Most edges in Markov random fields for white matter hyperintensity segmentation are worthless.
Christopher G. SchwarzEvan FletcherBaljeet SinghAmy LiuNoel SmithCharles DeCarliOwen T. CarmichaelPublished in: EMBC (2012)
Keyphrases
- markov random field
- white matter
- gray matter
- image segmentation
- cerebrospinal fluid
- mrf model
- corpus callosum
- energy function
- magnetic resonance images
- grey matter
- low level vision
- textured images
- diffusion tensor images
- graph cuts
- human brain
- higher order
- random fields
- mri data
- energy minimization
- parameter estimation
- brain mri
- diffusion tensor imaging
- medical images
- segmentation method
- diffusion tensor
- partial volume effects
- maximum a posteriori
- segmentation algorithm
- fiber bundles
- conditional random fields
- image restoration
- pairwise
- mr brain images
- medical imaging
- mr images
- level set
- multiscale
- image analysis
- shape prior
- brain structures
- brain mr images
- region growing
- brain tissue
- texture segmentation
- edge information
- partial volume
- watershed transform
- shape model
- similarity measure
- probabilistic model
- gray level
- image gradient
- tensor field