Information Hiding with Optimal Detector for Highly Correlated Signals

In this paper, a novel scaling based information hiding approach robust against noise and gain attack is presented. The host signal is assumed to be stationary Gaussian modeled with a first-order autoregressive process. For data embedding, the host signal is divided into two parts. One part is manipulated while the other part is kept unchanged for parameter estimation. The decoding scheme using the ratio of samples is suitable for highly correlated signals in which the decoding process is difficult. By calculating the distribution of the ratio, the performance of the maximum likelihood decoder is analytically studied. The proposed algorithm is applied to several artificial Gaussian autoregressive signals to verify the validity of our results.