Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net

Open paper

Benchmark	Model	Rank	Results
all-day-semantic-segmentation-on-all-day	IB-Net	#3	mIoU: 64.5
domain-generalization-on-gta-to-avg	IBN	#21	mIoU: 34.63
robust-object-detection-on-dwd	IBN-Net	#10	mPC [AP50]: 25.5