Unique Identification of a File

[nickl]

Hi All,

I'm pretty sure this is the right place to put this:

Take a situation in which I want to distribute image files to a number of people, and I have rules by which they may use the files. However, I find that someone (whose identity I don't know) has broken my rules, and I would like to identify them. How would I go about determining their identity. I would assume that this would require the addition of some unique identifier at the time that I give them the data. Optimally, the embedding of this data would be fast to perform and would not damage the image, and would remain intact if the image was cropped. I assume that if someone manipulated the image further, such as by changing hue or saturation or whatever, the data would be lost. Of course, it would need to be robust enough to ensure you can't do something simple to get rid of it.

 

I don't want code specifically, mainly just stuff to read up about. Would it be some kind of stegonography such as LSB or jsteg?

-Nick

[bgibbz]

are you looking for an automatic or manual way of doing this? You can simply put a serial number in the metatag that is attached to the image. There are many free meta editiors online. If you want to do this automatically, you would need to research how to edit meta data through your programming language, then have the program generate a serial number, log it to a table, then return the image with the modified tag.

[Andy Cheese]

Hi

I don’t believe there is a way to do this easily as every way I can think of can be bypassed unless you go down the advanced steganography route. Metadata is the easiest to remove with a simple right click, Embedded data can be distorted by compressing or as you said applying a colour filler of some kind.

 Even watermarks can be removed in a short amount of time. I did this recently after some friends went mountain biking and the person at the event wanted to charge them about 3 time the normal amount for photos (Note, I do not condone this and would NOT normally did it however if you going to be greedy then what do you expected!) It took me less them 2 hours to do 6 photos in photoshop by simply cloning pixels.

Some reading points ideas could be:

Disappearing Cryptography: Information Hiding: Steganography and Watermarking (Book covering DRM)

Digital Rights Management: Concepts, Methodologies, Tools, and Applications (Book also covering DRM but with sample chapters online at https://books.google.co.uk/books?isbn=1466621370)

https://www.controlledvocabulary.com/blog/embedded-metadata-wont-help-seo.html (webpage covering Metadata)

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4119663/ (webpage covering Digital Watermarking and Fingerprinting Techniques, also checkout their ‘similar articles’ on the right-hand side of the page)

http://scisweb.ulster.ac.uk/~kevin/spvol90.pdf (26-page pdf covering steganography)

Something to think about is how much is the value of the images against the amount of time you want to invest on a solution and also the cost of its implantation.   

All the best
Andy

[Mira Yurizaki]

You can kind of do this with cryptography.

 

I can't really think of how the system works exactly, but this is how people who receive/send messages know it's more or less legitimate

[nickl]

On 12/11/2016 at 9:34 PM, bgibbz said:

snip

 

On 12/12/2016 at 11:46 AM, Andy Cheese said:

snip

Thank you both for your responses. I'll look into these solutions.