# How can I retrieve data embedded in an image with steganography?

I want to know is there any way to retrieve data embedded in a steganalised image by just scanning the image with mobile scanner? Just like qr code scanning

Hmm. A QR code is designed to allow reliable embedded data extraction. They have significant (~30%) error correcting data to guarantee this. You might call it the antithesis of steganography.

However, there's a whole list of steganalysis techniques in A review of image steganalysis techniques for digital forensics (2018). I haven't gone through all of the techniques in this summary paper, but some might apply to scanned images. Don't forget that the scanner will introduce 'shmear' of the printed image and consequently any embedded data. You may even get a moiré pattern appearing that overwhelms that data. These are very common for certain scans like those of a halftone print. Greasy thumb prints might be a problem too, as well as all other cumulative imperfections arising from the physical manifestation of a virtual object and it's conversion back.

The following attacks are listed against both LSB and DCT embedding:-

• Stego only attack - only the stego object is available for analysis.
• Known cover attack - the cover and the stego object are both available for analysis.
• Known message attack - the message is known and can be compared with the stego object.
• Chosen stego attack - the stego object and the stego tool (algorithm) are available for analysis.
• Chosen message attack - the steganalyst generates stego-media from some steganography tool or algorithm from a known message. The goal in this attack is to determine corresponding patterns in the stego-media that may point to the use of specific steganography tools or algorithms.
• Known stego attack - the steganography tool (algorithm) is known and both the original and stego-object are available.

using these techniques:-

• visual steganalysis
• signature or specific steganalysis
• statistical steganalysis
• transform domain steganalysis
• universal or blind steganalysis

The best bit is that there are 119 references too if you want to wade in deeper. It's important to highlight what the steganography tag says though:-

Steganography is the art and science of writing hidden messages in such a way that no one, apart from the sender and intended recipient, suspects the existence of the message. While cryptography protects the contents of a communication, steganography aims to prevent others from learning that a communication even occurred.

So yes, 'data' may be retrieved. Whether the semantic message can is a matter for another question. But consider; an attacker just knowing that a message exists means that the attempt at hiding has failed and the steganography is broken.

In general, you can't.

The reason is this: the data might be encrypted on a way what makes it indistinguishable from random noise. Actually, any good quality encryption does it so.

Thus, if you try to separate any possibly steganographically hidden data, you have no way to know, if it is really random noise, or it is encrypted, but worthy content.

It might be possible if the hidden data is not encrypted on some reason (for example, for better data safety).