Upload
colleen-morton
View
226
Download
0
Tags:
Embed Size (px)
Citation preview
GrabCutGrabCut Interactive Foreground Extraction Interactive Foreground Extraction
using Iterated Graph Cutsusing Iterated Graph Cuts
Carsten RotherCarsten RotherVladimir Kolmogorov Vladimir Kolmogorov
Andrew BlakeAndrew Blake
Microsoft Research Cambridge-UKMicrosoft Research Cambridge-UK
GrabCutGrabCut Interactive Foreground Extraction Interactive Foreground Extraction
using Iterated Graph Cutsusing Iterated Graph Cuts
Carsten RotherCarsten RotherVladimir Kolmogorov Vladimir Kolmogorov
Andrew BlakeAndrew Blake
Microsoft Research Cambridge-UKMicrosoft Research Cambridge-UK
Graph Cuts modelling in imagesGraph Cuts modelling in images Graph Cuts modelling in imagesGraph Cuts modelling in images
GrabCut – Interactive Foreground Extraction 5
ImageImage Min CutMin Cut
Cut: separating source and sink; Energy: collection of edges
Min Cut: Global minimal enegry in polynomial time
Foreground Foreground (source)(source)
BackgroundBackground(sink)(sink)
Graph Cuts for foreground extractionGraph Cuts for foreground extraction Graph Cuts for foreground extractionGraph Cuts for foreground extraction
GrabCut – Interactive Foreground Extraction 5
Min CutMin Cut
Foreground Foreground (source)(source)
BackgroundBackground(sink)(sink)
Assume we know foreground is white and background is black
GrabCut – Interactive Foreground Extraction 5
Min CutMin Cut
Foreground Foreground (source)(source)
BackgroundBackground(sink)(sink)
Assume we know foreground is white and background is black
Data term = (cost of assigning label)
Regularization = (cost of separating neighbors)
Graph Cuts for foreground extractionGraph Cuts for foreground extraction Graph Cuts for foreground extractionGraph Cuts for foreground extraction
GrabCut – Interactive Foreground Extraction 5
Min CutMin Cut
Foreground Foreground (source)(source)
BackgroundBackground(sink)(sink)
Assume we know foreground is white and background is black
Data term = whiteness(cost of assigning label)
Regularization =(cost of separating neighbors)
Graph Cuts for foreground extractionGraph Cuts for foreground extraction Graph Cuts for foreground extractionGraph Cuts for foreground extraction
GrabCut – Interactive Foreground Extraction 5
Min CutMin Cut
Foreground Foreground (source)(source)
BackgroundBackground(sink)(sink)
Assume we know foreground is white and background is black
Data term = whiteness(cost of assigning label)
Regularization = color match (cost of separating neighbors)
Graph Cuts for foreground extractionGraph Cuts for foreground extraction Graph Cuts for foreground extractionGraph Cuts for foreground extraction
We are all set now !We are all set now ! We are all set now !We are all set now !
User Initialisation
Learn foreground color model
Graph cuts to infer the
foreground
?
GrabCut – Interactive Foreground Extraction 6
Iterated Graph CutsIterated Graph Cuts Iterated Graph CutsIterated Graph Cuts
User Initialisation
Learn foreground color model
Graph cuts to infer the
foreground
?
GrabCut – Interactive Foreground Extraction 6
1 2 3 4
Iterated Graph CutsIterated Graph Cuts Iterated Graph CutsIterated Graph Cuts
GrabCut – Interactive Foreground Extraction 7
Energy after each IterationResult
Guaranteed to
converge
Moderately straightforward Moderately straightforward
examples examples
Moderately straightforward Moderately straightforward
examples examples
… GrabCut completes automatically
GrabCut – Interactive Foreground Extraction 10