So it's long been known that there's a lot of motion which we can't see, or can't quite make out, with our eyes. Computers can though, using some clever analysis of pixels and surrounding pixels, the same kind of way tracking works in HitFilm. A team at MIT have taken this to the logical next step though - they've applied the computer's vision back to the video, basically making the invisible, visible:http://www.youtube.com/watch?v=3rWycBEHn3s
This is insanely cool. If this could be ported over to Google Glass for example, another step towards our future badass cyborg selves is complete!