If you’ve been using pixelation to redact sensitive information, you may want to use a different method instead as a security researcher has developed a new way to successfully recover pixelated words from redacted documents.
Lead researcher at the security firm Bishop Fox, Dan Petro has published a new blog post detailing how he was able to completely recover text from an image that was redacted suing the pixelation method.
While both news outlets and researchers often using pixelation or blurring to hide text in sensitive images published online, Petro has demonstrated that it is much safer to use opaque bars to hide text instead.
Petro began his research into how retrieve pixelated words from redacted documents after Jumpsec Labs put forth an open challenge for anyone to decipher the text in a pixelated image. After studying a number of pixelation and deobfuscation techniques, he came up with a solution for the challenge and sent his findings to Jumpsec Labs.
Unredacter
While there are a number of existing photo editing tools for enhancing pixelated images of people or landscapes, up until now there hasn’t been a tool capable of recovering text found in pixelated images.
For this reason, Petro and Bishop Fox have released a new open source tool on GitHub called Unredacter. The tool is able to correctly reconstruct a document’s original text in its entirety if it was redacted using pixelation.
Petro provided further insights on how users that need to redact text in a document should do so in his blog post, saying:
“The bottom line is that when you need to redact text, use black bars covering the whole text. Never use anything else. No pixelization, no blurring, no fuzzing, no swirling. Oh, and be sure to actually edit the text as an image. Don’t make the mistake of changing your Word document so that it has black background with black text.”
Thankfully, many of the best PDF editors such as Adobe Acrobat DC already provide the ability to redact text in sensitive business documents.
Via BleepingComputer