segmentation-driven-pose
segmentation-driven-pose copied to clipboard
about generate mask
hi, sorry, I have a problem about mask generate. I follow the following procedure to generate a mask:
- load the object vertexs from LINEMOD_vertex.npy(you provide)
- correct the vertexs using Transform_RT_to_OccLINEMOD_meshes.npy(like R*vertexs+t)
- then projection those 3d vertexs to 2d plane using pose Rt and linemod_k
- extract the projection area as mask
I do like the above, but the result looks strange:
this, I want that cat's mask can you help me, what problem?
@YinlinHu