Lifting the Convex Conjugate in Lagrangian Relaxations: A Tractable Approach for Continuous Markov Random Fields.
Hartmut BauermeisterEmanuel LaudeThomas MöllenhoffMichael MöllerDaniel CremersPublished in: CoRR (2021)
Keyphrases
- markov random field
- convex relaxation
- graph cuts
- belief propagation
- mrf model
- semidefinite
- parameter estimation
- energy function
- np hard
- higher order
- pairwise
- globally optimal
- image segmentation
- energy minimization
- maximum a posteriori
- random fields
- piecewise constant functions
- potential functions
- convex optimization
- labeling problems
- lower bound
- conditional random fields
- low level vision
- wavelet transform
- multi label
- potts model
- image restoration
- optimal solution
- map inference
- loopy belief propagation
- textured images
- iterative conditional
- map estimation
- markov networks
- discriminative random fields
- linear programming
- min cut
- message passing
- image processing
- partition function
- lp relaxation
- multiple kernel learning
- graphical models
- convex hull