0

Sometimes in documentaries, they have the face darkened out. Is their an effective way to undo this anoynmous effect? Obviously, the curves levels doesn't work as it just shows a bunch of square like patches of dark pixels and no detail as seen. Is their an algorithm that can guess as to how the darkening effect was done?

Below are examples from a documentary film ( It was the only one I could immediately think of as I couldn't find one on gangs / prisoners / murders out of the blue )

4
  • 1
    A black pixel will remain a black pixel. The information about the face is simply not there. You can change curves/levels/anything, if the information is not recored you can not retrieve it. Mar 6, 2013 at 19:54
  • 1
    You could just try saying, "computer, enhance!" ;) Mar 7, 2013 at 14:50
  • 1
    Sorry. I know that's not constructive. But seriously, as others have said, you can't retrieve what's not there. However, in the case that you had access to the RAW camera data (if it exists), information in the shadows would be retrievable. Mar 7, 2013 at 14:56
  • Why information do you mean the aspect ratio the lens or in opengl terms the front amd the back ortho or perspective Or the color data? If color data how does that store pixel color not in each frame of the avi wmv mpg or whatever format video version? Mar 8, 2013 at 2:13

1 Answer 1

1

If it is done properly, it should be impossible. If there is some amount of information there, then it might be possible to get a result by cranking up gamma or luminance, but done correctly, the information is set to black pixels for the entire area and there is no difference between pixels over that area to work from.

It is sometimes possible to reverse images that have been pixelated, blurred or distorted (or at least determine if it is the same person) but there is no way to make up information that isn't there.

2
  • Thanks and Thanks to both of you. This makes sense. I thought maybe the information what is visible ( the other part of the face ) could be used to "generate" an approximation using the darkened pixels of data, given that they aren't completely stored with complete / no saturation I.E. absolutely black. Mar 6, 2013 at 20:47
  • @ChrisOkyen - Just to be clear, if they are not uniform black, then there can possibly be ways to recover some of it, but if they are doing it right, they should be going to pure black.
    – AJ Henderson
    Mar 6, 2013 at 21:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.