Skip to Main Content
We propose a high capacity and imperceptible data hiding method in this paper. The hidden data can be easily to be embedded in a gray level cover image by simple least significant bits substitution, and to be extracting without using the original image. We divide the cover image into small 5 by 5 blocks with nonoverlapping boundaries, shrink them into four overlapping 3 by 3 blocks for each 5 by 5 block, and generate back to 5 by 5 blocks by interpolation technique. If the blocks own parts of edges of this cover image, they are the candidates of embedding positions. We check the middle of the four 3 by 3 overlapping blocks of each 5 by 5 block if the absolute value of the difference between the interpolation gray level and the original gray level is less than a threshold and is larger than zero, we may embed secret data in that position. We apply a pseudorandom number generator to the hidden data before embedding. Simply follow the embedding steps to extract the embedded data where the location is embedded then shuffle back the order to find the original hidden data. The experiment results demonstrate our method is quite simple and imperceptible.
Date of Conference: 18-20 June 2008