Efficient training for pairwise or higher order CRFs via dual decomposition.

Nikos Komodakis

Published in: CVPR (2011)

Keyphrases

higher order
pairwise
dual decomposition
markov random field
conditional random fields
structured prediction
high order
multi class
max margin
map inference
semi supervised
energy minimization
parameter learning
training set
supervised learning
parameter estimation
hidden markov models
graph cuts
shortest path