LTI-Lib: lti::regionGrowing Class Reference

It uses a set of seed pixels (given in the parameters) to start the region growing. To check the similarity, a comparition is done between the pixel value at the candidate position and the average value at the seed position. If the difference of these values is lower than a given threshold, the pixel will be considered as part of the growing region. It is possible to get automatically values for these thresholds if a known background region is given in the parameters.

 lti::loadBMP loader;

 lti::image img;     // the color image to be segmented
 lti::channel8 chnl; // the segmentation mask will be left here

 loader.load("img/leo040801_00_019.bmp",img); // read an image

 lti::regionGrowing segmenter;         // segmentation functor and
 lti::regionGrowing::parameters param; // its parameters

 param.useGaussKernel = true;            // use a gaussian kernel 5x5
 param.localStatisticsKernelSize = 5;
 param.localStatisticsKernelVariance = -1;

 param.mode = lti::regionGrowing::parameters::UseGivenThresholds;
 param.averageThresholds = trgbPixel<float>(0.05,0.05,0.05);
 param.edgesThreshold = 0.04;

 // do not use all corners as seed-points, just the three following...
 param.seedPoints.resize(3,tpoint<float>(0,0));
 param.seedPoints.at(1).y = 1.0f;
 param.seedPoints.at(2).x = 1.0f;
 param.seedPoints.at(2).y = 1.0f;

 // indicate the segmentation functor the user-parameters
 segmenter.setParameters(param);

 // segmentate the image and leave the segmentation in the channel
 segmenter.apply(img,chnl);

 view1.show(img);  // show original image
 view2.show(chnl); // show the segmentation mask

Constructor & Destructor Documentation

lti::regionGrowing::regionGrowing

(

const parameters &

par

)

Default constructor with parameters.

lti::regionGrowing::regionGrowing

(

const regionGrowing &

other

)

Copy constructor.

Parameters:

other

the object to be copied

Member Function Documentation

channel8& lti::regionGrowing::apply	(	const image &	src,
		channel8 &	dest
	)	const

Operates on a copy of the given parameters.

Parameters:

	src	image with the source data.
	dest	channel8 where the segmentation mask will be left. This mask contains 0 for a background point and (usually) 255 for an object point.

Returns:: a reference to the dest.

channel8& lti::regionGrowing::apply	(	const channel &	src,
		channel8 &	dest
	)	const

Operates on a copy of the given parameters.

Parameters:

	src	channel with the source data.
	dest	channel8 where the segmentation mask will be left. This mask contains 0 for a background point and (usually) 255 for an object point.

Returns:: a reference to the dest.

channel8& lti::regionGrowing::apply	(	const channel8 &	src,
		channel8 &	dest
	)	const

Operates on a copy of the given parameters.

Parameters:

	src	channel8 with the source data.
	dest	channel8 where the segmentation mask will be left. This mask contains 0 for a background point and (usually) 255 for an object point.

Returns:: a reference to the dest.

image& lti::regionGrowing::apply

(

image &

srcdest

)

const

Operates on the given parameter.

The background-pixels will be set to black, grey or white depending on the parameter "backgroundType".

Parameters:

srcdest

image with the source data. The result will be left here too.

Returns:: a reference to the srcdest.

channel& lti::regionGrowing::apply

(

channel &

srcdest

)

const

Operates on the given parameter.

The background-pixels will be set to 0, 0.5 or 1.0 depending on the parameter "backgroundType".

Parameters:

srcdest

channel with the source data. The result will be left here too.

Returns:: a reference to the srcdest.

channel8& lti::regionGrowing::apply

(

channel8 &

srcdest

)

const

Operates on the given parameter.

The background-pixels will be set to 0,128 or 255 depending on the parameter "backgroundType".

Parameters:

srcdest

channel8 with the source data. The result will be left here too.

Returns:: a reference to the srcdest.

virtual functor* lti::regionGrowing::clone ( ) const [virtual]

Returns a pointer to a clone of this functor.

Reimplemented from lti::segmentation.

regionGrowing& lti::regionGrowing::copy

(

const regionGrowing &

other

)

Copy data of "other" functor.

Parameters:

other

the functor to be copied

Returns:: a reference to this functor object

const parameters& lti::regionGrowing::getParameters

(

)

const

Returns used parameters.

Reimplemented from lti::functor.

virtual const char* lti::regionGrowing::getTypeName ( ) const [virtual]

Returns the name of this type ("regionGrowing").

Reimplemented from lti::segmentation.

template<class T>
void lti::regionGrowing::mask	(	matrix< T > &	img,
		const channel8 &	msk,
		const T &	background
	)	const `[inline, protected]`

All pixels at img for which the mask is 0 will be set to the given value.

The mask and img MUST have the same size!

channel8& lti::regionGrowing::segmentate	(	const channel &	src,
		channel8 &	dest
	)	const `[protected]`

Operates on a copy of the given parameters.

Parameters:

	src	image with the source data.
	dest	the mask will be 0 for background pixels

Returns:: a reference to the dest.

channel8& lti::regionGrowing::segmentate	(	const image &	src,
		channel8 &	dest
	)	const `[protected]`

Operates on a copy of the given parameters.

Parameters:

	src	image with the source data.
	dest	the mask will be 0 for background pixels

Returns:: a reference to the dest.

bool lti::regionGrowing::similar	(	const float &	neighbour,
		const float &	avrgAtSeed,
		const float &	avrgThreshold
	)	const `[inline, protected]`

The pixel at position "neighbour" will be compared with the average value corresponding to the position "seed".

The Codition for similarity proof if the absolute value of the difference between the channel value at "neighbour" and the local average on the "seed"-position is lower than the given avrgThreshold

bool lti::regionGrowing::similar	(	const ubyte &	neighbour,
		const ubyte &	avrgAtSeed,
		const int &	avrgThreshold
	)	const `[inline, protected]`

The pixel at position "neighbour" will be compared with the average value corresponding to the position "seed".

The Codition for similarity proof if the absolute value of the difference between the channel value at "neighbour" and the local average on the "seed"-position is lower than the given avrgThreshold


Public Member Functions
	regionGrowing ()
	regionGrowing (const parameters &par)
	regionGrowing (const regionGrowing &other)
virtual	~regionGrowing ()
virtual const char *	getTypeName () const
channel8 &	apply (channel8 &srcdest) const
channel &	apply (channel &srcdest) const
image &	apply (image &srcdest) const
channel8 &	apply (const channel8 &src, channel8 &dest) const
channel8 &	apply (const channel &src, channel8 &dest) const
channel8 &	apply (const image &src, channel8 &dest) const
regionGrowing &	copy (const regionGrowing &other)
virtual functor *	clone () const
const parameters &	getParameters () const
Protected Member Functions
bool	similar (const ubyte &neighbour, const ubyte &avrgAtSeed, const int &avrgThreshold) const
bool	similar (const float &neighbour, const float &avrgAtSeed, const float &avrgThreshold) const
template<class T>
void	mask (matrix< T > &img, const channel8 &msk, const T &background) const
channel8 &	segmentate (const image &src, channel8 &dest) const
channel8 &	segmentate (const channel &src, channel8 &dest) const
Classes
class	parameters
	The parameters for the class regionGrowing. More...

lti::regionGrowing Class Reference
[Segmentation Algorithms and related tools]

Public Member Functions

Protected Member Functions

Classes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

lti::regionGrowing Class Reference [Segmentation Algorithms and related tools]

Public Member Functions

Protected Member Functions

Classes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

lti::regionGrowing Class Reference
[Segmentation Algorithms and related tools]