Using Dynamic Programming Strategy to Find an Optimal Solution to Exploiting Modification Direction Embedding Method

In this paper, we use dynamic programming strategy to find an optimal solution to the EMD (exploiting modification direction) embedding method proposed by Zhang et al. to carry (2n+1)-ary secret digits. An embedding/extraction function defined as a weighted sum function modulo (2n+1) is put forward in the EMD embedding method for secret digit embedding/extraction to/from an image. Among all the different permutations of these (2n+1) numbers, there exists an optimal solution corresponding to an embedding/extraction function, which can be used to minimize the distortion of stego-image. Without doubt, finding the best permutation will take a lot of computational time. Therefore, we shall use dynamic programming strategy to get an optimal solution.