applicable to color images also. Several predictors are considered in this research and their relative performance are analyzed and discussed in detail. The performance of the method is compared against several existing reversible hiding techniques. The method has low computational complexity with fast execution times.
A. Rasmi and M. Mohanapriya, 2018. Reversible Data Hiding of Images using Trinary Representation. International Journal of Soft Computing, 13: 38-47.