Abstract: Building deep learning models that can rapidly segment whole slide images (WSIs) using only a handful of training samples remains an open challenge in computational pathology. The difficulty ...