Local_orientation_per_joint / Local_position_per_joint


I got the same issue as Håvard here (local_orientation_per_joint always equal or close to [0,0,0,1] · Issue #211 · stereolabs/zed-python-api · GitHub) that I am trying to track bodies and call their local_orientation_per_joint for each but all of the points are stuck at (0, 0, 0, 1). for most of the time and if they show other values they are barely off. Has anyone else experienced this problem so far or even got a solution for that?


Not all the keypoints have orientation data available. See the image below. That will change in the future.


1 Like

Hi Antoine,

thanks, that clarified everything a bit and those joints indeed give me values. The only exception is the hip/0, which stays at 0, 0, 0 all the time. Is there a reason for that?

I’m also not sure if I understand kp_3d and local_position_per_joint values and their difference correctly. What point is the child and what is the parent with local_position_per_joint values? I’m trying to get the sceleton data into a mediaserver but am not sure how or if I have to compute the keypoint and the local joint position together/with each other.

Hi again,

kp_3d are absolution position in the world. local_position_per_joint is the position of the child joint compared to its parent.
We may have an issue about the hip, it’s the spine (1) takes has rotation instead.

Thank you for reporting it !

1 Like