Creating point clouds or spalts from equirectangular videos

[Inshu32]

Hi,

I have a set of videos from which want to construct 3D models. Can your pipeline be used to do this if we have a lot of panaromas from the videos. Can the FreDSnet network be trained on custom dataset ?