Open-Vocabulary Semantic Segmentation with Mask-adapted CLIP.
Feng LiangBichen WuXiaoliang DaiKunpeng LiYinan ZhaoHang ZhangPeizhao ZhangPeter VajdaDiana MarculescuPublished in: CoRR (2022)
Keyphrases
- semantic segmentation
- street scenes
- superpixels
- label transfer
- conditional random fields
- scene classification
- weakly supervised
- object class
- object categories
- bounding box
- object classes
- image set
- image understanding
- object recognition
- bayesian networks
- multiscale
- long range
- image classification
- graphical models
- pixel level
- higher order
- co occurrence
- multi class