This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
Basically what the kinect SDK does, creating a skeleton of a person is what you would be needing. In openCV however it is not that easy to define. You could perform person detection, but this would not give you exact location of the limbs.
You could define a detector for hands and feet, but don't think that is optimal also.
Look at optical flow? It is a way of displaying movement between frames. Maybe the motion can be captured as a sequence of typical optical flow movements?
Just some suggestions here!
