Open vijayvuyyuru opened 3 years ago
In case it's helpful, a good starting point might be bayesian methods like ekf / ukf trackers, ie: https://github.com/LCAS/bayestracking
You'll have to solve for streaming point data from each kinect and tranforming to a single global space first though.
Is there any code publicly available that shows how to merge multiple body tracks of the same body from different angles into 1 unified track? I have the calibration data necessary to get the transforms between multiple cameras. I am aware of using joint confidence to do this but I was hoping there was already code for this application.