Tuesday, January 31, 2006

Porn recognition software?

The USPTO continues to amaze me. I mean, one would hope that any silly patent that can be awarded has already been granted. How could it get any worse? Regard US Patent 6,751,348 for protecting an "Automated detection of pornographic images ". The abstract states that the patent covers:

A method of detecting pornographic images, wherein a color reference database is prepared in LAB color space defining a plurality of colors representing relevant portions of a human body. A questionable image is selected, and sampled pixels are compared with the color reference database. Areas having a matching pixel are subjected to a texture analysis to determine if the pixel is an isolated color or if other comparable pixels surround it; a condition indicating possible skin. If an area of possible skin is found, the questionable image is classified as objectionable. A further embodiment includes preparation of a questionable image reference shape database defining objectionable shapes. An image with a detected area of possible skin is compared with the shape database, and depending on the results of the shape analysis, a predefined percentage of the images are classified for manual review.
I love that "If an area of possible skin is found, the questionable image is classified as objectionable". Good-bye to the well-loved "I know it when I see it" Casablanca test.

3 comments:

Anonymous said...

This might be a case of a good software patent.

Like patents on spamming methods, if the patent is on something you object to, all the patent's freedom restricting effects are working against that object of objection.

If you don't believe that machines should decide what is objectionable (I don't believe so either) you should rejoice that this particular field gets mined by patents.

Andres Guadamuz said...

I have several probelms with it. First is prior art. I have talked with a colleague who says that this has been discussed in AI fields for a while. The other problem that I have with it is that there seems to be a presumption that images are to be specific shades. And then there is the fact that computers find content objectionable.

But I agree that it would be best if that field has too many patents/

Anonymous said...

Prior art - I worked with designers in a city council about four years ago and they had troubles with images of children for their brochures, if there was X% of skin tone in the photo, then it would be rejected. Couldn't say what the software was, but I'd be surprised if it wasn't fairly common knowledge in software circles.