Questions about "Different Representation Learning Frameworks" in the paper

[LEONHWH]

Hi, @ethnhe. I notice that you have effect of representation learning framework on the
two modalities of data.

But I still don't know how to implement 3DC-D?

1 Which kind of network represents 3DC-D, 3D U-net or PVCNN?

2 Another more exciting question is whether the point-voxel based method is feasible in FFB6D？