Visual crypto and SIRDS, was Re: Call for stego ideas
From: Arthur J. O'Dwyer (ajo_at_nospam.andrew.cmu.edu)
Date: 07/15/04
- Next message: Tom St Denis: "Re: Surrogate factoring, update"
- Previous message: Lance Lamboy: "Re: Surrogate factoring, update"
- In reply to: John Bailey: "Re: Call for stego ideas"
- Next in thread: James Vanns: "Re: Call for stego ideas"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Wed, 14 Jul 2004 20:53:39 -0400 (EDT)
On Tue, 13 Jul 2004, John Bailey wrote:
>
> On Mon, 12 Jul 2004 17:22:39 -0400 (EDT), Arthur J. O'Dwyer wrote:
> > Yes, this would definitely fall on my list of "amazingly nifty
> >results in recreational crypto" if it were really possible to eliminate
> >the artifacts in the first image.
> >
> > Unfortunately, I myself know next to nothing about random dot
> >stereograms. Definitely post if you figure out how to remove the
> >artifacts, please!
>
> I could not resist such a pleasant challenge, check out:
> http://home.rochester.rr.com/jbxroads/interests/rec.puzzles/images/
> dtimanyd.gif is an ordinary stereogram
> tnt_key.gif is a (somewhat crude) stereogram with the message:
> *Explosive* displayed
> Superimposing the two images additively, reveals a third message.
[...]
> The technique to avoid artifacts is to use only a single vertical
> stripe of the stereogram as the random field. This stripe is
> modulated in polarity to produce the corresponding vertical stripe in
> the second stereogram. Superimposing the two stripes, whether as parts
> of full sterograms or in isolation results in the visual crypto
> message. Using standard stereogram technique, the phase locations of
> bits are modulated horizontally across the full stereogram field to
> encode the less hidden messages as variations in depth.
This afternoon I took some time to Google up some SIRDS resources
(that's Single-Image Random Dot Stereogram, for those not paying
attention ;) and now I think I have a reasonable grasp of the concept.
Which makes it somewhat less amazingly nifty, of course. :)
I've posted a couple of images of my own here:
http://www.contrib.andrew.cmu.edu/~ajo/images/sirds_donut.png
http://www.contrib.andrew.cmu.edu/~ajo/images/sirds_steps.png
and the program that made them (at least temporarily) lives here:
http://www.contrib.andrew.cmu.edu/~ajo/disseminate/sirds_hider.c
They're unfortunately a lot harder to "see" than your images,
because they have exactly balanced black-and-white pixels ---
no big irregularities for the eye to latch onto. I can make
them out with enough attempts, but then I know what to look
for. :(
Anyway, overlaying those two images produces a snippet of
a black-and-white image (of a cat, not that it's important).
The image gets repeated, with more and more distortion, across
the whole image. I don't have your images in front of me right
now, but ISTR that somehow your overlaid image didn't repeat.
How did you go about it?
-Arthur
- Next message: Tom St Denis: "Re: Surrogate factoring, update"
- Previous message: Lance Lamboy: "Re: Surrogate factoring, update"
- In reply to: John Bailey: "Re: Call for stego ideas"
- Next in thread: James Vanns: "Re: Call for stego ideas"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
|
