ST++: Make Self-trainingWork Better for Semi-supervised Semantic Segmentation.
Lihe YangWei ZhuoLei QiYinghuan ShiYang GaoPublished in: CVPR (2022)
Keyphrases
- semantic segmentation
- semi supervised
- weakly supervised
- street scenes
- supervised methods
- conditional random fields
- superpixels
- semi supervised learning
- label transfer
- scene classification
- unlabeled data
- object categories
- labeled data
- pairwise
- object classes
- supervised learning
- generative model
- multi view
- active learning
- image understanding
- image set
- pascal voc
- multiscale
- higher order
- object class
- markov random field
- object detection
- small number
- multi class