Answer the question
In order to leave comments, you need to log in
Steganography in pictures?
Actually the question is how to determine this and preferably exactly?
Answer the question
In order to leave comments, you need to log in
If steganography can be detected, then it is not steganography.
In other words: no way.
One of the simplest methods of steganography (but by no means the only one) is to change the least significant bit of each pixel in a row.
In fact, the professor was talking about extracting an encrypted message from an image. But the essence of steganography is that the very fact that it contains an encrypted message is not known, and even more so it is not known exactly how it was sewn there and in what format this message itself is.
Of course, if you extract the byte stream bitwise like this and it makes sense, for example, the string "Hello world!" - then this is hardly an accident, most likely it was actually hidden. (Bingo! You discovered steganography and extracted it). But this is possible only if:
1. You guessed the correct extraction algorithm, the same one that was used to sew information (from an infinite number of possible ones)
2. You can easily determine that the result is meaningful, and not just garbage.
But if used correctly, the sewn information will be encrypted, so it will be impossible to distinguish it from garbage. This can be checked with a simple gzip. If the stream is well compressed, there is a strong redundancy in it, it is some kind of text, structure, but not a stream of noise.
It's like decrypting a very securely encrypted file, because a fool encrypted it and didn't do everything right (for example, using a password). In a particular case it is possible, in a general case it is impossible.
You can use a program to extract the file from the picture. If it was possible to extract some file in a known format such as zip, then this is a very reliable proof that there is a shorthand.
That is, the answer is: you need to find the program by which the file was placed in the picture, or at least pick up an algorithm to write your own program and thereby check it.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question