> if you know the statistics of the communication channel, you can build secure steno
That's a pretty big if though. I mean, if you miss one nuance the NSA didn't, then your scheme could be toast. At least in the high data limit.
That's a pretty big if though. I mean, if you miss one nuance the NSA didn't, then your scheme could be toast. At least in the high data limit.
Also, most of the research in the field has been done on black and white image, so it not directly usefull.
And come to think on it, if we're talking about image stego, what prevents the NSA from controlling the statistical distributions of output of cameras in such a way that stego will be noticed(even after some image transforms) ?
It would be easy enough to make a one-click batch process, and would help provide cover noise for real communication channels.