Hard-aware Instance Adaptive Self-training for Unsupervised Cross-domain Semantic Segmentation.
Chuang ZhuKebin LiuWenqi TangKe MeiJiaqi ZouTiejun HuangPublished in: CoRR (2023)
Keyphrases
- cross domain
- semantic segmentation
- weakly supervised
- semi supervised
- superpixels
- transfer learning
- knowledge transfer
- conditional random fields
- training set
- object class
- text categorization
- scene classification
- unsupervised learning
- object categories
- object classes
- machine learning
- image understanding
- co occurrence
- image set
- semi supervised learning
- co training
- pascal voc
- learning algorithm
- data sets