Discovering Object Masks with Transformers for Unsupervised Semantic Segmentation.
Wouter Van GansbekeSimon VandenhendeLuc Van GoolPublished in: CoRR (2022)
Keyphrases
- semantic segmentation
- bounding box
- weakly supervised
- supervised methods
- superpixels
- object classes
- object class
- object segmentation
- label transfer
- object categories
- scene classification
- pascal voc
- object model
- conditional random fields
- object detection
- object detectors
- topic models
- object recognition
- training set
- pixel level
- d objects
- image processing
- supervised learning
- multi class
- long range
- probabilistic model
- multiscale