0 / 60 seg.

Much to our surprise, we only required 50 -- I repeat, only 50 -- of these composite images to train our algorithms to high efficiencies.