Graph cuts in computer vision: Difference between revisions

Revision as of 14:41, 21 June 2019

Revision as of 19:46, 21 June 2019

Line 33:

==== Likelihood / Color model / Regional term ====

<math>E_{\rm color}</math> — unary term describing the likelihood of each color.

* This term can be modeled using different local (e.g. texons{{not a typo}}) or global (e.g. histograms, GMMs, Adaboost likelihood) approaches that are described below.

===== Histogram =====

Line 46:

===== Texon =====

* A texon{{not a typo}} (or texton{{not a typo}}) is a set of pixels that has certain characteristics and is repeated in an image.

* Steps:

# Determine a good natural scale for the texture elements.

# Compute non-parametric statistics of the model-interior texons{{not a typo}}, either on intensity or on Gabor filter responses.

* Examples:

Line 60:

* Costs can be based on local intensity gradient, Laplacian zero-crossing, gradient direction, color mixture model,...

* Different energy functions have been defined:

** Standard [[Markov random field]] (MRF): Associate a penalty to disagreeing pixels by evaluating the difference between their segmentation label (crude measure of the length of the boundaries). See Boykov and Kolmogorov ICCV 2003

** [[Conditional random field]] (CRF): If the color is very different, it might be a good place to put a boundary. See Lafferty et al. 2001; Kumar and Hebert 2003

== Criticism ==