Classes Matter: A Fine-grained Adversarial Approach to Cross-domain Semantic Segmentation

Benchmark Model Rank Results
image-to-image-translation-on-synthia-to-cityscapesFADA (ResNet-101)#14mIoU (13 classes): 52.5
domain-adaptation-on-synthia-to-cityscapesFADA (ResNet-101)#26mIoU: 45.2
domain-adaptation-on-synthia-to-cityscapesFADA (VGG-16)#29mIoU: 39.5
synthetic-to-real-translation-on-synthia-to-cityscapesFADA(ResNet-101)#30MIoU (16 classes): 45.2MIoU (13 classes): 52.5
synthetic-to-real-translation-on-gtav-to-cityscapes-labelsFADA#47mIoU: 50.1