Using Wavelets to Embed Images in Images
James Nolie Hughes · 2004
The wavelet transform takes a signal (or in general, any square integrable function) and transforms it into approximation and wavelet spaces. The inverse wavelet transform recombines the information in these separate spaces into the original signal. Of interest, we can consider an image as our signal, and then note that the information in the wavelet spaces is often much less significant than that in the approximation spaces. With this observation in mind, we investigated the possibility of hiding one image in another by copying a smaller image into the wavelet spaces of another decomposed image. When scaled appropriately, this data does not have a strong visual impact on the reconstructed image.