A two-layer Conditional Random Field for the classification of partially occluded objects.
Sergey KosovPushmeet KohliFranz RottensteinerChristian HeipkePublished in: CoRR (2013)
Keyphrases
- conditional random fields
- partially occluded objects
- restricted boltzmann machine
- sequence classification
- decision trees
- probabilistic model
- long range
- feature vectors
- hidden markov models
- markov random field
- graphical models
- text classification
- maximum entropy
- class labels
- generative model
- machine learning
- image classification
- higher order
- multi class
- crf model
- pairwise
- support vector
- information retrieval
- unsupervised learning
- segmentation method
- decision rules
- supervised learning
- search space
- dictionary learning
- semantic segmentation
- hough space