SegDA: Maximum Separable Segment Mask with Pseudo Labels for Domain Adaptive Semantic Segmentation.
Anant KhandelwalPublished in: CoRR (2023)
Keyphrases
- semantic segmentation
- street scenes
- superpixels
- conditional random fields
- scene classification
- weakly supervised
- object classes
- object categories
- label transfer
- object class
- bounding box
- pascal voc
- learning algorithm
- image understanding
- higher order
- object detection
- natural images
- image classification
- domain specific
- training data