Score:1

Crypt-analysis for finding information hidden in images?

cn flag

I was wondering if anyone is familiar with any historical aspects (as to whether someone was able to) for discovering code using images on the web as a transport method ? As in hiding byte values in pixel data broken up between the component values ?

** Edit** Answer below adds to a good search tree of how hide text but doesn't relate to discovering the ciphertext.

SAI Peregrinus avatar
si flag
The phrase you want to search for is "steganography".
Paul Uszak avatar
cn flag
I think that this is one for @patriot – Paul Uszak
fgrieu avatar
ng flag
Are you after _historical_ facts (say, 20th century and before (there is a tag [tag:history] for that)? More recent _anecdotal_ facts ? Theoretical aspects ? Also: steganography often does not follow Kerchoff's principle, which makes it hard to discuss it's modern theoretical aspects. And there's enormous difference between practical steganography, and theory.
Score:0
cn flag

Steganography means hiding secret data/msgs inside the bits of the image matrix (as it would be harder to discover they r used as a transferring media in the first place)

The most naive approach would be to just store in the grades of a color accuracy like Black in a high resolution image for example; such naive approaches can be easily discovered by any resize/rotate of the image.

More complicated approaches started to appear in 2000-2005 maybe, I think this one is also about steganography although not in the title "Hiding Secret Messages in Huffman Trees" https://www.researchgate.net/publication/261021203_Hiding_Secret_Messages_in_Huffman_Trees

I even recall articles after 9/11 suspecting that Osama Ben Laden was sending encrypted msgs to his allies thru his broadcasted videos using steganography ( I remember an article entitled "Could Ben Laden be a Steganography Master"? https://www.wired.com/2001/02/bin-laden-steganography-master/ )

Anyways, I don't think the topic flourished as a research area afterwards. I just Googled it now, there are books & explaining dated to 2011, but almost nothing new except 2 papers One in 2019 "An efficient steganographic technique for hiding data" https://joems.springeropen.com/articles/10.1186/s42787-019-0061-6 One in 2015 "Steganography for inserting message on digital image using least significant bit and AES cryptographic algorithm" https://ieeexplore.ieee.org/document/7577468 (you can find more references inside those papers)

John Sohn avatar
cn flag
I like this subject better than studying cnn's which loses me !
John Sohn avatar
cn flag
As a POC I divided ciphertext up among the color components lowest sig 3 bits and spaced them out. At 3 bit changes, it is barely visible as different. Thanks for the quick reply !
ShAr avatar
cn flag
What CNN & media study has to do with Cryptography?! & I think I mentioned in my reply that these simplest 1st thought approaches become revealed with any change in scale or rotation or....in the image
John Sohn avatar
cn flag
yes I realize that that was more simplistic than what you were likely talking about since anything including huffman is general complex :P ish. however its what people are looking for which forms another part. sure if you post an image from a suspected account someone might inspect that, but the question is how would you find it in an automated fashion ? I don't think a bit difference makes much of a difference visually btw. you couldn't hide MUCH info in a single image, but you could hide some.
John Sohn avatar
cn flag
Delivery has something to do with it for sure. Enciphering would occur prior to tagging the image.
John Sohn avatar
cn flag
oh and i meant studying the theory behind CNN's (convolutional neural networks) has been a confusion reattempt since i'm not quite that focused of late.
John Sohn avatar
cn flag
This however doesn't actually answer the question as to how to discover that ciphertext is hidden. But it does give an example of how.
ShAr avatar
cn flag
The image is stored as a bit matrix of colors(each entry of the matrix represents the grade of color of the corresponding pixel, the msg is hidden in there). Now if someone scales the msg, rotate, change resolution? the inserted msg will be scattered and will affect the pixels of the image & thus be revealed. Check the book in the answer for examples.
John Sohn avatar
cn flag
using some statistical analysis correct ? see though the issue I see is how does a computer program differentiate between aberration, damage, or editing by a human ? A human eye can see what I'm talking about by looking closely but on different colored images with lots of noise, its very difficult to see at 1 bit data per 3-color tuple.
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.