I admit that I don’t know anything about the inner working of the skeleton tracking algorithm. Therefore, I apologize if my question is hopelessly naive.
I’m wondering if it could be (in principle) feasible to have Nuitrack perform skeleton tracking based on a point cloud / depth map that has been generated outside of Nuitrack. This would allow a user of the nuitrack SDK to perform some operations on the point cloud (such as thresholding) before skeleton tracking takes place.
I look forward to any feedback concerning this idea.
Thanks and best regards