Perturb-and-MPM: Quantifying Segmentation Uncertainty in Dense Multi-Label CRFs.
Raphael MeierUrspeter KnechtAlain JungoRoland WiestMauricio ReyesPublished in: CoRR (2017)
Keyphrases
- multi label
- graph cuts
- multi label classification
- text categorization
- segmentation algorithm
- conditional random fields
- image classification
- image annotation
- image segmentation
- semantic segmentation
- multi label learning
- label assignment
- max margin
- energy function
- binary classification
- segmentation method
- error correcting output coding
- text classification
- class labels
- scene classification
- multiple labels
- hierarchical text categorization
- object segmentation
- higher order
- information extraction
- web page prediction
- graphical models
- active learning
- similarity measure
- probabilistic model
- support vector machine
- semi supervised
- superpixels
- markov random field
- maximum a posteriori
- k nearest neighbor
- prediction accuracy