Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation

Benchmark Model Rank Results
image-to-image-translation-on-synthia-to-cityscapesProDA#10mIoU (13 classes): 62.0
semantic-segmentation-on-gtav-to-cityscapes-labelsProDA#13mIoU: 57.5
image-to-image-translation-on-gtav-to-cityscapes-labelsProDA#17mIoU: 57.5
synthetic-to-real-translation-on-synthia-to-cityscapesProDA(ResNet-101)#18MIoU (16 classes): 55.5MIoU (13 classes): 62.0
domain-adaptation-on-gta5-to-cityscapesProDA#22mIoU: 57.5
synthetic-to-real-translation-on-gtav-to-cityscapes-labelsProDA(with knowledge distillation)#27mIoU: 57.5