We need to render 3D model in python at running-time. For example in instance-level 3D pose tracking and result demonstration in 3D pose detection.
SIXD toolkit has implemented render scripts using glumpy. In the toolkit the model used is ply and both depth ang rgb image are rendered.
However, I only want to get the mask of the object or the depth image. So I only render depth image. Only vertex and faces information are need from 3D model files to render depth img. So I write a simple script to load vertex and faces from .obj files, The depth render part is modified from SIXD toolkit.
And the test code used RBot dataset.