So much work goes into adding virtual objects into real space with Augmented Reality, but what about removing unwanted objects? This impressive research work from folks at the Ilmenau University of Technology shows editing of video in real-time to remove unwanted objects using something similar to CS5’s Content Aware Fill.

The most recent research project in the area of Diminished Reality by Jan Herling and Wolfgang Broll, Ilmenau University of Technology, Department of Virtual Worlds / Digital Games, 2010.

It works by degrading the image into a low-resolution luminance version, and removing the object there (much simpler).  It then resamples the image back to near-original quality, and uses it as a mask/composite with the original.  And it does it all in 41ms, although the hardware used is unspecified. (CPU? GPU? FPGA? Who knows).

There are a few artifacts in some of the scenes, but the effect is amazingly good.  I am curious to see what would happen if you turn the camera completely away from the object.  Would it mistakenly remove something else?

It really does remind me of some of the hacks in Ghost In The Shell tho, hacking people’s vision to make people (thieves) invisible.