Labeling Where Adapting Fails: Cross-Domain Semantic Segmentation with Point Supervision via Active Selection.
Fei PanFrançois RameauJunsik KimIn So KweonPublished in: CoRR (2022)
Keyphrases
- semantic segmentation
- cross domain
- superpixels
- conditional random fields
- transfer learning
- weakly supervised
- knowledge transfer
- object categories
- object classes
- text categorization
- pascal voc
- scene classification
- object class
- active learning
- object detection
- image set
- multiscale
- similarity measure
- target domain
- computer vision
- object recognition
- image understanding
- learning algorithm
- bag of words
- feature vectors
- domain specific
- hidden markov models