Detecting mosaic censorship in images

Question

0.00/5 (No votes)

See more:

I'm trying to figure out what would be a good approach to detecting mosaic censorship in images. You know, where you see squares of a solid color to censor a part of a photograph or something, like the face of a criminal in a newspaper.

Now, I don't want to remove the censorship, I only want to detect if it's there. But it also has to be somewhat reliable. The last thing I want is false positives. I'd rather have false negatives (i.o.w. censorship in an image which isn't detected is better than censorship being detected when there is none).

I was thinking of somehow detecting when there are small squares with evenly filled color, but that would also return positive for large (non-censored) areas filled with a color, such as a part of the sky. I guess it comes down to some sort of pattern detection, but I have no idea how to approach this...

It's ok if at first the censorship is "perfect", i.e. no JPEG-artefacts or other malformings. But I do want to compensate for that, ultimately.

Any help will be eternally appreciated :)

Posted 22-Jul-10 11:38am

thany.nl

Add a Solution

2 solutions

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Christian Graus · Answer 1 · 2010-07-22T11:58:00

Solution 1

If you detected that there's a square, a perfect square, with colors within a range that compensates for JPEG compression, then that should work pretty well. It's pretty hard to have 100% success with NO false positives, but I reckon you could get close.

Posted 22-Jul-10 11:58am

Christian Graus

thany.nl · Answer 2 · 2010-07-22T13:20:00

That would be good for detecting *where* a mosaic is, when I already know it's somewhere in a given picture. If I feed my algorithm a totally never before seen image, it should basically say "yes" or "no". So detecting a square would generate a huge number of false positives for images that have larger areas of the same color. After all, even a circle (with solid color fill) could easily contain many such squares.

The trick is to detect if a square is part of a mosaic pattern, not if there's a square at all. One square a mosaic pattern doesn't make ;)

Even detecting multiple ones and requiring them to be of differing colors and requiring them to neighbour eigh other, wouldn't be a nice way to go. It'd be extremely slow given the megapixel-pictures these days, even apart from the fact that mosiac "tiles" could be of any size that's 2x2 or more. Plus it might generate a false positive for something as simple as a straight line that happens to be >=2px wide...