But the thing is, in that case the information contained in the images was actually much less than what we are meant to make believe.
So if we are reconstructing letters from a known font we essentially are extracting 8 bits of information from the image. I'm pretty certain that if you distort the image to an SNR equivalent of below 8 bits you will not be able to extract the information.
So if we are reconstructing letters from a known font we essentially are extracting 8 bits of information from the image. I'm pretty certain that if you distort the image to an SNR equivalent of below 8 bits you will not be able to extract the information.