Unsupervised Contrastive Domain Adaptation for Semantic Segmentation.
Feihu ZhangVladlen KoltunPhilip H. S. TorrRené RanftlStephan R. RichterPublished in: CoRR (2022)
Keyphrases
- domain adaptation
- semantic segmentation
- weakly supervised
- semi supervised
- superpixels
- labeled data
- conditional random fields
- semi supervised learning
- multiple sources
- cross domain
- supervised learning
- unsupervised learning
- scene classification
- unlabeled data
- pairwise
- transfer learning
- object classes
- target domain
- sentiment classification
- co training
- topic models
- active learning
- object categories
- machine learning
- multi view
- multiscale
- named entities
- image classification
- training set
- object recognition
- bayesian networks
- training data