Information embedding codes matched to locally stationary Gaussian image models
Mehmet K. Mihcak, Pierre Moulin · Proceedings - International Conference on Image Processing · 2003
We have designed a practical information embedding code for images. The method assumes a locally i.i.d. (independent identically distributed) Gaussian source model for wavelet image coefficients. The proposed algorithm is developed based on this model and recent information theoretic results on information hiding for parallel Gaussian channels. Promising results have been obtained on test images.