Embodied Visual Active Learning for Semantic Segmentation.
David NilssonAleksis PirinenErik GärtnerCristian SminchisescuPublished in: AAAI (2021)
Keyphrases
- semantic segmentation
- active learning
- street scenes
- label transfer
- superpixels
- scene classification
- weakly supervised
- conditional random fields
- semi supervised
- object categories
- visual features
- pascal voc
- supervised learning
- machine learning
- image classification
- object recognition
- image understanding
- object classes
- object class
- object classification
- long range
- pixel level
- markov random field
- image retrieval
- multiscale