@Dr_Muesli I know.
But it is unable to track rotations and fails with many motions.
Even at its best it still falls behind even the
#kinect, and we knew how that turned out.
If I do anything, it should be proper else I'm wasting time I feel
Furthermore camerabased solutions require very ecellent light conditions and especially much room I recon.
These requirements I cannot fullfil.
And also reflection and recording artifacts are a issue.
TL;DR
The resulting motionfiles would have too many artifacts, false positives and require
#cleanup lots. Of. Cleanup.
Camera solutions also lack one very crucial ingredient of proper tracking:
#DepthWith stations, imus and the like you have info of "where" in 3d space each element is positioned.
From a
#mathmatical standpoint it is a
#nocontestThe only alt using 3 cameras in a circle around thr object to record cresting a cylinder of sorts...
But being limited to a certain space yet again
That limitation is what I dislike the most tbh ๐
(And IK and FK feet n handbones fail hilariously with cameras too.
Its not even funny:
https://m.youtube.com/watch?v=I2ryXscndoo&t=0s)