Camera as Robot Controller
Fellow scholars,
I thought this video by Two Minute Papers was really interesting but I'm trying to figure out a real world application for it.
Of course it doesn't need a real world application for the research to be useful. The research is cool whether or not it's immediately useful or not.
I'm having a hard time thinking of a time when more conventional input device wouldn't be better than this camera AI combo.
Maybe there's a scenario where you'll have a robot operator who can't be trained to use a conventional input device. The camera method of control does appear to be intuitive. Maybe a robot which will have multiple short term controllers?
Anyway, I thought it was interesting and I'm curious if any of you can come up with an appropriate application for this tech.
Comments
Very cool, Boss! This is way beyond the figure 8 challenge!
What was that gizmo ~8 years ago, a black USB plinth on the table looking up that very accurately read hand gestures and finger motions? Haven't heard much from that "game changer" lately.
Edit: Ah yes, the Leap Motion.
Leap motion was alright, but it had a closed marketplace for the apps. Dooming it to obscurity.
The recommended direction the sensor was looking is up at the ceiling. That made it hard to see individual fingers unless you lifted up your elbow and rotated your hand so your palm was parallel with the desk. Which got tiring.