January 08, 2007

Decoding scrambled pixels

Removing data from a digital file is sometimes easier said than done. Redacting PDFs has sometimes proven tricky (something the Acrobat team has worked to address), and now a research report notes methods for unscrambling numbers or text that have been obscured via simple Photoshop tricks. [Via]

I asked a few Photoshop engieneers for comments & got some useful nuggets:

  • Gregg Wilensky says, "The ability to recover text that is blurred is limited by the amount of
    noise in the image (and knowledge of the blurring function). So, adding
    a bunch of noise to the image is better, but still not foolproof. I
    would suggest completely replacing the text with noise and blurring that
    (for aesthetics)."
  • Jerry Harris notes, "Having a known set of limited targets, OCR numbers in his example, makes the
    reconstruction task a bit more realistic in terms of useful results than
    most."
  • And Todor Georgiev writes, "Using a known set of blur kernels (those in Photoshop!),
    and a known set of targets, limits the set of possible outcomes
    and makes this technique work. But slightly change lighting
    and/or use custom blur filter, and your data is safe."
Posted by John Nack at 12:30 PM on January 08, 2007

Sorry, the comment form is closed at this time.

Copyright © 2019 Adobe Systems Incorporated. All rights reserved.
Terms of Use | Privacy Policy and Cookies (Updated)