Contextual unmixing of geospatial data based on Markov random fields and conditional random fields.
Ryuei NishiiTomohiko OzakiPublished in: WHISPERS (2009)
Keyphrases
- conditional random fields
- geospatial data
- markov random field
- spatial data
- random fields
- graph cuts
- higher order
- data collection
- pairwise
- data processing
- belief propagation
- maximum a posteriori
- maximum entropy
- data sources
- multiple sources
- energy function
- parameter learning
- crf model
- image segmentation
- parameter estimation
- data management
- efficient inference
- structured prediction
- potential functions
- loopy belief propagation
- message passing
- approximate inference
- graphical models
- superpixels
- energy minimization
- undirected graphical models
- image labeling
- markov networks
- computer vision
- partition function
- hidden markov models
- semi supervised