Question about warp field usage

[ShichenLiu]

Hi @mihaibujanca ,

Thanks for sharing the excellent code! I know this is an old repo so you might not remember all the details. But I still have one little question and hope you could answer me!

My understanding about dynamic fusion is that, in each step, we compute a global transformation with ICP, and then perform a non-rigid residual deformation via nodes in the warp fields. Therefore, it only makes sense to me that the warping field (dual quaternion blending) is initialized with the global transformation before optimization. Indeed, I can see you tried to use warp_to_live for this. However, I couldn't find this member variable being set anywhere in the repo. Please bare me if I am wrong, I think it should be set before calling the warp function here.

It would be very helpful if you can take a loot at this! Thanks again.