Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Hide Messages in Images with Steganography

mit.edu

Post reply on HN
Print view View original

11–18 of 18 posts

Re: Show HN: Hide Messages in Images with Steganography

#11

Post by davepeck »

DanBC wrote: post #6

I'm a bit drunk, so uh, yeah. LSB steganography is trivially easy to detect. Do not use LSB steganography. Steganography can be provably secure, with the correct amounts of plain text and "cover data". Here's one PDF. ( http://www.cs.cmu.edu/~biglou/PSS.pdf ) Some people need this kind of stuff to work to avoid torture or other serious consequences, but most implementations are toys to demonstrate proof of concept or…

I'm not a cryptographer.

That said: the PDF you link to seems to show that provably secure steganography (as defined by this paper) exists if, and only if, one-way functions exist. (See section 5.3)

It so happens that if one-way functions exist then, as a corollary, P != NP. In other words: if steganography is provably secure, P != NP.

So let's not get ahead of ourselves. ;-)

Re: Show HN: Hide Messages in Images with Steganography

#13

Post by atmz »

I wrote an essay/slides a while back which may be of interest to people wanting to know a bit more about basic steganographic image creation and detection: http://www.cl.cam.ac.uk/teaching/0910/R08/work/slides-at443-... http://www.cl.cam.ac.uk/teaching/0910/R08/work/essay-at443-s...

Quick improvement to the LSB method: Use JPEG and embed the data in the LSB of the DCT coefficients (still detectable, but not as trivially)

Cool demo though - I'd add a disclaimer though. It's easy to think that it's obvious that people shouldn't use this to store their password in, say, their Facebook profile photo, but..

Re: Show HN: Hide Messages in Images with Steganography

#15

Post by batgaijin »

davepeck wrote: post #11
DanBC wrote: post #6

I'm a bit drunk, so uh, yeah. LSB steganography is trivially easy to detect. Do not use LSB steganography. Steganography can be provably secure, with the correct amounts of plain text and "cover data". Here's one PDF. ( http://www.cs.cmu.edu/~biglou/PSS.pdf ) Some people need this kind of stuff to work to avoid torture or other serious consequences, but most implementations are toys to demonstrate proof of concept or…

I'm not a cryptographer. That said: the PDF you link to seems to show that provably secure steganography (as defined by this paper) exists if, and only if, one-way functions exist. (See section 5.3) It so happens that if one-way functions exist then, as a corollary, P != NP. In other words: if steganography is provably secure, P != NP. So let's not get ahead of ourselves. ;-)

Isn't a hashing function one-wayish enough?

Re: Show HN: Hide Messages in Images with Steganography

#16

Post by hk_kh »

liyanchang wrote: post #2

Actually a really cool way of storing passwords in plain sight. EDIT: Though if you were really secure, you shouldn't be typing your passwords into a third party website in plaintext. Alas, I'll have to write my own.

That's an interesting idea.

A big QR-code in your table is something obvious to scan, and aesthetically ugly. A framed picture not so much, and could serve a similar purpose.

Re: Show HN: Hide Messages in Images with Steganography

#18

Post by DanBC »

batgaijin wrote: post #15
davepeck wrote: post #11

Earlier quoted context omitted.

I'm not a cryptographer. That said: the PDF you link to seems to show that provably secure steganography (as defined by this paper) exists if, and only if, one-way functions exist. (See section 5.3) It so happens that if one-way functions exist then, as a corollary, P != NP. In other words: if steganography is provably secure, P != NP. So let's not get ahead of ourselves. ;-)

Isn't a hashing function one-wayish enough?

But is it provably mathematically one way?
Post reply on HN