I thought it was fairly readable, you can at least get the gist of how they use the database of gestures to estimate the pose of the hand on a per-frame basis.
Thanks for that. I was curious how they are getting depth information with only a single camera; it looks like they are using knowledge of the size of the users hand to infer depth.
Incredible. The problem is, they look like bowling shoes for your hands.
I'm betting you can fix this by putting some paint on there that reflects infrared or ultraviolet light at different frequencies. Might need to upgrade that webcam though.
That's a really excellent point. Displaying other contact indicators (surface deformation?) when virtual objects come into contact with other (real or virtual) objects would probably also be beneficial.
