Dataset & Code

Datasets

  1. Scribble KITTI for Scribble-Supervised LiDAR Semantic Segmentation
  2. ACDC Dataset: Semantic pixel-level annotations for 19 Cityscapes classes in adverse conditions
  3. Omni Auditory Perception Dataset: Binaural Sounds + 360 degree videos

Code:

  1. DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation (CVPR 2022)
  2. Scribble-Supervised LiDAR Semantic Segmentation (CVPR 2022)
  3. Continual Test-Time Domain Adaptation (CVPR 2022)