Currently view coordinates relative to the volume are required so you first have to solve the SLAM problem before you can optimize a network representation of a given volume.