Predict trajectory of hand based on 3D coordinates

asked 2019-07-20 08:01:41 -0500

Jérémy K gravatar image

updated 2019-07-20 08:04:03 -0500

Hi all,

I'm looking into a way to predict the trajectory of a hand that is grabbing towards an object in Python. I already have 3D coordinates of all joints of the hand (I'm using a Charuco board and stereo vision and thus everything is calibrated, beside that I'm using a pose estimation algorithm). I'm basically just looking into an efficient way to predict or estimate where the hand will be at frame 'x' in the future based on the 3D coordinates of the x number of previous frames. Which method would you recommend me for this? I'm sure there has to be an 'easy way' of doing this. I already took a look into Kalman filtering, Meanshift and CamShift, but I'm not sure if these are appropriate techniques.

Thanks a lot! Jérémy

edit retag flag offensive close merge delete