Steganography

anyone interested in? Currently i know all except wat to do key exchange. I find something but i'm just too dumb to gotcha. I send 3 IMO interest PDFs about it, pls ELI5. How does it work in practice and can it be implemented somehow? I only found one project that was using public key but without solving the problem of replace the key https://github.com/hfeeki/elliptic-curve-steganography > Bob creates a public key image and shares the image with Alice via some predetermined method. > via some predetermined method. how to hell i I'm to inform about it Alice?

Steganography is great for hiding that you sent a key in the first place. What else do you want?

>>14276 stego looks like a asymmetrical crypto - it is in its nature: stegotext - message cover - Public key embedding - Encryption extracting - Decryption steganokey - Private key one of the flaws of course is that anyone familiar with the embedding method can read the message. At least in classic stego. Thanks to the neural network, this problem can be avoided, by adding one ingredient: salt. Without knowing which input did you use when generating the effect with AI, it is impossible to detect the message. and finally I come to the key point. Everything's fine except this - it does not solve the problem the key exchange. In classic cryptography, you use DDH and you know that you want to achieve such communication in both sides. but how to achieve this in stego? one real life example which i know, it.. conjecture https://www.youtube.com/watch?v=7CqOYM7cCX8 you must pray that the good guys will notice the message and that the bad guy will not notice. It's very risky.

>>14296 the fuarrrk does AI have to do with using a salted encryption algorithm?

>>14296 You are autistic. Just do a public key exchange with images and you're good. But why man? Only thing this does is: - hide key was exchanged in the first place - prevent easy interception

>>14300 in fact, I already have a complete knowledge of how to build a complete system except for this thing: >Just do a public key exchange with images and you're good how do I inform the other person that I have left a message for them? I guess you can't, without secure channel eg. IRL, which makes all the operation is no point, like pic rel (Ward, in pic vidya, it can take over the key because it is symmetrical.) DDH it works with machine to machine but not when you don't have computers. I think that for this to occur, the entire procedure would have to be subject to it. I just don't know how to do it. It would have to be a one-way function that could be performed with basic tools like paper and pencil, something like that. It does not have to offer high security, i mean only the possibility of asymmetric key exchange. A unrealistic, stupid example, nonogram - you send rows, person solve and get key in the picture (it's almost like writing "play in winamp" in image where the song is hidden so that's why is stupid) >>14297 SteganoGAN it's protip

>>14464 >SteganoGAN thanks Kalyan

I recently found this about F5 https://github.com/matthewgao/F5-steganography I'm looking forward to understand better steganography as I don't understand well the lsb tricks behind it.

https://matrix.to/#/#steganography:matrix.org?via=matrix.org&via=privacytools.io&via=nordgedanken.dev matrix channel

Noob when it comes to steg. This subtopic in regards to steg seems to be lacking in information anywhere How feasible is it to reverse engineer or mitigate forensic watermarking used as content protection for live streams like sports or movie premiers? soykaf like: https://dtv.nagra.com/nexguard-forensic-watermarking https://www.ibc.org/thought-leadership/using-forensic-watermarking-to-protect-uhd-content/946.article https://www.red5pro.com/blog/why-it-makes-sense-to-watermark-live-streaming-video-in-real-time/ The forensic watermarking is generally deployed in such a way that it is unique to every client's stream, which in turn allows them to identify the pirate streams client via watermarking and cut them off mid stream. The companies that implement this sort of watermarking claim its not possible to reverse engineer and that its not possible to mitigate via reencoding, cropping the frame, altering color palette, or even recording with a video camera. If you reverse and identify the general method of how they watermark their content, would it not be possible to just repeat their method of watermarking at random to throw off their ability to correctly identify the stream? Was thinking something like: Official Stream -----> OBS / other video capture method -----> video capture is piped through 'obfuscation' layer with additional random watermarks thrown in at random intervals -----> streamed out to where ever you wish

first became aware of it when a professor who worked with CIA did a whole lecture on it. He said the key images are usually pornography, which I thought was pretty funny

rather than think about sneaking a message "in" to someone. think about sneaking a message out. say you were spying and trying to transmit information outbound stego is a good method to smuggle the information out if the receiver knows "i'm going to receive a picture and decode it with this key"

>>>16492 >The companies that implement this sort of watermarking claim its not possible to reverse engineer and that its not possible to mitigate of course they do... just like its "impossible" to break any other form of DRM right? The reality is you just need to figure out what they are doing to "watermark" it in the first place and their smug "non of these methods will work" is a good hint. You're on the right path. I assume its some form of metadata since it hides the data without effecting the content of the stream itself. >latency-free mark injection and identification of pirate sources within 5 to 15 seconds This tells me that there is a high-low value that makes up the decoding of the watermark which makes me think its some kind of UDP padding of some kind that would normally be discarded and ignored.

>>17856 >I assume its some form of metadata since it hides the data without effecting the content of the stream itself. blatant ignorance of how steganography works consider that it doesn't matter if the raw bitstream is changed, just if the end product (pixels on a screen) is indistinguishable from the original. Video files have a TON of entropy sitting around and properly designed stego systems use proper cryptography to make it hard to distinguish the watermark from random noise. OTOH... run it through an ffmpeg re-encoding and there is a high chance the watermark doesn't survive.

>>17857 also: most porn sites have discovered that just moving some text across a picture at random speed is pretty hard to remove algorithmically.

>>17855 Good point. You remind me about Solitaire: > Solitaire is only safe as long as its key is safe. [...] Here are some suggestions for ways to exchange keys >1. Shuffle the deck of cards. >2. use bridge notation >3. Organize your deck based on your password And IMO 2 method is probably best you can do. A full description of it can be found in the cryptonomicon. AFAIK looks like book cipher.

What's the point in 2022? Most network traffic is encrypted, so you have plausible deniability for using encryption. Once attackers can't distinguish regular content from suspicious content anymore, there is no need to put your information into an image or just use plain text. It's both not decryptable for the attacker.

>>19350 Steganography is a plausible deniability mechanism. It is not a replacement for encryption. Some observers (NSA) only store encrypted traffic (or at least claim to). Given sufficient time, they will break this encryption. By steganographically-embedding a ciphertext, you can convince said entities not to store it.

>>19353 >Given sufficient time, they will break this encryption. If it's AES encrypted, this is extremely unlikely.

>>19355 This, post quantum cryptography is a thing already even though quantum computers will probably take at least a decade to become somewhat useable