Algorithm that applies 3D camera pose transformation to a frame using meshes/pointclouds.
Each pixel + corresponding depth is being considered as a point in space. The pointcloud or mesh, composed from composition of neighboring points, is rendered to obtain the final image. Returned inpaint mask is rendered in the same fashion. Type of rendering for image and mask can differ.