Improved WOW adaptive image steganography method

Abstract

Currently, the most secure adaptive image steganographic methods for spatial domain are to design cost functions to minimize embedding distortion. Wavelet Obtained Weights (WOW) is one of these methods, which can adaptively embed secret message into cover image according to textural complexity. In this paper, an improved WOW adaptive image steganography method is proposed. We apply a binary stochastic matrix to preprocess the cover image to generate a new image firstly, which helps to reduce the probability of locating embedding change positions by stego image. Then, three directional filters are used to weigh the embedding cost of each pixel in the new image and the syndrome-trellis codes are applied to minimize the distortion. Experimental results demonstrate that the proposed method achieves a better performance on resisting the state-of-the-art steganalysis over prior works.