---

DistanceFMM Class Reference

#include <DistanceFMM.hpp>

Inheritance diagram for DistanceFMM:

[legend]Collaboration diagram for DistanceFMM:

[legend]List of all members.

---

Detailed Description

Distance based Fast Marching Method.

Author:

Bernard De Cuyper

Version:

0.05

Date:

15/06/2002

 
Purpose:        Define a Fast Marching Method applied for direct distance maps
                        and distance based data feeding. (e.g.: OUT/IN tactic, and more)
                        Extensions to three dimension is obvious. My computer is too small
                        for 3D.                                                 

Note:           We may inherit from AnImageOp, but we think there is much more
                        in the fast marching that simply filtering and distance maps.
                        It is simply a fantastic idea of Sethian & all.

Paper:          "Optimal Algorithm for Shape from Shading and Path Planning.", 
                        Ron Kimmel, James A. Sethian, Kluweer Academic Publishers 2001 , Netherlands.
                

Bug:

last pixel is not selected

@ Copyrights: Bernard De Cuyper & Eddy Fraiha 2002, Eggs & Pictures. MIT/Open BSD copyright model.

Public Methods
	DistanceFMM (int amaskFlag=1)
virtual	~DistanceFMM ()
virtual void	setMaskActivity (int amaskFlag)
virtual double	getDistance (int x, int y)
virtual StatePsi	getLabel (int x, int y)
virtual void	setData (int w, int h, int x0, int y0)
	You may want to load the image part too, for advanced Fast Marching.
virtual void	setData (int w, int h, int n, int *xMask, int *yMask)
virtual void	setData (AnImage *im)
virtual int	getMaxIterations ()
virtual int	numberOfIteration ()
virtual void	setIteration (int i)
virtual bool	iterate ()
	simple iteration
virtual bool	iterate (int *x, int *y)
	iteration
virtual bool	iterate (int *x, int *y, double *rate)
	iteration
virtual RImage *	compute (RImage *dest=0)
	multiple FMM iterations
virtual void	sortedArea (AnImage *mask, int *npoints, int **xposition, int **yposition, bool inside=true)
	sort all pixels from the border to inside the object area
virtual void	sortedArea (AnImage *mask, int *npoints, int **xposition, int **yposition, double **rating, bool inside=true)
	sort all pixels from the border to inside the object area
virtual AnImage *	filter (AnImage *mask, AnImage *dest=0)
	like an image op generate a variant of a distance map or more
virtual AnImage *	selectedFilter (AnImage *mask, AnImage *src, AnImage *dest=0)
	Local filtering in a mask area.
virtual FImage *	ffilter (AnImage *mask, FImage *dest=0, bool dmapFlag=false)
	measures
virtual RImage *	rfilter (AnImage *mask, RImage *dest=0, bool dmapFlag=false)
	measures
virtual AnImage *	singlePath (AnImage *src, int x0, int y0, int x1, int y1, AnImage *dest=0)
	euclidian path
virtual void	report ()
Protected Methods
virtual double	getPotential (int i, int j)
	return a local potential value
virtual Psi2DValue *	select ()
virtual void	getSelectionNeighbour (int x, int y)
virtual void	resolve (int x, int y)
Protected Attributes
int	maskFlag
int	iteration
double	selectedDistance
double	root2
double	slowness
NeigbourSet *	neigbourSet
CandidateHeapList *	candidateList
int	iterMax
int	width
int	height
DomainOfAction *	dmap
	kind of distance map
int	xSelected
	current selected point position
int	ySelected
	current selected point position

---

Member Function Documentation

virtual double DistanceFMM::getPotential (  int    i, int    j )  [inline, protected, virtual]

return a local potential value Parameters: i  is the x position j  is the y position Returns : is local value like a distance, an edge strength, ... A distance map increase always. Overload this function gives Wondering Power e..g: using edge indicators, images, ... ;-) MAGIC, Think now! DO NOT FORGET: for good result the returned value MUST BE > 0 all the time (in the tracked area) Reimplemented in SimilarPixelFMM.

virtual bool DistanceFMM::iterate (  int *    x, int *    y, double *    rate )  [inline, virtual]

iteration Parameters: x  is the x position selected FMM point y  is the y position selected FMM point rate  is the calculated value of the selected FMM point Returns : boolean , test succeeding iteration, point tracking used for animation from border to ... :-) MAKE BURNING STUFF 0. DistanceFMM outIn; 1. outIn.setData(...) 2. while( outIn.iterate(...) ){ <What you like> } or rc= DistanceFMM.iterate(&x,&y); // n times Reimplemented in PathTrackingFMM.

virtual bool DistanceFMM::iterate (  int *    x, int *    y )  [inline, virtual]

iteration Parameters: x  is the x position selected FMM point y  is the y position selected FMM point Returns : boolean , test succeeding iteration, point tracking used for animation from border to ... :-) MAKE BURNING STUFF 0. DistanceFMM outIn; 1. outIn.setData(...) 2. while( outIn.iterate(...) ){ <What you like> } or rc= DistanceFMM.iterate(&x,&y); // n times Reimplemented in PathTrackingFMM.

AnImage * DistanceFMM::selectedFilter (  AnImage *    mask, AnImage *    src, AnImage *    dest = 0 )  [virtual]

Local filtering in a mask area. Parameters: mask  is AnImage* is a ByteImage src  is AnImage* is AnImage source channel dest  is AnImage* is AnImage result/placeholder Returns : AnImage* as result, Reimplemented from AnImageOp. Reimplemented in DTransportFMM, and SimilarPixelFMM.

---

The documentation for this class was generated from the following files:

• DistanceFMM.hpp
• DistanceFMM.cpp

Restoreinpaint sourceforge project `C++/Java Image Processing, Restoration, Inpainting Project'.

Bernard De Cuyper: Open Project Leader: Concept, design and development. Bernard De Cuyper & Eddy Fraiha 2002, 2003. Bernard De Cuyper 2004. Open and free, for friendly usage only. Modifications on Belgium ground of this piece of artistic work, by governement institutions or companies, must be notified to Bernard De Cuyper.