----- Original Message ----- From: Matthew Brown morven@gmail.com
We recognise pictures all the time for Wikipedia content.
As a concrete example, whenever I see a cast photo in an article I make sure the caption includes a "from right to left: John Smith, Jane Doe, etc" description of who is in the image. I do this via a process of recognizing the faces of the people in the image and translating that into textual names. If this is "original research" I'll eat my hat (assuming I recognize it).