Hi,
Microsoft's CVPR paper of the kinect is very fancy. I am trying to produce a similar dataset.
I have read
http://www.mocapclub.com/Pages/Tutorials.htm and viewed
http://lesterbanks.com/2011/03/using-3ds-max-with-motionbuilder/I am not sure how I would accomplish the following:
1) Create a camera to record the animation.
2) During each frame, record the image the camera sees, the depth image of that scene, and the <x,y,z> marker positions.
Thank you for your time and help,
Phi