Online Domain Adaptation for Semantic Segmentation in Ever-Changing Conditions.
Theodoros PanagiotakopoulosPier Luigi DovesiLinus Härenstam-NielsenMatteo PoggiPublished in: CoRR (2022)
Keyphrases
- domain adaptation
- semantic segmentation
- conditional random fields
- semi supervised
- multiple sources
- labeled data
- transfer learning
- test data
- superpixels
- scene classification
- semi supervised learning
- object categories
- cross domain
- sentiment classification
- object classes
- target domain
- active learning
- object recognition
- co occurrence
- data mining
- reinforcement learning
- training data
- decision trees
- information retrieval