This paper introduced a reversible data hiding method based on pixel value ordering with the prediction-error expansion technique and the average value of end pixels'. A host image is first segmented into non-overlapping sub-blocks of three pixels and ordered them as ascending order. For each subblock maximum pixel value and the minimum pixel value is predicted by the middle pixel value and also the middle pixel value is predicted by the average of the maximum and minimum pixel values. Then by using prediction-error expansions, we can embed secret bits into maximum pixel and minimum pixel and also by using the average value of these two pixels we can embed secret bit into the middle pixel of every sub-block. All secret bits can be recovered and restored the cover image completely from watermarked image. Experimental result of this scheme demonstrates that the embedding capacity and average PSNR value is larger than another pixel value ordering and prediction error expansion based approach for relatively smooth images. Also, the visual quality of the obtained marked image is better than other Pixel Value Ordering and Prediction Error Expansion based method.

No abstract

This paper presents a better reversible data hiding method depending on pixel value ordering and prediction-error expansion technique. A host image is first segmented into non-overlapping sub-blocks of adjacent three pixels and ordered them as ascending order. For each sub-block, the maximum pixel value is predicted by the second maximum pixel value. Then the second maximum pixel value is predicted by the minimum pixel value or minimum pixel value is predicted by the second maximum pixel value. Then by using prediction-error expansions, we can insert one or two secret bits into every sub-block pixels and also we can recover the hidden secret bits and restore the cover image fully from watermarked image. Experimental results of this method demonstrate that the embedding capacity and PSNR value is larger than another pixel value ordering and prediction error expansion based approach. Also, the visual quality of the obtained marked image is better than other PVO and PEE based method.

No abstract