[2307.00371] Learning Content-enhanced Mask Transformer for Domain Generalized Urban-Scene Segmentation