Simulating data from a monocular camera or a RGB-D stream(3D) camera in ignition gazebo

I am working on a project that uses OpenCV for face tracking then reconstructs the human face after estimating positions of the eyes, ears, nose, etc. The reconstructed face is stored in the form of point clouds. Please see the image below for getting a better sense of the purpose of the project.


My query is that how can I simulate this face in the ignition gazebo?

Additionally, I have written a couple of plugins(visual) on PCL and ignition gazebo but nothing seems to work so far.

Any advice is appreciated.