Unsupervised domain adaptation for semantic segmentation via cross-region alignment.
Zhijie WangXing LiuMasanori SuganumaTakayuki OkataniPublished in: Comput. Vis. Image Underst. (2023)
Keyphrases
- semantic segmentation
- street scenes
- conditional random fields
- superpixels
- scene classification
- label transfer
- weakly supervised
- object categories
- input image
- image understanding
- image regions
- pascal voc
- object recognition
- region of interest
- object classes
- object class
- natural images
- image classification
- markov random field
- computer vision
- long range
- pixel level
- image features