The technology for this isn't that far out of reach.
Most video compression software does motion estimation, and I'm sure that's what this is based off of. Find the biggest glob of moving vectors over a certain threshold and shoot it. If you watch the video, it only tracks movement; it doesn't actually recognize a facemask.
Some of the hardware out there is capable of doing motion estimation in ... hardware. I think if you get that out of the way, the rest is less intensive. Until you get to recognition and range estimation...
Last edited by GoatBoy; 11-29-2013 at 01:36 PM.
"Accuracy by aiming."
Definitely not on the A-Team.