Graph Cuts Always Find a Global Optimum for Potts Models (With a Catch).
Hunter LangDavid A. SontagAravindan VijayaraghavanPublished in: ICML (2021)
Keyphrases
- graph cuts
- global optimum
- markov random field
- belief propagation
- energy function
- energy minimization
- image segmentation
- shape prior
- mean shift
- multi label
- optimization method
- graph cut optimization
- graph cut segmentation
- spatial smoothness
- object segmentation
- global convergence
- objective function
- conditional random fields
- parameter estimation
- step size
- simulated annealing
- random walker
- image processing
- genetic algorithm