Texture Challenge #2

Ok, here's the next test. I have supplied both the original image and one with 2 separate images embedded in it. The challenge is to find both hidden images. $L2000 goes to the first person to post the correct results for both images. There's no prize for just posting one, it's too easy. Don't post until you have both!

Usually, when this is done, no one gets the original image to comapre with the other, so all in all this is an easy task. I'm still interested in how quickly people find the images.

No hints this time! Refer to the previous thread if you need to.

I'm going off to bed, so I will look for replies tomorrow.

This was a lot of fun! Thanks Namssor

(Although surely this puzzle would be impossible without the original image?)

Congratulations Seifert!

The possibility of solving the puzzle without the original image is probably about as hard as factoring one of these numbers , but not impossible. There might also be an easy shortcut, but I don't know.

Well the first thing I did was take the difference between the original and the encoded image. Without both images there is surely no way to get at this information? It's like a one-time pad, which is provably impossible to crack.

You could be right. It would certainly be beyond my capability to reverse engineer an image encrypted this way without the original to compare it with.

Do you think anyone could solve a puzzle with just the carrier file and a copy of what the decrypted image looked like?

Isn't the decrypted image already the solution to the puzzle?

Yes, but what is the process that arrives at the decrypted image?

I suspect this is mostly impossible - the data isn't there. The changes made to the eye picture are so slight that they could just be natural variation in the photograph. Perhaps if the image had previously been encoded in some format that left recognisable artefacts, but the adding of the message distorted those artefacts then you could work something out.

If the encoding resulted in larger changes to the original photograph, to the extent that the encoded result doesn't look normal any more, then the way in which it doesn't look normal might give enough hints as to how to decode. In a sense this is Kurshie's puzzle, as the grid and the circles were visibly part of the encoding.