About 3D labels about vnect-tensorflow HOT 4 OPEN

same question！

from vnect-tensorflow.

@guker The question be solved. 2D heatmap represent x,y image coordinate. 3D heatmap use x,y,x directly. and than 3D position use 2D message.

from vnect-tensorflow.

@guker The question be solved. 2D heatmap represent x,y image coordinate. 3D heatmap use x,y,x directly. and than 3D position use 2D message.

for example , location map Xj of the joint J, what is the values of this map?
i think a location map have same values( root relative coordination), is right?

from vnect-tensorflow.

@guker The question be solved. 2D heatmap represent x,y image coordinate. 3D heatmap use x,y,x directly. and than 3D position use 2D message.

for example, I generate Location map like this:

loc_map[:, :, :, j, 0] = x_j
loc_map[:, :, :, j, 1] = y_j
loc_map[:, :, :, j, 2] = z_j

where loc_map \in R^{Batch \times H \times W \times J \times 3}, is it right?

from vnect-tensorflow.