Steganography in discrete wavelet transform based on human visual system and cover model

In this paper we present a model-based image steganography method in discrete wavelet transform (DWT). This method is based on the human visual system model. The proposed steganography method assumes a model for cover image statistics. In this algorithm, the DWT coefficients are used as the carrier of the hidden message. An unpleasant outcome of this algorithm is that its perceptual characteristic is degraded. The perceptual detectability weakness of this approach is improved by introducing another algorithm which is proposed based on the Watson visual system model to prevent visually perceptible changes during embedding. In the first step, the maximum tolerable change in each DWT coefficient is extracted using the human visual model. Then, a model is fitted to the histogram of low-precision coefficients and the message bits are encoded to this model. In the final step, the encrypted message bits are embedded in the coefficients whose maximum possible changes are visually imperceptible. Experimental results illustrate that changes occurred during data embedding by employing the human visual model leads to perceptually undetectable changes. The perceptual detectability is satisfied while the perceptual quality and the security usually increased. The perceptual quality is measured by structural similarity measure, and the security is measured by two well-known steganalysis methods.